a theory-based analysis of coercion in addiction treatment · addiction treatment through a...
TRANSCRIPT
A Theory-Based Analysis of Coercion in Addiction Treatment
by
Karen Anne Urbanoski
A thesis submitted in conformity with the requirements for the degree of Doctor of Philosophy
Dalla Lana School of Public Health University of Toronto
© Copyright by Karen Anne Urbanoski, 2010
ii
A Theory-Based Analysis of Coercion in Addiction Treatment
Karen Anne Urbanoski
Doctor of Philosophy
Dalla Lana School of Public Health University of Toronto
2010
Abstract
The use of coercion to induce entry to addiction treatment is controversial and a large body of
research has accumulated considering ethical issues, benefits, and repercussions. However,
development of evidence-based policy and practices is hampered by limitations of existing
literature. Theoretical and empirical work on self-determination suggests that perceptions of
coercion have negative implications for motivation, behaviour change, and psychological well-
being; however, these insights have not generally informed research on coerced treatment. The
present work seeks to further understandings of the meaning and effectiveness of coerced
addiction treatment through a theory-based, prospective study of coercion and treatment
processes. The sample includes 276 adults admitted to an outpatient counseling program for
alcohol- and drug-related problems. At admission, participants completed questionnaires on
motivation, perceived coercion, and pressures to enter treatment. Two months later, a second
questionnaire assessed engagement in treatment and substance problem severity (follow-up rate
= 74.3%). Retention was determined via self-report and agency records. Analysis was guided by
a conceptual model based on Self-Determination Theory. Perceived coercion at admission was
associated with greater pressures from legal and informal sources, and lower substance problem
severity. Fewer than half (45.7%) of participants were still attending treatment at 2-month
follow-up. Clients who reported greater coercion were more likely to leave treatment within the
iii
first 2 months, and to qualify that decision by statements indicating a lack of perceived need for
continued treatment. Greater autonomous motivation was associated with higher client
confidence in treatment, and lower perceived coercion and greater informal pressure were
associated with greater resolution of substance problems in the weeks following admission. This
work contributes empirical evidence to ongoing debates over the legitimacy of coerced addiction
treatment by reframing relevant concepts in terms of client perspectives and evaluating the
impact on treatment processes. Results raise questions about previous conclusions of the
effectiveness of coerced treatment and suggest many future avenues for research. In particular,
research is needed to evaluate the longer-term implications of coercion and the changing nature
of perceptions and motivation during treatment.
iv
Acknowledgments I would first like to thank Karen Parsons and the staff at PAARC, particularly Nolle, Robyn, and
Ryan, for generously opening their doors to this research and providing continued support
throughout the study. I am grateful for the financial support provided for my doctoral degree
from the Ontario Graduate Scholarships in Science and Technology and the Canadian Institutes
of Health Research Strategic Training Program in Research in Addictions and Mental Health
Policy and Services. I would also like to thank the Research Office at the Centre for Addiction
and Mental Health for providing the participant fees for this research. Thanks are extended to
Drs. Cameron Wild, Jurgen Rehm, Susan Bondy, and Edward Adlaf for their constructive
guidance in planning, conducting, and reviewing this work. Special thanks are extended to Joan
Urbanoski for her valuable editorial support and to Jay Larson for his unwavering patience and
encouragement. I wish to also express my sincere gratitude to Dr. Brian Rush, my thesis
supervisor and mentor, for past and continued support in this project and my career in general.
Your willingness to always help me out, or just listen to me, goes beyond the call of duty.
Finally, I sincerely thank all of the clients at PAARC who shared their time and experiences for
the purposes of this study.
v
Table of Contents Acknowledgments.......................................................................................................................... iv
Table of Contents............................................................................................................................ v
List of Tables ................................................................................................................................ vii
List of Figures .............................................................................................................................. viii
List of Appendices ......................................................................................................................... ix
Chapter 1 Introduction .................................................................................................................... 1
Chapter 2 Literature Review........................................................................................................... 5
2.1 External pressures to enter addiction treatment .................................................................. 5
2.1.1 Legal pressures........................................................................................................ 6
2.1.2 Non-legal formal pressures ................................................................................... 10
2.1.3 Informal social pressures ...................................................................................... 14
2.2 Perceived coercion ............................................................................................................ 19
2.3 Treatment process ............................................................................................................. 26
2.4 Theoretical perspectives on help-seeking and health behaviour change .......................... 35
2.4.1 An overview of Self-Determination Theory ......................................................... 36
2.4.2 Self-Determination Theory and coerced addiction treatment ............................... 41
2.5 Objectives ......................................................................................................................... 51
Chapter 3 Methods........................................................................................................................ 53
3.1 Participants........................................................................................................................ 53
3.2 Setting ............................................................................................................................... 57
3.3 Measures ........................................................................................................................... 58
3.4 Procedures......................................................................................................................... 62
3.5 Conceptual framework...................................................................................................... 65
3.6 Analysis............................................................................................................................. 66
vi
3.7 Sample size calculations ................................................................................................... 70
Chapter 4 Results .......................................................................................................................... 72
4.1 Admission process ............................................................................................................ 72
4.2 Early treatment process..................................................................................................... 80
4.2.1 Two-month retention ............................................................................................ 80
4.2.2 Treatment engagement.......................................................................................... 85
4.2.3 Client-reported reasons for leaving treatment....................................................... 89
4.3 Early progress in treatment ............................................................................................... 91
Chapter 5 Discussion .................................................................................................................... 96
5.1 Main Findings and Implications ....................................................................................... 98
5.1.1 Levels of pressures, coercion, and motivation at treatment entry......................... 98
5.1.2 The interplay of social pressures, coercion, and motivation............................... 100
5.1.3 Other predictors of perceived coercion............................................................... 102
5.1.4 Characteristics of the treatment process ............................................................. 104
5.1.5 Autonomy and the treatment process.................................................................. 107
5.1.6 Other factors relevant to the treatment process................................................... 111
5.2 Study limitations ............................................................................................................. 114
5.3 Conclusions..................................................................................................................... 116
References................................................................................................................................... 119
Appendices.................................................................................................................................. 142
vii
List of Tables Table 1. Characteristics of the study sample at admission ................................................ 55
Table 2. Descriptive statistics for admission process measures......................................... 73
Table 3. Spearman rank correlations (ρ) between admission variables ............................. 76
Table 4. Client characteristics by perceived coercion at admission .................................. 77
Table 5. Linear regression predicting perceived coercion ................................................. 79
Table 6. Admission process and client characteristics by 2-month retention.................... 81
Table 7. Logistic regression predicting 2-month retention ................................................ 83
Table 8. Reduced three-variable models predicting 2-month retention............................. 84
Table 9. Descriptive statistics for measures of 2-month treatment engagement ............... 85
Table 10. Admission process measures by 2-month engagement in treatment ................... 86
Table 11. Client characteristics by 2-month treatment engagement .................................... 87
Table 12. Linear regression predicting client confidence in treatment................................ 88
Table 13. Linear regression predicting counsellor rapport .................................................. 88
Table 14. Linear regression predicting client commitment to treatment ............................. 89
Table 15. Admission process measures by perceived need for treatment at 2-months among
former clients .............................................................................................................................. 91
Table 16. Admission process measures by 2-month substance problem resolution............ 92
Table 17. Client characteristics by 2-month substance problem change ............................. 93
Table 18. Regression predicting change in past-month substance-related problems ......... 95
viii
List of Figures Figure 1. TCU treatment model ................................................................................................... 30
Figure 2. Flow diagram of study and treatment ........................................................................... 54
Figure 3. Conceptual framework for an SDT-based analysis of addiction treatment process..... 65
Figure 4. Perceived coercion........................................................................................................ 74
Figure 5. Legal pressure............................................................................................................... 74
Figure 6. Other formal (non-legal) pressure ................................................................................ 75
Figure 7. Informal pressure ......................................................................................................... 75
ix
List of Appendices Appendix 1. Measures ............................................................................................................ 142
Appendix 2. Consent letter ..................................................................................................... 149
Appendix 3. Attrition analysis................................................................................................ 152
Appendix 4. Parametric tests for non-normally distributed variables .................................... 153
Appendix 5. Regression diagnostics and alternative models.................................................. 158
1
Chapter 1 Introduction
Substance use and its associated consequences pose a significant threat to population health. The
Global Burden of Diseases project by the World Health Organization identifies tobacco and
alcohol as the fourth and fifth leading causes of disease and disability, respectively (Ezzati,
Lopez, Rodgers, Vander Hoorn, & Murray, 2002). In addition to a host of threats to individual
physical and mental health and psychosocial functioning, substance use is associated with
consequences to societal functioning and threats to public safety from drug-related property
crime, the spread of AIDS and other infectious diseases, and impaired driving. Recent estimates
suggest that approximately $40 billion was spent in Canada on substance use in 2002 alone,
including direct health care and law enforcement costs and indirect productivity losses (Rehm et
al., 2006).
One response to these threats is formal treatment. Because of its dual function in potentially
aiding both the individual and society as a whole, it is a potent strategy for social control. There
are many reasons why people enter treatment for substance use; however, many will do so only
after external pressure. In 1980, Room wrote of the four L’s that bring people into treatment for
alcohol problems: liver, lover, livelihood and the law. This statement highlights the presumption
that, outside of feeling ill, people seek help because they are pressured to do so by their spouses,
employers, and/or the legal authorities.
The use of coercion in initiating addiction treatment is controversial, and a large body of
research has accumulated considering the ethical issues, benefits, and repercussions. Proponents
claim that, given the potential benefits to both the individual and society, formal use of the
coercive potential of the courts and other institutions to provide a motivating force to enter and
remain in treatment is legitimate (Fagan, 1999; Miller & Flaherty, 2000; Nace et al., 2007). To
the extent that treatment is more successful at rehabilitation than is incarceration, both the
individual and the public are theoretically better off if drug-abusing offenders are treated
(Gostin, 1991). The legitimacy of coercion has also been argued on the basis of neurobiological
changes that occur with addiction, compromising the individual’s ability to make decisions in
their best interest and warranting intervention by medical and legal authorities (Caplan, 2006;
2
Nace et al., 2007; Sullivan et al., 2008). In this sense, mandated or forced treatment is viewed as
a way to restore autonomy and the capacity for choice to those who are not acting with self-
determination (Caplan, 2006; Nace et al., 2007; Sullivan et al., 2008).
Criticisms of the use of coercion to initiate treatment include the limited ability of the system to
accommodate and retain a large number of involuntary clients – a relevant concern in the face of
limited resources (Gostin, 1991; Platt, Buhringer, Kaplan, Brown, & Taube, 1988; Sowers &
Daley, 1993). To the extent that coerced clients are less motivated to participate in treatment, it
is controversial to provide them with already limited spaces. Importantly, coercive strategies
may be successful in initiating treatment, but they do not ensure participation (Fagan, 1999).
Conflicting ideologies, values and goals may also hinder the ability of the treatment system to
effectively collaborate with the legal system (Fagan, 1999; Fischer, Roberts, & Kirst, 2002;
Wild, 1999a) and other social service systems (Drabble, 2007). Conflicts may arise in terms of
how to deal with confidentiality and occasional relapses when professionals not directly
involved in the treatment process are monitoring its success (Fagan, 1999; Weisner, 1987). Such
tensions were confirmed in a recent evaluation of the impact on treatment programs of a law
providing for expanded mandated treatment of drug-involved offenders in California (Niv,
Hamilton, & Hser, 2009). The treatment providers consulted in this study reported considerable
impact on clinical decision-making in their programs resulting from the increased collaboration
with the legal system, around issues such as assigning level of care, discharging clients because
of non-compliance, and treatment duration. Overall, providers perceived a reduction in the
flexibility they were afforded when responding to their clients’ needs.
The abstinence orientation of treatment based in or monitored by the criminal justice system is
arguably not reflective of a chronic disease model of addiction, which calls for a recognition of
the role of relapse and the potential for multiple treatment episodes over the course of recovery
(Covington, 2001; McLellan, Lewis, O'Brien, & Kleber, 2000; Werb et al., 2007). Abstinence-
based programs with punitive sanctions may not be suitable for all individuals with substance
use problems, especially those with more severe substance dependence. Such individuals may
be at a higher risk of failing, thereby incurring additional punishment rather than treatment.
More generally, an emphasis on the use of compulsory treatment to the benefit and protection of
society over that of the individual may result in the imposition of treatment even when it is
3
found to be ineffective, or over a longer period of time than is strictly necessary for treatment
purposes (Fagan, 1999; Gostin, 1991). It has been suggested that treatment provided in, or
mediated by, the legal system may be driven less by client need than by local practices and
policies for dealing with drug-using offenders (Stevens et al., 2005).
Implications of the widespread use of coercive strategies to bring people into treatment include
changes to the nature of the client base and the purpose of treatment (Schmidt & Weisner, 1993;
Weisner, 1987, 1990). When the system shifts from treating those with severe alcohol
dependence, who have “hit bottom” and recognize a need for treatment, to treating problem
drinkers and drug users who may not yet have made a connection between their substance use
and the problems they are experiencing, the focus of treatment programs must also shift to
overcoming denial and developing strategies to engage a larger and potentially more resistant
client population. Increases in the proportion of users with sociolegal problems but less severe
addiction problems also suggest that treatment is increasingly being used as a method of social
control. Importantly, research indicates that those referred or mandated by legal and social
systems or by their employers are younger, less severely addicted, and less likely to have
previous treatment experience than others entering the treatment system (Brecht, Anglin, &
Dylan, 2005; Copeland & Maxwell, 2007; Friedmann, Lemon, Stein, & D'Aunno, 2003;
Grichting, Uchtenhagen, & Rehm, 2002; Kelly, Finney, & Moos, 2005; Kline, 1997; Marshall
& Hser, 2002; Polcin & Beattie, 2007; Polcin & Weisner, 1999; Rush & Wild, 2003). This is
sometimes used to suggest that coercion is an effective early case-finding strategy, bringing
people into the treatment before their addiction and other health and social problems become
severe (Marlowe, Patapis, & DeMatteo, 2003; Sowers & Daley, 1993). However, the line
between individual and societal benefit is blurred. Balancing issues of public good with those of
individual rights is paramount to considerations of the prevention and alleviation of substance
problems through treatment.
Previous work has established that a variety of formal and informal external pressures are
commonly applied to induce treatment entry. However, the development of evidence-based
policy and treatment practices is hampered by the numerous limitations of this body of literature
(Wild, 2006). These limitations include: a focus on non-empirical arguments defending or
denouncing the use of coercive strategies; prioritization of legal strategies over other forms of
pressure; lack of recognition of the heterogeneity in the implementation of social pressures
4
across types and jurisdictions; neglect of iatrogenic effects of treatment; and exclusion of
stakeholder (i.e., client and service provider) perspectives on coercion.
Wild (2006) also highlights the limitations of commonly employed measures of coercion, which
are primarily informed by health services or objective measures that neglect client perceptions
and internal motivational processes. That is, the majority of work in this area has defined
coerced treatment in terms of referral source or the presence of monitoring conditions or
reinforcements, but has neglected the implications of these measures for client motivation,
interest, or intent in pursuing treatment. Theoretical and empirical work on self-determination
suggests that perceptions of coercion and threats to autonomy have negative implications for
motivation, successful and sustainable behaviour change, and psychological well-being;
however, these insights have not generally informed research on coerced addiction treatment.
This study seeks to further our understandings of the meaning and effectiveness of coerced
addiction treatment through a comprehensive critical review and consolidation of existing
empirical and theoretical literature and an original theory-based study of coercion, motivation,
and treatment processes.
5
Chapter 2 Literature Review
2.1 External pressures to enter addiction treatment There are a variety of ways in which individuals with substance use problems are coerced,
induced, or otherwise persuaded to enter addiction treatment. Wild (2006) separates these
external pressures or social control strategies into three broad types according to their source.
Legal pressures include civil commitment, court-ordered treatment, and diversion-to-treatment
programs, such as drug treatment courts. Formal non-legal pressures are those mitigated by
non-legal institutions or systems, including mandatory treatment referrals by employers,
schools, children’s aid or social assistance programs. Informal social pressures refer to forms of
interpersonal persuasion, including threats and ultimatums by friends and family. In line with
evidence of the important role played by these agents in the treatment entry process
(summarized below), this typology explicitly recognizes the coercive potential that exists both
inside and outside the legal system and offers a comprehensive framework for external pressures
that is useful in the study of coerced addiction treatment.
Studies of substance use treatment systems suggest that these strategies are frequently used to
induce or encourage treatment entry. In Ontario in 2008-09, 22% of clients receiving addiction
treatment (corresponding to over 23,000 people) reported a condition attached to treatment
entry, including treatment as a condition of probation or parole, child custody, receipt of social
assistance, continued school attendance or family contact (Drug and Alcohol Treatment
Information System, 2009). In addition, just over 8% (or roughly 9,100 people) reported being
referred by the legal system, with a similar proportion being referred by friends or family. Since
its inception in 1999, over 25,000 individuals have completed Back on Track, Ontario’s
remedial measures program for convicted impaired drivers (Mann et al., 2007). Drawing on
nation-wide data from 2002, another study estimated that approximately 17,500 individuals
participated in such remedial measures programs, which are mandatory for license
reinstatement, across Canada in that year (Smart, Mann, Stoduto, & Thomas, 2008). This
corresponds to only 56% of those convicted of drinking-related offences during this time,
6
suggesting that this figure has the potential to almost double if these programs included a greater
proportion of those convicted.
In the United States, the legal system plays an even greater role in the addiction treatment entry
process. In the latest national drug treatment evaluation initiative, the Drug Abuse Treatment
Outcome Study (DATOS), conducted in the early 1990s, 43% of clients were referred by the
legal system (Hubbard, Craddock, Flynn, Anderson, & Etheridge, 1997). This represented an
increase from 31% found in the previous national evaluation initiative conducted a decade
earlier. In 2006, a system-wide estimate of 38% of admissions to publicly-funded addiction
treatment programs were referred by the legal system (Substance Abuse and Mental Health
Services Administration, 2008).
The impact of legal, formal non-legal, and informal pressures to enter treatment is undoubtedly
large in shaping the clientele and focus of addiction services and supports around the world. The
nature and evidence base for each of these three sources of pressures is considered in more
detail below.
2.1.1 Legal pressures
Overall, clients who enter treatment under legal pressures show comparable responses to those
who are not mandated or referred by the legal system, and this is typically interpreted as
evidence of their effectiveness (Anglin, Brecht, & Maddahian, 1989; Brecht, Anglin, & Wang,
1993; Collins & Allison, 1983; Grichting et al., 2002; Kelly et al., 2005; Leukefeld & Tims,
1988; Polcin, 2001). However, what constitutes legal pressure across these studies is not
uniform, involving variable levels of initial coercion and ongoing monitoring and surveillance.
By extension, the treatment conditions of the comparison groups used in these studies are also
variable. Perhaps not surprisingly then, conflicting results are reported in the literature (Stevens
et al., 2005). Taking two recent studies as examples, clients who report legal concerns and
pressures at admission have shown both higher rates of abstinence (Burke & Gregoire, 2007)
and higher rates of drug use (Longshore & Teruya, 2006) at 6-month follow-up.
Evaluations of treatment under legal controls tend to use non-equivalent comparison groups
comprised of a heterogeneous mix of clients, who may or may not be pressured or mandated to
treatment by non-legal formal or informal sources. To the extent that mandated clients are
systematically younger and at an earlier stage in their addiction and treatment careers,
7
interpretations of differences in outcomes between such groups, and attributions of these
differences to treatment, should be made cautiously. Notably, a recent review by the Cochrane
Collaboration of drug treatment programs for offenders found little evidence supporting existing
interventions in either prison or community settings when only randomized experimental studies
were considered (Perry et al., 2006).
Retention is the outcome most commonly examined in evaluations of mandated treatment
(Wild, Roberts, & Cooper, 2002). There is evidence that legally referred clients are retained in
treatment longer (Brecht et al., 2005; Collins & Allison, 1983; Copeland & Maxwell, 2007;
Grichting et al., 2002; Knight, Hiller, Broome, & Simpson, 2000; Leukefeld & Tims, 1988;
Maglione, Chao, & Anglin, 2000; Young, 2002; Young & Belenko, 2002). In turn, duration of
treatment, at least for some minimum period of time, is consistently reported to be associated
with positive outcomes (Hubbard, Craddock, & Anderson, 2003; Moos & Moos, 2003;
Simpson, 2004; Stark, 1992; Zhang, Friedmann, & Gerstein, 2003). Again, however, the
research is equivocal: a number of studies have reported an inverse association between legal
pressure and retention (Beynon, Bellis, & McVeigh, 2006; Claus & Kindleberger, 2002;
Longshore & Teruya, 2006; Mertens & Weisner, 2000; Stevens et al., 2005), particularly in
methadone maintenance treatment (Desmond & Maddux, 1996; Joe, Simpson, & Broome, 1998,
1999). Within this modality specifically, it is possible that negative perceptions of methadone by
legal authorities in the United States may contribute to shorter retention (Knight et al., 2000).
There is also evidence to suggest that legal involvement is related to longer retention only when
it is perceived by clients to be a contributing factor related to treatment entry (Vickers-Lahti et
al., 1995). In this study, legal involvement that was recorded in the medical chart, but not
reported by clients as a reason for admission, was not associated with retention. Findings such
as these provide a potential explanation for the conflicting findings of previous studies and
highlight the critical importance of devising and specifying appropriate definitions.
Finally, relative to others in treatment, the long-term impact of treatment under legal controls is
largely unknown. There is evidence that initially beneficial outcomes of legally-initiated
treatment do not persist after the threat of sanctions is lifted (Anglin & Hser, 1991; Hser,
Yamaguchi, Chen, & Anglin, 1995; Weisner, 1990). An evaluation of addiction treatment
delivered within the correctional system of Canada highlighted the importance of treatment
continuity in the community to achieving the best outcomes related to criminal recidivism post-
8
release (Porporino, Robinson, Millson, & Weekes, 2002). Many outcome evaluations of
mandated treatment tend to include time on parole and probation in their post-treatment follow-
up periods, while the client is still being monitored by the legal system (Covington, 2001).
Covington suggests that, for programs in which there is a close link between treatment and
punishment, post-treatment outcomes can only be assessed once the period of supervision and
risk of punishment has ended.
Over the past 20 years, the role of the legal system in addiction treatment has become
increasingly realized through formalized referral and collaboration mechanisms for offenders;
most notably, mandatory programs for impaired driving offenders and an ever-expanding litany
of drug treatment and other problem-solving courts. As of 2002, with the exception of British
Columbia, all Canadian provinces had in place province-wide mandatory educational or
therapeutic interventions for license reinstatement for those convicted of impaired driving
offences (Health Canada, 2004). A meta-analysis of studies evaluating remedial measures
programs found a small but consistent reduction in drinking-driving recidivism and alcohol-
related crashes associated with treatment, relative to no treatment or punishment via license
suspension, fines or jail terms (Wells-Parker, Bangert-Drowns, McMillen, & Williams, 1995).
An updated review of the literature offered further support for the effectiveness of these
programs in reducing impaired driving recidivism (Health Canada, 2004), although both reviews
note the variable methodological rigour of evaluation work in this area. Ongoing evaluation of
the Back on Track remedial measures program in Ontario has also reported significant
reductions in alcohol and drug use between assessment and 6-month follow-up (Mann et al.,
2007; Sharpley, Flam Zalcman, Mann, Brands, & Stoduto, 2007).
Drug treatment courts represent a particularly interesting case example of treatment under legal
controls, because of the close and ongoing collaboration between the treatment and criminal
justice systems. Briefly, drug treatment courts are specialized courts offering closely supervised
treatment as an alterative to jail or prison time for offenders convicted of non-violent, often
drug-related, crimes (Werb et al., 2007). They were introduced in the United States in the late
1980s in response to the high volume of drug-related offences coming before the courts, and the
apparent lack of effectiveness of incarceration in deterring future criminal activity in this
population (Cooper, 2003; Harrison & Scarpitti, 2002). Despite a fairly weak evidence base,
drug treatment courts are highly popular, having grown to number over 1,600 in the United
9
States in the past 20 years, with programs now implemented in Australia, Canada, and
throughout the United Kingdom (Werb et al., 2007).
Evaluation work has largely concluded that these programs are successful in reducing drug use
and criminal activity for the duration of the program, although the impact on longer-term post-
treatment outcomes is less clear (Belenko, 2001, 2002; Sanford & Arrigo, 2005; Werb et al.,
2007). Further, the magnitude of the impact on criminal recidivism, the most commonly studied
outcome, is highly variable across programs and jurisdictions (Belenko, 2001, 2002; Sanford &
Arrigo, 2005). Although descriptive summaries of “key components” are available in the
literature (Hora, 2002), in reality drug treatment court programs vary widely in operational
procedures, both across jurisdictions and within jurisdictions over time (Belenko, 2002; Cooper,
2003; King & Pasquarella, 2009; Taxman & Bouffard, 2002). This complicates the task of
identifying critical elements central to program success through comparisons across evaluations.
Few empirical studies are available that evaluate and link specific activities and services with
program outcomes, in general and by specific types of clients (Garrity et al., 2008; Goldkamp,
White, & Robinson, 2001; Marlowe, Festinger, Dugosh, & Lee, 2005; Marlowe, Festinger, Lee,
Dugosh, & Benasutti, 2006; Roll, Prendergast, Richardson, Burdon, & Ramirez, 2005).
Similar to evaluation work on mandated treatment in general, evaluations of drug treatment
courts have been of variable quality. Methodological weaknesses include the use of non-
experimental designs and non-equivalent comparison groups, high attrition rates, short follow-
up periods, failure to collect post-treatment data on drug use and other outcomes, and the use of
incomplete costing data when drawing conclusions of cost-effectiveness (Fischer, 2003; Fischer
et al., 2002; Werb et al., 2007). The lack of randomized control studies is of particular concern.
Comparison groups tend to be comprised of non-participating offenders (i.e., pre-drug treatment
court offenders, those processed through traditional courts, and those who opted-out of the
program), or non-graduates (i.e., those who were expelled or withdrew from the program). This
potentially introduces serious biases due to systematic differences in baseline levels of substance
problem severity, criminal history, co-occurring problems, and motivation for behaviour change
– factors that are likely associated with outcomes such as recidivism and relapse to substance
use (Belenko, 2002; Fischer, 2003; Werb et al., 2007).
10
Reviewing the evidence for drug treatment courts in the Canadian context specifically, Werb et
al. (2007) conclude that evaluations conducted to date of the Vancouver and Toronto programs
have failed to demonstrate a positive impact on criminal recidivism or substance relapse. Both
programs exhibited high rates of program failure, with 60% in Vancouver and 84% in Toronto
either being expelled or withdrawing from the program over the 3 to 4 years of study. Generally,
higher program completion rates averaging approximately 50% are reported in the United States
(Belenko, 2001, 2002; Sanford & Arrigo, 2005). In this respect, cross-country differences in
sentencing practices for drug-related crimes and, therefore, in the offender populations served
by drug treatment courts, are relevant considerations for developing suitable policy options in
Canada (La Prairie, Gliksman, Erickson, Wall, & Newton-Taylor, 2002).
2.1.2 Non-legal formal pressures
Extant research has focused largely on pressures and mandates administered by legal authorities,
downplaying those mitigated by other formal and informal agents (Wild, 2006). A variety of
non-legal governmental and other institutions play an important role in encouraging or
mandating treatment entry in the current system. Non-legal formal pressures include mandatory
treatment as a condition of continued employment, school participation, child custody, or receipt
of social assistance, as well as persuasive techniques used by these individuals and institutions.
System-wide trends since the 1980s have contributed to growth in the application of these
formal referral mechanisms. Here, the roles of workplace and social assistance programs are
discussed in turn.
Recognition of the link between substance abuse and impaired job performance and risks to
workplace safety, as well as the high rate of employment among those with substance problems,
has legitimized and supported the growth of workplace alcohol and drug treatment programs,
which typically operate as part of broader behavioural health Employee Assistance Programs
(EAPs) (Levy Merrick, Volpe-Vartanian, Horgan, & McCann, 2007; White, McDuff, Schwartz,
Tiegel, & Judge, 1996). Widespread growth of the private treatment sector in the United States
through the 1980s is credited in part to workplace alcohol treatment programs (Schmidt &
Weisner, 1993). In Ontario, the proportion of workplaces with EAPs also increased significantly
over the 4-year period spanning 1989 to 1994 (Macdonald & Wells, 1994). The theory
underlying workplace referrals is that the threat of job loss can be used as leverage to promote
readiness for treatment and, ultimately, behaviour change (Trice & Sonnenstuhl, 1988).
11
Mixed findings are reported on the effectiveness of employer pressures and mandates to enter
treatment. Employer pressure has predicted treatment entry following an intake interview
(Weisner, Mertens, Tam, & Moore, 2001), as well as longer retention (Mertens & Weisner,
2000) and greater likelihood of program completion (Lawental, McLellan, Grissom, Brill, &
O'Brien, 1996). However, studies comparing clients required to complete a treatment program
for continued employment to others in treatment have found no group differences in retention
(Slaymaker & Owen, 2006) or substance-related outcomes at post-treatment follow-up
(Freedberg & Johnston, 1980; Lawental et al., 1996). A randomized trial comparing mandated
inpatient treatment, Alcoholics Anonymous meetings, and a choice of programs for employees
with alcohol-related problems found no difference between the groups in terms of work
performance, but some evidence of better drinking-related outcomes among those in the
inpatient treatment group over the 2-year follow-up period (Walsh et al., 1991). Although not
limited to alcohol and drug treatment, mandatory EAP participation was also associated with
significantly reduced absenteeism and costs related to health benefits (Keaton & Yamatani,
1993).
The methods used to identify substance-related disorders and evaluate candidacy for treatment
will have an impact on the findings of outcome evaluations. EAP literature highlights the
primacy of evaluations of job performance as the key factor to be used by supervisors in
determining need and formulating EAP referrals (Roman, 1988; Trice & Sonnenstuhl, 1988;
White et al., 1996)). Others have noted that random drug testing programs cannot distinguish
between casual or recreational drug use and substance use disorders, making them poor
indicators of treatment need when used in isolation (MacDonald et al., 2001; Pollack, Danziger,
Jayakody, & Seefeldt, 2002).
Many of the methodological limitations of work evaluating legally mandated or pressured
treatment similarly apply to the workplace literature. In addition to the lack of randomized study
designs and use of non-equivalent comparison groups, there is a lack of standard terminology
used to describe treatment that is pressured or monitored by employers. For instance, a
workplace referral may indicate anything from informal suggestions by coworkers or
supervisors that treatment be sought voluntarily, to formal conditions of treatment entry carrying
the threat of job loss (Weisner, 1990). Once in treatment, urinalyses and ongoing assessments of
12
work performance and treatment compliance may or may not be used as further leverage to
promote behaviour change.
In addition, the specific elements of treatment that are essential for success have yet to be
determined (Polcin, 2003). There is some empirical support for a balanced use of confrontation
and constructive advice by managers in discussions with employees experiencing alcohol-
related impaired work performance (Trice & Beyer, 1984). This balanced approach was found
to be most effective in facilitating treatment acceptance, which in turn led to improved work
performance; however, findings were based on managers’ retrospective evaluations of
discussions with employees and subsequent problem resolution. In interpreting findings of
reduced injuries following implementation of a program to detect and discipline substance-using
employees, another study speculated that threat of disciplinary action and dismissal, coupled
with random drug testing, was central to the success of work-based treatment programs (Miller,
Zaloshnja, & Spicer, 2007). However, the utility of taking disciplinary action, over and above
threats, was questionable. In the study mentioned above, written warnings and suspensions from
work were counterproductive and linked to worsened work performance, although the analysis
did not take into account problem severity – a likely confounder of this association (Trice &
Beyer, 1984).
Alongside the growth in workplace programming, the role of substance abuse as a barrier to
employment and economic self-sufficiency also figured largely into debates over welfare reform
in North America in the 1990s. Reform initiatives resulted in a model of social assistance in
which the receipt of income supports is contingent on participation in work and work-related
activities (Jayakody, Danziger, & Pollack, 2000; Morgenstern & Blanchard, 2006; Snyder,
2006). In jurisdictions with work-based social assistance programs, including Ontario, addiction
treatment may qualify as an employment readiness activity, with income benefits contingent
upon participation and success in treatment. Policy directives for Ontario’s work-based social
assistance program, Ontario Works (OW), state that addiction treatment can qualify as a work-
related activity if the recipient’s substance use and related problems are deemed sufficient to
interfere with his or her ability to participate in other program components and/or retain
employment (Ministry of Community and Social Services, 2001). However, the screening and
assessment process used in practice to identify and divert into treatment those who meet these
13
vague criteria is not well-documented, and likely varies in accordance with local resources and
OW program structure.
There is limited work evaluating outcomes among those who are mandated or required to
participate in addiction treatment as a condition of receiving social assistance. Much of the work
that has been done was conducted in the United States and focused on the subpopulation of
women and single mothers receiving public aid. These studies generally report positive impacts
of treatment on substance use (McLellan, Gutman et al., 2003; Morgenstern et al., 2006;
Morgenstern et al., 2009) and employment outcomes, including job rates and earned wages
(McLellan, Gutman et al., 2003; Metsch, Pereyra, Miles, & McCoy, 2003; Morgenstern et al.,
2009; Wickizer, Campbell, Krupski, & Stark, 2000). However, outcomes have not been gauged
against equivalent comparison groups of non-mandated or untreated individuals. One study
comparing those referred to treatment through the welfare system to self-referred clients found
no difference in rates of treatment completion (Brizer, Maslansky, & Galanter, 1990). In
addition to being a non-equivalent comparison group, however, the entire concept of self-
referral in studies of coercion is problematic, as will be considered in a later section.
While OW policy specifically states that benefits cannot be terminated if the participant does not
comply or succeed with treatment, it is also clear that benefits can be terminated where
conditions of the assistance contract signed by the recipient and caseworker are not met
(Ministry of Community and Social Services, 2001). The impact of continued drug use and
relapse on benefits termination in Ontario is not clear from the available literature. In some parts
of the province, the linkage between social assistance and addiction treatment has been more
formalized through a pilot project to enhance collaboration and cost-sharing of treatment for
OW participants. This initiative has received anecdotal praise from caseworkers and other
stakeholder groups (Matthews, 2004) and has been expanded since its introduction as a pilot
project in 2001. However, the degree to which the program has met its objectives of facilitating
access to treatment and promoting economic self-sufficiency is unknown in the absence of
formal evaluation.
With respect to the underlying rationale of this and other work-based social assistance programs,
it is notable that the relationship between substance abuse and social assistance is complex, and
more work is needed to evaluate the mechanisms through which substance abuse is presumed to
14
impact on employment readiness and economic self-sufficiency (Schmidt, Dohan, Wiley, &
Zabkiewicz, 2002). For instance, the rate of substance use disorders among welfare recipients is
not unequivocally higher than in the general population (Grant & Dawson, 1996).a In addition,
mixed results are reported on the association of substance use disorders with welfare
dependency and transitions from welfare to work (Chandler, Meisel, Jordan, Menees Rienzi, &
Naylor Goodwin, 2004; Schmidt et al., 2002; Schmidt et al., 1998; Schmidt, Zabkiewicz,
Jacobs, & Wiley, 2007); although a recent trial of mandated treatment among women receiving
social assistance found days of abstinence to be associated with greater subsequent employment
(Morgenstern et al., 2009). With social assistance programs increasingly focused on workfare
principles, it is certainly economically and socially advisable for governments to make addiction
treatment more accessible to those on social assistance, in conjunction with other employment
supports. However, the mechanisms through which this can be accomplished in an effective
manner have not been adequately studied, and the impact of current practices on long-term
economic self-sufficiency and recovery from substance use problems is unknown.
2.1.3 Informal social pressures
There is, overall, a dearth of research on the role that informal social networks play in
influencing entry to addiction treatment, and much of what is available is focused on alcohol
rather than psychoactive substances more generally. Three lines of evidence can be brought to
bear on the subject of informal social pressures to enter treatment: studies of drinkers and their
relations in the general population; studies of the reasons for help-seeking among clients in
treatment programs; and evaluations of manualized social network interventions used to
encourage or induce a loved one to commit to a course of treatment.
a Grant and Dawson (1996) offer a comprehensive analysis of rates of 12-month drug and alcohol use disorders across several types of social assistance programs in the US national general population in 1992. Rates varied from 1.3% to 8.2% depending on the disorder and program, and were not significantly elevated relative to the general population. Conversely, another study conducted in one California county found a significantly elevated rate of 12-month substance dependence among welfare recipients (ranging from 11-43%) relative to the county’s general population (2.8%) (Schmidt, Weisner, & Wiley, 1998). A study considering the role of barriers to self-sufficiency among single mothers receiving welfare found substance dependence to be among the least prevalent of the considered barriers, relatively less frequent than factors such as mental and physical illnesses, low education and poor job skills (Danziger, Kalil, & Anderson, 2000).
15
Studies of the general population have demonstrated the important role of social networks in
pressuring problem drinkers to change their behaviour and/or enter treatment (Hasin, 1994;
Room, Bondy, & Ferris, 1996; Room, Greenfield, & Weisner, 1991; Room, Matzger, &
Weisner, 2004). In 1993 in Ontario, 65% of adults in the general population indicated that they
had, at some point in their lives, said something to friends or family about their drinking and
34% had done so in the past year (Room et al., 1996). While these events typically represented
suggestions to cut down, 28% of those who said something to friends or family in the past year
(or 15% of the total adult population) reported suggesting or helping the target individual to
obtain treatment. Past-year suggestions to enter treatment were often precipitated by alcohol-
related job, legal or family problems. Based on his work, Room (1989) has suggested that
pressures from social networks to change behaviours and/or enter treatment are sufficiently
common that few people likely enter treatment without being spoken to or pressured by friends
or family.
In clinical samples, family and friends are among the most common sources of pressures to
enter alcohol and drug treatment (Cunningham, Sobell, Sobell, & Gaskin, 1994; Marlowe, Glass
et al., 2001; Marlowe et al., 1996; Marlowe, Merikle, Kirby, Festinger, & McLellan, 2001;
Polcin & Weisner, 1999). Encouragement from social networks has been linked with increased
likelihood of treatment entry among problem drinkers (George & Tucker, 1996) and illicit drug
users (Gyarmathy & Latkin, 2008). Others have found that alcohol-related social and functional
problems, including pressures to seek help, are rated as the most influential factors in treatment
seeking (Tucker, Vuchinich, & Rippens, 2004). Personal definitions of drinking problems and
need for treatment are likely to be shaped at least partly by the responses of the social network
to drinking behaviours (Tucker et al., 2004). Accordingly, having a greater number of social
network members in treatment has been shown to increase the likelihood of treatment entry
among urban illicit drug users (Davey, Latkin, Hua, Tobin, & Strathdee, 2007).
These findings are not surprising when examined in light of theoretical work on social control.
Social relations and family can serve as powerful sources of informal social control through
such devices as fostering a sense of mutual obligation and responsibility, or more overtly by
applying threats of ostracism and estrangement (Cuzzort & King, 1995; Tucker & Anders,
2001). They may be expected to be successful in discouraging deviant or undesired behaviour to
the extent that the target individual cares about the perceptions of the group or network
16
members and his or her own continued membership (Goode, 2008). Despite being largely
neglected by sociological research, it is contended that the majority of social controls placed on
behaviour are informally enacted by private individuals rather than by larger and more
formalized institutions or organizations (Goode, 2008). Orford (1992) takes a more integrative
stance, noting similarities in the dilemmas faced by all social control agents in responding to or
coping with excessive drinking, regardless of their formal or informal stature, calling them
“universal features of social systems” (p. 1524). He notes that most contemporary approaches to
studying substance-related problems focus on the behaviour and coping responses of the
affected individual rather than the social networks surrounding him or her, although significant
others may play an important and understudied role in both prevention and recovery.
Recognizing the power inherent in social networks, a number of intervention methods involving
family and friends aimed at encouraging or inducing treatment entry have been developed.
Perhaps the most well-known of these is the Johnson intervention (Johnson, 1986). Briefly,
members of the social network, potentially including employers and co-workers in addition to
friends and family, gather without the affected individual present to prepare and practice
statements outlining their issues with the individual’s substance use and how it has affected
them. They then confront the individual as a group and in the presence of a counsellor, with
specific consequences should the individual refuse treatment.
Mixed success rates are reported in the literature on Johnson interventions. A study based on a
retrospective chart review reported that, relative to legal or employer mandates and self-
referrals, the Johnson intervention was associated with greatest odds of treatment entry and was
comparable to legal mandates in its positive association with treatment completion (Loneck,
Garrett, & Banks, 1996a). However, it was also associated with the highest likelihood of alcohol
or drug use during treatment, which was itself inversely associated with treatment completion
(Loneck, Garrett, & Banks, 1996b). Logan (1983) reported high success rates in terms of
treatment entry (90%) and completion (83% of those who entered treatment), as well as
collateral-reported abstinence from alcohol following treatment (70% of the 62% who were
found for follow-up). High success rates have likewise been reported when the intervention
takes place; however, these are tempered by refusal rates by networks of typically 70% to
actually carry out the intervention (Liepman, Nirenburg, & Begin, 1989; Miller, Meyers, &
17
Tonigan, 1999). When all social networks are included in the calculation of “success” rates for
treatment entry, these decline to 24%-30%.
Partly in recognition of these figures, others have adapted the Johnson approach to serve as the
final stage or “last resort” in a graduated sequence of events centred on guiding treatment entry
through social network intervention (Barber & Crisp, 1995; Landau et al., 2000). The ARISE
method, an acronym for A Relational Sequence for Engagement, is a manualized intervention
approach consisting of three progressively confrontational and demanding stages involving
telephone and face-to-face contact with trained counsellors (Garrett, Landau-Stanton, Stanton,
Stellato-Kabat, & Stellato-Kabat, 1997; Garrett et al., 1999; Landau et al., 2000). It is only when
the less confrontational approaches have failed that the network is called upon for a Johnson-
style intervention. Also different is that the affected individual is invited to participate in
meetings at every stage. An evaluation of ARISE demonstrated an 83% success rate in terms of
promoting treatment entry among those with alcohol and drug problems. In addition, the vast
majority (98%) entered treatment at the earlier stages, such that only 2 of the 91 successful
interventions necessitated a Johnson-style approach. The refusal rate by networks was low (3%),
suggesting that the staged and less secretive approach was relatively more acceptable.
A different approach is used by the Community Reinforcement and Family Training (CRAFT)
method for engaging individuals in treatment through social networks (Meyers, Miller, Hill, &
Tonigan, 1998). This approach focuses on teaching network members skills for dealing with
their loved one’s substance use, including contingency management strategies to reinforce
desired behaviours, awareness and communication training, as well as strategies for motivating
and supporting treatment entry and participation. In an initial single-group evaluation, CRAFT
was found to have a 74% success rate in promoting treatment entry (Meyers et al., 1998). The
target individuals who entered treatment reported significant decreases in substance use and
improvements in the quality of their relationships with the concerned network member, while
the latter reported significant improvements in their own well-being. Two subsequent
randomized trials found CRAFT to be superior to other intervention approaches, including the
Johnson intervention and Al-Anon/Nar-Anon facilitation, in promoting treatment entry (Meyers,
Miller, Smith, & Tonigan, 2002; Miller et al., 1999).
18
There is evidence that the impact of attempts by informal social networks to control or influence
a loved one’s behaviours may be determined in part by the nature and type of the strategy
employed, as well as the target individual’s perceptions of these attempts. In a study of the
health-related social control strategies used by married couples, Tucker and Anders (2001)
found that individuals who were subject to positive social control strategies, such as rewards and
cooperation, were more likely to report both positive affect related to the control attempt and
positive behavioural responses (i.e., engaging the desired behaviour). Conversely, negative
strategies, such as nagging and threats, were associated with negative affect and behavioural
responses (i.e., ignoring the attempts at control or engaging in the opposite to the desired
behaviour). Further, the association between social control and health behaviours was partially
mediated by the affective response to the attempt at control. This research both supports the
potential effectiveness of informal social control strategies for behaviour change and allows for
the possibility that they may backfire and result in undesired consequences – particularly when
they are perceived as too confrontational or as rooted in disingenuous motives.
While supporting and involving significant others in the treatment entry and recovery process is
both intuitively beneficial and largely supported by research, many questions remain about the
specific role of informal social control in this process. The predictors, correlates, and outcomes
of informal conversations and suggestions relative to more formalized methods offered by
manual-driven interventions have not been considered. Research has tended to avoid assessing
perceptions of the target individual, as well as details on the positive/negative quality of control
strategies. With only a couple of notable exceptions (Meyers et al., 1998; Room et al., 1996),
studies have also failed to consider the relationship between the network and target individuals
as a correlate or outcome of the use of such strategies.
With respect to manualized interventions more specifically, the methodological rigour of
evaluations has not been uniformly high. Most either failed to use comparison groups (Landau
et al., 2004), or else used non-equivalent comparison groups such as networks that refused to
stage the intervention (Liepman et al., 1989) or treatment clients who were subject to other
referral strategies (Loneck et al., 1996a). Rigorous evaluation of outcomes past the point of
treatment entry is also uncommon. Some have employed telephone check-ins with collaterals to
obtain information on the identified individual’s status, without corroboration from other
sources or the use of psychometrically sound instruments (Liepman et al., 1989; Logan, 1983).
19
Finally, the impact of the above-reviewed interventions on longer-term outcomes and
functioning remains largely unknown.
2.2 Perceived coercion Despite these methodological and conceptual issues, positive (or null) empirical findings on
legal, other formal, and informal social pressures have been used to support the claim that
coercion works and is a viable strategy for initiating treatment among those who would not
otherwise voluntarily seek it (Leukefeld & Tims, 1988; Nace et al., 2007; Sullivan et al., 2008).
However, there are reasons to believe that alternative definitions of coercion, incorporating
client perceptions and a broader array of informal and formal sources, are critical to a full
understanding of the effectiveness of coerced treatment.
In his work, Wild (2006) differentiates between coerced addiction treatment and treatment under
social controls. The term coercion is reserved to describe situations in which clients perceive a
lack of control over the treatment entry process, while social controls refer to the range of
mandates and other pressures, described in previous sections, that are objectively applied to
ensure or encourage treatment entry. In this framework, coerced treatment refers to that which is
perceived by the client as a coercive imposition, regardless of the agent applying the pressure or
control. It also encompasses a wide range of actors from the legal system, social services, the
workplace, and social networks. This definition recognizes that coercion involves both an
objective act and a subjective perception, and is highly contextually dependent (Hoge et al.,
1993; Lidz et al., 1997). Others have taken similar stances toward the assessment and evaluation
of coercion in addiction treatment (Klag, Creed, & O'Callaghan, 2006; Longshore, Prendergast,
& Farabee, 2004; Polcin, 2001).
To be sure, the pressures discussed above have the potential to be experienced as coercive;
however, there is a wealth of empirical support for the lack of a direct correspondence between
objective strategies and client perceptions of coercion. In the DATOS evaluation, a reported
43% of clients who were legally mandated to enter treatment stated that they would have been
willing to enter treatment anyway (Farabee, Prendergast, & Anglin, 1998). This is consistent
with other findings that court-ordered treatment is not always experienced by clients as a
coercive imposition (Stevens et al., 2006). Despite reporting higher perceived pressure to enter
treatment than those who were not mandated, some court-ordered clients reported that they had
20
been waiting to get into treatment or had intended to enter treatment in the future anyway. The
converse was also true: some clients who were not mandated reported entering treatment
because they believed it would assist their legal situations. Also of interest is the finding of
individual-level heterogeneity in motivation for treatment, such as would be expected in any
treatment population, among offenders mandated to treatment programs (Hiller, Knight,
Leukefeld, & Simpson, 2002; Prendergast, Greenwell, Farabee, & Hser, 2009). In other words,
it does not appear to be the case that mandated clients uniformly lack motivation to change their
substance use behaviours and participate in treatment.
In a study of non-mandated clients, one-third of the reasons for entering treatment still involved
escaping or avoiding a negative external contingency (Marlowe et al., 1996), demonstrating that
the absence of a mandate does not equate with voluntarism. Anglin et al. (1989) found that fully
80% of their sample who were on probation or parole at the time of admission to methadone
treatment did not report legal troubles as a reason for entering treatment, while 27% of those
who were not on probation or parole did report legal troubles as a major reason for entering
treatment. Classifying clients according to self-reported legal problems as a reason for
admission versus chart-documented legal involvement, Vickers-Lahti et al. (1995) found the
latter to be twice as common. That is, approximately half of those who had legal system
involvement at admission did not perceive it to be an important factor in their decision to enter
treatment. Alternatively, Marlowe, Glass et al. (2001) found that many clients who reported
legal pressure to enter treatment were not actually currently involved with the legal system, but
reported fears of potential legal problems. When client-reported pressures are ranked according
to their importance in the decision to enter treatment or to change behaviours, there is some
evidence that legal pressures do not play as great a role as other reasons related to health and
social functioning (Downey, Rosengren, & Donovan, 2001; Marlowe, Glass et al., 2001;
Marlowe et al., 1996; Marlowe, Merikle et al., 2001; McBride et al., 1994).
The presumed equivalency of social pressures and coercion is conceptually problematic for a
number of reasons. First, it can not be assumed that external pressures are always tied to the
decision to enter addiction treatment (Weisner & Matzger, 2002; Weisner et al., 2001). People
seeking treatment are often experiencing multiple internal and external pressures (Cunningham
et al., 1994; Marlowe et al., 1996; Marlowe, Merikle et al., 2001; Polcin & Weisner, 1999), and
the importance of any given one is likely dependent on a host of individual and contextual
21
factors. Further, the client’s attitude toward the pressure is likely to be more influential than its
objective presence (Marlowe, Glass et al., 2001). Second, it is not necessarily coercive for a
social service to identify and refer an individual to addiction treatment. Many individuals with
substance problems initially seek help from services outside the addiction sector, depending on
how they and those around them define their problems (Weisner, 1987).
The self-referral construct is another common and highly problematic one for research on help-
seeking and pathways to treatment. Notably, self-referred clients may nonetheless be avoiding
legal or employer sanctions and still perceive a great deal of coercion to enter treatment. In
reviews of work-based programming for alcohol use problems, a trend is noted toward a greater
focus on self-referred clients in later programs (Macdonald & Wells, 1994; Roman, 1988).
Roman (1988) speculates on the underlying causes of this shifting emphasis, including the
growing publicity and acceptance of EAPs among employees, a preference among counsellors
for non-confrontational approaches, and the evolution of such programs to treat non-substance
related psychiatric and personal problems. However, employer pressure can still be applied in
the absence of a mandatory referral (Roman, 1988; Trice & Sonnenstuhl, 1988). It has been
suggested that the high prevalence of self-referred clients in treatment reflects only the
importance placed upon self-regulation, independence, and responsibility in North American
culture (Trice & Sonnenstuhl, 1988). These authors argue that because the term does not reflect
a common set of circumstances and processes leading to treatment that can be accurately
described by the term self-referral, it is not a valid category to use in outcome evaluations.
Despite evidence documenting the potential for misclassification that results from equating
objectively coercive acts with experiences of coercion, studies evaluating client perceptions of
coercion at admission are rare. The most commonly used measure of perceived coercion in
addiction treatment settings has been the MacArthur Perceived Coercion Scale (MPCS),
developed by researchers with the MacArthur Research Network to assess the perceptions of
psychiatric inpatients of hospital admission processes (Gardner et al., 1993). This work is briefly
summarized, along with applications to the addiction field.
Initial exploratory analysis by the MacArthur group identified exclusion from the admission
process, including the lack of opportunity to express opinions and to “have a say” in events, as
the most coercive aspect of entering the hospital (Bennett et al., 1993). Other factors included
22
not being treated with equality and respect and being deceived. The legitimacy of the actions of
others during admission was judged by patients according to role expectations and perceptions
that the underlying motives concerned the patient’s well-being. Drawing upon interviews with
consumers, family members, and clinicians and reviews of the literature, five concepts were
identified as relevant to perceptions of coercion. These were operationalized as true or false
statements to form the five-item MPCS:
influence - “I had more influence than anyone else on whether I came to the hospital”
control - “I had a lot of control over whether I went to the hospital”
choice - “I chose to come into the hospital”
freedom - “I felt free to do what I wanted about coming into the hospital”
idea - “It was my idea to come into the hospital”
The MPCS offers a brief, self-reported instrument appropriate for assessing client perceptions of
the admission process without restriction to particular sources (i.e., legal, other formal, or
informal). Studies using the MPCS in psychiatric populations have demonstrated high internal
consistency, inter-rater and test-retest reliability, and construct validity through its associations
with theoretically related factors such as involuntary status and procedural justice (Cascardi &
Poythress, 1997; Cascardi, Poythress, & Ritterband, 1997; Gardner et al., 1993; Hiday, Swartz,
Swanson, & Wagner, 1997; Hoge et al., 1997; Hoge et al., 1998; Ivar Iversen, Hoyer, Sexton, &
Gronli, 2002; Lidz et al., 1995; Nicholson, Ekenstam, & Norwood, 1996). The MPCS has also
been adapted and evaluated for use with community mental health care consumers (Farabee,
Shen, & Sanchez, 2002; Steadman et al., 2001; Swartz, Wagner, Swanson, Hiday, & Burns,
2002), mental health court participants (Poythress, Petrila, McGaha, & Boothroyd, 2002), non-
psychiatric medical and surgical inpatients (Taborda, Bapista, Gomes, Nogueira, & Chaves,
2004), sex offenders in prison-based treatment (Rigg, 2002), and clients in addiction treatment
(Marlowe, Glass et al., 2001; Prendergast et al., 2009; Wild, Cunningham, & Ryan, 2006; Wild,
Newton-Taylor, & Alletto, 1998).
Few studies have provided information on the determinants and correlates of perceived coercion
at admission to addiction treatment. Marlowe, Glass et al. (2001) examined the relationship
between MPCS scores and reasons for seeking treatment among methadone clients. Subjects
were clustered according to the source and nature of pressures that were most influential in their
decision to enter treatment. There was a trend towards higher perceived coercion and lower
23
perceived input into the decision to enter treatment among those who reported legal pressures to
enter treatment. However, the small sample size of this study raises concerns about the stability
and generalizability of these results.
The most comprehensive analysis of perceived coercion during treatment admission is offered
by Wild et al. (1998). Among clients entering outpatient treatment for alcohol and drug
problems, MPCS scores were associated with treatment mandates from employers and legal
authorities, as well as ratings of pressure from family and friends to enter treatment. However,
notably 35% of the court-mandated, and 61% of those mandated by their employers or another
source, reported no coercion to enter treatment, while 37% of those who identified as being self-
referred did report coercion (i.e., endorsed one or more of the items). Perceived coercion was
also inversely associated with perceived addiction severity: the less clients believed they were
addicted, the more coercion they reported. Sociodemographic variables, perceived need for
abstinence, pressures from friends and family to quit using substances, previous criminal
convictions, and objective addiction severity, were not significant predictors of perceived
coercion in a multivariable regression analysis. In another analysis of these data, MPCS scores
were found to be positively associated with external motivation and inversely associated with
internal motivation for treatment (Wild et al., 2006).
A third, more recent study examined the link between perceived coercion and motivation in
greater detail in a sample of offenders mandated to substance use treatment (Prendergast et al.,
2009). MPCS scores demonstrated small but statistically significant inverse correlations with
motivation, assessed in terms of problem recognition, ambivalence, and steps already taken
toward behaviour change. The authors take this as evidence that perceived coercion and
motivation represent related but distinct constructs. Another notable finding from this study was
the relatively low average level of perceived coercion among these offenders mandated to
treatment (i.e., the mean MPCS score was 1.6 out of a possible 5.0).
Such findings are broadly supported by the MacArthur and later studies, which demonstrated
that involuntary inpatient and outpatient commitment is generally perceived as more coercive
than voluntary admission, although the correspondence is not direct and many of those who are
officially voluntarily-admitted perceive the admission process as coercive to some degree
(Bindman et al., 2005; Hiday et al., 1997; Hoge et al., 1997; Ivar Iversen et al., 2002; Kjellin,
24
Hoyer, Engberg, Kaltiala-Heino, & Sigurjonsdottir, 2006; McKenna, Simpson, & Coverdale,
2006; Monahan et al., 1995). In addition, studies have not found sociodemographic
characteristics to play a major role in perceptions of coercion (Cascardi & Poythress, 1997;
Hiday et al., 1997; Hoge et al., 1998; Lidz et al., 1995).
With respect to more specific elements of the admission process, perceptions of being treated
unfairly, of not having a voice, and the use of deception, threats and force are consistently the
strongest predictors of perceived coercion among psychiatric inpatients (Bindman et al., 2005;
Cascardi & Poythress, 1997; Hiday et al., 1997; Hoge et al., 1998; Lidz et al., 1995).
Perceptions of being treated with fairness and respect are strong determinants of perceived
legitimacy and satisfaction in research on general dispute resolution, with demonstrated benefits
for emotional well-being among those mandated to mental health treatment (La Fond & Srebnik,
2002; Monahan et al., 1995; Poythress et al., 2002). There are no evaluations of specific aspects
of the admission process in relation to perceived coercion in the context of addiction treatment;
although studies have documented perceptions of violation and of not being able to influence the
course of care among clients in legally mandated addiction treatment (Larsson-Kronberg,
Ojehagen, & Berglund, 2005; Sallmen, Berglund, & Bokander, 1998). Other work has also
emphasized the importance of fostering personal responsibility and control among clients in
addiction treatment, including the provision of choice regarding goals and modes of treatment
(Miller, 1987; Miller & Rollnick, 2002).
Only one study has been published to date on the association between perceived coercion and
outcomes following addiction treatment. Among offenders mandated to treatment, perceived
coercion at admission was not associated with either treatment completion or re-arrest in the
approximately 8 months following treatment (Prendergast et al., 2009). More work is needed on
the role of perceived coercion in the treatment process and post-treatment outcomes. This single
study involved substance-using offenders mandated to treatment in one American state, and did
not assess outcomes related to substance use, employment or psychological well-being. The
outcomes selected (i.e., treatment completion and re-arrest) were obtained from administrative
data sources, such that the study did not involve a personal follow-up of the clients themselves.
As will be discussed later, to the degree that treatment completion was mandated by the legal
system, the meaning of this measure as an indicator of outcome is perhaps limited.
25
With respect to treatment process and outcomes in the mental health care system, studies have
reported conflicting findings on the role of perceived coercion. Scores on the MPCS have shown
stability between admission and discharge, even when the perceived need for hospitalization
changes (Bindman et al., 2005; Gardner et al., 1999). This suggests that perceptions of coercion
may have an enduring effect and, as such, may influence outcomes in the long-term. One study
found perceived coercion at admission to be predictive of negative perceptions of staff and
treatment at both intake and 6 months post-discharge, as well as being associated with a lower
likelihood of using medication and services as directed and greater psychiatric symptom severity
at follow-up (Kaltiala-Heino, Laippala, & Salokangas, 1997). It is likely that poor treatment
adherence mediated the relationship between perceived coercion and psychiatric status;
however, this was not directly tested. In a sample of consumers with severe mental illness and
co-occurring substance use disorders, perceived coercion was associated with a more negative
consumer-provider relationship and poorer evaluations of community service contacts
(Stanhope, Marcus, & Solomon, 2009). Other studies of individuals receiving psychiatric care
have failed to find significant associations between perceived coercion at admission and
perceived benefit of hospitalization (Nicholson et al., 1996) or adherence to community
treatment following discharge (Bindman et al., 2005; Rain, Steadman, & Robbins, 2003; Rain,
Williams et al., 2003). There is evidence of a negative association between perceived coercion
at entry to outpatient mental health treatment and quality of life assessed over the subsequent
year (Link, Castille, & Stuber, 2008; Swanson, Swartz, Elbogen, Wagner, & Burns, 2003). At
the same time, however, these analyses reveal positive associations between outpatient
commitment and follow-up measures of psychiatric symptoms and functioning, highlighting the
complexity of the issue of coercion in psychiatric care (Link et al., 2008; Swanson et al., 2003).
Further studies are needed to replicate and extend these findings and assess their stability and
generalizability. Additional work is also needed in defining and evaluating the concept of
perceived coercion in addiction treatment samples. The MacArthur construct of perceived
coercion has been criticized for its global character, in that it does not allow for distinctions to
be drawn between the sources or agents of coercion (Klag et al., 2006). Klag et al. (2006) have
taken steps in this direction, through the development of the Perceived Coercion Questionnaire
(PCQ). In addition to assessing coercion from specific sources, the PCQ offers the advantage of
having been developed specifically for use with addictions treatment populations. Initial
psychometric evaluation supports the utility of the PCQ; however, it appears to define perceived
26
coercion in terms of perceptions of being pressured into treatment from a variety of internal and
external sources. As such, items focus predominantly on the occurrence of objectively coercive
tactics, as has been the norm in previous work on coercion in the addiction field, rather than
with client perceptions or internal experiences of such events. Further, pressures from the self
comprise one domain covered by the scale; however, the precise nature of the relationship
between these pressures and the experience of autonomy during the admission process is not
immediately clear. Considering the potentially large role that coercion plays in current addiction
treatment systems in North America and elsewhere, improved understanding of this concept in
both formally mandated and non-mandated populations is critical.
2.3 Treatment process Considerable evidence supports the benefits of alcohol and drug treatment for substance use and
other outcomes (Gossop, Marsden, Stewart, & Kidd, 2003; Hubbard et al., 2003; McLellan et
al., 2000; Miller & Wilbourne, 2002; Prendergast, Podus, Chang, & Urada, 2002). Less is
known about the processes of treatment, that is, the specific elements or components of
treatment programs and the early stage recovery processes that support or impede positive
outcomes in the long-term. In-treatment measures of client engagement and progress have high
value in terms of further treatment planning, as well as providing important information for
evaluations of treatment effectiveness and organizational functioning (Simpson & Joe, 2004).
Calls have been made for more research focused on in-treatment progress and other factors, to
allow for a better understanding of treatment-assisted recovery (Simpson, 2004; Wild, 2006).
Retention in drug treatment is a consistent predictor of positive outcomes, regardless of
modality (i.e., residential programs, outpatient counselling, and methadone maintenance
treatment) (Hser, Evans, Huang, & Anglin, 2004; Hubbard et al., 2003; Simpson, 2004; Zhang
et al., 2003). While brief interventions for alcohol abuse and high-risk drinking are well-
supported (Kaner et al., 2007; Miller & Wilbourne, 2002; Moyer, Finney, Swearingen, &
Vergun, 2002), an association between treatment duration or completion and outcome has
likewise been demonstrated in more intensive alcohol treatment settings (Bottlender & Soyka,
2005; Moos & Moos, 2003; Stark, 1992).
From a conceptual perspective, it is logical that remaining in treatment for at least some
minimum period of time is essential to accruing the benefits of treatment. Beyond assessing the
27
impact of time in treatment, Joe, Simpson and colleagues have defined minimum retention
thresholds for optimal clinical outcomes among clients in drug treatment. Specifically, they have
found that treatment benefits begin to accrue following 90 days in outpatient or residential care,
or one year in methadone maintenance treatment (Joe et al., 1998; Joe, Simpson, & Broome,
1999). Clients who remained in treatment for shorter periods of time were found to have a
significantly lower probability of improved outcomes at follow-up. Others, however, have noted
smoother linear relationships between retention and drug-related outcomes across modalities,
with improvement dropping off at durations of over a year in non-methadone treatment settings
(Zhang et al., 2003).
Regardless of the specific nature of the association between retention and recovery, high rates of
drop-out in the initial weeks of treatment present a challenge to addiction treatment providers. In
an often cited review of attrition from addiction treatment, Stark (1992) reported that a
minimum of half of clients leave treatment within the first month. Rates of attrition vary widely
across studies and modalities; however, studies continue to document the typically short
duration of treatment, with the majority of outpatient clients in particular receiving fewer than
three or four appointments (Claus & Kindleberger, 2002; Pulford & Wheeler, 2007). In the
multisite DATOS project, 19% of the total sample was retained in treatment for less than one
week (Hubbard et al., 1997) and, overall, clients typically remained in treatment approximately
one-third to one-half of the duration recommended by program staff (Simpson, Joe, Broome et
al., 1997). In a study exploring homeless men’s reasons for dropping out of treatment, clients
tended to offer multiple explanations rather than a single cause, with reasons commonly
involving a lack of readiness for treatment, conflict or a lack of rapport with staff, and mismatch
between their needs and program characteristics (Stahler, Cohen, Shipley, & Bartelt, 1993). In
addition to the potentially negative prognosis for clients, high early attrition rates can carry high
costs to programs, in terms of resources dedicated to assessment and treatment planning in the
early stages of treatment (Simpson, Joe, Broome et al., 1997; Walker, 2009).
A wealth of client characteristics are associated with longer retention, including older age, being
married, having greater social support, and having fewer or less severe substance-related
problems (Curran, Stecker, Han, & Booth, 2009; Joe et al., 1998; Klein, di Menza, Arfken, &
Schuster, 2002; Knight et al., 2000; Orlando, Chan, & Morral, 2003; Simpson & Joe, 1993;
Stark, 1992). Greater pre-treatment motivation, assessed in terms of problem recognition, desire
28
for help, and treatment readiness, has also been consistently linked with longer retention (Evans,
Li, & Hser, 2009; Joe et al., 1998; Knight et al., 2000; Orlando et al., 2003; Simpson & Joe,
1993; Simpson, Joe, & Rowan-Szal, 1997). Equivocal findings are reported with respect to the
role of prior treatment experience, including null and negative associations (Cacciola, Dugosh,
& Camilleri, 2009; Stark, 1992), as well as positive associations (Joe et al., 1998; Simpson &
Joe, 1993) reported with retention in an index episode. As noted earlier, current legal system
involvement and pressure have also shown equivocal relationships with retention, depending on
the definition used and the service modality (Beynon et al., 2006; Claus & Kindleberger, 2002;
Joe et al., 1998; Knight et al., 2000; Maglione et al., 2000; Vickers-Lahti et al., 1995). Extent
and history of criminal activity and legal system involvement appear to be more consistently
associated with poorer retention (Evans et al., 2009; Joe et al., 1998; Knight et al., 2000;
Simpson & Joe, 1993; Young, 2002; Young & Belenko, 2002). At the program-level,
availability of a greater range of services, including medical, mental health, employment and
financial services, has been linked with longer retention (Walker, 2008). Highlighting the
complexity of this indicator, adjusting for client-level differences, service intensity, and the
range of services available, significant program-level variability remains in retention rates
(Simpson, Joe, Broome et al., 1997).
Improvements in substance use and other behaviours measured at points in time closely
following treatment are a necessary first step in demonstrating treatment effectiveness (Hubbard
et al., 1997). However, the importance of study designs that allow for the linkage of treatment
processes with post-treatment outcomes is increasingly recognized. Progress in treatment is an
important determinant of substance-related outcomes in the longer term (McKay & Weiss,
2001). The incorporation of in-treatment performance indicators is also consistent with a
chronic disease perspective on addiction (McLellan, McKay, Forman, Cacciola, & Kemp,
2005). Approaches to evaluation that involve monitoring outcomes concurrently with treatment
and recovery (as opposed to conducting follow-up interviews months or years after admission)
are used to determine the effectiveness of interventions for other chronic health conditions, such
as diabetes or hypertension (McLellan et al., 2005). It has been argued that the impact of any
given treatment episode is best indicated by functioning at the conclusion of that episode, with
longer term outcomes assessed within the broader context of clients’ lives, including social,
psychological, and environmental circumstances as well as interactions with treatment and other
service systems (Hser, Longshore, & Anglin, 2007; McKay & Weiss, 2001). These perspectives
29
frame treatment as one part of a complex process of change, rather than as a single event that
directly and solely results in a measurable outcome in the longer term (Simpson, 2004).
There is no commonly accepted set of measures for client involvement or engagement in the
treatment process. Many studies are limited to retention- or attendance-based measures, defining
engagement in terms of combinations of intensity and duration of service (Claus &
Kindleberger, 2002; Fiorentine, Nakashima, & Anglin, 1999). Recent efforts by a
multidisciplinary panel of researchers, clinicians, and policy representatives in the United States,
referred to as the Washington Circle, convened to develop performance measures for substance
use treatment, have attempted to define such benchmarks in representing treatment process
(Garnick et al., 2002). Using insurance claims data from health maintenance organizations, this
group defined and pilot tested two performance indicators corresponding to the treatment
process. Treatment initiation refers to the proportion of individuals who receive an alcohol or
drug service within 14 days of receiving a diagnosis or having an initial visit to an alcohol or
drug service. Treatment engagement refers to the proportion that receives an additional two
sessions or services within 30 days of treatment initiation. These performance measures have
been adopted by a national health care quality assurance organization in the United States and
have been adapted for use in the public treatment sector (Garnick, Lee, Horgan, & Acevedo,
2008). However, limited evaluation of their relationship with outcomes has been published to
date. Engagement, thus defined, showed statistically significant, but clinically modest,
associations with improvements in substance use and legal problems 7 months following
admission (Harris, Humphreys, Bowe, Tiet, & Finney, 2008), and an inverse association with
likelihood of arrest or incarceration assessed one year post-treatment (Garnick et al., 2007).
Aggregated to the facility-level, engagement was not associated with substance-related
outcomes, failing to support its utility in distinguishing facility performance (Harris,
Humphreys, & Finney, 2007). Although preliminary, these mixed results may speak to the
difficulties in identifying benchmarks for retention that are clinically meaningful and common
across clients with different expectations and needs for services.
More generally, however, measures of treatment engagement defined solely on the basis of
retention or attendance can make only a limited contribution to understandings of treatment
processes. While physical presence in treatment may reflect one aspect of client engagement, it
does not necessarily guarantee meaningful participation or material learned (Schacht Reisinger,
30
Bush, Colom, Agar, & Battjes, 2003; Sung, Belenko, Feng, & Tabachnick, 2004). As is evident
from the above summary, retention is impacted by a number of individual, environmental, and
programmatic features, which may interact with each other in complex ways (Simpson, 2004;
Walker, 2009). Teasing apart these variable influences and their contributions to recovery
remains an important area for further study.
Simpson and colleagues at the Texas Christian University (TCU) have made important
contributions to current understandings of treatment processes. Their work draws from multisite
evaluation projects conducted in a variety of treatment settings since the early 1990s. Most
notably, these include the Drug Abuse Treatment for AIDS-Risk Reduction (DATAR) project
aimed at improving services for opiate users (Simpson, 1993; Simpson, Joe, Dansereau, &
Chatham, 1997) and the afore-mentioned DATOS project, for which the TCU served as the
coordinating centre for a substudy of treatment processes (Flynn, Craddock, Hubbard,
Anderson, & Etheridge, 1997). Based on findings from these projects, these researchers have
outlined a conceptual model of treatment-assisted recovery, incorporating client characteristics,
contextual factors, and interim and longer-term functioning and recovery indicators (Figure 1;
Simpson, 2004). Measures have been defined and tested for the various components of the
model (Joe, Broome, Rowan-Szal, & Simpson, 2002), contributing to its utility.
Figure 1. TCU treatment model
Reproduced from Simpson, 2004.
Early Engagement
Early Recovery
Stabilized Recovery Client
attributes Readiness Program
participationBehavioural change
Therapeutic relationship
Psychosocial change
Sufficient retention
Severity Drug use Program
attributes Criminal activitySocial relations
Post-treatment
Resources Staff Climate Mgmt info
31
In the TCU treatment model, the treatment process contains three sequential phases:
early engagement: program participation (i.e., session attendance and cognitive
engagement, such as commitment to the therapeutic process) and development of the
therapeutic relationship (i.e., counsellor rapport) in the initial weeks of treatment
early recovery: changes in behaviour (i.e., substance use) and psychosocial functioning
(i.e., depression, risk-taking, self-esteem) measured a few months into treatment
stabilized recovery: retention beyond minimum thresholds required for treatment
effectiveness and transition to post-treatment services and supports
These three treatment phases are impacted by client factors such as motivation and problem
severity, and program factors such as resources, policies and service compliment. Once in
treatment, clients differ in the extent to which they become engaged in a treatment-assisted
recovery process. The components of early engagement are sequential and interacting, such that
session attendance contributes positively to development of the counselling relationship, which
in turn has a reciprocal effect on continued attendance. An engaged client, by this definition, is
one who attends sessions, becomes cognitively involved and committed to the program, and
develops a positive relationship with counsellors over the first few weeks of treatment.
Following early engagement, clients progress to early recovery, during which psychosocial and
behavioural changes start to emerge. Finally, recovery becomes stabilized as clients are retained
beyond the minimum threshold for treatment effectiveness and begin to prepare for transitioning
out of treatment. These process indicators are linked with positive post-treatment outcomes,
such as continued involvement with support networks, and sustained behavioural and
psychosocial improvements.
The cognitive components of early engagement represent particularly significant contributions
of the TCU model, having received comparatively little attention in previous studies of
treatment processes (Joe, Simpson, & Broome, 1999). Among pre-treatment client factors,
motivation and readiness for treatment stand out as particularly salient predictors of session
attendance and cognitive engagement (i.e., perceived helpfulness and satisfaction, interest and
confidence in treatment, commitment to treatment, and counsellor rapport). This is the case
whether engagement is rated by clients (Broome, Simpson, & Joe, 1999; Joe et al., 1998; Joe,
Simpson, & Broome, 1999; Knight et al., 2000; Sia, Dansereau, & Czuchry, 2000; Simpson,
Joe, Greener, & Rowan-Szal, 2000; Simpson, Joe, Rowan-Szal, & Greener, 1995) or counsellors
32
(Joe, Simpson, Greener, & Rowan-Szal, 1999; Simpson & Joe, 2004; Simpson, Joe, Rowan-
Szal, & Greener, 1997). The association between motivation and early cognitive engagement
has also been shown among adolescents (Broome, Joe, & Simpson, 2001). Finally, interventions
designed to promote treatment readiness have been shown to enhance client engagement (Sia et
al., 2000). Greater problem severity at admission has been linked with greater treatment
motivation (Boyle, Polinsky, & Hser, 2000), as well as with the formation of better therapeutic
relationships with counsellors and peers in the initial weeks of treatment (Broome, Knight,
Knight, Hiller, & Simpson, 1997).
Another pre-treatment factor related to attendance and cognitive engagement is legal pressure.
Among clients in residential treatment, legal pressure at admission was not significantly
associated with cognitive engagement, despite being associated with longer retention (Joe,
Simpson, & Broome, 1999; Knight et al., 2000). Among clients in outpatient counselling
programs, legal pressure was positively associated with session attendance and retention, but
negatively impacted on subjective engagement (i.e., client ratings of confidence in treatment,
treatment commitment and counselling rapport) (Joe, Simpson, & Broome, 1999). Further
studies have replicated the positive associations between cognitive engagement in treatment and
measures of problem recognition and treatment motivation in corrections-based treatment
settings (Broome et al., 1997; Hiller et al., 2002). These findings are notable as they highlight
the variability and relevance of treatment motivation and engagement in settings where all
clients are legally mandated to participate.
Empirical work has demonstrated positive associations between the cognitive and objective
components of engagement (Broome et al., 2001; Broome et al., 1999; Joe, Simpson, &
Broome, 1999; Joe, Simpson, Greener et al., 1999; Simpson & Joe, 2004; Simpson et al., 2000;
Simpson et al., 1995; Simpson, Joe, Rowan-Szal et al., 1997). In structural equation models,
subjective ratings of cognitive engagement and session attendance have exhibited positive
significant reciprocal relationships across modalities of treatment (Joe, Simpson, & Broome,
1999; Joe, Simpson, Greener et al., 1999; Simpson & Joe, 2004; Simpson et al., 2000; Simpson,
Joe, Rowan-Szal et al., 1997). Indicators of early engagement in treatment also have important
positive impacts on early recovery or progress in treatment, including in-treatment substance use
and psychosocial functioning, as well as on longer-term retention (De Weert-Van Oene,
Schippers, De Jong, & Schrijvers, 2001; Fiorentine et al., 1999; Joe, Simpson, & Broome, 1999;
33
Joe, Simpson, Greener et al., 1999; Simpson & Joe, 2004; Simpson et al., 2000; Simpson et al.,
1995; Simpson, Joe, & Rowan-Szal, 1997; Simpson, Joe, Rowan-Szal et al., 1997). Finally, the
elements of treatment process comprising early engagement, early recovery, and retention have,
in turn, been associated with improved family and peer relationships, reduced substance use, and
reduced criminal activity assessed at 12-month follow-up (Griffith, Knight, Joe, & Simpson,
1998; Hser, Grella, Hsieh, Anglin, & Brown, 1999; Joe, Simpson, Dansereau, & Rowan-Szal,
2001; Simpson & Joe, 2004; Simpson et al., 2000). Related work in a sample of offenders
mandated to residential treatment showed that favourable perceptions of counsellors were
associated with reduced re-arrest rates up to 2 years later (Broome et al., 1997; Broome, Knight,
Hiller, & Simpson, 1996).
By incorporating both behavioural and cognitive factors, the TCU concept of treatment
engagement is more comprehensive than other uses of the term in recent literature. A major
advantage of defining engagement based on retention is that such information can typically be
obtained from administrative data sources, while cognitive involvement in a program and
perceptions of the counselling relationship need to be assessed in follow-up surveys or
interviews with clients and/or counsellors. However, measures of cognitive involvement carry
their own advantages, particularly in terms of improved conceptual specification of the
treatment process. First, defining retention based on clinical records is not always a
straightforward process, particularly in outpatient treatment settings (Simpson, Joe, & Brown,
1997). Outpatient treatment is often characterized by periodic contacts, broken by gaps in
sessions or appointments, and uncertainty can arise as to the date when treatment is terminated.
While measures based on shorter defined benchmarks of retention or session attendance can
potentially improve reliability over measures of length of stay (Simpson, Joe, & Brown, 1997),
the evidence base and clinical meaning of such benchmarks is not always apparent (Walker,
2009). Further, because in some residential services, as well as in mandated treatment, session
attendance can be compulsory, incorporating additional measures of cognitive engagement may
be critical to accurately describing client involvement in the treatment process (Simpson, 2004).
Additional research supports the importance of capturing both cognitive and behavioural aspects
of treatment process by describing client and counsellor definitions of treatment success, and by
documenting the discord between retention and positive post-treatment outcomes. In a study
evaluating an adolescent treatment program, Schacht Reisinger et al. (2003) describe two
34
perspectives of the treatment process: engagement versus navigation. Counsellors preferentially
endorsed the importance of engagement, defining treatment success in terms of commitment to
the program and motivation to change. Conversely, adolescent clients tended to define success
in terms of navigating the program, by complying with rules and fulfilling requirements with
minimal commitment to meaningful recovery. According to clients, the main objective of
treatment involved completing the steps required by external parties who referred or mandated
their attendance in treatment. “Going through the motions” of treatment without actively
engaging or participating in the therapeutic process was also one of the most commonly reported
incidents of non-compliance among adult offenders diverted from the legal system to residential
treatment (Sung, Belenko, & Feng, 2001).
In a second qualitative study, this one involving homeless men in treatment primarily for crack
cocaine addiction, clients and counsellors did not vary in their definitions of treatment success:
both cited progress in recovery, involving sobriety and personal goal achievement, rather than
program attendance or completion (Stahler, Cohen, Greene, Shipley, & Bartelt, 1995). Elements
identified as leading to success were similar to those outlined in the TCU model of treatment
process, including motivation and program structure and culture, as well as support from peers
and external social networks. A process akin to navigation (i.e., acting in accordance with
program rules and culture without internalizing the value of recovery) was offered as a potential
explanation for the perception by both groups that program retention did not necessarily
constitute success (Stahler et al., 1995).
With respect to the potential discord between retention and positive behaviour change, Schacht
Reisinger et al. (2003) describe the treatment process for two adolescent clients with opposing
post-treatment outcomes. The first client left the program early, but had engaged in the process
of behaviour change and achieved positive substance-related outcomes after leaving treatment.
The second client completed the program as required, but remained uncommitted throughout
and returned to substance use after discharge. Many of the homeless men who left treatment
prior to program completion similarly relayed positive evaluations of the program and staff and
reported current abstinence from crack cocaine (Stahler et al., 1993). Some felt that they had
received what they needed from treatment in the time they were there and did not feel that
additional services were required. Such perspectives have been echoed elsewhere by clients who
fail to complete a scheduled program of treatment (Leigh, Ogborne, & Cleland, 1984).
35
Studies employing retention-based outcome measures make implicit assumptions about the
effectiveness and necessity of treatment of a given duration for successful recovery (Walker,
2009). However, it has been noted that cognitive engagement in treatment, including
internalization and personal valuation of treatment and recovery, is likely to be more closely
related to treatment effectiveness and is a relevant outcome measure in and of itself (Schacht
Reisinger et al., 2003). Personal valuation of the recovery process and lifestyle change is
ultimately more integral to positive outcomes in the long-term.
There is reason to believe that evaluations of mandated or coerced addiction treatment would
particularly benefit from the incorporation of more comprehensive measures of cognitive
engagement (Wild, 2006). To the degree that session attendance is mandatory, retention-based
measures provide limited additional information on cognitive involvement in treatment, and
may be a poor proxy for the internalization of treatment content and behaviour change. There is
also evidence that, while being associated with higher session attendance and longer retention,
legal pressure at admission is, at the same time, associated with poorer cognitive engagement in
treatment (Joe, Simpson, & Broome, 1999). No studies of treatment process to date have
incorporated more comprehensive measures of social pressures at treatment (i.e., including also
non-legal formal or informal sources) or considered client perceptions of coercion. Further,
retention remains the most common outcome assessed in studies of mandated treatment (Wild et
al., 2002). Given the above concerns, it is not clear that studies linking social pressures to
retention are well-suited to provide evidence of the effectiveness of coerced treatment. Research
is needed examining the relative impacts and interactions of social pressures, perceived
coercion, and motivation on cognitive engagement in treatment and the recovery process.
2.4 Theoretical perspectives on help-seeking and health behaviour change
Most existing frameworks for the study of help-seeking and health behaviour change involve
perceptions of threats to health as the central precursors (Aday & Andersen, 1974; Mechanic,
1968; Rosenstock, 1974; Suchman, 1970). Others focus on expectations of personal capability
to successfully perform a given behaviour (Ajzen, 1991; Bandura, 1977). However, due to the
large role played by external pressures in providing the driving force for entry into addiction
treatment, it has been suggested that help-seeking in this context is more accurately framed in
terms of the individual’s greater or lesser success in resisting social control (Weisner, 1987).
36
In answering questions of behaviour change or treatment effectiveness, the role played by client
autonomy is particularly central to the study of coerced treatment. The consequences of threats
to psychological well-being through constraints placed on autonomy by coerced treatment have
the potential to be especially great given the already marginalized position of those with alcohol
and drug use problems. Among social-psychological models for studying health behaviour
change, Self-Determination Theory (SDT) is unique in its consideration of autonomy as the
central concept. It provides a useful framework for coerced addiction treatment by addressing
how social events are perceived and how those perceptions affect motivational processes and
behavioural outcomes.
2.4.1 An overview of Self-Determination Theory
Self-Determination Theory (SDT) is a broad theory of human motivation that incorporates
social and environmental contexts and basic needs satisfaction in explaining behaviour and
psychological health (Deci & Ryan, 2000, 2002; Ryan & Deci, 2000). It proposes that humans
have basic needs for autonomy, competence, and relatedness; that is, to feel that their activities
are self-organized and self-endorsed, that they are able to bring about desired outcomes, and that
they are close and connected with significant others. To the extent that these psychological
needs are satisfied, people experience effective functioning and healthy development, while
threats to these needs result in poor psychological functioning and ill health (Deci & Ryan,
2000; Reis, Sheldon, Gable, Roscoe, & Ryan, 2000; Ryan & Deci, 2000). Indeed, SDT research
implicates persistent need deprivation in the aetiology of some forms of psychopathology (Ryan
& Deci, 2008).
Within SDT, motivation takes on several distinct forms based on the relative levels of self-
determination of behaviour initiation and regulation (Deci & Ryan, 2000; Ryan & Deci, 2002).
The greatest degree of self-determination is found in intrinsically motivated behaviours. These
have an internal perceived locus of causality, meaning that the individual sees him or herself as
providing the original impetus for action (Ryan & Connell, 1989). Intrinsically motivated
behaviours are inherently interesting and are performed in the absence of external contingencies,
such as rewards, punishment, or evaluation. Events that offer information and choice and
support the individual’s needs for autonomy, competence and relatedness, tend to facilitate
intrinsic motivation. Conversely, pressures, rewards, punishments and other controlling events,
such as deadlines, surveillance, and evaluation tend to reduce perceptions that the behaviour is
37
being internally initiated and are experienced as limiting freedom and choice. Such controlling
events, even when positive in nature (i.e., rewards), tend to undermine intrinsic motivation by
causing the perceived locus of causality to shift from internal to external (Deci, Koestner, &
Ryan, 1999; Deci & Ryan, 1987).
At the opposite end of the self-determination continuum lies amotivation. In states of
amotivation, any behaviour that occurs is neither regulated nor intentional, reflecting an
impersonal locus of causality. This concept resembles learned helplessness, which is
characterized by depression and impaired learning and performance (Garber & Seligman, 1980).
Persistent negative feedback about performance, repeated failures, and experiences of outcomes
that are perceived to be independent of one’s behaviour can bring about perceptions of
incompetence and a lack of personal causation. Perceptions of competence and the existence of
a link between behaviour and a desired outcome are required for intentional behaviour in other
related theories of behaviour change (Ajzen, 1991; Bandura, 1977).
Between these extremes lies extrinsic motivation, which characterises much of human
behaviour. Extrinsically motivated behaviours are undertaken because satisfaction is derived
from the outcome rather than from the activity itself. However, within extrinsic motivation,
behaviours can vary greatly in the degree to which they are self-determined. External regulation
characterises behaviours that are fully governed by outside forces. These are initiated to satisfy
external demands, pressures, or social norms that do not form part of the actor’s personal value
system; that is, they have an external locus of causality. External regulation represents the least
self-determined form of extrinsic motivation; however, it is differentiated from amotivation in
that behaviours are still to some degree regulated and intentional. Introjected regulation pertains
to behaviours that are initiated and governed by internal contingencies, such as the relief or
avoidance of anxiety or guilt, approval from others, or the enhancement of self-esteem. Like
external regulation, introjection is experienced as controlling and external to the actor’s sense of
self. Identified regulation characterises behaviours that are consciously valued and personally
endorsed. As behaviours are accepted for their personal importance, identification represents a
relatively internal form of regulation. The most self-determined form of extrinsically motivated
behaviour falls under integrated regulation. Integrated behaviours form part of one’s self-
concept and are enacted entirely free of pressure. They are still extrinsically motivated as they
38
are performed to attain a specific outcome rather than out of pure enjoyment; however, the
outcome is congruent with other self-endorsed and valued goals.
The initiation and regulation of behaviours have important implications for learning and
performance, creativity, flexibility, and persistence, as well as for overall psychological well-
being (Deci & Ryan, 2000; Ryan & Deci, 2000). Behaviours and goals which are undertaken
and pursued with a greater degree of autonomy are found to be more effective and persistent
than those which are selected and regulated externally. Integrated behaviours are hypothesized
to share many of the positive qualities of intrinsically motivated behaviours (Ryan & Deci,
2000). Conversely, controlled forms of behaviours tend to persist only until the external
contingencies are removed, such that any initial compliance may be time-limited and dependent
on the presence of rewards and sanctions.
Accounting for psychological well-being in the face of a multitude of daily extrinsically
motivated behaviours, SDT posits that humans have a natural inclination toward behaviour
integration (Deci & Ryan, 2000; Ryan & Deci, 2002). With appropriate supporting conditions,
people will tend to integrate behaviours that are initially externally regulated, thereby
experiencing them as increasingly autonomous over time. The active process by which this
occurs is called internalization. As with intrinsic motivation, this process is dependent on the
degree to which social contexts offer support for the basic psychological needs of autonomy,
competence, and relatedness. Contexts that are supportive of internalization offer autonomy
support or the experience of choice and freedom from external pressure, structure or the
facilitation of realistic goals and expectations as well as the provision of positive feedback, and
involvement, which refers to trust, understanding and genuine investment by significant others
(Markland, Ryan, Tobin, & Rollnick, 2005). While support for relatedness and competence are
necessary, autonomy is considered to be crucial for full integration of behaviours (Deci & Ryan,
2000; Ryan & Deci, 2002). Through internalization, therefore, SDT specifies a process model of
behaviour change in which support for psychological needs mediates the relationship between
social context and positive outcomes (Deci & Ryan, 2000; Williams, 2002).
While social context therefore plays a vital role in supporting behaviour change processes,
motivation is impacted by the psychological meaning or functional significance that individuals
assign to such external factors, rather than being directly influenced by them (Deci & Ryan,
39
1987). Certain types of events and contexts are typically found to be amotivating, controlling, or
autonomy-supportive when averaged over groups of people; however, there are also individual
differences in meaning assigned to particular situations. In addition, the meaning of specific
events may vary according to the broader social context in which they occur. For instance,
events that may otherwise be construed as controlling, such as rewards, may not necessarily
undermine intrinsic motivation when delivered in a context that is overall supportive of
autonomy. Finally, people can be controlled by internal states, dispositions, and general
orientations toward their surroundings, in the same way as external events and contexts, with
impacts on self-determination and functioning.
Due to the central role played by autonomy in SDT, as a basic need and distinguishing feature of
healthy goal-setting and behaviour regulation, a more detailed explication of its properties is
warranted. The SDT concept of autonomy “refers to volition – the organismic desire to self-
organize experience and behaviour and to have activity be concordant with one’s integrated
sense of self” (Deci & Ryan, 2000, p. 231). In this view, autonomy is not synonymous with
independence or a lack of reliance on others, but refers instead to an ability to self-govern and
make informed choices that are in line with personal interests (Deci & Ryan, 2000; Deci &
Ryan, 2008; Koestner & Losier, 1996). Importantly, SDT specifies that both autonomy and
relatedness (i.e., social connectedness to others within one’s network or community) are basic
psychological needs, and empirical work supports a positive link between them (Hodgins,
Koestner, & Duncan, 1996; La Guardia & Patrick, 2008; Sheldon & Bettencourt, 2002). The
self-integration specified by autonomy and the social-integration that comes with relatedness
are, therefore, seen as compatible and parallel processes and both are required for optimal
growth and development (Deci & Ryan, 2000). This view of the autonomous self as embedded
within a broader social-relational context, as opposed to existing in isolated self-sufficiency, is
recognized in other contemporary accounts of autonomy, such as that of relational autonomy
developed in feminist scholarship (Ho, 2008; MacKenzie & Stoljar, 2000).
The distinction between autonomous and controlled forms of motivation represents the major
distinguishing factor of SDT from other theories of motivation and behaviour change. Other
commonly used frameworks consider only the strength of motivation for intentional behaviours;
that is, people are either motivated to act or they are not. Perhaps the best example of this is
found in the stages of change construct of the Transtheoretical Model (TTM), which has
40
immense popularity within the addiction field (Prochaska & DiClemente, 1986; Prochaska,
DiClemente, & Norcross, 1992). Briefly, individuals engaged in a process of behaviour change
are hypothesized to proceed through the sequential stages of precontemplation, contemplation,
preparation to action and finally longer-term maintenance. Their level of motivation is reflected
in the extent to which they recognize they have a problem, exhibit or cite a desire for change,
and take steps toward changing (for instance, by entering a formal treatment program or
achieving specified behavioural goals).b Particularly in the area of addiction treatment research,
stage-based measures of motivation remain the norm in empirical studies (De Weert-Van Oene,
Schippers, De Jong, & Schrijvers, 2002; Project MATCH Research Group, 1997; Rollnick,
Heather, Gold, & Hall, 1992; Simpson & Joe, 1993).
Differences in the behaviour change process hypothesized by SDT versus TTM stem mainly
from the ways that motivation is formed and expected to change over time. By its own
admission, the stages of change construct does not account for why some people undertake
behaviour change while others do not (Prochaska & DiClemente, 1986). An individual who
engages in an activity because of perceptions that it is required by others or to maintain self-
esteem would not be differentiated in level of motivation from another who engages in an
activity out of a sense of personal commitment. In SDT, the reasons for undertaking a given
behaviour or activity are seen to have a direct impact on the quality of motivation and outcomes.
Like stage-based conceptualizations, SDT recognizes the dynamic nature of motivation through
the process of internalization. However, internalization does not represent a series of stages
through which people sequentially progress over the process of behaviour change (Ryan & Deci,
2000; Vallerand, Pelletier, & Koestner, 2008). It is theoretically possible for an activity to be
adopted anywhere along the behavioural continuum, and one need not progress through each
form of regulation as behaviour becomes internalized. For instance, it is not necessary or
expected that one would experience introjection as behaviour moves from external to identified
regulation.
b While a full critical evaluation of the TTM is beyond the scope of the present review, it should be noted that the stages of change, which have proved both theoretically problematic and difficult to operationalize in practice, have received a fair degree of negative criticism in the published literature (Bandura, 1997; Joseph, Breslin, & Skinner, 1999; Littell & Girvin, 2002; Sutton, 1996; West, 2005).
41
A large body of empirical evidence has accumulated to support the central predictions of SDT.
This work has addressed self-determination across a wide range of domains and settings,
including education, parenting, religion, sports and exercise, work, political activism, and health
care (Deci & Ryan, 2008; Deci & Ryan, 2002). Specific to the study of health behaviour change,
the benefits of autonomy-supportive therapeutic contexts and autonomous motivation has been
documented for a diverse range of behaviours, including weight loss, diet, exercise, prescription
medication adherence, smoking cessation and addiction recovery, and glycemic control among
those with diabetes (Levesque et al., 2007; Wild et al., 2006; Williams, Freedman, & Deci,
1998; Williams, Grow, Freedman, Ryan, & Deci, 1996; Williams, McGregor et al., 2006;
Williams, McGregor, Zeldman, Freedman, & Deci, 2004; Williams, Rodin, Ryan, Grolnick, &
Deci, 1998; Zeldman, Ryan, & Fiscella, 2004). Empirical application of SDT has been greatly
facilitated by the development of measures for many of its theoretical concepts, including
autonomous and controlled forms of motivation (Levesque et al., 2007; Ryan, Plant, &
O'Malley, 1995; Wild, 1999b) and the autonomy-supportiveness of health care and interpersonal
social contexts (Williams et al., 1996; Williams, Lynch et al., 2006). The following section
considers the theoretical implications of SDT for the study of coerced addiction treatment more
specifically, as well as empirical support for its use in this context.
2.4.2 Self-Determination Theory and coerced addiction treatment
As a guiding theoretical framework for the study of coerced addiction treatment, SDT offers a
number of helpful propositions. First, it specifies several distinct types of motivation with
different consequences for the therapeutic response, the robustness of behaviour change, and
psychological well-being. With the exception of those who seek psychotherapy for the purposes
of self-discovery (Pelletier, Tuson, & Haddad, 1997), it is unlikely that intrinsic motivation
plays a significant role in therapeutic change, as few people will embark upon a process of
changing harmful health behaviours out of a sense of intrinsic enjoyment (Markland et al., 2005;
Vansteenkiste & Sheldon, 2006). Nonetheless, behaviours undertaken during the process of
recovery from addiction may still be personally significant and have value in terms of need
satisfaction. Clients who report autonomous reasons for entering treatment may experience
better and more stable outcomes than those for whom treatment is externally regulated or
introjected. Conversely, clients who report controlled motivation for treatment may initially
comply with the program, but be more prone to relapse once the controlling factors are lifted.
42
Excessive pressures and controls placed on treatment admission and participation are expected
to influence clients’ relative levels of autonomous versus controlled motivation. Those who
perceive that they were treated fairly and respectfully throughout the admission process can be
expected to exhibit greater autonomous motivation for treatment.
SDT also allows for individual differences in the ways that people respond to events and
contexts. For instance, past treatment experience may affect how a given client will respond to
new pressures or mandates to enter treatment. This recognition of client heterogeneity highlights
the inadequacy of considering only external circumstances when addressing coercion.
Objectively coercive acts, such as a treatment mandate, may promote either an internal or an
external locus of causality with respect to treatment entry, depending on the functional
significance that the client assigns to it. A mandated client with prior positive experiences with
treatment, or an internal desire to change his or her level of substance use, may exhibit a fair
degree of autonomous motivation for treatment. Conversely, a mandated client with repeated
failed attempts at treatment in the past may perceive that a decision on their part to enter (or to
resist) treatment will have no effect anyway. This client may be amotivated rather than
externally motivated for treatment, and may respond to pressure with passivity and helplessness.
SDT therefore offers a theoretical rationale for measuring client perceptions of events in
addition to situational factors when evaluating the presence and consequences of coercion.
Finally, through internalization, SDT offers a mechanism for therapeutic change and outlines the
conditions that are likely to support versus impede this process. This has high relevance in
clinical settings, where new behaviours are likely to be initially regulated to at least some degree
by external or introjected motivation. To the extent that people are naturally inclined to integrate
new behaviours of adaptive significance, as hypothesized by SDT, treatment interventions that
more readily demonstrate the value of recovery in terms of satisfaction of psychological needs
will be more likely to meet with success in the long-run. Accordingly, the key to therapeutic
success with all clients, regardless of initial motivation, lies in creating a therapeutic
environment that actively supports, rather than obstructs, the individual’s natural progression
toward integration and growth (Ryan & Deci, 2008). Responsibility for the creation of this
environment lies not only with the health professionals directly involved in service provision,
but extends also to family, friends, employers, legal authorities, and any others involved in the
treatment process.
43
If the purpose of coerced treatment involves sustained recovery from addiction rather than
punishment, as is claimed by proponents of legal diversion strategies, then the promotion and
support of internalization is likely highly relevant to the treatment process. If treatment and
behaviour change are perceived to have adaptive significance in terms of promoting autonomy,
competence, and relatedness, they may shift from being externally regulated to integrated. This
has implications for the use of pressures, sanctions, and rewards throughout treatment.
Treatment that is respectful of clients’ autonomy, by offering information and a rationale for
behaviour change, a choice of programs or goals, and limited use of controls and pressures, is
hypothesized to be more successful at helping clients internalize and integrate the value of
reducing or abstaining from substance use. Conversely, initial autonomous motivation for
treatment may be equally “externalized” if treatment entry and participation comes to be
associated with external contingencies.
A number of studies have applied SDT to the study of recovery from addiction, amassing
evidence of the correlates and consequences of autonomous and controlled motivation. Scales
developed to differentiate between motivational subtypes in addiction treatment populations
differ somewhat in their ability to distinguish among the types of extrinsic motivation. Of the
two most commonly used scales, the Treatment Entry Questionnaire (TEQ) provides separate
subscale scores for identified, introjected and external motivation (Wild, 1999b), while the
Treatment Motivation Scale (TMQ) differentiates only between external motivation and internal
motivation, the latter covering both introjected and identified motivation (Ryan et al., 1995).c
These differences are reflected in the terminology used throughout this section.
Among clients entering outpatient treatment for alcohol use, legal problems at admission were
associated with lower internal motivation and higher external motivation (Ryan et al., 1995).
Similarly, treatment mandates from the legal system, employers, or social services were
associated with lower identified motivation and higher external motivation in a mixed drug and
alcohol treatment sample (Wild et al., 2006). The association between motivation and informal
pressures may be more complicated. Client ratings of social network pressures to quit or reduce
substance use and to enter treatment were positively associated with external and introjected
c The term internal connotes that regulation of treatment entry and/or substance-related behaviour change has been internalized to some degree, although it does not necessarily qualify as autonomously motivated.
44
motivation for treatment, although they were unrelated to identified motivation (Wild et al.,
2006). In another study, however, the number of significant others reported to be opposed to the
client’s use of substances was positively associated with internal motivation for abstinence
(Downey, Rosengren, & Donovan, 2000). Acting in accordance with the norms and expectations
of one’s social network may support internal motivation through satisfying the need for
relatedness, such that clients may not necessarily experience pressures from this source
threatening their autonomy.
Substance problem severity was also positively associated with internal motivation and
negatively associated with external motivation (Ryan et al., 1995; Wild et al., 2006). This is
consistent with research suggesting that those who are objectively externally pressured to enter
treatment are less severely impaired by their substance use. Clients who are more severely
impaired may perceive a greater need for treatment and be more autonomously motivated to
enter and participate in treatment. As noted previously, perceived problem severity has also
been inversely associated with perceived coercion (Wild et al., 1998).
Studies that have compared strength- or stage-based measures of motivation with motivational
quality report mixed findings. In a population-based sample of tobacco smokers, internal
motivation for quitting smoking was higher among those in the preparation stage, while external
motivation was higher among precontemplators (Curry, Grothaus, & McBride, 1997). Among
clients entering treatment for drug and alcohol related problems, there was a positive association
between internal motivation for abstinence and the degree to which clients had taken action
toward changing their substance use (Downey et al., 2000). In another study, those in a pre-
action stage of change (i.e., precontemplation, contemplation, or preparation) scored higher on
introjected and identified motivation than did those in the action or maintenance stages (Wild,
1999b). In other words, those who had already taken steps to change their behaviour reported
lower autonomous motivation to enter treatment. Stage of change, defined in this way, was
unrelated to external motivation. However, because the study did not differentiate between the
pre-action stages of change, it is not possible to determine if clients in the early stages of change
differed in their levels of autonomous versus controlled motivation. Although more research is
needed, these initial mixed results are not unexpected: SDT would hypothesize inconsistent
relationships between quality and quantity-based measures of motivation, due to the latter’s
45
inability to distinguish the reasons for behaviour change or treatment entry among those with
“high” motivation.
Others have examined admission levels of client interest in and attitudes toward treatment in
light of internal and external motivation. Internalized motivation has been linked with more
positive attitudes toward treatment, including greater perceived benefits of reducing substance
use and higher ratings of interest in treatment (Conner, Longshore, & Anglin, 2009; Wild et al.,
2006). External motivation, on the other hand, has been associated with both more positive
client attitudes toward treatment (Conner et al., 2009) and with lower counsellor ratings of client
interest in treatment (Wild et al., 2006). In the latter study, controlling for demographic
variables and objective measures of social pressure to enter treatment, external motivation was
unrelated to client-perceived costs and benefits of quitting, while introjected motivation was
associated with both perceived benefits and costs of behaviour change. This pattern of findings
suggests that externally motivated clients may not have yet been contemplating the need for
change, while those reporting introjected motivation for treatment may be somewhat ambivalent
about their need to change. Notably, objective measures of legal, formal non-legal, and informal
pressures to reduce substance use or enter treatment were not associated with interest in or
commitment to treatment in the multivariable analysis. In line with SDT, measures that reflected
clients’ reported reasons for seeking treatment and the personal meaning assigned to social
events were more important predictors of orientation toward treatment than were the objective
measures of social events themselves.
There is limited evidence on the association between motivation and perceived coercion at
treatment entry, as only one study to date has included assessments of both constructs in the
same sample. Consistent with SDT, perceived coercion to enter treatment was positively related
to external motivation for treatment, unrelated to introjected motivation, and inversely related to
autonomous motivation (Wild et al., 2006). Autonomous motivation was also positively related
to client perceptions of the fairness of the admission process (Wild, 1999b).
A small number of studies have assessed the associations of autonomous and controlled forms
of motivation with measures of addiction treatment process and outcomes. Internal motivation
for treatment is associated with treatment retention (De Leon, Melnick, Kressel, & Jainchill,
1994; Ryan et al., 1995), increased session attendance (Ryan et al., 1995; Zeldman et al., 2004),
46
and lower rates of drug use in the months following treatment entry (Downey et al., 2001;
Zeldman et al., 2004). Elsewhere, the positive association between internal motivation for
abstinence and 90-day self-reported abstinence was independent of whether or not the client
remained in treatment during this time (Downey et al., 2000). External motivation has been
linked with reduced likelihood of abstinence (Downey et al., 2001) and poorer session
attendance (Zeldman et al., 2004), although it has also been shown to positively predict
retention (De Leon et al., 1994; Ryan et al., 1995). While not explicitly guided by SDT,
ethnographic research has highlighted the relevance of internalization of the process of
behaviour change to successful treatment process and outcomes (Schacht Reisinger et al., 2003;
Stahler et al., 1995).
To date, studies guided by SDT that address longer-term, post-treatment outcomes are limited.
However, autonomous motivation at admission has been linked with lower frequency of
drinking 9 to 12 months after discharge from alcohol treatment (Staines et al., 2003) and
increased smoking cessation up to 30 months following a brief intervention (Curry et al., 1997;
Williams, Gagne, Ryan, & Deci, 2002).
Similar findings regarding the impact of autonomous motivation have been reported in the
context of therapy for mental health problems. Autonomous motivation for psychotherapy was
correlated with more positive attitudes toward therapy, greater satisfaction, and stronger
intentions to continue, while the reverse was found for external motivation (Pelletier et al.,
1997). Greater autonomous motivation was also correlated with lower levels of in-treatment
symptomatology and impairment, as well as more positive client ratings of their progress in
subsequent sessions (Michalak, Klappheck, & Kosfelder, 2004). In another study, which
surveyed clients 3 weeks into treatment, those with higher autonomous motivation provided
more positive ratings of the therapeutic alliance and subsequently had better post-treatment
outcomes, including remission from depression and symptom reduction (Zuroff et al., 2007).
Controlled motivation, on the other hand, was not associated with either the therapeutic alliance
or post-treatment outcomes. The authors concluded that promoting autonomous motivation
through the initial weeks of treatment may be a more important focal point for therapists than
attempting to reduce controlled motivation.
47
The scales assessing autonomous and controlled forms of motivation are not constructed to
necessarily yield bipolar motivation profiles, such that an individual may score high on both
dimensions. In analyses, the motivational subscales can either be treated separately or
mathematically combined into a relative autonomy index (i.e., by weighting the subscales
according to their relative levels of autonomy and subtracting the controlled from the
autonomous subscale scores; Vallerand et al., 2008). An advantage of the former approach is
that it allows for the study of the impact of specific configurations of motivation on outcomes.
Two studies provide empirical support for the relevance of motivational configuration for
outcomes of addiction treatment, reporting significant statistical interactions between internal
and external motivation. Categorizing clients according to high and low levels of each type of
motivation based on a median split of subscale scores, Zeldman et al. (2004) reported that the
group defined by high external and low internal motivation had the lowest rates of session
attendance and highest rates of in-treatment drug use detected by urinalysis, relative to those
with other motivational configurations. Using the same statistical procedure, Ryan et al. (1995)
found that the high external and high internal combination stood out in terms of longer retention
and better session attendance.
While these studies produced somewhat divergent results, it is relevant that they differed in both
setting and length of follow-up. The study by Zeldman et al. (2004) was conducted in a
methadone maintenance treatment program and subjects were followed for a minimum of 6
months, while subjects in the study by Ryan et al. (1995) were attending an outpatient alcohol
counselling program and were followed over a 2-month period. There is reason to expect
differences across these treatment modalities. As noted earlier, legal pressure is typically found
to be associated with longer retention in generalized treatment programs, but has also been
associated with shorter retention in methadone maintenance treatment. To the extent that legal
pressure is associated with lower internal and/or higher external motivation for treatment, the
above results may be expected. More research is needed to determine the optimal configuration
of motivation and the role played by contextual factors in determining the nature and impact of
different combinations (Vallerand et al., 2008). There is, however, a convergence across these
studies, and others reviewed above, of the primacy of autonomous motivation for achieving
positive treatment outcomes.
48
Another area of SDT research in addiction treatment settings has examined contextual factors
that support internalization of new behaviours. As noted above, the optimal environment for
internalization is one characterized by structure, involvement and support for autonomy. More
specifically, this involves providing choice and a meaningful rationale for any specific requests,
minimizing pressures and controls, acknowledging clients’ feelings, offering personalized
feedback, and providing unconditional support (Deci, Eghari, Patrick, & Leone, 1994; Ryan &
Deci, 2008). Given the particular requirement of autonomy support for successful integration of
new behaviours and values, the majority of research in this area has focused on elucidating the
elements and consequences of this particular aspect of therapeutic settings.
In order to provide evidence of an association between autonomy support and internalization,
motivation for the target behaviour has to be measured at more than one point during the
treatment episode, as it is the change in motivation that is of interest. However, in most
applications of SDT to addiction treatment, motivation is measured at a single point in time (i.e.,
treatment entry) and correlated with later treatment processes or outcomes (Ryan et al., 1995;
Staines et al., 2003; Zeldman et al., 2004). One study to date has examined changes in
motivation over the initial weeks of an outpatient program for alcohol and drug use (Simoneau
& Bergeron, 2003). While there was no significant change in motivation between admission and
6-week follow-up in the sample as a whole, there were significant changes at the level of
individual clients (i.e., both increases and decreases in autonomous motivation were
documented, which in effect cancelled each other out). In line with SDT predictions, perceptions
of competence and autonomy support predicted increases in the relative levels of autonomous
motivation over the first few weeks of treatment.
Randomized trials of brief smoking cessation interventions have also reported increases in
autonomous motivation associated with participant ratings of autonomy support, with
concomitant positive outcomes among adults (Williams, McGregor et al., 2006) and adolescents
(Williams, Cox, Kouides, & Deci, 1999). In a trial evaluating the relative efficacy of an
autonomy-supportive smoking cessation intervention against a control condition representing
standard care for adult smokers (Williams, McGregor et al., 2006), outcomes were superior in
the intervention condition. However, the underlying SDT-hypothesized process of behaviour
change (i.e., linking perceived autonomy support with internalized motivation and better
outcomes) held for participants in both groups. Further, the benefits of perceived autonomy
49
support, in terms of increased use of smoking cessation aids and quit rates, were seen regardless
of whether or not participants intended to quit smoking at study enrolment, highlighting the
potential of SDT-based interventions for those who are initially ambivalent to behaviour change
(Williams, McGregor et al., 2006).
While not providing evidence on internalization per se, additional studies have confirmed the
relevance of autonomy support itself for positive outcomes. Client ratings of autonomy support
have been shown to predict lower drug use in a methadone maintenance program (Zeldman et
al., 2004) and higher ratings of the therapeutic alliance in psychotherapy (Zuroff et al., 2007). In
both studies, autonomous motivation and autonomy support were positively associated, as
predicted. However, because motivation was assessed at only one point in time, it is not possible
to determine the specific nature or direction of this association.
The context of autonomy support is similar in many ways to that promoted by Motivational
Interviewing (MI) techniques (Miller & Rollnick, 2002), which have also met with success in
treating substance use problems (Burke, Arkowitz, & Menchola, 2003; Carroll et al., 2006;
Vasilaki, Hosier, & Cox, 2006). MI employs the language of both the TTM and SDT in its aims
to support progress from precontemplation or contemplation to preparation and action and to
develop a personal commitment for change (Miller & Rollnick, 2002). The link between SDT
and MI has been made explicit in the literature through suggestions that SDT provides a
theoretical framework for explaining the mechanism underlying the technique of MI (Foote et
al., 1999; Markland et al., 2005; Vansteenkiste & Sheldon, 2006). This work reframes the
methods of MI in terms of psychological need satisfaction. For instance, SDT provides a
theoretical justification for the importance of fostering personal endorsement of the need for
change and avoiding direct demands on clients through its concept of autonomy support.
Relative to a usual care group, clients randomized to an intervention group based on MI
principles rated their therapeutic climate as more autonomy-supportive, which was in turn
associated with better subsequent treatment attendance (Foote et al., 1999).
More research is needed evaluating the salient factors in creating a therapeutic context that is
supportive of autonomy, as well as the consequences for stabilized recovery in the long-term.
Notably, the social context in which one recovers from addiction problems is not comprised
solely of the treatment setting. Recent research has begun to look beyond health professionals to
50
the role of autonomy support from family and significant others in health behaviour change
(Williams, Lynch et al., 2006). In this initial study, autonomy support from significant others
was associated with increases in autonomous motivation and perceived competence for quitting
smoking and improving diet, as well as with smoking cessation and dietary outcomes assessed
at a 6-month follow-up. In outcomes research more generally, social support, and particularly
social network support for abstinence, has been linked with several aspects of treatment and
recovery, including treatment entry (Davey et al., 2007), compliance with the treatment process
(Sung et al., 2004), and positive outcomes (Broome, Simpson, & Joe, 2002; Flynn, Joe, Broome,
Simpson, & Brown, 2003; McKay et al., 2005; Moos, 2007; Project MATCH Research Group,
1997). SDT research on close relationships suggests that the quality of relationships and their
impact on health and well-being are functions of the extent to which the relationship is
supportive of the partners’ basic psychological needs (La Guardia & Patrick, 2008). As such, the
benefits of social support for the therapeutic process are expected to be optimized when the
client perceives it as supportive of autonomy, rather than as controlling (Ryan & Solky, 1996).
More research is needed to examine this particular role of family and friends in the addiction
treatment process.
While promising in providing initial support for SDT as a framework for evaluating coerced
addiction treatment, the work to date is limited and requires replication in other samples. More
research is required to study the interplay between social pressures, perceptions of coercion to
enter treatment, motivation quality, and the associations with treatment process and post-
treatment outcomes (Wild, 2006). Over and above measures of retention and session attendance,
studies are needed that address active and meaningful engagement in the treatment process. The
concepts of therapeutic alliance and counsellor rapport fit easily into the theoretical framework
of SDT through their potential role in satisfying the need for relatedness (Ryan & Deci, 2008).
However, these concepts have received limited attention in SDT research conducted to date in
addiction treatment settings. Work that has examined post-treatment outcomes and the stability
of gains made in treatment is also particularly limited.
Longitudinal study designs employing statistical techniques such as path analysis and structural
equation modeling, which allow for the modeling of relationships between multiple theoretical
constructs over time, are needed to test the assumptions of SDT for successful behaviour change
in addiction treatment contexts. Ryan et al. (1995) used path analysis to test an SDT model of
51
treatment process, in which motivation mediated the associations between admission
characteristics, including problem severity and legal pressure, and program involvement over
the initial weeks of treatment. However, the internal validity and generalizability of their
findings are potentially limited by a relatively low sample size (n=109), and by their use of a
series of data-driven composite measures to represent key variables (i.e., legal pressure and
treatment process), which were not independently validated and are likely sample-specific.
Others offer a more robust test of SDT, using validated measures and randomized study designs;
however, these have been limited to evaluations of smoking cessation interventions offered in
primary care settings (Williams et al., 2002; Williams, McGregor et al., 2006). The
generalizability of these findings to the specialized alcohol and drug treatment system, with its
formalized roles for legal and other authorities, has yet to be explored.
In summary, the theoretical framework provided by SDT is highly relevant to the study of
coercion and social pressure in addiction treatment entry. Its explication of the process of health
behaviour change offers a potential explanation for many of the conflicting findings reported to
date on mandated or pressured treatment, and highlights many avenues for future work in this
area.
2.5 Objectives The overall aim of the present work is to contribute to understandings of the nature and meaning
of coerced addiction treatment, through a short-term prospective study grounded in SDT. The
objectives are two-fold:
to explain perceptions of coercion and autonomous motivation for treatment among
adults entering outpatient addiction treatment, and
to determine the impact of perceived coercion and autonomous motivation on retention,
cognitive engagement, and substance problem severity in the initial weeks of treatment.
It is hypothesized that greater social pressure to enter treatment will be associated with greater
perceived coercion and lower autonomous motivation for treatment (H1). Although these
associations are expected to hold regardless of the source of pressure, it is expected that they
will be stronger for legal and other formal pressures than for informal pressures. Among other
client factors, substance problem severity at admission is expected to be associated with lower
perceived coercion and higher autonomous motivation (H2). Next, it is hypothesized that lower
52
perceived coercion and higher autonomous motivation for treatment at admission will be
associated with a greater likelihood of continued attendance beyond 2 months (H3) and greater
cognitive engagement in the treatment process (H4). Finally, clients who are still retained in
treatment after 2 months are expected to exhibit greater reduction in substance problem severity
over the study period (H5).
No study to date provides a comprehensive account of the associations between social pressures,
perceived coercion, and treatment motivation, or its consequences for the treatment process and
behaviour change. Through a focus on treatment entry and early treatment processes, this study
is designed to provide evidence on which further outcome evaluations can be based. This work
has further implications for the design and delivery of services and broader system planning
initiatives. More specifically, prospective studies of the impact of coercion on recovery
processes and outcomes are needed to inform optimal approaches toward intervening and
responding to substance use problems in the community and throughout treatment. Given the
significant burden of substance use problems on society, combined with the currently high
prevalence of potentially coercive pressure tactics for influencing admission to specialized
treatment, there is an urgent need for evidence-based policies in this area.
53
Chapter 3 Methods
A short-term prospective study of clients accessing outpatient counselling for alcohol and drug
use problems was conducted to address the above objectives. The following sections describe
the study procedures, measures, and analysis.
3.1 Participants The study sample includes 276 adults admitted to the Peel Addiction and Assessment Referral
Centre (PAARC) over a 20-month period, from July 2007 to February 2009. In total, 371 clients
were assessed for eligibility (Figure 2). Of these, 5 individuals (1.3%) did not speak or
understand English well enough to participate in the study. Of those who were eligible to
participate (n=366), 276 (75.4%) were enrolled in the study. Of the remaining 89 clients who
were not recruited, 67 refused, typically because they did not have enough time or were not
interested in the study. Another 23 stated initially that they were interested, but did not have
time to complete the survey at the time approached; later attempts to recruit them were
unsuccessful due to missed appointments or other failures to connect.
The final sample was predominantly male and young to middle-aged (Table 1). The mean age of
participants at admission was 36.3 years (SD=10.6). Almost half were currently married or in a
significant relationship. Approximately two-thirds were employed either full- or part-time,
although over one-fifth were unemployed and looking for work. Over half of the sample had a
high school education or less. Most clients reported personal employment as their current source
of income. Despite the fact that over 20% were unemployed, less than 10% were currently
receiving support through Ontario Works (OW). Seven percent reported no current source of
income at admission.
54
Figure 2. Flow diagram of study and treatment
TREATMENT TIMELINE
STUDY RECRUITMENT
Recruited (n=276)
Followed up (n=205)
Assessed for eligibility (n=371)
Lost to follow-up (n=71) Not reached (n=25) Number not in service or
wrong number (n=8) Scheduled, but not
completed (n=35) Refused (n=3)
Analyzed (by dependent variable) 1. Perceived coercion (n=276) 2. Retention (n=276) 3. Engagement (n=112); excluded: lost to follow-up
(n=14), not in treatment (n=150) 4. Reasons for leaving treatment (n=93); excluded: lost to
follow-up (n=57), still in treatment (n=126) 5. Change in substance problem severity (n=205);
excluded: lost to follow-up (n=71)
Excluded (n=95) Refused (n=67) Missed (n=23) Ineligible, poor English
skills (n=5)
Admission
Clinical assessment, 2 sessions (letter of attendance provided, if desired)
Counselling
Baseline
Dropped out(n=150)
2 Months Retained (n=126)
55
Table 1. Characteristics of the study sample at admission (n=276)
Client characteristics n % Gender:
Male 216 78.3 Female 60 21.7
Age: < 20 yrs 6 2.2 20-29 yrs 85 30.8 30-39 yrs 78 28.3 40-49 yrs 72 26.1 50-59 yrs 32 11.6 60 yrs and older 3 1.1
Marital status: Married, partnered, common-law 123 45.1 Single (never married) 107 39.2 Widowed, separated, divorced 43 15.8 Missing 3
Employment status: Employed (full- or part-time) 179 65.6 Unemployed (looking for work) 59 21.6 Not in the labour force or on leave a 35 12.8 Missing 3
Education: Less than high school 68 26.1 High school completed 74 28.4 Some post-secondary 38 14.6 Post-secondary completed 81 31.0 Missing 15
Income source: Employment 172 63.9 Insurance b 13 4.8 Supported by family or friends 19 7.1 None 19 7.1 Ontario Disability Support Program (OSDP) 9 3.3 Ontario Works (OW) 24 8.9 Other c 13 4.8 Missing 7
Previous treatment: g No 176 64.5 Yes 97 35.5 Missing 3
Treatment mandate: d No 170 61.6 Yes 106 38.4
Legal problems: None 130 47.3 Awaiting trial or sentencing 56 20.4 Probation or parole 88 32.0 Under investigation 1 0.4 Missing 1
Referral source: Addiction service 29 10.8 Family or friends 32 11.9 Social services 10 3.7 Legal system 92 34.3 Mental or general health services 43 16.0 Workplace or school 7 2.6 Provincial helpline and registry 24 9.0 Self 31 11.6 Missing 8
56
Table 1. Continued
Client characteristics n % Problem substance(s): e
Alcohol 146 52.9 Cocaine or crack 83 30.1 Cannabis 31 11.2 Opiates 14 5.1 Methamphetamine 7 2.5 Ecstasy or ketamine 3 1.3 Non-opiate prescription drugs f 5 1.8 None 64 23.2
a includes students, those in retraining programs, homemakers, retirees, and disabled individuals b includes employment, disability, and other unspecified insurance c includes retirement income, savings, and unspecified other income d includes mandates from the legal system and Children’s Aid Society (CAS) e multiple problem substances were permitted per client; percentages do not sum to 100 f includes benzodiazepines, barbiturates, and amphetamines g excludes self-help or mutual aid groups (e.g., Alcoholics Anonymous)
Approximately two-thirds of subjects were new to specialized addiction treatment, excluding
participation in mutual aid groups such as Alcoholics/Narcotics Anonymous (AA/NA). A
sizable minority (38%) was classified as mandated to the current treatment episode. This
information was recorded at intake during the first telephone contact with clients. Consultation
with the intake worker revealed that mandates typically stemmed from legal authorities, but
occasionally also from the Children’s Aid Society (CAS). Over half of clients reported legal
problems at admission, ranging from awaiting trial or sentencing to being on probation or
parole. Not surprisingly given this high level of legal system involvement, the courts
represented the most common referral source, reported by over a third of clients. Sizable
proportions were also referred by mental or general health services or other addiction treatment
services. Only 12% of clients in this sample identified as self-referred.
During intake, clients listed up to five problem substances that contributed to the current
treatment episode. Because multiple substances per client were allowed, the proportions
endorsing each problem substance do not sum to 100%. Over half identified problems with
alcohol and just under a third reported problems with cocaine or crack. Cannabis was reported
as a problem substance for approximately 11%. Other drugs, including opiates,
methamphetamine, and synthetic drugs, were reported by small numbers of clients. The majority
(55.8%, n=154) identified only one problem substance at admission; 15.6% identified two
(n=43), 4.7% (n=13) identified three, and 0.7% (n=2) identified five problem substances.
Overall, 34.8% (n=96) reported problems with alcohol only, 23.9% (n=66) reported problems
57
with other drugs only, and 18.1% (n=50) reported problems with both alcohol and drugs. The
remaining clients (n=64, 23.2%) did not acknowledge current substance problems at admission.
3.2 Setting PAARC, located in a densely populated suburban region of Ontario, offers alcohol and drug
assessment, referral, education, counselling, and after-care services on an outpatient basis to
residents of Peel County (estimated population in 2006 = 1,159,405) (Peel Data Centre, 2006).
The agency is funded by the Ontario Ministry of Health and Long-Term Care (MOHLTC) and
services are provided free-of-charge to residents. Over the course of the study, average wait
times varied from 1 week to 4 months, due to significant staffing shortages experienced
periodically throughout the 20 months.d
The core counselling program at PAARC consists of individual hourly sessions with a trained
addiction counsellor, on a monthly to weekly basis for several months. The first two sessions are
dedicated to assessment and treatment planning, after which the client may remain at PAARC
for ongoing counselling support on an outpatient basis or receive a referral to a residential or
other service of their choice. The frequency and number of sessions is dictated by client need
and counsellor availability; typically, clients receive between 2 and 10 sessions of 1 hour each.
Counsellors use motivational enhancement and psychotherapeutic techniques to aid clients in
achieving their personally selected goals of treatment. PAARC ascribes to principles of harm
reduction, and a goal of abstinence is not required for admission or continued participation. A
total of eight counsellors were employed at the agency during the study period, although the
total number working at any one time ranged from two to four.
PAARC operates on a self-referral basis; that is, a physician referral is not required to access
their services. Individuals interested in receiving support for their own or another person’s
substance problems call a central agency number and complete a brief intake interview, after
which they are booked to meet individually with a counsellor. PAARC also accepts individuals
who indicate during the intake interview that they do not believe they have current problems
with alcohol or drugs. It is typically the case that such individuals are attempting to access
d The potential implications of agency instability during data collections are discussed in the section on Study Limitations.
58
services to satisfy a legal or other mandate, or to obtain a letter of attendance for legal or other
purposes. PAARC is not affiliated with a drug treatment court program, however, such that none
of the legally mandated clients are participating in this type of formalized diversion program.
Over the course of the study, a change occurred in the type of service offered to clients who
denied current substance-related problems. For the first 12 months of recruitment, such clients
were offered essentially the same service as others. That is, they completed the agency’s
standardized assessment and met individually with a counsellor to discuss the results and plan
their next steps. A minimum of two sessions was required in order to obtain a letter of
attendance, although all clients were invited to continue with the service if desired. Typically,
the letter confirming completion of the assessment was all that was formally required by
probation or parole officers, lawyers, social services, or other agencies; particularly if the client
maintained the absence of substance-related problems. For the final 8 months of study
recruitment, due to ongoing insufficient staffing, the agency changed its policy to offer a 6-hour
educational group to those who denied current substance-related problems. The group covered
risk behaviours, population levels of substance use, safe-level drinking guidelines, and the
stages of change. In groups of up to 16, clients attended either a single 6-hour session or three 2-
hour sessions offered over consecutive weeks. Completion of the full 6 hours was required in
order to receive a letter confirming attendance; however, clients again were offered the option of
continuing support with an individual counsellor if desired. As before, the letter confirming
completion of the educational group was typically all that was required by the legal or other
authorities, particularly if clients maintained the absence of substance-related problems and
declined the need for further treatment.
3.3 Measures Copies of all scales are available in Appendix 1.
Index of social pressures:
Legal, other (non-legal) formal, and informal pressures to enter treatment were assessed using a
modified version of an index developed by Polcin and Weisner (1999). Respondents rated the
level of pressure to enter treatment from nine potential sources, including spouses or partners,
other family members, friends, employers or colleagues, legal authorities, CAS workers, OW or
59
Ontario Disability Support Program (ODSP) case-workers, physicians or other health care
workers, and “other” (for which respondents indicate the source). Contrary to the original index,
which documented only the presence or absence of pressure from each source, the version used
in the present study allowed respondents to rate the level of pressure along a 5-point scale from
none (0) to extreme (4). This approach was used previously in another study of adults entering
outpatient treatment for substance-related problems (Wild et al., 2006).
Separate summary scores were calculated for each of the three types of pressures: legal, other
formal, and informal. Because the number of ratings contributing to each type of pressure
differed (i.e., one pertains to legal pressures, four to other formal pressures, and three to
informal pressures), the maximum rating within each type was used as the summary score. This
avoids the potential for dilution of extreme pressure ratings from other formal and informal
sources. The summary score for legal pressure was based on the single rating for legal
authorities, while that for formal pressures was based on the maximum rating for pressure from
employers or colleagues, CAS workers, OW or ODSP case-workers, and physicians or health
care workers. Finally, the summary score for informal pressure was based on the maximum
rating for pressure from spouses or partners, family members, or friends. Responses to the
“other” category were incorporated into the appropriate pressure type. Each of the three
summary scores, therefore, retained the original item scaling (i.e., 0-4, with higher scores
indicating greater pressure).
MacArthur Perceived Coercion Scale (MPCS):
The five-item MPCS assesses the extent to which clients perceive they had influence and control
over the decision to enter treatment (Gardner et al., 1993). Each item is scored as true (0) or
false (1) yielding a final score out of 5, with higher scores indicating greater perceived coercion
at admission. Originally developed for psychiatric inpatients, the MPCS has been adapted and
evaluated in several different clinical populations, including clients entering substance use
treatment (Marlowe, Glass et al., 2001; Prendergast et al., 2009; Wild, 1999b; Wild et al., 2006;
Wild et al., 1998).
Treatment Entry Questionnaire (TEQ):
60
The 12-item version of the TEQ was used to assess motivation to enter treatment (Wild, 1999b;
Wild et al., 2006). This scale addresses the reasons for entering treatment, grounded in an SDT
framework. It consists of three subscales, representing external, introjected and identified
motivation for treatment entry. Each item is scored on a 7-point Likert scale ranging from
“strongly disagree” (1) to “strongly agree” (7). These are summed to provide scores for each
subscale, ranging from 4-28 for identified motivation (4 items), 3-21 for introjected motivation
(3 items), and 4-28 for external motivation (4 items). Support for the three-factor structure of the
12-item scale has been found in a multisite study of adults entering outpatient treatment
(Urbanoski & Wild, 2009). The original 30-item version has also shown high internal
consistency and construct validity in a sample similar to that of the present study (Wild, 1999b;
Wild et al., 2006).
The TEQ subscale scores were transformed into a Relative Autonomy Index (RAI) for analysis
(Vallerand et al., 2008). This is a measure of the level of a client’s autonomous motivation,
accounting for the presence of any accompanying externalized motivation. To calculate the RAI,
the three subscales are weighted according to their relative level of autonomy and the less
internalized forms are subtracted from the identified motivation score:
RAI = (identified motivation * 2) + (introjected motivation * -1) + (external motivation * -2)
Treatment engagement:
Cognitive engagement was assessed with a modified version of a 15-item scale developed and
used in the DATOS evaluation initiative (Joe et al., 1998). The self-reported measure consists of
three factors describing client confidence in treatment, rapport with counsellors, and
commitment to treatment. Through its use in DATOS, the scale has been used to represent
therapeutic engagement and involvement across drug treatment modalities (Broome et al., 2001;
Broome et al., 1999; Joe et al., 1998; Joe, Simpson, & Broome, 1999; Knight et al., 2000).
The subscale assessing counsellor rapport consists of five items, each scored on a 3-point scale.
Total scores range from 3-15, with higher scores indicating greater rapport. The subscale
assessing client commitment to treatment also consists of five items, each scored on a 4-point
scale. Total scores range from 5-20, again with higher scores indicating greater commitment to
the treatment process.
61
Two adaptations were made to the scoring procedure for the subscale assessing client
confidence in treatment, due to item interdependence and unequal item scaling in the original
subscale. Specifically, two items covered 1) whether treatment had helped the client stop or cut
down on their alcohol or drug use, followed by 2) how much they felt treatment had helped.
Because valid responses to the second item logically depend on a positive response to the first,
these two questions were combined into a single item representing client ratings of the degree to
which treatment had helped them cut down or stop using (1 = not at all; 2 = a little bit; 3 = a lot).
Second, the final item on this subscale contains four response categories in contrast to the other
items, which contain only three categories. In order for the items to contribute equally to the
subscale score, the final item was scored out of three by rescaling the response categories to be
separated by an interval of 0.75 instead 1.0. The adapted total score for the subscale assessing
client confidence in treatment (four items) therefore ranged from 3-12, with higher scores
indicating greater confidence.
An error was made in reconstructing the treatment engagement scale from previously published
studies, whereby two items used in the DATOS evaluation were omitted and substituted with
similar items that also formed part of the DATOS questionnaire. On the treatment commitment
subscale, the item “Anyone can talk about changing, but I’m actually doing something about it”
was substituted with “Sometimes I still find myself struggling with problems that I thought I had
taken care of once and for all”. On the treatment confidence subscale, the item “Maybe this
place will be able to help me” was substituted with “Overall, how helpful has this treatment
been?” The error was discovered partway through the study and corrected. Of those who were
still attending treatment 2 months after admission and completed the follow-up survey (n=112),
46 subjects (41.1%) were administered the incorrect items. Response substitution from the
initially used items was used to replace the missing data from the earlier surveys. A split-sample
comparison of subscale totals confirmed that the presence of the substituted item did not
significantly impact the mean level of client confidence (10.4 versus 10.2; t106=0.87, p=.387) or
client commitment (18.5 versus 18.0; t108=1.76, p=.081).
Substance Problem Scale (SPS):
The 16-item SPS assesses the experience of alcohol and drug use problems. It comprises part of
the Global Appraisal of Individual Needs (GAIN) standardized assessment instrument (Dennis,
62
White, Titus, & Unsicker, 2006). The SPS consists of four items that tap into symptoms of
substance abuse, seven that correspond with symptoms of dependence, and five that assess other
common problems related to substance use. Problems are collapsed across substances, such that
no substance-specific attributions of problems are made. Each item is scored on a 3-point scale
according to the recency of the last occurrence of the problem (i.e., in the past month, 2 to 12
months ago, over 1 year ago or never). The abuse and dependence items map onto DSM-IV
diagnostic criteria (American Psychiatric Association, 1994), such that the scale can be used to
differentiate between non-problematic and problematic use, abuse, and dependence, across
different time frames; although the lack of substance-specific attributions complicates diagnoses
among poly-drug users. Counts of problems and symptoms experienced over different time
frames can also be taken to yield dimensional measures of severity. The present study used the
count of problems experienced in the year prior to treatment as the baseline measure of
substance problem severity. In addition, change in the count of problems experienced in the
month prior to admission and follow-up was used in the final analysis. The SPS has
demonstrated reliability and validity among clients in substance use treatment (Dennis et al.,
2006) and has shown sensitivity to change over time (Godley, Godley, Dennis, Funk, & Passetti,
2002; Shane, Jasiukaitis, & Green, 2003).
3.4 Procedures The study involved two client surveys, completed at admission and 2-months later, as well as
review of clients’ clinical charts. All clients assessed for the core program or educational group
at PAARC were invited assessed for study eligibility. The only criterion for exclusion was an
inability to speak and understand English at a level sufficient to allow for completion of the
surveys. Prospective participants who were interested in the study but unable to read English
were offered the option of completing the surveys in an interview format. This option was
selected by a small number of subjects (n=5).
All study procedures were explained to clients by the researcher during the process of obtaining
written informed consent (see Appendix 2 for the consent letter). To minimize response bias,
participants were assured that their information would not be shared with program staff, except
at an aggregated level at the conclusion of the study and in cases requiring disclosure (i.e.,
threatened violence against self or others). Participants were offered $5 in cash following the
first survey and $10 in cash following the second (total = $15). Participant fees were paid by a
63
stipend from the Research Office at the Centre for Addiction and Mental Health (CAMH). The
study was approved by the Research Ethics Boards of the University of Toronto and CAMH.
The first survey was self-administered and included the measures of social pressures, perceived
coercion, and substance problem severity (completion time = 10 minutes). This survey was
completed onsite before or after the first or second counselling session or group meeting. The
measures of treatment motivation and other information, including sociodemographic
characteristics, problem substances, referral source, and legal problems, were abstracted from
clients’ clinical charts.
Approximately 2 months after admission, clients were contacted to complete a second survey
assessing past-month substance problem severity and, for those who were still accessing
services, treatment engagement (completion time = 20 minutes). Those who were not still
receiving services at PAARC or another treatment agency were asked to provide the reasons
underlying their decision to not pursue further treatment at that time. Of the original sample of
276 clients, 205 (74.3%) completed the follow-up survey (Figure 2, p.54). The majority
completed the survey by telephone (n=144, 52.2%); 60 subjects (21.8%) completed it as a self-
administered questionnaire at PAARC and one completed it electronically and emailed it to the
researcher. The interval between admission and completion of the second survey averaged 63
days, ranging from a minimum of 35 days to a maximum of 133 in one case.
Of the 71 subjects who were lost to follow up, approximately half (n=35) were reached by
telephone, but requested to complete the survey at another time and ultimately failed to connect
(Figure 2). Only 3 subjects refused to complete the second survey when approached. The
remainder were not successfully contacted for follow-up following several telephone messages
(n=25), or provided a telephone number that was incorrect or not in service at the time of
follow-up (n=8).
On the basis of subject report and agency records, 126 subjects (45.7%) were retained in
treatment at the time of follow-up. Where subject report was unavailable or not corroborated by
the agency record, the agency record was used. Those who indicated that they were only
accessing mutual help groups, such as Alcoholics or Narcotics Anonymous (AA/NA), were not
classified as retained. The 2-month point was selected as it represents a suitable time point for
assessing early treatment process at the study site (see Figure 2 for a timeline of treatment in
64
relation to study procedures). Although there is variation in the frequency of scheduled sessions,
2 months is a sufficiently long period of time to allow for completion of the assessment
procedures and initiation of the therapeutic relationship. Among clients whose primary
motivation was to obtain a letter of attendance, the required portion of their treatment will have
been completed within this length of time. Conversely, those who still have sessions booked at
the 2-month point are likely to have elected to continue with their sessions beyond the
assessment stage. This is an important consideration, as the validity of indicators of retention
and session attendance as outcome measures has been criticised when used in contexts where
treatment attendance is formally obligated (Simpson, 2004). Assessing retention past the point
of the mandated portion of treatment allows for a more meaningful consideration of the
associations between pressures, coercion, and treatment participation. Finally, although the
program does not have a designated length, it is unlikely that clients would have completed a
full course of treatment prior to the completion of the second month.
An attrition analysis was conducted to identify any variables that differentiate between clients
who completed the second survey versus those who were lost to follow-up (Appendix 3).
Several variables emerged as significant predictors of follow-up status. Being younger, having
less than a high school education, being mandated to treatment, having legal problems or being
referred by the legal system, and reporting fewer substance-related problems at admission all
increased the likelihood of attrition. With respect to the study variables, there were no
differences in perceived coercion or other types of social pressures by follow-up status.
However, those who were lost to follow-up reported greater legal pressure (p<.05) and
somewhat lower treatment motivation (p<.1). Those who were still attending treatment at the 2-
month point were more likely to complete the follow-up survey (n=112/126, 88.9%) than were
those who dropped-out of treatment (n=93/150, 62.0%), as subjects who were still attending
sessions at PAARC were more easily located for follow-up. As such, there is a potential attrition
bias in the analysis of change in substance-problem severity, due to the preferential follow-up of
those with less legal pressure and higher motivation, who remained in treatment for at least 2
months. However, this will not impact the analysis of retention, which was available for all
subjects from agency records (Figure 2, p.54). It is also unlikely to impact the analysis of
engagement, which was based only on those who were retained, among which 88.9% were
successfully followed up.
65
3.5 Conceptual framework Figure 3 depicts the SDT-based conceptual framework used to guide the analysis of study data.
The focal relationships of the study (Aneshensel, 2002; shaded in the diagram) involve the
relationships between client autonomy at admission, operationalized by perceived coercion and
treatment motivation, and early treatment processes, measured with retention and engagement.
Perceived coercion and treatment motivation are situated in the model as companion constructs,
negatively associated with each other and together offering a description of client decision-
making around entering a treatment program. They are influenced by pre-treatment events and
experiences, including social pressures and substance problem severity. Social pressures to enter
treatment, not limited to legal sources but encompassing a broader array of social agents, are
expected to increase perceptions of coercion and decrease autonomous treatment motivation
(H1), as they represent social events that are likely to be perceived, on average, as impacting
negatively on client autonomy. Substance problem severity is expected to decrease perceived
coercion and increase autonomous motivation for treatment (H2), inasmuch as clients who
acknowledge a greater number of symptoms and negative consequences related to their use of
substances will be less likely to view treatment as a coercive imposition.
Figure 3. Conceptual framework for an SDT-based analysis of addiction treatment process
Substance problem severity at admission
Perceived coercion and autonomous motivation, in turn, influence the treatment process. Both
objective and subjective indicators of treatment process are used in order to capture physical
presence and cognitive engagement. Lower perceived coercion and higher autonomous
Treatment process • Retention • Engagement
• Commitment • Confidence • Counsellor rapport
Change in substance problem severity
Client autonomy at treatment entry • Perceived coercion • Treatment motivation
Social pressures • Legal • Non-legal • Informal
66
motivation at admission is expected to increase the likelihood of being retained in treatment for
at least 2 months (H3) and, among those retained, increase the level of cognitive engagement in
treatment (H4). The impacts of social pressures and substance problem severity on retention and
engagement are transmitted through perceived coercion and autonomous motivation.
Specifically, those who experience social pressures as being coercive are expected to stay only
as long as required to satisfy the external conditions placed on their participation and to
demonstrate lower meaningful engagement in treatment. Finally, longer retention and greater
engagement are expected to lead to greater progress in resolving substance-related problems in
the short-term (H5). The impacts of perceived coercion and autonomous motivation are
transmitted through retention and engagement. Situating the focal relationships within this larger
theoretical framework, including antecedent and consequent variables, strengthens
interpretations that any empirical associations represent meaningful relationships, as opposed to
simply covariation (Aneshensel, 2002).
3.6 Analysis The conceptual framework formed the basis for the model-building strategy used in the analysis.
The analytic strategy is based on Anenshesel’s (2002) elaboration model, which emphasizes a
focal relationship that is of primary interest to the guiding theory. Following demonstration of
an empirical association, the analysis proceeds to consider causality through the systematic
addition of third variables (i.e., confounding, intervening, antecedent and consequent variables).
The framework outlined in Figure 3 was broken down into its component parts and
systematically evaluated using a series of nested multivariable regression models. In preparation
for the modeling analysis, bivariate tests were used to identify significant associations between
the dependent variables and other sociodemographic characteristics and treatment-related
variables. An alpha level of .05 was used to determine statistical significance in all analytical
procedures. Analyses were conducted in SPSS 15 and Stata 11.
With one exception, the number of subjects with missing data on study variables was low and
these cases were excluded from the analysis. Due to a lapse in the study protocol, 49 subjects
(17.8% of 276) did not complete the measure of treatment motivation that formed part of the
agency’s standardized assessment and the study’s admission assessment. In the face of staffing
shortages, remaining counsellors strove to shorten and adapt assessment procedures and, as a
result, the full battery of assessment tools was not administered in all cases. Following detection
67
of this change in the agency’s routine procedure, administration of the TEQ was ensured by the
researcher. For the modeling analysis, missing values for the RAI version of TEQ scores were
estimated using regression imputation. The imputation model included all other dependent and
independent variables involved in the modeling analysis, as well as additional variables
associated with motivation (i.e., gender, age and marital status). The missing values were
presumed to be missing at random with respect to the dependent variables, as they resulted from
agency-related rather than client-related factors. As such, in this case, imputation is not expected
to bias the results. A t-test confirmed that the difference between means of the observed and
imputed values was not significant (79.11 versus 83.96, t274=-1.45, p=.147) However, common
to regression imputation (Garson, 2008), the standard deviation of imputed values was
considerably lower than that of observed values (11.12 versus 22.77). Because this reduced
variance can contribute to an increased likelihood of detecting false associations (i.e., Type I
error), all significant findings involving treatment motivation were re-tested excluding the
imputed values.
First, a correlational analysis assessed the presence of hypothesized empirical associations
between the measures of admission process: social pressures, perceived coercion (MPCS),
treatment motivation (TEQ-RAI), and substance problem severity (SPS). A more detailed
analysis of perceived coercion was then conducted to determine the types and characteristics of
clients who felt forced to participate in a treatment program. MPCS scores were regressed on the
ratings of social pressures and problem severity, controlling for sociodemographic
characteristics and other treatment-related factors identified in the bivariate analysis (Model 1).
Given the paucity of empirical work evaluating the determinants and correlates of perceived
coercion, this part of the analysis plays a critical role in establishing a foundation for evaluating
coerced treatment. If the conceptual framework is accurate, the coefficients corresponding to
social pressures and substance problem severity should be significantly associated with
perceived coercion, even after controlling for other client characteristics.
A series of nested logistic regression models was then used to address the null hypothesis that
client autonomy at admission is unrelated to 2-month retention. In the first step, 2-month
retention (yes/no) was regressed on the focal independent variables: perceived coercion and
treatment motivation (Model 2A). Second, social pressures, substance problem severity, and any
other factors identified in the bivariate analysis were entered into the equation (Model 3A). The
68
significance of this block of variables was tested with a likelihood ratio test (Aneshensel, 2002).
Changes in the magnitude and significance level of the odds ratios for perceived coercion and
treatment motivation between Models 2A and 3A was also examined to ascertain the degree to
which their empirical associations with retention can be accounted for by the other independent
variables. If the conceptual framework is accurate, perceived coercion and treatment motivation
will exhibit significant associations with retention independent of social pressures and problem
severity, as the impact of the latter variables is expected to be transmitted through the measures
of autonomy (i.e., social pressures and problem severity are antecedent to the focal independent
variables in the causal chain). A third, reduced model containing all of the independent variables
except for treatment motivation and perceived coercion (i.e., the ratings of social pressures,
problem severity, and any other sociodemographic and treatment-related factors) was also tested
to evaluate the magnitude and significance of the unique association between the focal
independent variables and retention.
This model-building strategy was then repeated to evaluate the impact of treatment entry
experiences on engagement, using the three subscale scores for client confidence, counsellor
rapport, and commitment to treatment as dependent variables (Models 2B-D and 3B-D). This
analysis involved the subset of clients who were retained in treatment at the 2-month follow-up.
In the remaining clients who were not still attending treatment at the 2-month follow-up,
responses to the open-ended question in the follow-up survey of why they decided not to
continue with treatment were examined. Types of responses were examined by admission levels
of treatment motivation, perceived coercion, social pressures and substance problem severity.
The aim of this part of the analysis was to explore decision-making among those who failed to
engage in the treatment process, incorporating perceptions and experiences of the admission
process.
The final stage of the analysis evaluated the impact of the focal study variables on early
recovery from substance-related problems. The dependent variable was the change in substance
problem severity between admission and follow-up (i.e., count of problems experienced in the
past month at admission minus the count of past-month problems at follow-up). The change
score was first regressed on perceived coercion and treatment motivation (Model 4). This
allowed for a quantification of the full empirical association between the focal independent
variables and early progress in treatment. Next, social pressures and other variables identified in
69
the bivariate analysis were added to the equation (Model 5). As above, if the conceptual
framework is accurate, perceived coercion and treatment motivation will exhibit significant
associations with the change in problem severity independent of social pressures and other
control variables. Finally, the indicator for continued treatment attendance was added to the
model (Model 6). This measure was chosen to represent the early treatment process as, unlike
the measures of cognitive engagement, it does not require that the analysis be restricted to only
those who were retained in treatment. This model allowed for a broader examination of recovery
and maximized the sample size available for analysis. Being a factor more proximal to recovery,
retention is expected to attenuate the coefficients for perceived coercion and treatment
motivation (i.e., retention is expected to act as a mediator).
Regression diagnostics were performed to ensure the adequacy of the models. Prior to running
linear regression models, the fractional polynomial regression command in Stata was used to fit
polynomial terms for continuous independent variables and statistically evaluate their fit relative
to a linear term. Also for linear regression models, the absence of multicollinearity (i.e., high
multiple correlations between independent variables) was verified by calculating variance
inflation factors (VIF). A VIF of less than 10 is typically used to indicate the absence of
multicollinearity (Norman & Streiner, 2000). The absence of influential outliers was verified by
examining maximum absolute values of the standardized residuals and by calculating DfFit and
DfBeta statistics. Standardized residuals of greater than 3.0 (i.e., larger than 3 standard
deviations from the mean [0]) indicate outlying cases (Cohen & Cohen, 1983). Absolute values
of DfFit statistics (which correspond to the change in the predicted values of the dependent
variable when a given subject is excluded from the analysis) and DfBeta statistics (which
correspond to the change in regression coefficients that occurs when a given subject is excluded)
greater than 2.0 indicate the presence of potentially influential outliers (Garson, 2008). The
model was re-tested excluding outlying cases to determine their impact on results. Standardized
residuals were also plotted using histograms to check the assumption of normally-distributed
error terms. Shapiro-Wilk’s W test was used to offer a quantitative test for deviation of the
distribution of standardized residuals from normality.
For the logistic regression model, the Hosmer-Lemeshow Goodness of Fit (GOF) test, which
assesses the significance of differences between observed and expected frequencies of the
outcome denoted by the dependent variable, was used to assess overall fit of the model (Hosmer
70
& Lemeshow, 2000). The test statistic follows a chi-square (χ2) distribution, with nonsignificant
values indicating good fit between the observed data and model estimates. In addition,
standardized residuals were calculated to quantify differences between observed and model-
fitted values. Absolute values for standardized residuals of greater than 2.0 indicate the presence
of covariate patterns (i.e., specific combinations of the independent variables) that are poorly
predicted by the model (Garson, 2008). Finally, DBeta statistics were calculated to assess the
change in regression coefficients with the exclusion of a given case. Absolute values of greater
than 1.0 indicate the presence of potentially influential outliers. The model was re-tested
excluding outlying cases to determine their impact on results.
3.7 Sample size calculations Sample size calculations were conducted for multivariable linear and logistic regression models
consisting of eight independent variables. The models described above contain between four and
six independent variables; therefore, room is left for the addition of other factors identified as
significantly associated with the dependent variable through preliminary bivariate analysis. Both
calculations focus on the minimum number of subjects required to detect the significance of a
partial coefficient of primary interest in the study, using an alpha level of .05 and desired power
of 80%.
For linear regression, the calculation generated the minimum number of subjects required to test
the significance of a partial coefficient for an independent variable given that variable’s unique
contribution to the total amount of variance explained by the model (Cohen & Cohen, 1983).
This method takes into account the alpha level and desired level of power, the number of
variables in the model, the total amount of variance explained in the dependent variable (R2),
and the unique contribution of a given focal independent variable. For a model explaining an
estimated 30% of the variance in the dependent variable with eight independent variables, a
sample size of 192 yields an 80% probability of detecting a significant (alpha=.05) partial
coefficient for an independent variable that uniquely explains 3% of the total variance.
The calculation for logistic regression proceeded by a similar logic. The method takes into
account the probability of the outcome event (y=1) at the mean value of the given independent
variable and the desired effect size for detection, specified in terms of the odds ratio
corresponding to a single standard deviation increase in the independent variable. Sample size is
71
then adjusted to account for the variance that the independent variable is expected to share with
other covariates in the model (Hsieh, 1989). Given 50% retention at the mean level of the focal
independent variable (i.e., either perceived coercion or treatment motivation), and a multiple
correlation of .5 between these and the remaining covariates, an estimated 219 subjects are
required to detect an odds ratio of 1.5 associated with a single standard deviation increase (or,
alternatively, an odds ratio of 0.67 associated with a single standard deviation decrease) in the
focal independent variable.
Also relevant, the subsample used for each part of the analysis varied with the dependent
variable (Figure 2, p.54). The models for perceived coercion and 2-month retention utilized the
full sample, while the linear model predicting change in substance problem severity used only
those who completed the follow-up survey. In addition, the models predicting engagement used
only those who were retained at the 2-month follow-up. To account for these restrictions, while
maintaining feasibility within the timelines for the study, the target sample size was adjusted
upwards from 192 by approximately half to yield target sample size of 275.
72
Chapter 4 Results
4.1 Admission process Table 2 summarizes descriptive statistics for the scales administered at admission, including the
measures of treatment motivation (TEQ), perceived coercion (MPCS), social pressure ratings
from legal, other (non-legal) formal, and informal sources, and substance problem severity
(SPS). Four summary scores are presented for the TEQ, corresponding to the three subscales for
identified, introjected, and external motivation, as well as their weighted combination with the
Relative Autonomy Index (RAI). On average, clients in this sample reported a high degree of
identified motivation for entering treatment. Despite covering the full range of potential values,
the distribution was negatively skewed, with 45.4% of clients (n=103) scoring the highest
possible value of 28. The distribution for introjected motivation scores was bimodal, with spikes
of 22.9% (n=52) at the lowest score of 3 and 11.0% (n=25) at the highest score of 21.
Conversely, the distribution for external motivation was positively skewed, with a concentration
of values (20.3%, n=46) at the lowest score of 4. The RAI presented a more balanced
representation of autonomous motivation, with reduced skew and a more centred mean.
The majority of clients (57.1%, n=153/268) endorsed one or more of the five MPCS items.
Scores on the MPCS covered the full potential range of the scale, although the distribution
showed a concentration of values at the low end of the scale followed by a smaller increase at
the high end (Figure 4). Previous studies have likewise found MPCS scores to be positively
skewed or even bimodally distributed with concentrations of respondents at either end of the
range of possible values (Bindman et al., 2005; Gardner et al., 1993; Hiday et al., 1997; Rain,
Williams et al., 2003; Swartz et al., 2002).
The vast majority of clients (87.3%, n=241) reported one or more sources of pressures to enter
treatment. Across the three sources, informal pressures (i.e., from friends and family) were most
common, reported by 64.5% of the sample (n=178). Just under half of clients (45.3%, n=125)
reported pressure from the legal system and 40.6% (n=112) reported pressures from other
formal sources, such as employers or coworkers, school authorities, or health or social service
73
providers. Overall, 28.6% (n=79) reported pressures from one source only, 21.0% (n=58)
reported two sources, and 37.7% (n=104) reported three or more. The ratings for degree of legal
pressure showed a bimodal distribution, with over three-quarters of clients indicating either no
pressure (54.7%, n=151) or maximum pressure (24.3%, n=67; Figure 5). Non-legal formal
pressures, on the other hand, showed a large spike at no pressure (59.4%, n=164), followed by
smaller proportions (8-13%) at each subsequent level (Figure 6). The distribution for informal
pressures was similar to that of other formal pressures (Figure 7).
Finally, on average, clients in this sample reported experiencing a high number of substance-
related problems in the year prior to entering treatment (median = 12 per client). A minority
(7.6%, n=21) reported having experienced no substance-related problems during this time.
Observed distributions for these admission process variables all exhibited highly significant
deviations from normality, as indicated by significant Shapiro-Wilk tests (W statistics ranged
from .704 to .946, all p’s<.001). As such, non-parametric tests were used throughout the
analysis to evaluate associations between study variables. Corresponding parametric analyses
are summarized in Appendix 4 and any discrepancies are highlighted in the text.
Table 2. Descriptive statistics for admission process measures
Scales n Mean ± SD Median Min-Max Treatment Motivation (TEQ):
Identified a 227 22.34 ± 8.38 27.0 4-28 Introjected a 227 10.91 ± 6.28 10.0 3-21 External a 227 11.83 ± 6.64 11.0 4-28 Relative Autonomy Index (RAI) b 227 79.11 ± 22.77 85.0 30-114
Perceived coercion (MPCS) 268 1.38 ± 1.61 1.0 0-5 Social Pressures:
Legal 276 1.43 ± 1.72 0.0 0-4 Other formal 276 0.97 ± 1.40 0.0 0-4 Informal 276 1.71 ± 1.55 2.0 0-4
Substance problem severity (SPS-12m) 276 9.79 ± 4.99 12.0 0-16
Figure 4.
74
Perceived coercion (n=268)
42.9%
19.8%17.2%
5.6% 6.0%8.6%
0.0%
10.0%
20.0%
30.0%
40.0%
50.0%
0 1 2 3 4 5
MPCS score
Figure 5. Legal pressure (n=276)
54.7%
4.0%
9.4%7.6%
24.3%
0.0%
10.0%
20.0%
30.0%
40.0%
50.0%
60.0%
0 1 2 3 4
Legal pressure
75
Figure 6. Other formal (non-legal) pressure (n=276)
59.4%
13.4%
8.7% 8.0%10.5%
0.0%
10.0%
20.0%
30.0%
40.0%
50.0%
60.0%
0 1 2 3 4
Other formal pressure
Figure 7. Informal pressure (n=276)
35.5%
11.6%
18.8%
14.5%
19.6%
0.0%
10.0%
20.0%
30.0%
40.0%
50.0%
60.0%
0 1 2 3 4
Informal pressure
76
Bivariate associations between admission process variables were examined using Spearman
rank correlation tests (Table 3). As expected, greater autonomous motivation for treatment,
represented with the RAI, was inversely correlated with perceived coercion. Not shown in the
table, all three subscales of the TEQ were individually associated with perceived coercion in the
expected direction. That is, perceived coercion was strongly inversely correlated with identified
motivation (ρ=-.652, p<.001), weakly inversely correlated with introjected motivation (ρ=-.319,
p<.001), and positively correlated with external motivation (ρ=.397, p<.001).
Table 3. Spearman rank correlations (ρ) between admission variables
TEQ-RAI MPCS SP-L SP-OF SP-I SPS TEQ-RAI 1.000 MPCS -.549** 1.000 SP-L -.385** .505** 1.000 SP-OF .096 -.070 -.090 1.000 SP-I .015 -.019 -.120* .337** 1.000 SPS .281** -.218** -.203** .269** .443** 1.000 TEQ-RAI: Treatment Entry Questionnaire – Relative Autonomy Index MPCS: MacArthur Perceived Coercion Scale SP-L: Social Pressures – Legal SP-OF: Social Pressures – Other Formal SP-I: Social Pressures – Informal SPS: Substance Problem Severity * p<.05; ** p<.01
Legal pressures to enter treatment were inversely correlated with autonomous motivation and
positively correlated with perceived coercion. Conversely, neither of these variables was
correlated with pressures from other formal or informal sources. Interestingly, ratings for other
formal and informal pressures were positively associated with each other, and both showed
weak inverse correlations with legal pressures. That is, higher ratings of pressure from the legal
system did not correspond to higher ratings of pressures from other sources. Finally, higher
substance problem severity at admission was associated with greater autonomous motivation
and lower perceived coercion, as expected. Greater problem severity was also associated with
lower pressures from legal sources, but greater pressures from other formal and informal
sources. Results were equivalent with parametric correlation tests (Table A4-1, Appendix 4).
A more in-depth analysis of perceived coercion was then conducted to determine the types and
characteristics of clients who feel forced to participate in treatment. First, a series of bivariate
tests was conducted of the associations between perceived coercion and sociodemographic
characteristics and other treatment-related factors (Table 4). Non-parametric Mann-Whitney and
77
Kruskal Wallis tests were used because of the asymmetric distribution of MPCS scores. Results
were identical for a corresponding parametric analysis (Table A4-2, Appendix 4).
Table 4. Client characteristics by perceived coercion at admission (n=268)
Client characteristics Mean ± SD Median Statistic Gender:
Male 1.55 ± 1.68 1.0 U=4476.0, p=.001 Female 0.78 ± 1.20 0.0
Marital status: Married 1.42 ± 1.62 1.0 χ2
2=0.31, p=.858 Single 1.34 ± 1.62 1.0 Widowed, separated, divorced 1.38 ± 1.65 1.0
Education: Less than secondary school 1.27 ± 1.51 1.0 χ2
3=1.06, p=.787 Secondary school completed 1.48 ± 1.58 1.0 Some post-secondary 1.35 ± 1.74 1.0 Post-secondary completed 1.43 ± 1.69 1.0
Employment: Employed 1.42 ± 1.66 1.0 χ2
2=0.35, p=.839 Not employed 1.21 ± 1.50 1.0 Not in the labour force or on leave 1.30 ± 1.57 1.0
Previous treatment: No 1.63 ± 1.77 1.0 U=6356.5, p=.002 Yes 0.89 ± 1.16 0.0
Treatment mandate: No 0.74 ± 1.12 0.0 U=3595.5, p<.001 Yes 2.42 ± 1.74 2.0
Legal problems: No 0.65 ± 0.96 0.0 U=4782.0, p<.001 Yes 2.04 ± 1.79 2.0
Referral source: Addiction service 0.52 ± 0.83 0.0 χ2
5=74.46, p<.001 Family or friends 1.25 ± 1.34 1.0 Legal system 2.55 ± 1.72 2.0 Mental or general health services 0.76 ± 1.31 0.0 Workplace, school, or social services 1.24 ± 1.52 1.0 Self 0.53 ± 0.86 0.0
Problem substances: None 2.89 ± 1.66 3.0 χ2
3=68.72, p<.001 Alcohol only 1.03 ± 1.37 0.0 Drugs only 0.86 ± 1.14 0.0 Both alcohol and drugs 0.76 ± 1.28 0.0
Of the sociodemographic characteristics examined, only gender was significantly associated
with perceived coercion, with men reporting higher coercion than women. In addition to the
variables presented in Table 4, there was no significant association between perceived coercion
and age (ρ=-.082, p=.181). Conversely, all of the other treatment-related variables, including
treatment mandates, legal problems, referral source, and type of problem substances, were
significantly associated with perceived coercion. Specifically, those with no previous history of
treatment, as well as those with current mandates and/or legal problems, reported greater
coercion to enter treatment. Those referred by the courts reported higher levels of coercion than
78
others, particularly compared to those referred by other addiction, mental or general health
services and those who were self-referred. In addition, reporting any type of problem substance
was strongly associated with lower perceived coercion, relative to not acknowledging problem
substances at admission. However, there were no differences in perceived coercion between
those who reported alcohol, other drugs, or both as problem substances. This was confirmed by
a second Kruskal-Wallis test excluding those without presenting problem substances (χ22=2.28,
p=.320).
Next, multiple regression was used to predict perceived coercion. The primary purpose of the
model was to assess the independent and relative influences of social pressures and substance
problem severity on perceived coercion, in accordance with the conceptual framework (Figure
3). A preliminary check revealed the lack of significantly improved model fit with polynomial
terms for the continuous independent variables. As indicated in the bivariate analysis, the model
also controlled for gender (reference category = female) and previous treatment experience
(reference category = none). The final sample size for the model was 266, excluding 10 subjects
with missing values on model variables.
Despite being significantly associated with perceived coercion at bivariate level, the chart-
recorded variables representing treatment mandates, legal problems, and referral source are not
entered into the model as control variables. Conceptually, client ratings of social pressures are
used in place of these variables, such that both are not expected to be needed (i.e., they are not
expected to account for a significant amount of unique covariance with the dependent variable).
Bivariate non-parametric comparisons with ratings of legal pressure confirmed the overlap
between these variables (mandates: U=2687.0, p<.001; legal problems: U=2422.0, p<.001; and
referral source: χ25=119.81, p<.001). Finally, the variable denoting problem substance types was
not added to the model as the association between it and perceived coercion appeared to be due
to the category of clients who did not acknowledge substance-related problems at admission.
This overlaps conceptually and empirically with clients’ scores on the SPS, the dimensional
measure of past-year experience of substance-related problems, which forms part of the
conceptual framework guiding the analysis. Therefore, to prevent problems resulting from
multicollinearity between independent variables, and to preserve the conceptual integrity of the
model, the regression equation predicting retention did not control for treatment mandates, legal
problems, referral source, or type of problem substance.
79
Table 5 displays the results of the multiple linear regression predicting perceived coercion. The
model F-test was highly significant, indicating that the model explained a significant amount of
variance in perceived coercion. Together, the independent variables predicted approximately
31% of the variance in perceived coercion. Of the variables representing social pressures to
enter treatment, only legal pressure was independently associated with perceived coercion. For
each single-point increase in the rating of legal pressure, there was an increase of just under half
a point along the scale of perceived coercion. The inverse association between perceived
coercion and past-year problem severity also persisted in the multivariable model. There was a
trend toward lower perceived coercion among clients with previous treatment experience,
although this fell short of significance (p<.10). In accordance with the conceptual framework,
then, perceived coercion was significantly influenced by social pressures, in the form of legal
pressures, and substance problem severity after adjusting for other client characteristics.
Table 5. Linear regression predicting perceived coercion (n=266)
Model 1 F6,259=19.62, p<.001 R2=.313
Independent variables
b SE p Social pressures:
Legal 0.431 0.051 <.001 Other formal -0.056 0.064 .381 Informal 0.083 0.063 .186
Substance problem severity (SPS-12m) -0.050 0.020 .013 Previous treatment a -0.328 0.184 .075 Gender b 0.297 0.207 .153
a reference category = none b reference category = female
Regression diagnostics revealed an absence of multicollinearity among independent variables,
and low DfBeta and DfFit values suggested the absence of influential outliers (Appendix 5).
However, one of the standardized residuals exceeded 3.0 (value = 3.50), and the overall
distribution was significantly non-normal. Because of the restricted range of MPCS scores, the
model was re-tested using ordinal logistic regression. Similar results were obtained, with one
exception: the coefficient for informal pressures was significant. Specifically, greater informal
pressures to enter treatment was associated with greater perceived coercion (OR=1.28,
se=0.027, p=.003).
80
4.2 Early treatment process The next stage of the analysis involved a detailed examination of early treatment processes, with
indicators covering both objective participation and cognitive involvement on the part of clients,
as well as early recovery from substance-related problems.
4.2.1 Two-month retention
As noted earlier, 45.7% of subjects (n=126) were still attending treatment 2 months following
admission. Of those who were admitted, 24.6% (n=68) attended only the one appointment. An
additional 28.3% (n=78) attended only two appointments, while the remainder (47.1%, n=130)
attended three or more during the first 2 months. The maximum number of sessions attended
during this time was seven (mean = 2.6, median = 2.0). Among those who left treatment within
the first 2 months, the majority (66%, n=99) attended only one or two appointments in total;
notably, the median number of appointments in this subgroup was 2.0, compared to 3.0 among
those who were retained beyond 2 months.
A series of tests was conducted to evaluate the bivariate associations between 2-month retention
and the admission process measures, sociodemographic characteristics, and other treatment-
related factors (Table 6). The results of non-parametric Mann-Whitney U tests are presented for
associations with admission process variables, although a series of t-tests yielded the same
results (Table A4-3, Appendix 4). T-tests and chi-square tests are used to evaluate the
associations with the other continuous and categorical client characteristics, respectively.
As hypothesized, mean levels of autonomous treatment motivation were higher among those
who were retained. In addition, perceived coercion was significantly higher among those who
dropped out. Specifically, of the 23 subjects who reported the highest level of coercion at
admission (MPCS=5.0), only 5 (21.7%) were still attending treatment 2 months later. Ratings of
legal pressure were similarly inversely associated with retention, with those who were retained
reporting significantly lower levels of such pressure at admission. Conversely, both other formal
pressure and informal pressure ratings were higher among those who remained in treatment. The
association between informal pressures and retention fell just short of statistical significance
(p=.052). Finally, those who were retained scored higher on the SPS at admission, endorsing a
greater number of substance-related problems in the year prior to entering treatment.
81
Table 6. Admission process and client characteristics by 2-month retention (n=276)
2-month retention Admission process and client characteristics No
(n=150) Yes
(n=126) Statistic
Treatment motivation (TEQ-RAI) 74.99 ± 23.39 83.73 ± 21.23 U=4972.0, p=.003 Perceived coercion (MPCS) 1.71 ± 1.72 0.98 ± 1.37 U=6650.0, p<.001 Social pressures:
Legal 1.74 ± 1.78 1.06 ± 1.57 U=7523.5, p=.001 Other formal 0.77 ± 1.28 1.21 ± 1.49 U=7712.5, p=.003 Informal 1.55 ± 1.50 1.90 ± 1.58 U=8208.0, p=.052
Substance problem severity (SPS-12m) 8.91 ± 5.34 10.84 ± 4.31 U=7610.0, p=.005 Gender:
Male 56.0% (121) 44.0% (95) χ21=1.11, p=.290
Female 48.3% (29) 51.7% (31) Age 35.45 ± 10.48 37.29 ± 10.62 t274=-1.44, p=.150 Marital status:
Married 47.2% (58) 52.8% (65) χ22=4.85, p=.089
Single 57.0% (61) 43.0% (46) Widowed, separated, divorced 65.1% (28) 34.9% (15)
Education: Less than secondary school 54.4% (37) 45.6% (31) χ2
3=4.25, p=.235 Secondary school completed 58.1% (43) 41.9% (31) Some post-secondary 39.5% (15) 60.5% (23) Post-secondary completed 58.0% (47) 42.0% (34)
Employment: Employed 56.4% (101) 43.6% (78) χ2
2=2.22, p=.330 Not employed 52.5% (31) 47.5% (28) Not in the labour force or on leave 42.9% (15) 57.1% (20)
Previous treatment: No 61.9% (109) 38.1% (67) χ2
1=11.89, p=.001 Yes 40.2% (39) 59.8% (58)
Treatment mandate: No 41.8% (71) 58.2% (99) χ2
1=28.25, p>.001 Yes 74.5% (79) 25.5% (27)
Legal problems: No 43.1% (56) 56.9% (74) χ2
1=13.08, p>.001 Yes 64.8% (94) 35.2% (51)
Referral source: Addiction service 44.8% (13) 55.2% (16) χ2
5=29.19, p>.001 Family or friends 56.3% (18) 43.8% (14) Legal system 75.0% (69) 25.0% (23) Mental or general health services 46.3% (31) 53.7% (36) Workplace, school, or social services 35.3% (6) 64.7% (11) Self 29.0% (9) 71.0% (22)
Problem substances: None 85.9% (55) 14.1% (9) χ2
3=33.73, p>.001 Alcohol only 45.8% (44) 54.2% (52) Drugs only 42.4% (28) 57.6% (38) Both alcohol and drugs 46.0% (23) 54.0% (27)
Of the sociodemographic characteristics examined, none were significantly associated with
retention. There was a trend toward a greater likelihood of retention among married individuals,
relative to those who were single and, particularly, widowed, separated or divorced (p<.10). In
contrast, other treatment-related variables, including previous treatment experience, current
mandates, legal problems, referral source, and type of substance problems, were all highly
significantly associated with retention. Having previous treatment experience and being self-
82
referred were associated with a greater likelihood of 2-month retention, while treatment
mandates, legal problems, and legal system referrals were associated with a lower likelihood of
retention. Reporting alcohol, other drugs, or both as problem substances at admission was
associated with a greater likelihood of staying in treatment, relative to not acknowledging
problem substances. A chi-square test of the association between type of problem substance and
retention excluding those who denied current problem substances revealed no difference in
retention across substance types (χ22=0.22, p=.895).
Next, multiple logistic regression was used to evaluate predictors of 2-month retention. In
addition to the focal independent variables (treatment motivation and perceived coercion) and
antecedent variables (social pressures and past-year problem severity), the analysis controlled
for previous treatment experience (reference group = none). As before, the variables
representing mandates, legal problems, referral source, and type of problem substance are not
included in the model to preserve conceptual integrity and prevent problems resulting from
multicollinearity. Preliminary curve estimation for the continuous independent variables
revealed that no non-linear terms improved the fit of the model. The final sample size was 266,
excluding 10 subjects missing values on the model variables. The sample is invariant across
steps to allow for nested model comparisons.
Results are summarized in Table 7. In the first step (Model 2A), 2-month retention was
regressed on the two focal independent variables alone. The model chi-square test indicates that
the overall model is significant in predicting retention. However, of the two variables, only
perceived coercion is significant. Specifically, for each single-point increase in the scale
assessing perceived coercion, there is a 20% reduction in the odds of being retained at the 2-
month follow-up. By contrast, the odds ratio for treatment motivation was not significantly
different from 1.00. In a simple model containing only treatment motivation, the odds ratio for
motivation was only 1.02, but was statistically significant (p=.001). In other words, on its own, a
10-point increase on the RAI is associated with a 20% increase in the odds of retention.
However, adjusting for perceived coercion, this increase in treatment motivation is associated
with only a 10% increase in the odds of retention and is no longer significant. The magnitude of
the impact of treatment motivation, therefore, is small and its shared covariance with perceived
coercion is large enough to render it nonsignificant in the three-variable model. With respect to
83
the conceptual framework, only the expected empirical association between perceived coercion
and retention was demonstrated with these data.
With the addition of social pressures, problem severity, and other client characteristics, the odds
ratio for perceived coercion was decreased by 11.0% and failed to maintain statistical
significance (Model 3A). That is, the empirical association found between perceived coercion
and retention could be attributed to the shared covariation with other model variables. This was
confirmed by a likelihood ratio test comparing the full model (Model 3A) with a reduced
version that excludes treatment motivation and perceived coercion. The nonsignificant test
(χ22=4.69, p>.05) indicates that the focal independent variables do not explain a significant
amount of unique variance in retention.
Table 7. Logistic regression predicting 2-month retention (n=266)
Model 2A χ2
2=17.60, p<.001 logL=-174.30
Model 3A χ2
7=30.58, p<.001 logL=-167.81
Reduced model A χ2
5=25.89, p<.001 logL=-167.81
Independent variables
OR SE p OR SE p OR SE p Treatment motivation (TEQ-RAI) 1.01 0.008 .095 1.01 0.008 .219 ------ ------ ------ Perceived coercion (MPCS) 0.80 0.081 .028 0.89 0.099 .279 ------ ------ ------ Social pressures:
Legal ------ ------ ------ 0.89 0.080 .183 0.81 0.064 .008 Other formal ------ ------ ------ 1.16 0.114 .130 1.17 0.114 .101 Informal ------ ------ ------ 1.01 0.101 .893 0.98 0.094 .795
Substance problem severity (SPS-12m) ------ ------ ------ 1.03 0.034 .394 1.05 0.033 .138 Previous treatment a ------ ------ ------ 1.73 0.490 .053 1.76 0.492 .043
a reference category = none
Added together as a block, the other independent variables made a significant contribution to the
overall explanatory power of the model (likelihood ratio test: χ25=12.98, p<.05). However, none
of the variables independently predicted retention. Having attended treatment previously almost
doubled a client’s odds of still being in treatment after 2 months, and this association just failed
to reach statistical significance. The association between legal pressure and retention, significant
in the reduced model, was not significant in the full model containing the focal independent
variables.
Overall, fit of the full model is good, as indicated by a nonsignificant Hosmer-Lemeshow GOF
test (χ28=10.96, p=.204), and the residual analysis suggested an absence of influential outliers
(maximum |DBeta| = 0.19). Two values of standardized residuals exceeded 2.0 (both values =
84
2.17). Excluding these cases from the analysis did not impact on the magnitude of the odds
ratios, however, suggesting that they were not influencing the results.
With the simultaneous addition of multiple independent variables to a regression equation, the
impact of any single variable on the focal relationship can not be ascertained (Aneshensel,
2002). Therefore, to further explicate the focal relationship and obtain a clearer sense of how
each independent variable affects it, an additional set of reduced, three-variable models were
tested in which each variable was added separately to base models containing only perceived
coercion (Table 8). Because treatment motivation was not significant in the above modeling
analysis, it was not assessed at this stage. Percent changes in the odds ratio for perceived
coercion, as well as changes in statistical significance, were evaluated relative to a model
containing only perceived coercion (first row in Table 8).
Table 8. Reduced three-variable models predicting 2-month retention (n=266)
Independent variables OR SE p %∆ ORMPCS a
Perceived coercion (MPCS) 0.74 0.063 <.001 ----- Perceived coercion (MPCS) 0.79 0.077 .017 6.8% Legal pressures 0.86 0.075 .082 Perceived coercion (MPCS) 0.74 0.064 <.001 ----- Other formal pressures 1.22 0.112 .029 Perceived coercion (MPCS) 0.73 0.063 .001 1.4% Informal pressures 1.13 0.093 .152 Perceived coercion (MPCS) 0.76 0.067 .002 2.7% Substance problem severity (SPS-12m) 1.07 0.029 .022 Perceived coercion (MPCS) 0.76 0.067 .002 2.7% Previous treatment b 1.97 0.531 .012
a refers to the percent change in the OR for perceived coercion; %∆ = (OR2 – OR1)/OR1, where OR1=0.74 b reference category = none
The addition of legal pressure ratings to the bivariate model resulted in the greatest reduction in
the odds ratio for perceived coercion, which nonetheless retained its statistical significance in
the three-variable model. In fact, assessed separately, no single independent variable accounted
entirely for the empirical association between perceived coercion and retention. In addition to
legal pressure, substance problem severity, previous treatment experience, and informal
pressures each accounted for small proportions of the focal association. In the end, the addition
of these three variables together fully accounted for the empirical association observed between
perceived coercion and retention.
85
In summary, greater perceived coercion and lower treatment motivation were associated with
lower odds of 2-month retention, and this appeared to be accounted for by legal pressures to
enter treatment, combined with the client’s history of treatment, and problem severity.
4.2.2 Treatment engagement
As noted earlier, cognitive engagement in treatment was assessed in the subset of clients who
were both retained at the 2-month point and successfully reached for the second survey (n=112).
Table 9 displays descriptive statistics for the three subscale scores. As would be expected, the
subscales were positively associated with each other: confidence-rapport ρ=.545, p<.001;
confidence-commitment ρ=.284, p=.003; rapport-commitment ρ=.191, p=.051. However, all
three demonstrated pronounced ceiling effects, with high negative skew. Although the values for
confidence in treatment potentially range from 3-12, observed values in this sample ranged only
from 7-12, with a median of 10. Over one-third of subjects (35.2%, n=38) received the highest
possible score of 12. On the scale assessing counsellor rapport, although observed values
covered the full potential range of the scale, the median was equivalent to the highest possible
score. Specifically, 54.2% of clients (n=58) received a score of 15. Scores for commitment to
treatment covered an especially restricted range of values. Although they can potentially range
from 5-20, observed scores ranged only from 14-20. A total of 28 clients (25.5%) obtained the
highest value. On all three subscales, scores exhibited limited variability across clients, with
standard deviation values of less than 2 points along the scales.
Table 9. Descriptive statistics for measures of 2-month treatment engagement (n=112)
Subscales Mean ± SD Median Min-Max Confidence in treatment 10.31 ± 1.59 10.0 7-12 Counsellor rapport 14.07 ± 1.61 15.0 5-15 Commitment to treatment 18.27 ± 1.45 18.0 14-20
As with retention, a series of non-parametric tests was conducted to evaluate the bivariate
associations between engagement and admission process variables, sociodemographic
characteristics, and other treatment-related factors. Table 10 summarizes the Spearman rank
correlations between engagement subscales and admission process variables. Treatment
motivation showed a small but significant positive correlation with confidence in treatment and
trends toward positive correlations with the other two subscales (p≤.10). The only other
significant association was an inverse correlation between perceived coercion and commitment
86
to treatment. Parametric Pearson correlations between these variables were similar, with the
exception that the association between perceived coercion and commitment to treatment was not
significant (r=-.153, p=.118; Table A4-4, Appendix 4).
Table 10. Admission process measures by 2-month engagement in treatment (n=112)
Confidence in treatment
Counsellor rapport Commitment to treatment
Admission scales
ρ p ρ p ρ p Treatment motivation (TEQ-RAI) .262 .012 .173 .101 .197 .060 Perceived coercion (MPCS) -.158 .110 -.046 .643 -.216 .026 Social pressures:
Legal .072 .459 .043 .657 -.098 .307 Other formal .043 .662 .188 .052 .005 .956 Informal -.075 .438 -.076 .437 .011 .909
Substance problem severity (SPS-12m) .043 .660 .010 .922 -.117 .224
With one exception, none of the sociodemographic characteristics or other treatment-related
factors was associated with the measures of engagement (Table 11). This is perhaps not
surprising, as the low variability in engagement scores means that the analysis is considering
only very small differences across clients grouped by characteristics or other factors. In addition
to the variables summarized in Table 11, Spearman correlation tests between engagement and
age were likewise nonsignificant (confidence ρ=.160, p=.099; rapport ρ=.118, p=.227;
commitment ρ=-.023, p=.813). The one exception to the pattern of non-significance is the
bivariate association between confidence and referral source. Specifically, those who were self-
referred reported the highest levels of confidence in treatment relative to others. The differences
across these client groups were not, however, large. Results from the corresponding parametric
tests were identical (Table A4-5, Appendix 4).
Next, multivariable regression was used to predict 2-month engagement. The same model-
building strategy used above for retention was used with each of the subscales assessing client
confidence in treatment (Table 12), counsellor rapport (Table 13), and client commitment to
treatment (Table 14), respectively. Although not significant in the bivariate analysis, the ratings
for social pressures and problem severity were entered as control variables because of their
conceptual relevance. A small number of clients was excluded from each model because of
missing values on one or more of the independent variables. The final sample size varied
slightly across models, from 108 for confidence in treatment, to 107 for counsellor rapport, and
110 for commitment to treatment. For each dependent variable, the sample is invariant across
87
models to allow for direct comparisons of explanatory power. Preliminary curve estimation for
the continuous independent variables revealed that no non-linear terms improved model fit.
Table 11. Client characteristics by 2-month treatment engagement (n=112)
Confidence in treatment
Counsellor rapport
Commitment to treatment
Client characteristics Mean ± SD Statistic Mean ± SD Statistic Mean ± SD Statistic Gender:
Male 10.3 ± 1.5 U=1121.0 14.0 ± 1.8 U=1092.5 18.3 ± 1.5 U=1095.0 Female 10.3 ± 1.8 p=.862 14.2 ± 1.0 p=.916 18.1 ± 1.5 p=.471
Marital status: Married 10.5 ± 1.5 χ2
2=1.76 14.3 ± 1.5 χ22=3.54 18.2 ± 1.6 χ2
2=0.27 Single 10.1 ± 1.6 p=.415 13.9 ± 1.6 p=.171 18.4 ± 1.3 p=.874 Widowed/separated/divorced 10.0 ± 2.0 13.7 ± 1.8 18.4 ± 1.6
Education: Less than secondary school 10.5 ± 1.5 χ2
3=0.73 13.7 ± 2.3 χ23=3.57 18.5 ± 1.3 χ2
3=1.85 Secondary school completed 10.2 ± 1.6 p=.866 13.8 ± 1.8 p=.312 18.4 ± 1.3 p=.605 Some post-secondary 10.3 ± 1.7 14.3 ± 1.2 17.9 ± 1.6 Post-secondary completed 10.1 ± 1.6 14.4 ± 0.9 18.3 ± 1.5
Employment: Employed 10.2 ± 1.6 χ2
2=1.45 14.2 ± 1.2 χ22=1.04 18.1 ± 1.5 χ2
2=3.05 Not employed 10.7 ± 1.5 p=.484 13.8 ± 2.7 p=.594 18.5 ± 1.3 p=.217 Not in labour force, on leave 10.4 ± 1.5 14.1 ± 1.0 18.7 ± 1.4
Previous treatment: No 10.3 ± 1.5 U=1445.5 13.9 ± 2.0 U=1340.5 18.3 ± 1.4 U=1484.0 Yes 10.3 ± 1.7 p=.960 14.3 ± 0.9 p=.579 18.3 ± 1.5 p=.900
Treatment mandate: No 10.3 ± 1.6 U=873.5 14.1 ± 1.7 U=924.0 18.3 ± 1.4 U=960.5 Yes 10.3 ± 1.5 p=.750 14.1 ± 1.3 p=.926 18.2 ± 1.6 p=.954
Legal problems: No 10.2 ± 1.6 U=1194.5 14.1 ± 1.6 U=1279.5 18.3 ± 1.4 U=1416.0 Yes 10.5 ± 1.5 p=.265 14.1 ± 1.7 p=.596 18.3 ± 1.5 p=.985
Referral source: Addiction service 10.3 ± 1.5 χ2
5=11.45 14.1 ± 1.3 χ25=9.56 18.6 ± 1.1 χ2
5=1.82 Family or friends 9.8 ± 2.0 p=.043 13.8 ± 0.9 p=.089 17.9 ± 1.4 p=.874 Legal system 10.3 ± 1.6 14.1 ± 1.4 18.4 ± 1.5 Mental/general health service 9.9 ± 1.7 13.7 ± 2.3 18.3 ± 1.4 Workplace/school/CAS 10.9 ± 1.2 14.6 ± 0.7 18.4 ± 2.0 Self 11.3 ± 1.0 14.7 ± 0.7 18.1 ± 1.7
Problem substances: None 9.7 ± 1.6 χ2
3=1.29 14.1 ± 1.5 χ23=1.00 18.3 ± 1.1 χ2
3=3.31 Alcohol only 10.4 ± 1.6 p=.732 14.2 ± 1.3 p=.802 18.1 ± 1.6 p=.347 Drugs only 10.3 ± 1.6 13.7 ± 2.4 18.1 ± 1.4 Both alcohol and drugs 10.3 ± 1.5 14.2 ± 0.9 18.7 ± 1.3
In the first step, the three engagement subscale scores were regressed on the focal independent
variables (Models 2B-2D). The model F-test was significant only for the equation predicting
confidence in treatment. Greater autonomous motivation at admission predicted higher client
confidence at 2 months (Model 2B, Table 12). The effect was small: for each 10-point increase
on the RAI, there was a 0.3-point increase in level of confidence in treatment. Neither perceived
coercion nor treatment motivation was significantly associated with either counsellor rapport or
commitment to treatment. Therefore, contrary to expectations, perceived coercion and treatment
88
motivation were not consistently associated with indicators of early cognitive engagement in the
treatment process.
With the addition of social pressures and problem severity in the second step (Models 3B-3D),
the association between treatment motivation and confidence was unchanged. Results were
identical when the model was re-tested excluding imputed values of the TEQ (n=87), suggesting
that this small association was not attributable to the reduced variance resulting from regression
imputation.
Added together as a block, the control variables did not make a significant contribution to the
overall explanatory power of any of the three models (confidence: F4,97=1.01, p>.05; rapport:
F4,96=1.26, p>.05; commitment: F4,99=0.58, p>.05), and none of the other model variables was
associated with engagement. Together, the full models predicted approximately 14% of the
variance in client confidence, 5% of the variance in counsellor rapport, and 7% of the variance
in client commitment to the treatment process. Only the first full model, predicting client
confidence, was significant.
Table 12. Linear regression predicting client confidence in treatment (n=104)
Model 2B F2,101=5.86, p=.004 R2=.104
Model 3B F6,97=2.63, p=.021 R2=.140
Independent variables
b SE p b SE p Treatment motivation (TEQ-RAI) 0.030 0.009 .002 0.031 0.010 .002 Perceived coercion (MPCS) 0.035 0.123 .778 -0.028 0.128 .824 Social pressures:
Legal ------ ------ ------ 0.140 0.106 .190 Other formal ------ ------ ------ 0.123 0.106 .250 Informal ------ ------ ------ -0.073 0.114 .520
Substance problem severity (SPS-12m) ------ ------ ------ 0.028 0.040 .485
Table 13. Linear regression predicting counsellor rapport (n=103)
Model 2C F2,100=0.06, p=.941 R2=.001
Model 3C F6,96=0.86, p=.530 R2=.051
Independent variables
b SE p b SE p Treatment motivation (TEQ-RAI) 0.003 0.010 .761 0.006 0.010 .556 Perceived coercion (MPCS) 0.037 0.127 .774 0.020 0.132 .880 Social pressures:
Legal ------ ------ ------ 0.020 0.113 .854 Other formal ------ ------ ------ 0.196 0.111 .081 Informal ------ ------ ------ 0.082 0.121 .498
Substance problem severity (SPS-12m) ------ ------ ------ -0.053 0.043 .225
89
Table 14. Linear regression predicting client commitment to treatment (n=106)
Model 2D F2,103=2.51, p=.087 R2=.046
Model 3D F6,99=1.20, p=.313 R2=.068
Independent variables
b SE p b SE p Treatment motivation (TEQ-RAI) 0.014 0.009 .117 0.015 0.009 .099 Perceived coercion (MPCS) -0.076 0.111 .494 -0.066 0.117 .574 Social pressures:
Legal ------ ------ ------ -0.018 0.100 .856 Other formal ------ ------ ------ 0.037 0.099 .707 Informal ------ ------ ------ 0.084 0.107 .434
Substance problem severity (SPS-12m) ------ ------ ------ -0.053 0.038 .166
Regression diagnostics revealed an absence of multicollinearity, and low DfBeta and DfFit
values suggested the absence of influential outliers (Appendix 5). However, for the model
predicting counsellor rapport, two of the standardized residuals exceeded 3.0 (values = -4.76
and -5.47), and the distributions of standardized residuals were significantly non-normal for
each of the three full models. Because of the extreme skew in the dependent variables, these
were dichotomized using median splits and renalysed using logistic regression. Results were
identical. Because of the lack of impact of the other independent variables on the focal
associations, reduced three-variable models were not tested as for retention.
In summary, only the analysis of client confidence in treatment yielded a significant regression
equation. Greater autonomous motivation for treatment at admission predicted greater
confidence in treatment 2 months later. None of the other admission process factors (i.e.,
perceived coercion, social pressures, or substance problem severity) was significant in
predicting client engagement. It is not possible to rule out the impact of low variability in
engagement scale scores across clients on these results.
4.2.3 Client-reported reasons for leaving treatment
As part of the follow-up survey, participants who indicated that they were not still attending
treatment at PAARC or any other addiction treatment agency were asked to state, in their own
words, their reasons for not continuing with treatment. Their responses provide further context
for interpreting the results of the modeling analyses of retention and engagement. Among the
150 subjects who were not still attending treatment at follow-up, 62.0% (n=93) were
successfully recontacted for the second survey.
90
Most commonly, these former clients stated simply that their treatment was completed, or that
additional services were not needed (n=41, 44.1%). Some qualified this further by stating that
they do not have a substance abuse problem, are currently abstinent, or are doing well on their
own, while others noted that they only needed to have an assessment completed (i.e., two
sessions, typically to satisfy a probation order). Second to a lack of perceived need, being too
busy was commonly reported as a reason for leaving treatment (n=16, 17.2%). Other than time
restraints, a small number cited barriers related to travel distance, transportation, or language
difficulties that complicated or prohibited continued attendance (n=5, 5.4%). Eight former
clients (8.6%) mentioned that they had not been currently attending but planned on returning in
the very near future. Six individuals (6.5%) who were not still connected with a formal program
stated that they were instead attending mutual aid groups (i.e., AA, NA, or CA). Only a small
number voiced their dissatisfaction with the agency, their counsellor, or the program, or stated
that they did not think the treatment would help (n=5, 5.4%). Two former clients also
specifically cited the departure of their counsellor as the reason why they were no longer
connected with a treatment program.
Not all former clients articulated clear reasons for discontinuing treatment. A number did not
provide a reason, stated that they had just lost contact after missing appointments, that they were
waiting for the service to reconnect with them, or that they were unsure about whether they
would continue at this time (n=12, 12.9%).
Within the subset of clients who left treatment within the first 2 months, those with a lack of
perceived need for treatment (n=41) were compared to others (n=52) in terms of treatment
motivation, perceived coercion, social pressures and problem severity at admission (Table 15).
This particular grouping strategy was used because, given the objectives of the study, it is of
interest to examine elements of the admission process among those who felt that treatment was
not needed such a short time after seeking out a treatment program. Table 15 summarizes results
from the non-parametric Mann-Whitney U tests; although results of corresponding parametric t-
tests were identical (Table A4-6, Appendix 4).
Not surprisingly, those who left treatment because of a lack of perceived need scored
approximately 20-points lower in terms of their initial motivation for treatment (p<.001). These
former clients also perceived greater coercion and greater legal pressure to enter treatment.
91
Conversely, they reported significantly lower pressures to enter treatment from other formal and
informal sources, and significantly fewer substance-related problems in the year prior to
treatment.
Table 15. Admission process measures by perceived need for treatment at 2-months among former clients (n=93)
Reason for leaving treatment = lack of perceived need
Admission scales
No (n=52) Yes (n=41) Statistic Treatment motivation (TEQ-RAI) 85.89 ± 18.85 61.95 ± 19.88 U=289.0, p<.001 Perceived coercion (MPCS) 1.00 ± 1.30 2.44 ± 1.82 U=557.0, p<.001 Social pressures:
Legal 0.94 ± 1.55 2.56 ± 1.57 U=542.5, p<.001 Other formal 0.85 ± 1.23 0.32 ± 0.88 U=793.5, p=.008 Informal 2.04 ± 1.46 0.80 ± 1.15 U=565.5, p<.001
Substance problem severity (SPS-12m) 11.37 ± 4.21 5.66 ± 5.44 U=455.5, p<.001
4.3 Early progress in treatment The final stage of the analysis considers client-reported change in the experience of substance-
related problems over the initial weeks of treatment. The number of substance-related problems
experienced in the month prior to the follow-up survey was subtracted from the number of past-
month problems reported at admission to yield a change score. The score potentially ranged
from -16-16, with negative values indicating that more problems were reported during treatment
than before it, zero indicating no change, and positive values indicating that fewer problems
were experienced following admission. All subjects who completed the follow-up survey were
included in the analysis, regardless of whether or not they were still attending treatment
(n=205). Eleven subjects (5.4%) were excluded because of missing items on the follow-up
version of the SPS (final n=194).
Observed scores varied from -13-16, with mean of 1.68 (SD=4.80). The average change among
clients was, therefore, in the positive direction. However, there was a large spike at zero, with
30.9% (n=60) reporting no change in their past-month substance problem severity. A total of
28.9% (n=56) endorsed more substance-related problems at follow-up than at admission, while
the remainder (40.2%, n=78) indicated positive change. Among the 60 subjects who reported no
change in their symptoms, 76.7% (n=46) had reported no past-month substance-related
problems at admission and an additional 10.0% (n=6) had reported experiencing one such
problem in the month prior to entering treatment. The remaining 13.3% of clients (n=8) reported
larger numbers of persistent unresolved problems over the initial weeks of treatment. In other
92
words, a floor effect was observed, such that consideration of improvement in problems related
to substance use was hampered by the large number of clients who reported no change, mostly
because of an ongoing perceived lack of such problems.
Table 16 displays the Spearman rank correlations between admission process measures and
substance problem change scores. As hypothesized, greater autonomous motivation for
treatment was associated with greater early improvement, while perceived coercion was
associated with lower improvement. There was also an inverse correlation between legal
pressure and change in substance-related problems that just failed to reach statistical
significance. By contrast, greater pressure from informal sources was associated with greater
substance-related improvement. The results were similar in a corresponding parametric analysis,
with the exception that the correlation between legal pressure and change in substance problems
was significant (r=-.144, p=.045; Table A4-7, Appendix 4).
Table 16. Admission process measures by 2-month substance problem resolution (n=194)
Change in past-month substance problems a
Admission scales
ρ p Treatment motivation (TEQ-RAI) .178 .022 Perceived coercion (MPCS) -.299 <.001 Social pressures:
Legal -.138 .055 Other formal .096 .183 Informal .163 .023
a count of substance-related problems at admission – discharge
None of the sociodemographic characteristics was significantly associated with the change in
substance-related problems (Table 17), including age (ρ=.071, p=.323). Of the other treatment-
related factors, mandates and the type of problem substance exhibited significant associations
with the change score. Specifically, those who were mandated to treatment reported
significantly lower early improvement than others. A post-hoc Kruskal-Wallis test of differences
in early improvement excluding those who reported no problem substances at admission was
nonsignificant (χ22=2.03, p=.363), indicating that there was no difference in levels of change
based on whether clients had problems with alcohol, other drugs, or both. Results from the
corresponding parametric analysis yielded the same results (Table A4-8, Appendix 4), with the
exception that previous treatment experience exhibited a significant association with the change
score (t119= -2.13, p=.035). Specifically, those who had prior treatment experience reported
93
greater early improvement than others. Finally, mean change scores for past-month substance
problems were only slightly higher among those who were retained (1.93 ± 5.06) relative to
those who left treatment within the first 2 months (1.38 ± 4.49), and a Mann-Whitney U test
revealed that this difference was not statistically significant (U=4273.5, p=.282).
Table 17. Client characteristics by 2-month substance problem change (n=194)
Client characteristics Mean ± SD Median Statistic Gender:
Male 1.62 ± 4.68 0.0 U=3242.5, p=.520 Female 1.85 ± 5.21 0.0
Marital status: Married 1.91 ± 4.38 0.0 χ2
2=3.59, p=.167 Single 1.06 ± 5.25 0.0 Widowed, separated, divorced 2.06 ± 4.52 0.0
Education: Less than secondary school 0.45 ± 4.15 0.0 χ2
3=1.13, p=.770 Secondary school completed 2.30 ± 5.34 0.0 Some post-secondary 1.10 ± 3.72 0.0 Post-secondary completed 1.69 ± 5.04 0.0
Employment: Employed 1.70 ± 4.99 0.0 χ2
2=0.67, p=.716 Not employed 1.83 ± 5.12 0.0 Not in the labour force or on leave 1.52 ± 3.44 1.0
Previous treatment: No 1.08 ± 4.14 0.0 U=3795.0, p=.114 Yes 2.70 ± 5.62 0.0
Treatment mandate: No 2.32 ± 5.33 0.0 U=3416.0, p=.022 Yes 0.45 ± 3.29 0.0
Legal problems: No 2.18 ± 5.47 0.0 U=4353.0, p=.427 Yes 1.08 ± 3.89 0.0
Referral source: Addiction service 4.84 ± 6.27 2.0 χ2
5=9.86, p=.079 Family or friends 0.94 ± 3.09 0.0 Legal system 0.53 ± 3.49 0.0 Mental or general health services 1.60 ± 4.70 0.0 Workplace, school, or social services 0.85 ± 6.73 0.0 Self 2.67 ± 5.22 0.0
Problem substances: None -0.66 ± 2.69 0.0 χ2
3=16.57, p=.001 Alcohol only 1.94 ± 4.36 0.0 Drugs only 2.07 ± 5.99 0.0 Both alcohol and drugs 3.18 ± 5.04 1.5
Next, linear regression was used to evaluate the change in substance-related problems. The final
sample size was 190, excluding 4 additional subjects with missing values on the independent
variables. The sample is invariant across equations to allow for direct comparisons. Preliminary
curve estimation for the continuous independent variables revealed that no non-linear terms
improved model fit. Consistent with the analytical approach taken thus far, mandates and type of
problem substance were not entered into the equation as control variables.
94
Results are summarized in Table 18. In the first step, change scores were regressed on the focal
independent variables: perceived coercion and treatment motivation (Model 4). The model F-
test was significant, indicating that the variables explained a significant amount of the variance
in substance problem change (R2=7.0%). Similar to retention, perceived coercion was the sole
significant predictor at this stage. That is, despite being associated at the bivariate level,
treatment motivation was not significantly associated in the three-variable model. In a simple
model containing only treatment motivation, the regression coefficient (b) was 0.046 (se=.017,
p=.007), explaining 3.8% of the variance in the substance problem change score. On its own,
therefore, a 10-point increase on the RAI was associated with just under a half-point increase in
problem resolution. However, this effect was reduced by roughly two-thirds with the addition of
perceived coercion and was no longer significant. Again, it appears that the magnitude of the
impact of treatment motivation was small and its shared covariance with perceived coercion
rendered it redundant in this sense.
The association between perceived coercion and change in substance-related problems was
reduced by 11.9% with the addition of the social pressure ratings, but retained significance
(Model 5). A partial F-test comparing Model 5 with a reduced model (not shown) excluding the
focal independent variables revealed that the latter uniquely explained 4.4% of the variance in
substance problem change scores, a significant amount (F2,184=4.54, p<.05).
Added together as a block, the social pressure variables did not make a significant contribution
to the overall explanatory power of the model relative to one containing only the focal
independent variables (F3,184=2.25, p>.05). However, informal pressure was itself significantly
associated with early substance problem change. Specifically, for each single-point increase in
informal pressure experienced at admission, there was an approximately half-point increase in
the change score, indicating substance problem improvement. Neither of the other sources of
social pressures was significantly associated with the change score this model.
Two-month retention was then added to the model as an indicator of the early treatment process
(Model 6). It was expected that a more positive treatment process, indicated by continued
attendance 2 months following admission, would be associated with greater improvement.
Because of the downstream association specified between the focal independent variables and
retention, it was expected that retention would mediate any effects of the former variables.
95
However, the regression coefficient for perceived coercion was essentially unchanged with the
addition of retention, and the coefficient for retention was itself nonsignificant. The association
between informal pressures and change in substance-related problems was also left intact
following the addition of retention.
Table 18. Regression predicting change in past-month substance-related problems (n=190)
Model 4 F2,187=7.06, p=.001 R2=.070
Model 5 F5,184=4.22, p=.001 R2=.103
Model 6 F6,183=3.56, p=.002 R2=.104
Independent variables
b SE p b SE p b SE p Treatment motivation (TEQ-RAI) 0.017 0.020 .410 0.017 0.020 .396 0.019 0.021 .351 Perceived coercion (MPCS) -0.657 0.260 .012 -0.587 0.274 .033 -0.592 0.275 .032 Social pressures:
Legal ------ ------ ------ 0.003 0.233 .990 -0.006 0.234 .980 Other formal ------ ------ ------ 0.192 0.259 .459 0.218 0.263 .409 Informal ------ ------ ------ 0.504 0.234 .033 0.510 0.235 .031
2-month retention ------ ------ ------ ------ ------ ------ -0.404 0.722 .557
Diagnostic analysis of the full model (Model 6) revealed a lack of both multicollinearity
between independent variables and influential outlying cases (Appendix 5). However, the
distribution of standardized residuals was again significantly non-normal, with a concentration
of values around the mean (0). Again because of extreme skew in the dependent variable, the
model was reanalysed with multinomial logistic regression, with a three-level dependent
variable representing negative, positive, and no change in substance-related problems. Positive
change scores served as the reference category for comparisons. Results were similar. In the
final model (Model 6), greater perceived coercion was associated with higher odds of negative
change (OR=1.42, se=0.232, p=.030), and informal pressures were associated with a lower odds
of no change, relative to positive change (OR=0.65, se=0.088, p=.001).
In summary, perceived coercion was linked with increased experience or acknowledgement of
substance-related problems following admission, while informal pressures appeared to protect
against a lack of progress through the initial weeks of treatment. Also relevant, the majority of
those who reported no change in the number of past-month substance-related problems reported
an absence of problems at both time points.
96
Chapter 5 Discussion
This study offered a prospective examination of coercion in addiction treatment, including its
determinants, correlates and short-term impact. The primary objectives were to:
explain perceptions of coercion and autonomous motivation for treatment among adults
entering outpatient addiction treatment, and
determine the impact of perceived coercion and autonomous motivation on retention,
cognitive engagement, and substance problem severity in the initial weeks of treatment.
Based on previous work and guided by Self-Determination Theory (SDT), a conceptual
framework was devised to outline the expected mechanisms by which coercion would impact on
the treatment process (Figure 3, p.65). The analytical strategy involved the detection and
elaboration of empirical associations between client perceptions of autonomy during admission
and indicators of early treatment processes, through the systematic addition and removal of
other variables selected to rule out alternative explanations and situate the focal relationships
within a broader set of theoretically and empirically related variables (Aneshensel, 2002). This
elaborative approach is well-suited to theory-based analyses of observational data. While it is
not possible to test all potential alternative explanations or confounding variables, assertions of
causality are strengthened to the extent that empirical associations stand up to such tests.
According to SDT, perceptions of being coerced to enter a treatment program results in an
external orientation toward treatment and a lack of personal valuation of program goals and
substance-related behaviour change. Internalized or autonomous motivation for treatment and
behaviour change plays an integral role in the recovery process and is required for sustained
recovery over the long-term. Further, not simply the presence of social pressures, but the way
they are perceived and interpreted by individual clients determines their impact on this process.
External contingencies, ultimatums, and other social interactions that are perceived to restrict
autonomy are thought to undermine autonomous motivation and hinder long-term behaviour
change.
97
Given these theoretical propositions, it was hypothesized greater social pressures from all
sources, but particularly from legal authorities, would contribute to perceptions of coercion and
result in lower autonomous motivation for treatment (H1). Substance problem severity was also
hypothesized to play a role in determining client perceptions of autonomy during the admission
process, inasmuch as those who perceive greater need for treatment would see it as less of a
coercive imposition and would exhibit more autonomous motivation for participating (H2). In
turn, perceived coercion and autonomous motivation were expected to influence the treatment
process. Specifically, clients who exhibited lower perceived coercion and greater autonomous
motivation at admission were hypothesized to be more likely to still be attending treatment
sessions after 2 months (H3), to have greater cognitive engagement in the treatment process
(H4) and, in turn, to display greater reductions in substance-related problems (H5).
It should be noted that, strictly speaking, SDT does not propose that high perceived coercion
and low autonomous motivation for treatment would absolutely prohibit positive outcomes, as
the process of behaviour change can become internalized and personally valued over time given
the correct conditions (Deci & Ryan, 2000; Ryan & Deci, 2002). At the same time, however,
they do not present an optimal starting point for the process of behaviour change.
Prior to discussing the results and implications of the study, consideration should be given to the
characteristics of the study sample. Specifically, these results are based on a sample comprised
largely of young to middle-aged males who are new to formal addiction treatment. Over half
reported legal system involvement at admission and just under 40% reported being mandated to
seek treatment by legal authorities or children’s aid services. The most common problem
substances were alcohol and cocaine or crack; although roughly one-quarter did not
acknowledge any problem substances at admission. Taken as a whole, the proportions of
legally-involved and mandated clients within Ontario’s specialized treatment system in fiscal
year 2008-09 were 32% and 25%, respectively (Drug and Alcohol Treatment Information
System, 2009). In the same year, the provincial rate of clients reporting no problem substances
at admission was 3.3%. These figures suggest that the study sample contains a disproportionate
number of clients who are being objectively pressured or pushed into the treatment system.
Treatment agencies in Ontario vary in their admission policies for those who indicate during the
intake process that they do not feel they have a substance-related problem. Some, including
PAARC, have a service specifically designed for such clients, while others do not extend their
98
services to these individuals. As such, certain agencies may be expected to serve a
disproportionately larger number of clients who perceive little need for treatment. The
prevalence of external pressures and perceived coercion in this sample should be interpreted
with this in mind. The degree to which they are representative of other clinical samples will
depend on local admission policies and service complement. That said, because of PAARC’s
admission policies of accepting clients for the purposes of satisfying court-ordered and other
assessments, it is a particularly suitable agency in which to study coercion and its impact on
early treatment processes.
5.1 Main Findings and Implications
5.1.1 Levels of pressures, coercion, and motivation at treatment entry
Not surprisingly, pressures to enter treatment from a variety of sources were common and many
clients reported pressures from multiple sources. Notably, only 13% of clients reported no
pressures to enter treatment from any source. Two-thirds cited pressures from friends and
family, making informal social networks the most common source of pressures to enter
treatment. A number of previous studies have likewise noted the predominance of informal
social pressures among clients entering treatment (Cunningham et al., 1994; Marlowe, Glass et
al., 2001; Marlowe et al., 1996; Marlowe, Merikle et al., 2001; Polcin & Weisner, 1999).
Sizable proportions reported pressures from more formal sources, including legal authorities
(45%) and others such as employers or co-workers, school authorities, or social or health service
providers (41%). Distributions of pressures rated from minimal to extreme revealed that legal
pressures, when present, were more likely to receive an extreme rating than were pressures from
other sources. In contrast, among those who reported informal or other (non-legal) formal
sources of pressure, there was a more even distribution of ratings ranging from minimal to
extreme.
A fair amount of coercion was also perceived by clients at admission, and there was variation in
the degree to which clients were autonomously motivated to enter treatment. A majority of
clients (57%) reported at least some coercion, endorsing one or more of the items on the
perceived coercion scale. The mean on the scale was 1.4 out of a maximum of 5.0. Relative to
other studies that have used the MPCS in an addiction treatment context, this value is
comparable to the mean score observed in a sample of offenders mandated to treatment (1.6;
99
Prendergast et al., 2009), but approximately twice that of another study of clients entering
generalized outpatient addiction treatment (0.70; Wild et al., 1998). The comparable levels seen
between the mandated offenders and clients in this sample may be related to the relatively high
proportion that was legally mandated to have an assessment completed (38%, relative to
approximately 8% in Wild et al., 1998). However, mean scores obscure the concentrations of
clients reporting either an absence or extreme level of coercion. The distribution of MPCS
scores followed an approximate reverse-J shape: the majority reported no (43%) or low levels
(37%) of coercion, with a smaller increase at the high end of the scale (9%).
Noting a relative absence of intermediate values on the MPCS in their sample of psychiatric
patients entering hospital, the developers of the scale speculated on the underlying reasons for
the pattern of responses (Gardner et al., 1993). They suggested that the five items may have
similar thresholds of perceived coercion leading to endorsement, such that all items are either
endorsed or absent. They also suggested that, if coercion were experienced as highly aversive or
upsetting, clients would tend to either reformulate their admission experiences in order to deny
its role, or they would have a highly negative reaction to it. Thus, the scale would tend to yield
values concentrated on either extreme of the range of possible scores; that is, coercion would be
largely perceived as either absent or present. While a bimodal distribution of MPCS scores was
not observed in the present study, fully half of the sample received either the minimum or
maximum scores.
SDT-based measures of motivation are at a comparatively early stage of development in
addiction treatment research. The mean subscale scores observed in the present study are
comparable to that obtained in previous work using this scale (Wild, 1999b). Most notable is the
ceiling effect that tends to be obtained for identified motivation. More work is needed to
evaluate whether and to what degree clients have a tendency to overstate their personal
valuation of treatment at admission. Given the apparent primacy of this particular form of
motivation for positive outcomes in SDT-based evaluation research (Ryan et al., 1995; Zeldman
et al., 2004), studies are needed to determine the optimal methods for representing autonomous
motivation in addiction treatment settings. The present study opted to model treatment
motivation using a relative autonomy index: a mathematical representation of the relative level
of autonomous motivation for treatment, down-weighted to account for the presence of
100
externalized motivation (Vallerand et al., 2008). The validity of this method should be evaluated
against other options.
5.1.2 The interplay of social pressures, coercion, and motivation
The first objective of the study was addressed with an analysis of the correlations between the
admission process measures, followed by multiple linear regression predicting perceived
coercion. Overall, support was found for hypothesized relationships at this stage of the analysis.
In addition, findings support much of the previous work in this area and add to a growing body
of research examining issues of client autonomy during the admission process.
Consistent with SDT and with previous empirical study (Wild et al., 2006), greater identified
motivation for treatment corresponded to lower perceptions of coercion to enter treatment, while
greater external motivation corresponded to higher perceived coercion. Introjected motivation
(i.e., motivation arising from a desire to alleviate feelings of guilt, increase self-esteem, or
alleviate other negative internal states) also corresponded to lower perceptions of coercion,
albeit to a lesser degree than did identified motivation. Although a previous study of SDT-based
motivation found no correlation between introjected motivation and perceived coercion (Wild et
al., 2006), the present finding is nonetheless consistent with the theory. Occupying a position on
the self-determination continuum that falls between identified and external motivation, its
association with measures of autonomy would be expected to fall between those motivational
subtypes in terms of magnitude. Also, consistent with previous work (Prendergast et al., 2009),
the magnitudes of the correlations between motivation and perceived coercion (i.e., |ρ| = .30-
.70) suggest that they represent related but distinct constructs.
Findings relating social pressures to coercion and motivation highlight the complexity of
relationships between these constructs. First, pressures from different sources were not equally
experienced as impinging on client autonomy. Pressures from legal sources were more likely to
be experienced as coercive and corresponded to a lower level of autonomous motivation for
treatment. However, this study joins others in demonstrating that these constructs are not
equivalent (Farabee et al., 1998; Prendergast et al., 2009; Stevens et al., 2006; Wild et al., 1998).
That is, the moderate correlation between legal pressure ratings and perceived coercion (ρ=.51)
reveals that not all who rated their legal pressure as extreme felt coerced, and vice versa.
Similarly, rather than an absence of coercion, low average levels of coercion were perceived by
101
those who were not mandated and/or had no legal system involvement at admission. Second,
pressures from other formal sources were unrelated to perceived coercion and autonomous
motivation. In other words, collapsed together, pressures from employers, health and social
service providers, and others, were not typically experienced as impinging on client autonomy
during admission. Although this is contrary to expectations, it highlights the potential
inaccuracy of equating pressures from formal institutions to coercion.
Third, results of the ordinal logistic regression analysis suggested that, independent of shared
covariation with pressures from other sources and substance problem severity, informal
pressures from members of one’s social network corresponded to greater perceptions of
coercion. The association was not as strong as that observed for legal pressures. It is intuitive
that pressures to enter treatment would be perceived as relatively less coercive when coming
from friends and family with whom one has a personal relationship. Considered within an SDT
framework, the idea of basic needs satisfaction for both autonomy and relatedness also supports
the lower relative coerciveness of pressures stemming from close relations (Deci & Ryan,
2000).
However, it is also likely that more specific aspects of the methods through which informal
pressures are applied, as well as the nature of the relationship between those involved in the
exchange, are important determinants of perceived coercion. Early studies of psychiatric
inpatients found that experiences of coercion were judged in accordance with role expectations
and perceptions of underlying motives (Bennett et al., 1993). Further, affective responses to
health-related social control strategies within married couples are found to vary with the types of
strategies employed (Tucker & Anders, 2001). Summarized earlier, there is a range of options
available to social networks who are concerned with a significant other’s drinking or drug use
and wish, individually or collectively, to induce entry into a treatment program (Johnson, 1986;
Landau et al., 2000; Meyers et al., 1998). Such social network interventions vary widely in
approach, and would not be expected to correspond to equivalent perceptions of coercion. The
present study did not examine aspects of relational quality between clients and the friends and
family members who pressured them to enter treatment, nor did it consider the method of
application or the perceived positive or negative quality of the pressure itself. Future research is
needed to disentangle the relationships between the strategies employed by social networks to
induce treatment entry and the internal and outward responses of the target individual. Such
102
work may help to define strategies that are most effective in terms of behavioural responses,
with minimal stress or harm to the social or familial relationship.
Work in this area in the addiction field has focused largely on the source of pressures that are
reported by clients (Polcin & Weisner, 1999; Weisner & Matzger, 2002; Weisner et al., 2001).
However, frameworks are also available that offer classification based on other aspects, such as
positive versus negative quality (Lidz et al., 1995; Tucker & Anders, 2001). Positive pressures
include persuasion and inducements, through which attempts are made to convince the target
individual of the need for treatment. Conversely, negative pressures are those more explicitly
involving threats and force. Among individuals assessed at admission to psychiatric hospitals,
negative pressures were predictive of perceived coercion, while positive pressures were
unrelated (Lidz et al., 1995). Such a framework may have utility in the context of addiction
treatment and should be considered further. The additional guidance that is provided to
respondents by greater specification of what is meant by “pressures” in measurement scales may
also have utility. It is possible, for instance, that differential interpretations of what constitutes
pressure may have attenuated associations between variables in the present analysis.
In summary then, and consistent with expectations, of the three types of social pressures
examined in this study, those from legal sources were most closely associated with perceived
coercion. However, legally pressured treatment is not always perceived by clients to be a
coercive imposition, and treatment that is not legally pressured is not necessarily void of
coercion. Pressures from other sources also do not necessarily imply coercion. This is the first
study to provide a comprehensive examination of client autonomy across different types of
social pressures.
5.1.3 Other predictors of perceived coercion
As in previous work conducted in psychiatric and addiction treatment settings, perceived
coercion was largely unrelated to sociodemographic characteristics (Cascardi & Poythress,
1997; Hiday et al., 1997; Hoge et al., 1998; Lidz et al., 1995; Wild et al., 1998). Adding to the
construct validity of perceived coercion in addiction treatment settings, clients who were
mandated to treatment, those with legal system involvement, and those referred by legal
authorities reported significantly greater coercion at admission. In addition, those with a
previous history of treatment reported lower coercion to enter treatment, on average; although
103
this finding did not persist in a multivariable model containing social pressures and substance
problem severity. Clients with numerous prior treatment episodes may be more impaired by
their substance use (Cacciola et al., 2009), such that the association between treatment history
and perceived coercion may be confounded by substance problem severity.
In addition, past experiences with treatment are likely to shape current decision-making around
treatment entry and, therefore, also perceptions of coercion. In a recent study of mental health
care consumers attending outpatient treatment, those who were mandated to the current
treatment episode did not score higher than others on the MPCS (Link et al., 2008). However, a
treatment history characterized by multiple involuntary hospital admissions was significantly
related to present perceived coercion. The authors speculated that perceptions of coercion were
less shaped by isolated experiences than by a history of such experiences. Past experiences with
hospitalization and incarceration were elsewhere positively associated with perceived coercion
among mental health care consumers under outpatient commitment (McKenna et al., 2006).
These studies suggest that, over and above studies of single episodes of treatment, work is
needed that takes a life course approach to the study of coercion and other admission
experiences over the span of clients’ substance use and treatment careers (Hser, Anglin, Grella,
Longshore, & Prendergast, 1997; Hser et al., 2007). By explicitly incorporating aspects of the
social context, including social capital and critical external and internal events (Hser et al.,
2007), a life course approach to the study of coerced treatment may also help to clarify the
relationship between informal pressures and coercion, as well as the circumstances under which
pressures and coercion in general are most likely to achieve the desired outcomes. While this
study identified a link between perceived coercion and any previous history of treatment,
information was not available on the number or types of previous treatment episodes, their
effectiveness, or past adverse treatment experiences. Past history of mandated treatment, in
particular, would have been beneficial in this study.
Those who reported experiencing fewer substance-related problems prior to treatment and those
who did not acknowledge problem substances at intake perceived greater coercion. Greater
problem severity also correlated with greater autonomous motivation for treatment. These
findings echo previous work on perceived coercion and autonomy in addiction treatment
settings (Ryan et al., 1995; Wild et al., 2006; Wild et al., 1998). Also consistent with previous
104
work, greater legal pressure to enter treatment was associated with lower substance problem
severity (Brecht et al., 2005; Copeland & Maxwell, 2007; Friedmann et al., 2003; Grichting et
al., 2002; Kelly et al., 2005; Kline, 1997; Marshall & Hser, 2002; Polcin & Beattie, 2007; Polcin
& Weisner, 1999; Rush & Wild, 2003). In this respect, and given the variable link between
perceived coercion and other types of social pressures, it is relevant that the associations
between problem severity and ratings of pressures from other formal or informal sources were
positive. That is, legal pressure to enter treatment was more extreme among those with lower
problem severity, while pressures from all other examined sources were associated with greater
problem severity. Together, these findings highlight the uniqueness of the role played by the
legal system in shaping the process and client experiences of addiction treatment admission.
5.1.4 Characteristics of the treatment process
In addressing the second objective of examining the early treatment process, this study drew
from the TCU model of treatment-assisted recovery, incorporating both attendance-based
indicators and client-reported evaluations of their involvement in the program (Simpson, 2004).
Consistent with this model, the present study defined an engaged client as one who attends
sessions, becomes cognitively involved and committed to the program, and develops a positive
relationship with counsellors over the initial weeks of treatment. Consideration is then given to
early progress in resolving substance-related problems. No study to date has conducted a
comprehensive analysis of addiction treatment processes that incorporates measures of client-
perceived autonomy during admission, as well as addressing a range of social pressures to enter
treatment.
Prior to discussing the results of the modeling analysis, consideration needs to be given to the
process indicators themselves (i.e., those which constitute dependent variables in the regression
analysis). Fewer than half of study participants were still attending treatment at the study site or
another formal addiction treatment program at the 2-month follow-up, deduced from both self-
report and agency records. Approximately one-quarter of the study sample attended a single
appointment and another quarter left after the second session. While the service delivered at the
study site is not of a pre-determined length, it is relevant that the first two sessions are typically
dedicated to assessment and treatment planning purposes, with actual counselling beginning in
the third session. As such, clients who attended only two appointments received little, if any,
counselling intervention. High rates of early attrition from outpatient addiction treatment are
105
widely documented (Claus & Kindleberger, 2002; Garnick et al., 2002; Pulford & Wheeler,
2007; Stark, 1992). Noting that 40% of clients in their multisite study of outpatient alcohol and
drug treatment attended only a single session and three-quarters attended fewer than four
appointments before discontinuing the treatment episode, Pulford and Wheeler (2007) call for
an explicit recognition of the time-limited period of outpatient services and the development of
brief interventions, not unlike those for at-risk or harmful alcohol use, but suited to addressing
substance dependence. In essence, the assessment and educational services developed and
offered at PAARC are meant to achieve this purpose, with realistic appraisal of client levels of
commitment given perceived addiction severity. However, the effectiveness of such brief
interventions relative to other approaches remains unevaluated.
Pronounced ceiling effects were observed for the three subscales of treatment engagement.
Clients who were still attending sessions 2 months following admission fairly uniformly
declared high engagement in treatment. Prior to administering the scale assessing treatment
engagement, subjects were assured that their answers were confidential, and that the researcher
and the research process were independent of the agency and treatment. However, client
misunderstandings of the role of the researcher or the process of research at the agency, and
resultant concerns over confidentiality and the impact on their treatment, may nonetheless have
affected response patterns. Others may have felt compelled to provide socially desirable
responses to these questions, depicting themselves as an active participant, particularly if
external parties were continually monitoring the treatment episode.
There is a need for improved measurement specification and validation of methods for assessing
client-level cognitive engagement in addiction treatment settings. A number of previous studies
using the same or similar engagement measures as the present study have used standard linear
regression techniques to predict engagement, typically without providing information on
variable distributions or descriptive statistics (Broome et al., 2001; Broome et al., 1999; Hiller et
al., 2002; Joe et al., 1998; Joe, Simpson, & Broome, 1999; Knight et al., 2000). Others have
dichotomized scores for analysis, typically using high cut-points that are suggestive of ceiling
effects (Broome et al., 1997; Broome et al., 1996; Joe et al., 2001; Simpson, Joe, & Rowan-Szal,
1997). Still other studies, conducted with larger samples, have elected to model the engagement
measures as ordinal, again suggesting the presence of non-normal distributions and/or non-linear
relationships (Joe, Simpson, Greener et al., 1999; Simpson et al., 2000; Simpson, Joe, Rowan-
106
Szal et al., 1997). Relative to peer and counsellor ratings of client engagement, self-ratings have
been shown to be higher; that is, skewed to a greater degree toward positive evaluations
(Czuchry, Dansereau, Sia, & Simpson, 1998; Joe et al., 2001). Inconsistencies have also been
noted in the associations between client versus counsellor ratings of engagement and subsequent
measures of participation and progress (Czuchry et al., 1998). Aside from distributional
concerns, the scale contains item interdependence and unequal item scaling, neither of which
has been addressed explicitly in previously published work. Both of these issues were remedied
in the present study by adapted scoring procedures; however, additional work is required to
refine and evaluate measurement of engagement.
Early treatment process factors have been empirically linked to later outcomes (Simpson, 2004).
In his review of personal- and treatment-related factors associated with iatrogenic effects of
substance use treatment, Moos (2005) identifies a poor working alliance between clients and
counsellors as one treatment factor associated with deterioration of symptoms over treatment.
The clinical relevance of treatment process evaluation is further supported by the fact that such
process measures are actionable and can support ongoing quality improvement of services
(Garnick et al., 2007). This work speaks to the need for improved measurement specification of
treatment processes.
Finally, evaluation of past-month substance problem severity prior to the first and second
surveys revealed relatively equivalent proportions of clients endorsing negative (29%), positive
(40%), and no change (31%) in the experience of substance-related problems. There are
numerous potential explanations for the “negative change” detected by the SPS, such that it
should not be taken to necessarily indicate a worsening of clinical status. Importantly, because
problems were self-reported, the data are subject to response biases. To the degree that clients
were in denial or did not attribute their problems to their substance use, the SPS would not
detect substance-related impairment. Equally, to the degree that clients experienced greater
problem recognition over initial weeks of treatment, the measure would exhibit negative change.
The follow-up period was also fairly short and most subjects received a minimal amount of
treatment within this time. Longer follow-up, as well as additional objective measures of
substance use, such as urinalysis, or collateral reports of use and problems, were not possible
within the confines of the present study.
107
5.1.5 Autonomy and the treatment process
Overall, partial support was received for the theoretical model of treatment process outlined in
the conceptual framework. As hypothesized, lower perceived coercion and greater autonomous
treatment motivation were both associated with higher odds of 2-month retention; although
perceived coercion emerged as the stronger predictor. Contrary to expectations, however, the
effect of perceived coercion was rendered nonsignificant in the final model containing the
hypothesized antecedent variables (i.e., social pressures and substance problem severity), and
controlling for previous treatment experience. Reduced three-variable models revealed that,
although no single independent variable fully accounted for the observed empirical association
between perceived coercion and retention, their combined influences together accounted for the
association.
Among those who were still attending treatment 2 months after admission, greater autonomous
motivation for treatment at admission predicted subsequent client confidence in the treatment
process. Specifically, there was a small 0.3-point increase in client confidence in treatment for
each 10-point increase in treatment motivation. Possibly because of the low variability across
clients in levels of engagement and the resulting lack of discriminating power, the modeling
strategy was successful in predicting only one of the three indicators of engagement. In addition,
contrary to expectations, social pressures and substance problem severity, hypothesized to be
antecedent to the measures of client autonomy, were not associated with any of the indicators of
engagement.
Over and above risk factors for early attrition and poor engagement in treatment, it is of interest
to examine client perspectives on why they leave treatment. The present study offered a
valuable opportunity for examining client perspectives on attrition in relation to measures of
admission process, as all subjects completed a 2-month follow-up survey regardless of whether
or not they returned for even a second treatment session. Studies of attrition from addiction
treatment have reported that many clients who leave treatment early indicate they obtained what
they needed or resolved their problems and did not see a reason to remain (Leigh et al., 1984;
Stahler et al., 1993). Other commonly reported reasons include a lack of motivation for
treatment and a mismatch between the program and client preferences (Ball, Carroll, Canning-
Ball, & Rounsaville, 2006; Evans et al., 2009; Stahler et al., 1993).
108
In the present study, conflicts with staff and dissatisfaction with the program, which have been
reported in other research as common reasons for leaving treatment (Ball et al., 2006), were
extremely rare among former clients. Rather, client-reported reasons for leaving within the first
2 months largely involved a lack of perceived need for continuing with treatment. It was
common for clients to phrase this reason in terms of treatment completion; for instance, by
stating that their course of treatment or their assessment was finished. Such a statement implies
that a lengthy course of treatment, or at least one persisting past the point of the two-session
assessment or educational group, was not their intention from the time of admission. In other
words, from the perspective of many clients, lack of retention at the 2-month follow-up often
appeared to be less a matter of having “dropped out” as much as having completed the planned
course of treatment. Such statements are consistent with the previously discussed concept of
navigation; that is, “going through the motions” of participating in treatment without intention to
meaningfully engage in the process (Schacht Reisinger et al., 2003). A subset of former clients
was identified that appeared to have successfully navigated the program in this manner, and a
quantitative analysis confirmed their lower autonomous motivation and greater perceived
coercion at admission.
Finally, as hypothesized, greater perceived coercion and lower autonomous motivation were
associated with lower progress in resolving substance-related problems in the initial weeks of
treatment; although perceived coercion again emerged as the stronger predictor in the modeling
analysis. It is not possible to attribute this finding to iatrogenic effects of being coerced into
treatment. Three-quarters of those who reported no change in the past-month experience of
substance-related problems reported no substance-related problems at either time point. Those
who had an interest in describing themselves as abstinent or as using substances in a moderate
fashion may have been less likely to admit to having experienced substance-related problems.
The follow-up survey took place at a time when many of the legally pressured or mandated
clients would still have been under surveillance by the legal authorities (e.g., prior to the end of
their probation term).
Contrary to expectations, retention was not itself associated with early improvement in
substance problems. Specifically, it was expected that the association between perceived
coercion and early change in substance-related problems would be transmitted through
retention. It is possible that the study period was too short to allow for a demonstration of the
109
importance of retention in treatment in achieving positive change. Numerous studies have linked
retention with positive outcomes in both alcohol and drug treatment (Bottlender & Soyka, 2005;
Hser et al., 2004; Hubbard et al., 2003; Moos & Moos, 2003; Simpson, 2004; Stark, 1992;
Zhang et al., 2003). However, positive outcomes are also observed among those who receive an
extremely limited amount of treatment (Hubbard et al., 1997). These authors speculated that
high initial motivation for change may account for the improvement seen in such individuals,
rather than the impact of treatment. The present study did not consider the nature of broader
recovery processes occurring outside of the treatment context; although this is a potential
avenue for future research.
In summary, this study suggests a potential role for client perceptions of the admission process
and their personal valuation of treatment for subsequent involvement and behaviour change;
although the mechanisms through which these processes operate require further specification.
Clients who felt coerced to enter treatment were more likely to leave treatment soon after
admission, and to qualify that decision by statements indicating a lack of perceived need for
continued treatment. While this may appear intuitive, such associations had yet to be
demonstrated empirically. In addition to leaving treatment earlier, coerced clients reported less
resolution of substance-related problems in the weeks following admission.
Research on the impact of perceived coercion and SDT-based motivation on outcomes during
and following addiction treatment is at an early stage. In the only published study to date
examining the predictive utility of perceived coercion, the measure was not found to be a
significant predictor of either treatment completion or re-arrest in the months following
treatment (Prendergast et al., 2009). The sample consisted of offenders mandated to participate
in treatment as part of their probation. Outcome studies conducted in more generalized addiction
treatment settings have yet to appear in the literature. Internal motivation for treatment has been
previously linked with indicators of retention and attendance (De Leon et al., 1994; Ryan et al.,
1995; Zeldman et al., 2004). Others have also noted an inverse association with in-treatment
substance use (Downey et al., 2001; Zeldman et al., 2004) and post-treatment alcohol use
(Staines et al., 2003). On a bivariate level, the findings of the present study agree with previous
work by demonstrating inverse associations between autonomous treatment motivation and both
retention and reduction in substance problems in the early stage of treatment.
110
In light of findings relating autonomy to the treatment process, the associations between
admission process variables and substance problem severity deserve further discussion. Without
discounting the potential role of early treatment in preventing more serious and entrenched
substance problems and dependence down the road, there are a number of reasons to suggest
that intervening using formal treatment with those experiencing minimal substance-related
impairment may be suboptimal. First, it presupposes that treatment interventions are always
beneficial and ignores the potential for iatrogenic effects (Wild, 2006). It has been estimated that
between 7% to 15% of those who participate in addiction counselling programs show
deterioration in their substance use and psychosocial well-being during or shortly after treatment
(Moos, 2005). The association between iatrogenic effects and coercion has not been adequately
addressed to date, but presents a potential area for further research that is consistent with an
SDT-based perspective of recovery. In the present study, within the limitations of the measures
used, perceived coercion was associated with the endorsement of more substance-related
problems following admission than prior to the treatment episode. At a minimum, the statistic
reported above for iatrogenic effects highlights the important point that, although it may be
beneficial for the majority, formal treatment may not always be in the individual’s best interests.
Second, at least for alcohol, brief interventions delivered in primary care settings are rated
among the most effective (Miller & Wilbourne, 2002), particularly for those with alcohol-
related problems that fall short of dependence (Kaner et al., 2007; Moyer et al., 2002). Given
that the vast majority of Canadians visit a general practitioner over the course a year
(Nabalamba & Millar, 2007), it seems logical to expand efforts to reach problem drinkers in the
community through this method, rather than linking them with more formal services. Such
approaches to early intervention may prove superior in preventing future health and social
problems, and from a cost-benefit perspective.
A third relevant issue, related to brief intervention, is that of natural recovery, or recovery from
addiction problems without the help of formal treatment (Sobell, Sobell, & Toneatto, 1992).
While the effectiveness of formal treatment is not in question, studies using population survey
data have found that the majority of individuals who experience addiction problems recover
without it (Dawson et al., 2005; Sobell, Cunningham, & Sobell, 1996). Most of this research has
been conducted on problem drinkers in the community; however, the relevance of natural
recovery for illicit drug problems has also been demonstrated (Cunningham, 2000).
111
These issues raise important questions on the best uses of limited formal treatment resources.
Although system-wide estimates of the prevalence of coercion are unknown, given the high
prevalence with which clients come into the system through external referrals, it has the
potential to be high. The superiority of coerced treatment relative to other options, in the present
and over the long-term should be established prior to its acceptance as a viable strategy.
5.1.6 Other factors relevant to the treatment process
This work offers some insight into the roles played by other client characteristics and admission
process factors in contributing to the treatment process. First, this study adds to a growing body
of literature documenting an inverse association between legal pressure and retention in
treatment (Beynon et al., 2006; Claus & Kindleberger, 2002; Longshore & Teruya, 2006;
Mertens & Weisner, 2000; Stevens et al., 2005). Adjusting for substance problem severity, the
odds of remaining in treatment for longer than 2 months were reduced by roughly 20% for each
single-point increase in the level of legal pressure. This finding raises an important issue;
namely, the impact of legal pressure on continued participation in treatment beyond the point of
what is strictly mandated or required for legal purposes. In studies reporting a positive
association between legal pressure and retention, it remains unclear to what extent clients
remain in treatment following the legal mandate.
The present study was able to explicitly address the association between legal pressure and
retention beyond the mandated period of treatment, as this period was typically very short in the
study agency (i.e., consisting of two sessions or a 6-hour educational group). However, all
clients were provided with information on the counselling service offered at PAARC and were
invited to continue with treatment. Although legal pressure was not absent among those who
remained in treatment for longer than 2 months, it was significantly elevated among those who
left. Further, fully 75% of those who were mandated to treatment had left prior to the 2-month
point. As with coercion, this pattern of attendance is consistent with navigation in the absence of
meaningful engagement (Schacht Reisinger et al., 2003). The inclusion of cognitive indicators
of engagement was meant to shed further light on this issue; however, low variability in levels
of engagement across clients complicated this analysis. Contrary to hypotheses, legal pressures,
as well as pressures from other sources and substance problem severity, were not associated
with the measures of cognitive engagement, even at the bivariate level.
112
Nonetheless, the present study offers evidence that legally pressured clients may be likely to
participate in treatment only as long as they are required to by the legal authorities. One policy
implication of this is that the length of treatment required by legal mandates should be
sufficiently long to allow for the possibility of therapeutic engagement and the accrual of
treatment gains. However, it remains that the degree to which legally mandated clients engage in
treatment in an active and meaningful way has not been adequately addressed to date. Likewise,
the extent to which treatment gains are sustained over the long-term, once the period of legal
monitoring is complete, is also unclear. In the present study, pressures from legal sources were
weakly associated with a lack of positive change in the experience of substance problems in the
weeks following admission; although interpretation of this finding is complicated by the
substantial proportion of clients reporting an ongoing lack of any substance-related problems
over the study period.
Pressures from friends and family members were also linked to the treatment process. Greater
informal pressures were associated with positive change in substance-related problems over the
weeks following admission, independent of perceived coercion and other client characteristics.
Rather than representing an enduring role for informal pressures administered at admission, it is
possible that this indicator served as a proxy for ongoing social support for abstinence and/or
reduced substance use. The present study did not collect information on social support;
however, others have demonstrated its relevance for positive treatment processes and outcomes
(Broome et al., 2002; Flynn et al., 2003; McKay et al., 2005; Moos, 2007; Project MATCH
Research Group, 1997; Sung et al., 2004). This highlights the relevance of social network
involvement in the recovery process. Strategies for involving significant others more explicitly
in the treatment process, such as with the CRAFT approach discussed earlier (Meyers et al.,
1998), provide an important avenue for further exploration.
This study did not document much of a role for formal pressures in contributing to either
perceived coercion at admission or later treatment processes. It is possible that pressures from
the different sources that were collapsed into this category have different implications, such that
their combination attenuated any relationships and prevented detection. Specifically, although
they were not defined differently in this study, pressures from health care providers may be
perceived differently than pressures from employers or those from social service agencies. The
present study was not powered to examine the impacts of further refined categories of pressures.
113
However, the lack of statistical associations in the present study may indicate a shortcoming in
the framework applied here to the study of pressures. Again, the salient factor may be less one
of source than of relational quality and the methods of application. Future work is needed to
illuminate the role that non-legal social institutions play in addiction treatment and the
mechanisms through which they can influence the course of recovery.
The only other client factor associated with the outcomes measured in the present study was
prior treatment history, which was associated with greater odds of 2-month retention. Factors
such as age and the severity of problem substances can be expected to confound the association
between prior treatment history and retention (Stark, 1992). Here, while the association between
age and retention was not significant, net of perceived coercion, motivation, pressures, and
substance problem severity, having a previous treatment admission (excluding participation in
mutual aid groups like AA or NA) almost doubled the odds of remaining in treatment for at least
2 months.
As noted earlier, equivocal findings are reported with respect to the role of prior treatment
experience in retention (Stark, 1992). In a recent prospective study, the number of previous
treatment admissions was associated with longer length of stay in residential treatment settings,
but was not associated with the rate of treatment completion (Cacciola et al., 2009). Further,
clients with more previous treatment experience, although they made comparable gains in
treatment to others, exhibited greater problem severity at both admission and 6-month post-
treatment. Elsewhere, the number of months of previous treatment has been associated with a
small, but significant, decline in the odds of leaving treatment within 2 months of admission
(Simpson & Joe, 1993). A life course approach to the study of substance use and treatment
careers emphasizes the cumulative impact of treatment episodes on achieving and maintaining
positive outcomes in the long term (Hser et al., 1997; Hser et al., 2007); although the empirical
evidence for this is also equivocal (Cacciola, Dugosh, Foltz, Leahy, & Stevens, 2005; Hser et
al., 1999). Interpretation of the present finding linking prior treatment experience and retention,
and its potential role in determining longer-term outcomes, is complicated by lack of
information on number, duration and timing of previous treatment episodes, as well as their
mandated or coerced status.
114
5.2 Study limitations The impact on the study of instability in the treatment agency caused by staffing challenges and
changes to clinical programming over the course of data collection is difficult to ascertain;
however, it is likely that such events have impacted the data to some degree. For instance,
fluctuations in the agency’s ability to maintain an adequate complement of clinical staff to
match demand for treatment resulted in widely varying wait times for treatment over the 20
months of data collection (i.e., from 1 week up to 4 months). Studies conducted in outpatient
addiction treatment settings have suggested that longer wait times between initial assessment
and first treatment appointment increase early attrition (Claus & Kindleberger, 2002; Leigh et
al., 1984). These same staffing issues at the study site also affected the frequency of
appointments that could be offered to clients once they were admitted, and resulted in a
programming change for clients with low problem recognition (specifically, a move from
individual to group treatment). Unplanned counsellor departures from the agency unfortunately
resulted in a number of clients having to switch to a new counsellor partway through their
treatment episode. These agency-related factors may have impacted on client motivation,
retention, engagement, and other measures. The findings of this study would be strengthened by
replication at other clinical sites.
High turnover of clinical and administrative staff and difficulties in filling vacant positions at
addiction treatment agencies have been reported in the literature (Gallon, Gabriel, & Knudsen,
2003; McLellan, Carise, & Kleber, 2003), suggesting that the challenges experienced in
recruiting and maintaining counselling staff are not unique to the study site. Other studies have
documented a positive link between staff perceptions of the workplace environment and client
ratings of counsellor rapport and treatment satisfaction (Broome, Flynn, Knight, & Simpson,
2007; Greener, Joe, Simpson, Rowan-Szal, & Lehman, 2007). That is, unhealthy workplace
environments may have a detrimental impact on client engagement and, ultimately, on treatment
effectiveness. The TCU model of treatment (Simpson, 2004) puts specific emphasis on these
issues by incorporating program attributes, including measures of resources, staff, program
climate, and management information. More work is needed to evaluate the impacts of high
staff turnover and counsellor burn-out on agency functioning, staff morale, client engagement,
and treatment effectiveness.
115
It is interesting that, within the context of this relatively unstable treatment environment, clients
nonetheless reported very high levels of engagement in the treatment process. A positive
response bias was observed for the measures of both treatment engagement and identified
motivation for treatment. The nature of the observed distributions of these scale scores raises
questions regarding the validity of current measurement approaches for aspects of treatment
entry and participation. Asymmetric scale distributions and low variability across subjects
placed limitations on the analysis, and may have restricted the study’s ability to discriminate
between clients and detect important relationships. While self-reported data are always
potentially limited by a desire to present oneself in a certain light or to offer socially appropriate
responses, treatment motivation and engagement may be particularly prone to such a bias due to
the nature of the questions. It is notable in this respect that the measure of motivation (TEQ) was
administered by counsellors as part of routine clinical assessment, rather than as part of the
study protocol. It is possible that clients may exaggerate their personal interest in treatment to an
even greater degree when asked by a counsellor.
More generally, self-reported data are subject to biases related to recall difficulties and item
miscomprehension. In addition to attempting to ensure that clients were aware that the study did
not constitute part of their treatment, precautions were taken to minimize miscomprehension.
The surveys were completed in the presence of the researcher, and respondents were invited to
request assistance and clarification as needed. However, additional information from other
sources, including counsellor or collateral reports, or biological measures, would have been
useful in substantiating the findings.
The follow-up rate was acceptable at 74%, although there was evidence of lower legal pressure
and higher motivation in the follow-up sample. As noted earlier, this attrition bias will not
impact on the analysis of retention, as information on attendance was available on all subjects.
Bias is also minimized in the analysis of engagement among those still attending treatment, as a
higher rate of follow-up was achieved with these clients (89%). An attrition bias may have
impacted the analysis of client-reported reasons for leaving treatment however, as the follow-up
rate was significantly lower among these former clients (62%). These findings should be
interpreted keeping this in mind.
116
In addition, the assessment of recovery at follow-up was limited to substance-related problems.
A broader consideration of recovery involving other life domains, such as employment,
financial and psychosocial well-being, and social relationships, was not undertaken. However, it
is notable that the study period was short, with a focus placed on processes in the initial weeks
of treatment. Resolution of problems in these other domains, while certainly highly relevant to
recovery overall, may be expected to require lengthier periods of time. The types of problems
related to substance use assessed by the SPS, however, may be expected to decline more
immediately with reductions in use. Nonetheless, the present study involved a limited evaluation
of the recovery construct, and further study is need to address the role of coercion in broader
psychosocial and relational health.
Finally, rather than the elaboration modeling strategy used to analyse these study data
(Aneshensel, 2002), more sophisticated statistical methods are available for modeling complex
relationships. Path analysis and latent variable modeling techniques are well-suited to this kind
of study, allowing for multiple dependent variables, modeling of error terms and unobserved
phenomena, and simultaneous testing of multiple associations. Sample size requirements
prohibited a robust assessment of the theoretical model with such statistical techniques in the
present study, particularly in the face of non-normal variable distributions. As such, the model
was reduced into its component parts for analysis. Although this approach was enlightening in
terms of providing an in-depth look at the focal relationships through strategic variable
manipulation, larger studies with greater statistical power are needed to evaluate the framework
as a whole.
5.3 Conclusions The concepts of self-determination and autonomy have not traditionally played a large role in
outcome evaluations of addiction treatment. Discussions of the role of coercion in addiction
treatment have instead tended to centre on public health concerns of addiction, economic
productivity, crime and infectious disease (Anglin, 1988; Gostin, 1991; Leukefeld & Tims,
1988). For these and other reasons discussed earlier, it remains unclear to what extent many of
the commonly employed methods for getting people into treatment may be detrimental to the
treatment process and, by extension, longer-term outcomes.
117
The ethical issues posed by coerced addiction treatment are complex from a public health
perspective. Substance abuse poses real threats to public health and societal well-being, and this
provides a strong impetus for government and other formal institutions to intervene in the lives
of those with addiction problems. Ethical frameworks for the justification of public health
intervention cite factors such as effectiveness and necessity of the measure in promoting and/or
protecting the health of the public, and safe-guarding a balance between positive and negative
effects of the intervention, as relevant concerns in the debate over whether to infringe upon
individual autonomy and liberty (Childress et al., 2002; Upshur, 2002). Applied to the case of
coerced addiction treatment, evidence would have to be brought to bear on whether the
proposed course of treatment is likely to be successful in alleviating current harm and
preventing future harm to the public that stems from the individual’s use of substances and
whether it is a necessary means to achieve these ends. Once demonstrated, it also needs to be
considered whether the benefits outweigh any negative consequences resulting from the
infringement of the individual’s right to make their own decisions relating to treatment. It is not
at all clear that past research has satisfied these conditions. Ethical arguments do not preclude
the legitimacy of using coercive means or restricting individual rights in the name of public
health, but they do offer guidance to those charged with devising and implementing policies in
this regard.
The present study sought to contribute empirical evidence to ongoing debates over the
legitimacy of using coercion to influence addiction treatment entry by reframing the concept of
coercion in terms of client perspectives and evaluating the impact these have on the process of
treatment. The results raise questions about previous conclusions of the effectiveness of coerced
treatment and suggest many avenues for future research. Coercion was indeed commonly
reported by clients entering treatment; although, when viewed in dimensional terms, was present
in fairly low amounts on average. Partial support was obtained for its role in influencing
treatment processes over the short-term. The potential shortcomings of coercion and legal
pressure were highlighted in the analyses of retention and client-reported reasons for leaving
treatment, in which many such clients apparently failed to plan for, or take part in, a longer
course of treatment, and only completed what was required of them.
It remains to be demonstrated whether even the relatively limited exposure to treatment among
those who were coerced is ultimately beneficial or harmful in the long-run for the individual and
118
for the public. One important future avenue for research on coercion would involve randomizing
clients with legal involvement to receive brief and/or lengthier interventions versus no
treatment. Related to this, from a system-level perspective, the issue of opportunity costs and
optimal dedication of treatment resources also needs to be addressed. On the one hand,
particularly with the move to group treatment, these clients individually accounted for a
relatively small amount of agency resources. Collectively, however, their numbers were large
enough that they accounted for a substantial proportion of counselling hours, even with the
limited duration of most episodes. The interplay of legal pressures, substance problem severity,
and perceived need for continued services also raises complex questions about optimal
indicators for use in needs-based system planning and forecasting models.
Also relevant to the topic of therapeutic benefit in coerced treatment is the SDT process of
internalization. Specifically, the changing nature of motivation over time and the provision of
support for autonomy in therapeutic settings are understudied and potentially valuable avenues
for future research. To the extent that pressures and coercion already constitute established
pathways through which people will continue to enter the treatment system, fostering and
supporting the SDT process of internalization may be the preferred option for promoting
sustainable recovery. If so, further study is needed to determine the optimal methods and
therapeutic conditions for promoting internalization of initially externally motivated treatment
entry and behaviour change. In particular, methods to offset any negative impact of coercion at
admission on later treatment and recovery should be developed and evaluated.
The present study constitutes one part of a larger research agenda aimed to address the
prevalence and implications of coercion at treatment entry, some of which has been informed by
the present study. Additional work is underway testing the SDT model of treatment processes in
larger samples and other treatment modalities, including residential care and outpatient services
designed specifically for mandated offenders. Qualitative work is also underway to explore
client and counsellor experiences of coercion in more depth. This work is expected to yield
valuable insights on the role of pressure strategies, over and above their sources, and the
consequent impact on perceived coercion and the treatment process. Despite its limitations, the
present study, in conjunction with this other ongoing work, offers valuable insight into a rarely-
addressed aspect of addiction treatment.
119
References Aday, L. A., & Andersen, R. (1974). A framework for the study of access to medical care.
Health Services Research, 9, 208-220.
Ajzen, I. (1991). The theory of planned behavior. Organizational Behavior and Human Decision Processes, 50, 179-211.
American Psychiatric Association. (1994). Diagnostic and Statistical Manual of Mental Disorders (4th ed.). Washington, DC: Author.
Aneshensel, C. S. (2002). Theory-based data analysis for the social sciences. Thousand Oaks, CA: Sage.
Anglin, M. D. (1988). A social policy analysis of compulsory treatment for opiate dependence. Journal of Drug Issues, 18.
Anglin, M. D., Brecht, M. L., & Maddahian, E. (1989). Pre-treatment characteristics and treatment performance of legally coerced versus voluntary methadone maintenance admissions. Criminology, 27, 537-557.
Anglin, M. D., & Hser, Y. I. (1991). Criminal justice and the drug-abusing offender: Policy issues of coerced treatment. Behavioral Sciences and the Law, 9, 243-267.
Ball, S. A., Carroll, K. M., Canning-Ball, M., & Rounsaville, B. J. (2006). Reasons for dropout from drug abuse treatment: Symptoms, personality, and motivation. Addictive Behaviors, 31, 230-330.
Bandura, A. (1977). Self-efficacy: Toward a unifying theory of behavioral change. Psychological Review, 84, 191-215.
Bandura, A. (1997). The anatomy of stages of change. American Journal of Health Promotion, 12, 8-10.
Barber, J. G., & Crisp, B. R. (1995). The 'pressures to change' approach to working with the partners of heavy drinkers. Addiction, 90, 269-276.
Belenko, S. (2001). Research on drug courts: A critical review. 2001 update. New York: National Centre on Addiction and Substance Abuse.
Belenko, S. (2002). The challenges of conducting research in drug treatment court settings. Substance Use and Misuse, 37, 1635-1664.
Bennett, N. S., Lidz, C. W., Monahan, J., Mulvey, E. P., Hoge, S. K., Roth, L. H., et al. (1993). Inclusion, motivation, and good faith: The morality of coercion in mental hospital admission. Behavioral Sciences and the Law, 11, 295-306.
Beynon, C. M., Bellis, M. A., & McVeigh, J. (2006). Trends in drop-out, drug free discharge and rates of re-presentation: A retrospective cohort study of drug treatment clients in the North West of England. BMC Public Health, 6, 205.
Bindman, J. P., Reid, Y., Szmukler, G., Tiller, J., Thornicroft, G., & Leese, M. (2005). Perceived coercion at admission to psychiatric hospital and engagement with follow-up: A cohort study. Social Psychiatry and Psychiatric Epidemiology, 40, 160-166.
120
Bottlender, M., & Soyka, M. (2005). Outpatient alcoholism treatment: Predictors of outcome after 3 years. Drug and Alcohol Dependence, 80, 83-89.
Boyle, K., Polinsky, M. L., & Hser, Y. I. (2000). Resistance to drug abuse treatment: A comparison of drug users who accept or decline treatment referral assessment. Journal of Drug Issues, 30, 555-574.
Brecht, M. L., Anglin, M. D., & Dylan, M. (2005). Coerced treatment for methamphetamine abuse: Differential patient characteristics and outcomes. American Journal of Drug and Alcohol Abuse, 31, 337-356.
Brecht, M. L., Anglin, M. D., & Wang, J. C. (1993). Treatment effectiveness for legally coerced versus voluntary methadone maintenance clients. American Journal of Drug and Alcohol Abuse, 19, 89-106.
Brizer, D. A., Maslansky, R., & Galanter, M. (1990). Treatment retention of patients referred by public assistance to an alcoholism clinic. American Journal of Drug and Alcohol Abuse, 16, 259-264.
Broome, K. M., Flynn, P. M., Knight, D. K., & Simpson, D. D. (2007). Program structure, staff perceptions, and client engagement in treatment. Journal of Substance Abuse Treatment, 33, 149-158.
Broome, K. M., Joe, G. W., & Simpson, D. D. (2001). Engagement models for adolescents in DATOS-A. Journal of Adolescent Research, 16, 608-623.
Broome, K. M., Knight, D. K., Knight, K., Hiller, M. L., & Simpson, D. D. (1997). Peer, family, and motivational influences on drug treatment process and recidivism for probationers. Journal of Clinical Psychology, 53, 387-397.
Broome, K. M., Knight, K., Hiller, M. L., & Simpson, D. D. (1996). Drug treatment process indicators for probationers and prediction of recidivism. Journal of Substance Abuse Treatment, 13, 487-491.
Broome, K. M., Simpson, D. D., & Joe, G. W. (1999). Patient and program attributes related to treatment process indicators in DATOS. Drug and Alcohol Dependence, 57, 127-135.
Broome, K. M., Simpson, D. D., & Joe, G. W. (2002). The role of social support following short-term inpatient treatment. American Journal on Addictions, 11, 57-65.
Burke, A. C., & Gregoire, T. K. (2007). Substance abuse treatment outcomes for coerced and noncoerced clients. Health and Social Work, 32, 7-15.
Burke, B. L., Arkowitz, H., & Menchola, M. (2003). The efficacy of Motivational Interviewing: A meta-analysis of controlled clinical trials. Journal of Clinical and Consulting Psychology, 71, 843-861.
Cacciola, J. S., Dugosh, K., Foltz, C., Leahy, P., & Stevens, R. (2005). Treatment outcomes: First time versus treatment-experienced clients. Journal of Substance Abuse Treatment, 28, S13-S22.
Cacciola, J. S., Dugosh, K. L., & Camilleri, A. C. (2009). Treatment history: Relationship to treatment outcomes. Substance Use and Misuse, 44, 305-321.
Caplan, A. L. (2006). Ethical issues surrounding forced, mandated, or coerced treatment. Journal of Substance Abuse Treatment, 31, 117-120.
121
Carroll, K. M., Ball, S. A., Nich, C., Martino, S., Frankforter, T. L., Farentinos, C., et al. (2006). Motivational interviewing to improve treatment engagement and outcome in individuals seeking treatment for substance abuse: A multisite effectiveness study. Drug and Alcohol Dependence, 81, 301-312.
Cascardi, M., & Poythress, N. G. (1997). Correlates of perceived coercion during psychiatric hospital admission. International Journal of Law and Psychiatry, 20, 445-458.
Cascardi, M., Poythress, N. G., & Ritterband, L. (1997). Stability of psychiatric patients' perceptions of their admission experience. Journal of Clinical Psychology, 53, 1-7.
Chandler, D., Meisel, J., Jordan, P., Menees Rienzi, B., & Naylor Goodwin, S. (2004). Substance abuse, employment, and welfare tenure. Social Service Review, 78, 628-651.
Childress, J. F., Faden, R. R., Gaare, R. D., Gostin, L. O., Kahn, J., Bonnie, R. J., et al. (2002). Public health ethics: Mapping the terrain. Journal of Law, Medicine, and Ethics, 30, 170-178.
Claus, R. E., & Kindleberger, L. R. (2002). Engaging substance abusers after centralized assessment: Predictors of treatment entry and dropout. Journal of Psychoactive Drugs, 34, 25-31.
Cohen, J., & Cohen, P. (1983). Applied multiple regression/correlation analysis for the behavioral sciences (2nd ed.). Hillsdale, NJ: Lawrence Erlbaum Associates.
Collins, J. J., & Allison, M. (1983). Legal coercion and retention in drug abuse treatment. Hospital and Community Psychiatry, 34, 1145-1149.
Conner, B. T., Longshore, D., & Anglin, M. D. (2009). Modeling attitude towards drug treatment: The role of internal motivation, external pressure, and dramatic relief. Journal of Behavioral Health Services and Research, 36, 150-158.
Cooper, C. S. (2003). Drug courts: Current issues and future perspectives. Substance Use and Misuse, 38, 1671-1711.
Copeland, J., & Maxwell, J. C. (2007). Cannabis treatment outcomes among legally coerced and non-coerced adults. BMC Public Health, 7, 111.
Covington, J. (2001). Linking treatment to punishment: An evaluation of drug treatment in the criminal justice system. In C. Manski, J. Pepper & C. Petrie (Eds.), Informing America's policy on illegal drugs: What we don't know keeps hurting us (pp. 349-381). Washington, DC: National Academy of Sciences.
Cunningham, J., Sobell, L. C., Sobell, M. B., & Gaskin, J. (1994). Alcohol and drug abusers' reasons for seeking treatment. Addictive Behaviors, 19, 691-696.
Cunningham, J. A. (2000). Remissions from drug dependence: Is treatment a prerequisite? Drug and Alcohol Dependence, 59, 211-213.
Curran, G. M., Stecker, T., Han, X., & Booth, B. M. (2009). Individual and program predictors of attrition from VA substance use treatment. Journal of Behavioral Health Services and Research, 36, 25-34.
Curry, S. J., Grothaus, L., & McBride, C. (1997). Reasons for quitting: Intrinsic and extrinsic motivation for smoking cessation in a population-based sample of smokers. Addictive Behaviors, 22, 727-739.
122
Cuzzort, R. P., & King, E. W. (1995). Twentieth-century social thought (5th Ed.). Orlando, FL: Harcourt Brace College.
Czuchry, M., Dansereau, D. F., Sia, T. L., & Simpson, D. D. (1998). Using peer, self, and counselor ratings to evaluate treatment progress. Journal of Psychoactive Drugs, 30, 81-87.
Danziger, S. K., Kalil, A., & Anderson, N. J. (2000). Human capital, physical health, and mental health of welfare recipients: Co-occurrence and correlates. Journal of Social Issues, 56, 635-654.
Davey, M. A., Latkin, C. A., Hua, W., Tobin, K. E., & Strathdee, S. (2007). Individual and social network factors that predict entry to drug treatment. American Journal on Addictions, 16, 38-45.
Dawson, D. A., Grant, B. F., Stinson, F. S., Chou, P. S., Huang, B., & Ruan, W. J. (2005). Recovery from DSM-IV alcohol dependence: United States, 2001-2002. Addiction, 100, 281-292.
De Leon, G., Melnick, G., Kressel, D., & Jainchill, N. (1994). Circumstances, motivation, readiness, and suitability (the CMRS scales): Predicting retention in therapeutic community treatment. American Journal of Drug and Alcohol Abuse, 20, 495-515.
De Weert-Van Oene, G. H., Schippers, G. M., De Jong, C. A., & Schrijvers, G. A. (2002). Motivation for treatment in substance-dependent patients. Psychometric evaluation of the TCU motivation for treatment scales. European Addiction Research, 8, 2-9.
De Weert-Van Oene, G. H., Schippers, G. M., De Jong, C. A., & Schrijvers, G. J. (2001). Retention in substance dependence treatment: The relevance of in-treatment factors. Journal of Substance Abuse Treatment, 20, 253-261.
Deci, E. L., Eghari, H., Patrick, B. C., & Leone, D. R. (1994). Facilitating internalization: The self-determination theory perspective. Journal of Personality, 62, 119-142.
Deci, E. L., Koestner, R., & Ryan, R. M. (1999). A meta-analytic review of experiments examining the effects of extrinsic rewards on intrinsic motivation. Psychological Bulletin, 125, 627-668; discussion 692-700.
Deci, E. L., & Ryan, R. M. (1987). The support of autonomy and the control of behavior. Journal of Personality and Social Psychology, 53, 1024-1037.
Deci, E. L., & Ryan, R. M. (2000). The "what" and "why" of goal pursuits: Human needs and the self-determination of behaviour. Psychological Inquiry, 11, 227-268.
Deci, E. L., & Ryan, R. M. (2008). Facilitating optimal motivation and psychological well-being across life's domains. Canadian Psychology, 49, 14-23.
Deci, E. L., & Ryan, R. M. (Eds.). (2002). Handbook of self-determination research. Rochester, NY: University of Rochester Press.
Dennis, M. L., White, M., Titus, J. C., & Unsicker, J. (2006). Global Appraisal of Individual Needs: Administration guide for the GAIN and related measures (Version 5). Bloomington, IL: Chestnut Health Systems.
Desmond, D. P., & Maddux, J. F. (1996). Compulsory supervision and methadone maintenance. Journal of Substance Abuse Treatment, 13, 79-83.
123
Downey, L., Rosengren, D. B., & Donovan, D. A. (2001). Sources of motivation for abstinence: A replication analysis of the Reasons for Quitting Questionnaire. Addictive Behaviors, 26, 79-89.
Downey, L., Rosengren, D. B., & Donovan, D. M. (2000). To thine own self be true: Self-concept and motivation for abstinence among substance abusers. Addictive Behaviors, 25, 743-757.
Drabble, L. (2007). Pathways to collaboration: Exploring values and collaborative practice between child welfare and substance abuse treatment fields. Child Maltreatment, 12, 31-42.
Drug and Alcohol Treatment Information System. (2009). Substance abuse statistics. Retrieved August 27, 2009, from http://www.datis.ca/
Evans, E., Li, L., & Hser, Y. I. (2009). Client and program factors associated with dropout from court mandated drug treatment. Evaluation and Program Planning, 32, 204-212.
Ezzati, M., Lopez, A., Rodgers, A., Vander Hoorn, S., & Murray, C. J. (2002). Selected major risk factors and global and regional burden of disease. Lancet, 360, 1342-1343.
Fagan, R. (1999). The use of required treatment for substance abusers. Substance Abuse, 20, 249-261.
Farabee, D., Prendergast, M. L., & Anglin, D. M. (1998). The effectiveness of coerced treatment for drug-abusing offenders. Federal Probation, 62, 3-10.
Farabee, D., Shen, H., & Sanchez, S. (2002). Perceived coercion and treatment need among mentally ill parolees. Criminal Justice and Behavior, 29, 76-86.
Fiorentine, R., Nakashima, J., & Anglin, M. D. (1999). Client engagement in drug treatment. Journal of Substance Abuse Treatment, 17, 199-206.
Fischer, B. (2003). 'Doing good with a vengeance': A critical assessment of the practices, effects and implications of drug treatment courts in North America. Criminal Justice, 3, 227-248.
Fischer, B., Roberts, J. V., & Kirst, M. (2002). Compulsory drug treatment in Canada: Historical origins and recent developments. European Addiction Research, 8, 61-68.
Flynn, P. M., Craddock, S. G., Hubbard, R. L., Anderson, J., & Etheridge, R. M. (1997). Methodological overview and research design for the Drug Abuse Treatment Outcome Study (DATOS). Psychology of Addictive Behaviors, 11, 230-243.
Flynn, P. M., Joe, G. W., Broome, K. M., Simpson, D. D., & Brown, B. S. (2003). Recovery from opioid addiction in DATOS. Journal of Substance Abuse Treatment, 25, 177-186.
Foote, J., DeLuca, A., Magura, S., Warner, A., Grand, A., Rosenblum, A., et al. (1999). A group motivational treatment for chemical dependency. Journal of Substance Abuse Treatment, 17, 181-192.
Freedberg, E. J., & Johnston, W. E. (1980). Outcome with alcoholics seeking treatment voluntarily or after confrontation by their employer. Journal of Occupational Medicine, 22, 83-86.
124
Friedmann, P. D., Lemon, S. C., Stein, M. D., & D'Aunno, T. A. (2003). Community referral sources and entry of treatment-naive clients into outpatient addiction treatment. American Journal of Drug and Alcohol Abuse, 29, 105-115.
Gallon, S. L., Gabriel, R. M., & Knudsen, J. R. (2003). The toughest job you'll ever love: A Pacific Northwest Treatment Workforce Survey. Journal of Substance Abuse Treatment, 24, 183-196.
Garber, J., & Seligman, M. E. P. (1980). Human helplessness: Theory and applications. New York: Academic Press.
Gardner, W., Hoge, S. K., Bennett, N., Roth, L. H., Lidz, C. W., Monahan, J., et al. (1993). Two scales for measuring patients' perceptions for coercion during mental hospital admission. Behavioral Sciences and the Law, 11, 307-321.
Gardner, W., Lidz, C. W., Hoge, S. K., Monahan, J., Eisenberg, M. M., Bennett, N. S., et al. (1999). Patients' revisions of their beliefs about the need for hospitalization. American Journal of Psychiatry, 156, 1385-1391.
Garnick, D. W., Horgan, C. M., Lee, M. T., Panas, L., Ritter, G. A., Davis, S., et al. (2007). Are Washington Circle performance measures associated with decreased criminal activity following treatment? Journal of Substance Abuse Treatment, 33, 341-352.
Garnick, D. W., Lee, M. T., Chalk, M., Gastfriend, D., Horgan, C. M., McCorry, F., et al. (2002). Establishing the feasibility of performance measures for alcohol and other drugs. Journal of Substance Abuse Treatment, 23, 375-385.
Garnick, D. W., Lee, M. T., Horgan, C. M., & Acevedo, A. (2008). Adapting Washington Circle performance measures for public sector substance abuse treatment systems. Journal of Substance Abuse Treatment, 36, 265-277.
Garrett, J., Landau-Stanton, J., Stanton, M. D., Stellato-Kabat, J., & Stellato-Kabat, D. (1997). ARISE: A method for engaging reluctant alcohol- and drug-dependent individuals in treatment. Albany-Rochester Interventional Sequence for Engagement. Journal of Substance Abuse Treatment, 14, 235-248.
Garrett, J. A., Stanton, M. D., Landau, J., Baciewicz, G., Brinkman-Sull, D., & Shea, R. (1999). The "concerned other" call: Using family links and networks to overcome resistance to addiction treatment. Substance Use and Misuse, 34, 363-382.
Garrity, T. F., Prewitt, S. H., Joosen, M., Tindall, M. S., Webster, J. M., & Leukefeld, C. G. (2008). Baseline subjective stress predicts 1-year outcomes among drug court clients. International Journal of Offender Therapy and Comparative Criminology, 52, 346-357.
Garson, G. D. (2008). Statnotes: Topics in multivariate analysis. Retrieved August 24, 2009, from http://faculty.chass.nscu.edu/garson/pa765/statnote.htm
George, A. A., & Tucker, J. A. (1996). Help-seeking for alcohol-related problems: Social contexts surrounding entry into alcoholism treatment or Alcoholics Anonymous. Journal of Studies on Alcohol, 57, 449-457.
Godley, M. D., Godley, S. H., Dennis, M. L., Funk, R., & Passetti, L. (2002). Preliminary outcomes from the assertive continuing care experiment for adolescents discharged from residential treatment. Journal of Substance Abuse Treatment, 23, 21-32.
125
Goldkamp, J., White, M., & Robinson, J. (2001). Do drug courts work? Getting inside the drug court black box. Journal of Drug Issues, 31, 27-72.
Goode, E. (2008). Deviant behaviour (8th Ed.). Upper Saddle River, NJ: Pearson Prentice Hall.
Gossop, M., Marsden, J., Stewart, D., & Kidd, T. (2003). The National Treatment Outcome Research Study (NTORS): 4-5 year follow-up results. Addiction, 98, 291-303.
Gostin, L. O. (1991). Compulsory treatment for drug-dependent persons: Justifications for a public health approach to drug dependency. Milbank Quarterly, 69, 561-593.
Grant, B., & Dawson, D. A. (1996). Alcohol and drug use, abuse, and dependence among welfare recipients. American Journal of Public Health, 86, 1450-1454.
Greener, J. M., Joe, G. W., Simpson, D. D., Rowan-Szal, G. A., & Lehman, W. E. (2007). Influence of organizational functioning on client engagement in treatment. Journal of Substance Abuse Treatment, 33, 139-147.
Grichting, E., Uchtenhagen, A., & Rehm, J. (2002). Modes and impact of coercive inpatient treatment for drug-related conditions in Switzerland. European Addiction Research, 8, 78-83.
Griffith, J. D., Knight, D. K., Joe, G. W., & Simpson, D. D. (1998). Implications of family and peer relations for treatment engagement and follow-up outcomes: An integrative model. Psychology of Addictive Behaviors, 12, 113-126.
Gyarmathy, V. A., & Latkin, C. A. (2008). Individual and social factors associated with participation in treatment programs for drug users. Substance Use and Misuse, 43, 1865-1881.
Harris, A. H., Humphreys, K., Bowe, T., Tiet, Q., & Finney, J. W. (2008). Does meeting the HEDIS substance abuse treatment engagement criterion predict patient outcomes? Journal of Behavioral Health Services and Research. Electronic publication ahead of print.
Harris, A. H., Humphreys, K., & Finney, J. W. (2007). Veterans Affairs facility performance on Washington Circle indicators and casemix-adjusted effectiveness. Journal of Substance Abuse Treatment, 33, 333-339.
Harrison, L. D., & Scarpitti, F. R. (2002). Introduction: Progress and issues in drug treatment courts. Substance Use and Misuse, 37, 1441-1467.
Hasin, D. S. (1994). Treatment/self-help for alcohol-related problems: Relationship to social pressure and alcohol dependence. Journal of Studies on Alcohol, 55, 660-666.
Health Canada. (2004). Best practices: treatment and rehabilitation for driving while impaired offenders. Ottawa, ON: Author.
Hiday, V. A., Swartz, M. S., Swanson, J., & Wagner, H. R. (1997). Patient perceptions of coercion in mental hospital admission. International Journal of Law and Psychiatry, 20, 227-241.
Hiller, M. L., Knight, K., Leukefeld, C. G., & Simpson, D. (2002). Motivation as a predictor of therapeutic engagement in mandated residential substance abuse treatment. Criminal Justice and Behavior, 29, 56-75.
126
Ho, A. (2008). Relational autonomy or undue pressure? Family's role in medical decision-making. Scandinavian Journal of Caring Sciences, 22, 128-135.
Hodgins, H. S., Koestner, R., & Duncan, N. (1996). On the compatibility of autonomy and relatedness. Personality and Social Psychology Bulletin, 22, 227-237.
Hoge, S. K., Lidz, C., Mulvey, E., Roth, L., Bennett, N., Siminoff, L., et al. (1993). Patient, family, and staff perceptions of coercion in mental hospital admission: An exploratory study. Behavioral Sciences and the Law, 11, 281-293.
Hoge, S. K., Lidz, C. W., Eisenberg, M., Gardner, W., Monahan, J., Mulvey, E., et al. (1997). Perceptions of coercion in the admission of voluntary and involuntary psychiatric patients. International Journal of Law and Psychiatry, 20, 167-181.
Hoge, S. K., Lidz, C. W., Eisenberg, M., Monahan, J., Bennett, N., Gardner, W., et al. (1998). Family, clinician, and patient perceptions of coercion in mental hospital admission. A comparative study. International Journal of Law and Psychiatry, 21, 131-146.
Hora, P. F. (2002). A dozen years of drug treatment courts: Uncovering our theoretical foundation and the construction of a mainstream paradigm. Substance Use and Misuse, 37, 1469-1488.
Hosmer, D., & Lemeshow, S. (2000). Applied logistic regression (2nd Ed.). New York, NY: John Wiley & Sons.
Hser, Y. I., Anglin, M. D., Grella, C., Longshore, D., & Prendergast, M. L. (1997). Drug treatment careers. A conceptual framework and existing research findings. Journal of Substance Abuse Treatment, 14, 543-558.
Hser, Y. I., Evans, E., Huang, D., & Anglin, D. M. (2004). Relationship between drug treatment services, retention, and outcomes. Psychiatric Services, 55, 767-774.
Hser, Y. I., Grella, C. E., Hsieh, S. C., Anglin, M. D., & Brown, B. S. (1999). Prior treatment experience related to process and outcomes in DATOS. Drug and Alcohol Dependence, 57, 137-150.
Hser, Y. I., Longshore, D., & Anglin, M. D. (2007). The life course perspective on drug use: A conceptual framework for understanding drug use trajectories. Evaluation Review, 31, 515-547.
Hser, Y. I., Yamaguchi, K., Chen, J., & Anglin, M. D. (1995). Effects of interventions on relapse to narcotics addiction: An event-history analysis. Evaluation Review, 19, 123-140.
Hsieh, F. Y. (1989). Sample size tables for logistic regression. Statistics in Medicine, 8, 795-802.
Hubbard, R. L., Craddock, S. G., & Anderson, J. (2003). Overview of 5-year followup outcomes in the drug abuse treatment outcome studies (DATOS). Journal of Substance Abuse Treatment, 25, 125-134.
Hubbard, R. L., Craddock, S. G., Flynn, P. M., Anderson, J., & Etheridge, R. M. (1997). Overview of 1-year follow-up outcomes in the Drug Abuse Treatment Outcome Study (DATOS). Psychology of Addictive Behaviors, 11, 261-278.
Ivar Iversen, K., Hoyer, G., Sexton, H., & Gronli, O. K. (2002). Perceived coercion among patients admitted to acute wards in Norway. Nordic Journal of Psychiatry, 56, 433-449.
127
Jayakody, R., Danziger, S., & Pollack, H. (2000). Welfare reform, substance use, and mental health. Journal of Health Politics and Law, 25, 623-651.
Joe, G. W., Broome, K. M., Rowan-Szal, G. A., & Simpson, D. D. (2002). Measuring patient attributes and engagement in treatment. Journal of Substance Abuse Treatment, 22, 183-196.
Joe, G. W., Simpson, D. D., & Broome, K. M. (1998). Effects of readiness for drug abuse treatment on client retention and assessment of process. Addiction, 93, 1177-1190.
Joe, G. W., Simpson, D. D., & Broome, K. M. (1999). Retention and patient engagement models for different treatment modalities in DATOS. Drug and Alcohol Dependence, 57, 113-125.
Joe, G. W., Simpson, D. D., Dansereau, D. F., & Rowan-Szal, G. A. (2001). Relationships between counseling rapport and drug abuse treatment outcomes. Psychiatric Services, 52, 1223-1229.
Joe, G. W., Simpson, D. D., Greener, J. M., & Rowan-Szal, G. A. (1999). Integrative modeling of client engagement and outcomes during the first 6 months of methadone treatment. Addictive Behaviors, 24, 649-659.
Johnson, V. (1986). Intervention: How to help someone who doesn't want help. Minneapolis: Johnson Institute Books.
Joseph, J., Breslin, C., & Skinner, H. A. (1999). Critical perspectives on the transtheoretical model and the stages of change. In J. A. Tucker, D. A. Donovan & G. A. Marlatt (Eds.), Changing addictive behaviour (pp. 160-190). New York: Guilford Press.
Kaltiala-Heino, R., Laippala, P., & Salokangas, R. K. R. (1997). Impact of coercion on treatment outcome. International Journal of Law and Psychiatry, 20, 311-322.
Kaner, E. F., Beyer, F., Dickinson, H. O., Pienaar, E., Campbell, F., Schlesinger, C., et al. (2007). Effectiveness of brief alcohol interventions in primary care populations. Cochrane Database of Systematic Reviews, CD004148.
Keaton, B., & Yamatani, H. (1993). Benefits of mandatory EAP participation: A study of employees with last chance contracts. Employee Assistance Quarterly, 9, 67-77.
Kelly, J. F., Finney, J. W., & Moos, R. H. (2005). Substance use disorder patients who are mandated to treatment: Characteristics, treatment process, and 1- and 5-year outcomes. Journal of Substance Abuse Treatment, 28, 213-223.
King, R. S., & Pasquarella, J. (2009). Drug courts: A review of the evidence. Washington, DC: The Sentencing Project.
Kjellin, L., Hoyer, G., Engberg, M., Kaltiala-Heino, R., & Sigurjonsdottir, M. (2006). Differences in perceived coercion at admission to psychiatric hospitals in the Nordic countries. Social Psychiatry and Psychiatric Epidemiology, 41, 241-247.
Klag, S., Creed, P., & O'Callaghan, F. (2006). Development and initial validation of an instrument to measure perceived coercion to enter treatment for substance abuse. Psychology of Addictive Behaviors, 20, 463-470.
Klein, C., di Menza, S., Arfken, C., & Schuster, C. R. (2002). Interaction effects of treatment setting and client characteristics on retention and completion. Journal of Psychoactive Drugs, 34, 39-50.
128
Kline, A. (1997). Profiles of criminal-justice clients in drug treatment: Implications for interventions. Addictive Behaviors, 22, 263-268.
Knight, K., Hiller, M. L., Broome, K. M., & Simpson, D. D. (2000). Legal pressure, treatment readiness, and engagement in long-term residential programs. Journal of Offender Rehabilitation, 31, 101-115.
Koestner, R., & Losier, G. F. (1996). Distinguishing reactive versus reflective autonomy. Journal of Personality, 64, 465-494.
La Fond, J. Q., & Srebnik, D. (2002). The impact of mental health advance directives on patient perceptions of coercion in civil commitment and treatment decisions. International Journal of Law and Psychiatry, 25, 537-555.
La Guardia, J. G., & Patrick, H. (2008). Self-determination theory as a fundamental theory of close relationships. Canadian Psychology, 49, 201-209.
La Prairie, C., Gliksman, L., Erickson, P. G., Wall, R., & Newton-Taylor, B. (2002). Drug treatment courts--a viable option for Canada? Sentencing issues and preliminary findings from the Toronto court. Substance Use and Misuse, 37, 1529-1566.
Landau, J., Garrett, J., Shea, R. R., Stanton, M. D., Brinkman-Sull, D., & Baciewicz, G. (2000). Strength in numbers: The ARISE method for mobilizing family and network to engage substance abusers in treatment. A Relational Intervention Sequence for Engagement. American Journal of Drug and Alcohol Abuse, 26, 379-398.
Landau, J., Stanton, M. D., Brinkman-Sull, D., Ikle, D., McCormick, D., Garrett, J., et al. (2004). Outcomes with the ARISE approach to engaging reluctant drug- and alcohol-dependent individuals in treatment. American Journal of Drug and Alcohol Abuse, 30, 711-748.
Larsson-Kronberg, M., Ojehagen, A., & Berglund, M. (2005). Experiences of coercion during investigation and treatment. International Journal of Law and Psychiatry, 28, 613-621.
Lawental, E., McLellan, A. T., Grissom, G. R., Brill, P., & O'Brien, C. (1996). Coerced treatment for substance abuse problems detected through workplace urine surveillance: Is it effective? Journal of Substance Abuse, 8, 115-128.
Leigh, G., Ogborne, A. C., & Cleland, P. (1984). Factors associated with patient dropout from an outpatient alcoholism treatment service. Journal of Studies on Alcohol, 45, 359-362.
Leukefeld, C. G., & Tims, F. M. (Eds.). (1988). Compulsory treatment of drug abuse: Research and clinical practice. (NIDA Research Monograph No. 86). Rockville, MD: US Department of Health and Human Services.
Levesque, C. S., Williams, G. C., Elliot, D., Pickering, M. A., Bodenhamer, B., & Finley, P. J. (2007). Validating the theoretical structure of the Treatment Self-Regulation Questionnaire (TSRQ) across three different health behaviors. Health Education Research, 22, 691-702.
Levy Merrick, E. S., Volpe-Vartanian, J., Horgan, C., & McCann, B. (2007). Revisiting employee assistance programs and substance use problems in the workplace: Key issues and a research agenda. Psychiatric Services, 58, 1262-1264.
129
Lidz, C. W., Hoge, S. K., Gardner, W., Bennett, N. S., Monahan, J., Mulvey, E. P., et al. (1995). Perceived coercion in mental hospital admission. Pressures and process. Archives of General Psychiatry, 52, 1034-1039.
Lidz, C. W., Mulvey, E. P., Hoge, S. K., Kirsch, B. L., Monahan, J., Bennett, N. S., et al. (1997). The validity of mental patients' accounts of coercion-related behaviors in the hospital admission process. Law and Human Behavior, 21, 361-376.
Liepman, M. R., Nirenburg, T. D., & Begin, A. M. (1989). Evaluation of a program designed to help family and significant others to motivate resistant alcoholics into recovery. American Journal of Drug and Alcohol Abuse, 15, 209-221.
Link, B., Castille, D. M., & Stuber, J. (2008). Stigma and coercion in the context of outpatient treatment for people with mental illnesses. Social Science and Medicine, 67, 409-419.
Littell, J. H., & Girvin, H. (2002). Stages of change: A critique. Behavior Modification, 26, 223-273.
Logan, D. G. (1983). Getting alcoholics to treatment by social network intervention. Hospital and Community Psychiatry, 34, 360-361.
Loneck, B., Garrett, J. A., & Banks, S. M. (1996a). A comparison of the Johnson Intervention with four other methods of referral to outpatient treatment. American Journal of Drug and Alcohol Abuse, 22, 233-246.
Loneck, B., Garrett, J. A., & Banks, S. M. (1996b). The Johnson Intervention and relapse during outpatient treatment. American Journal of Drug and Alcohol Abuse, 22, 363-375.
Longshore, D., Prendergast, M., & Farabee, D. (2004). Coerced treatment for drug-using criminal offenders. In P. Bean & T. Nemitz (Eds.), Drug treatment: What works? (pp. 109-121). New York: Routledge.
Longshore, D., & Teruya, C. (2006). Treatment motivation in drug users: A theory-based analysis. Drug and Alcohol Dependence, 81, 179-188.
MacDonald, S., Bois, C., Brands, C., Dempsey, D., Erickson, P., Marsh, D., et al. (2001). Drug testing and mandatory treatment for welfare recipients. International Journal of Drug Policy, 12, 249-257.
Macdonald, S., & Wells, S. (1994). The prevalence and characteristics of employee assistance, health promotion and drug testing programs in Ontario. Employee Assistance Quarterly, 10, 25-55.
MacKenzie, C., & Stoljar, N. (Eds.). (2000). Relational autonomy: Feminist perspectives on autonomy, agency and the social self. New York: Oxford University Press.
Maglione, M., Chao, B., & Anglin, M. D. (2000). Correlates of outpatient drug treatment drop-out among methamphetamine users. Journal of Psychoactive Drugs, 32, 221-228.
Mann, R. E., Flam Zalcman, R., Stoduto, G., Nochajski, T., Koski-Jannes, A., Herie, M., et al. (2007). Evaluation of the Back on Track program: A regression-discontinuity analysis of program assignment procedures. (Back on Track Report 8). Toronto, ON: Centre for Addiction and Mental Health.
Markland, D., Ryan, R. M., Tobin, V. J., & Rollnick, S. (2005). Motivational interviewing and self-determination theory. Journal of Social and Clinical Psychology, 24, 811-831.
130
Marlowe, D. B., Festinger, D. S., Dugosh, K. L., & Lee, P. A. (2005). Are judicial status hearings a "key component" of drug court? Six and twelve months outcomes. Drug and Alcohol Dependence, 79, 145-155.
Marlowe, D. B., Festinger, D. S., Lee, P. A., Dugosh, K. L., & Benasutti, K. M. (2006). Matching judicial supervision to clients' risk status in drug court. Crime and Delinquency, 52, 52-76.
Marlowe, D. B., Glass, D. J., Merikle, E. P., Festinger, D. S., DeMatteo, D. S., Marczyk, G. R., et al. (2001). Efficacy of coercion in substance abuse treatment. In F. M. Tims, C. G. Leukefeld & J. J. Platt (Eds.), Relapse and recovery in addictions (pp. 208-227). New Haven, CT: Yale University Press.
Marlowe, D. B., Kirby, K. C., Bonieskie, L. M., Glass, D. J., Dodds, L. D., Husband, S. D., et al. (1996). Assessment of coercive and noncoercive pressures to enter drug abuse treatment. Drug and Alcohol Dependence, 42, 77-84.
Marlowe, D. B., Merikle, E. P., Kirby, K. C., Festinger, D. S., & McLellan, A. T. (2001). Multidimensional assessment of perceived treatment-entry pressures among substance abusers. Psychology of Addictive Behaviors, 15, 97-108.
Marlowe, D. B., Patapis, N. S., & DeMatteo, D. S. (2003). Amenability to treatment of drug offenders. Federal Probation, 67, 40-46.
Marshall, G. N., & Hser, Y. I. (2002). Characteristics of criminal justice and noncriminal justice clients receiving treatment for substance abuse. Addictive Behaviors, 27, 179-192.
Matthews, D. (2004). Review of employment assistance programs in Ontario Works and Ontario Disability Support Program. Toronto, ON: Ministry of Community and Social Services. Retrieved March 20, 2009 from http://www.mcss.gov.on.ca/mcss/english/ publications.
McBride, C. M., Curry, S. J., Stephens, R. S., Wells, E. A., Roffman, R. A., & Hawkins, J. D. (1994). Intrinsic and extrinsic motivation for change in cigarette smokers, marijuana smokers, and cocaine users. Psychology of Addictive Behaviors, 8, 243-250.
McKay, J. R., Foltz, C., Stephens, R. S., Leahy, P. J., Crowley, E. M., & Kissin, W. (2005). Predictors of alcohol and crack cocaine use outcomes over a 3-year follow-up in treatment seekers. Journal of Substance Abuse Treatment, 28, S73-S82.
McKay, J. R., & Weiss, R. V. (2001). A review of temporal effects and outcome predictors in substance abuse treatment studies with long-term follow-ups. Preliminary results and methodological issues. Evaluation Review, 25, 113-161.
McKenna, B. G., Simpson, A. I., & Coverdale, J. H. (2006). Outpatient commitment and coercion in New Zealand: A matched comparison study. International Journal of Law and Psychiatry, 29, 145-158.
McLellan, A. T., Carise, D., & Kleber, H. D. (2003). Can the national addiction treatment infrastructure support the public's demand for quality care? Journal of Substance Abuse Treatment, 25, 117-121.
McLellan, A. T., Gutman, M., Lynch, K., McKay, J. R., Ketterlinus, R., Morgenstern, J., et al. (2003). One-year outcomes from the CASAWORKS for Families intervention for substance-abusing women on welfare. Evaluation Review, 27, 656-680.
131
McLellan, A. T., Lewis, D. C., O'Brien, C. P., & Kleber, H. D. (2000). Drug dependence, a chronic medical illness: Implications for treatment, insurance, and outcomes evaluation. Journal of the American Medical Association, 284, 1689-1695.
McLellan, A. T., McKay, J. R., Forman, R., Cacciola, J., & Kemp, J. (2005). Reconsidering the evaluation of addiction treatment: From retrospective follow-up to concurrent recovery monitoring. Addiction, 100, 447-458.
Mechanic, D. (1968). Medical sociology. New York: Free Press.
Mertens, J. R., & Weisner, C. M. (2000). Predictors of substance abuse treatment retention among women and men in an HMO. Alcoholism: Clinical and Experimental Research, 24, 1525-1533.
Metsch, L. R., Pereyra, M., Miles, C. C., & McCoy, C. B. (2003). Welfare and work outcomes after substance abuse treatment. Social Service Review, 77, 237-254.
Meyers, R. J., Miller, W. R., Hill, D. E., & Tonigan, J. S. (1998). Community reinforcement and family training (CRAFT): Engaging unmotivated drug users in treatment. Journal of Substance Abuse, 10, 291-308.
Meyers, R. J., Miller, W. R., Smith, J. E., & Tonigan, J. S. (2002). A randomized trial of two methods for engaging treatment-refusing drug users through concerned significant others. Journal of Consulting and Clinical Psychololgy, 70, 1182-1185.
Michalak, J., Klappheck, M. A., & Kosfelder, J. (2004). Personal goals of psychotherapy patients: The intensity and the "why" of goal-motivated behavior and their implications for the therapeutic process. Psychotherapy Research, 14, 193-209.
Miller, N. S., & Flaherty, J. A. (2000). Effectiveness of coerced addiction treatment (alternative consequences): A review of the clinical research. Journal of Substance Abuse Treatment, 18, 9-16.
Miller, T. R., Zaloshnja, E., & Spicer, R. S. (2007). Effectiveness and benefit-cost of peer-based workplace substance abuse prevention coupled with random testing. Accident: Analysis and Prevention, 39, 565-573.
Miller, W. R. (1987). Motivation and treatment goals. Drugs and Society, 1, 133-151.
Miller, W. R., Meyers, R. J., & Tonigan, J. S. (1999). Engaging the unmotivated in treatment for alcohol problems: A comparison of three strategies for intervention through family members. Journal of Consulting and Clinical Psychology, 67, 688-697.
Miller, W. R., & Rollnick, S. (2002). Motivational interviewing. New York: Guilford Press.
Miller, W. R., & Wilbourne, P. L. (2002). Mesa Grande: A methodological analysis of clinical trials of treatments for alcohol use disorders. Addiction, 97, 265-277.
Ministry of Community and Social Services. (2001). Ontario Works Policy Directives. Toronto, ON: Author. Retrieved October 13, 2008 from http://www.mcss.gov.on.ca/mcss/english/ pillars/social/directives/ow_policy_directives.htm.
Monahan, J., Hoge, S. K., Lidz, C., Roth, L. H., Bennett, N., Gardner, W., et al. (1995). Coercion and commitment: Understanding involuntary mental hospital admission. International Journal of Law and Psychiatry, 18, 249-263.
132
Moos, R. H. (2005). Iatrogenic effects of psychosocial interventions for substance use disorders: Prevalence, predictors, prevention. Addiction, 100, 595-604.
Moos, R. H. (2007). Theory-based processes that promote the remission of substance use disorders. Clinical Psychology Review, 27, 537-551.
Moos, R. H., & Moos, B. S. (2003). Long-term influence of duration and intensity of treatment on previously untreated individuals with alcohol use disorders. Addiction, 98, 325-337.
Morgenstern, J., & Blanchard, K. A. (2006). Welfare reform and substance abuse treatment for welfare recipients. Alcohol Research and Health, 29, 63-67.
Morgenstern, J., Blanchard, K. A., McCrady, B. S., McVeigh, K. H., Morgan, T. J., & Pandina, R. J. (2006). Effectiveness of intensive case management for substance-dependent women receiving temporary assistance for needy families. American Journal of Public Health, 96, 2016-2023.
Morgenstern, J., Neighbors, C. J., Kuerbis, A., Riordan, A., Blanchard, K. A., McVeigh, K. H., et al. (2009). Improving 24-month abstinence and employment outcomes for substance-dependent women receiving temporary assistance for needy families with intensive case management. American Journal of Public Health, 99, 328-333.
Moyer, A., Finney, J. W., Swearingen, C. E., & Vergun, P. (2002). Brief interventions for alcohol problems: A meta-analytic review of controlled investigations in treatment-seeking and non-treatment-seeking populations. Addiction, 97, 279-292.
Nabalamba, A., & Millar, W. J. (2007). Going to the doctor. Health Reports, 18, 23-35.
Nace, E. P., Birkmayer, F., Sullivan, M. A., Galanter, M., Fromson, J. A., Frances, R. J., et al. (2007). Socially sanctioned coercion mechanisms for addiction treatment. American Journal on Addictions, 16, 15-23.
Nicholson, R. A., Ekenstam, C., & Norwood, S. (1996). Coercion and the outcome of psychiatric hospitalization. International Journal of Law and Psychiatry, 19, 201-217.
Niv, N., Hamilton, A., & Hser, Y. I. (2009). Impact of court-mandated substance abuse treatment on clinical decision making. Journal of Behavioral Health Services and Research, 36, 505-516.
Norman, G. R., & Streiner, D. L. (2000). Biostatistics: The bare essentials. (2nd ed.). Hamilton, ON: BC Decker.
Orford, J. (1992). Control, confront or collude: How family and society respond to excessive drinking. British Journal of Addiction, 87, 1513-1525.
Orlando, M., Chan, K. S., & Morral, A. R. (2003). Retention of court-referred youths in residential treatment programs: Client characteristics and treatment process effects. American Journal of Drug and Alcohol Abuse, 29, 337-357.
Peel Data Centre. (2006). 2006 Census. Retrieved September 8, 2009, from http://www.peelregion.ca/planning/pdc/data/
Pelletier, L. G., Tuson, K. M., & Haddad, N. K. (1997). Client Motivation for Therapy Scale: A measure of intrinsic motivation, extrinsic motivation, and amotivation for therapy. Journal of Personality Assessment, 68, 414-435.
133
Perry, A., Coulton, S., Glanville, J., Godfrey, C., Lunn, J., McDougall, C., et al. (2006). Interventions for drug-using offenders in the courts, secure establishments and the community. Cochrane Database of Systematic Reviews, CD005193.
Platt, J. J., Buhringer, G., Kaplan, C. D., Brown, B. S., & Taube, D. O. (1988). The prospects and limitations of compulsory treatment for drug addiction. Journal of Drug Issues, 18, 505-525.
Polcin, D. L. (2001). Drug and alcohol offenders coerced into treatment: A review of modalities and suggestions for research on social model programs. Substance Use and Misuse, 36, 589-608.
Polcin, D. L. (2003). Rethinking confrontation in alcohol and drug treatment: Consideration of the clinical context. Substance Use and Misuse, 38, 165-184.
Polcin, D. L., & Beattie, M. (2007). Relationship and institutional pressure to enter treatment: Differences by demographics, problem severity, and motivation. Journal of Studies on Alcohol and Drugs, 68, 428-436.
Polcin, D. L., & Weisner, C. (1999). Factors associated with coercion in entering treatment for alcohol problems. Drug and Alcohol Dependence, 54, 63-68.
Pollack, H. A., Danziger, S. K., Jayakody, R., & Seefeldt, K. S. (2002). Drug testing welfare recipients: False positives, false negatives, unanticipated opportunities. Women's Health Issues, 12, 23-31.
Porporino, F. J., Robinson, D., Millson, B., & Weekes, J. (2002). An outcome evaluation of prison-based treatment programming for substance users. Substance Use and Misuse, 37, 1047-1077.
Poythress, N. G., Petrila, J., McGaha, A., & Boothroyd, R. (2002). Perceived coercion and procedural justice in the Broward mental health court. International Journal of Law and Psychiatry, 25, 517-533.
Prendergast, M., Greenwell, L., Farabee, D., & Hser, Y. I. (2009). Influence of perceived coercion and motivation on treatment completion and re-arrest among substance-abusing offenders. Journal of Behavioral Health Services and Research, 36, 159-176.
Prendergast, M. L., Podus, D., Chang, E., & Urada, D. (2002). The effectiveness of drug abuse treatment: A meta-analysis of comparison group studies. Drug and Alcohol Dependence, 67, 53-72.
Prochaska, J. O., & DiClemente, C. C. (1986). Toward a comprehensive model of change. In W. R. Miller & N. Heather (Eds.), Treating addictive behaviors: Processes of change (pp. 3-27). New York: Plenum.
Prochaska, J. O., DiClemente, C. C., & Norcross, J. C. (1992). In search of how people change: Applications to addictive behaviors. American Psychologist, 47, 1102-1114.
Project MATCH Research Group. (1997). Matching alcoholism treatments to client heterogeneity: Project MATCH posttreatment drinking outcomes. Journal of Studies on Alcohol, 58, 7-29.
Pulford, J., & Wheeler, A. (2007). Documenting client attendance norms: Raw data and implications for treatment practice. Journal of Substance Use, 12, 95-102.
134
Rain, S. D., Steadman, H. J., & Robbins, P. C. (2003). Perceived coercion and treatment adherence in an outpatient commitment program. Psychiatric Services, 54, 399-401.
Rain, S. D., Williams, V. F., Robbins, P. C., Monahan, J., Steadman, H. J., & Vesselinov, R. (2003). Perceived coercion at hospital admission and adherence to mental health treatment after discharge. Psychiatric Services, 54, 103-105.
Rehm, J., Baliunas, D., Brochu, S., Fischer, B., Gnam, W., Patra, J., et al. (2006). The costs of substance abuse in Canada 2002. Ottawa, ON: Canadian Centre on Substance Abuse.
Reis, H. T., Sheldon, K. M., Gable, S. L., Roscoe, J., & Ryan, R. M. (2000). Daily well-being: The role of autonomy, competence, and relatedness. Personality and Social Psychology Bulletin, 26, 419-435.
Rigg, J. (2002). Measures of perceived coercion in prison treatment settings. International Journal of Law and Psychiatry, 25, 473-490.
Roll, J. M., Prendergast, M., Richardson, K., Burdon, W., & Ramirez, A. (2005). Identifying predictors of treatment outcome in a drug court program. American Journal of Drug and Alcohol Abuse, 31, 641-656.
Rollnick, D., Heather, N., Gold, R., & Hall, W. (1992). Development of a short "readiness to change" questionnaire for use in brief, opportunistic interventions among excessive drinkers. British Journal of Addiction, 87, 743-754.
Roman, P. M. (1988). Growth and transformation in workplace alcoholism programming. In M. Galanter (Ed.), Recent developments in alcoholism (Vol. 6, pp. 131-158). New York: Plenum Press.
Room, R. (1980). Treatment-seeking populations and larger realities. In G. Edwards & M. Grant (Eds.), Alcoholism treatment in transition (pp. 205-224). Baltimore, MD: University Park Press.
Room, R. (1989). The U.S. general population's experiences of responding to alcohol problems. British Journal of Addiction, 84, 1291-1304.
Room, R., Bondy, S., & Ferris, J. (1996). Determinants of suggestions for alcohol treatment. Addiction, 91, 643-655.
Room, R., Greenfield, T. K., & Weisner, C. (1991). "People who might have liked you to drink less": Changing responses to drinking by U.S. family members and friends, 1979-1990. Contemporary Drug Problems, 18, 573-595.
Room, R., Matzger, H., & Weisner, C. (2004). Sources of informal pressure on problematic drinkers to cut down or seek treatment. Journal of Substance Use, 9, 280-295.
Rosenstock, I. M. (1974). Historical origins of the health belief model. Health Education Monographs, 2, 328-335.
Rush, B. R., & Wild, T. C. (2003). Substance abuse treatment and pressures from the criminal justice system: Data from a provincial client monitoring system. Addiction, 98, 1119-1128.
Ryan, R. M., & Connell, J. P. (1989). Perceived locus of causality and internalization: Examining reasons for acting in two domains. Journal of Personality and Social Psychology, 57, 749-761.
135
Ryan, R. M., & Deci, E. L. (2000). Self-determination theory and the facilitation of intrinsic motivation, social development, and well-being. American Psychologist, 55, 68-78.
Ryan, R. M., & Deci, E. L. (2002). Overview of self-determination theory: an organismic dialectical perspective. In E. L. Deci & R. M. Ryan (Eds.), Handbook of self-determination research (pp. 3-36). Rochester, NY: University of Rochester Press.
Ryan, R. M., & Deci, E. L. (2008). A Self-Determination Theory approach to psychotherapy: The motivational basis for effective change. Canadian Psychology, 49, 186-193.
Ryan, R. M., Plant, R. W., & O'Malley, S. (1995). Initial motivations for alcohol treatment: Relations with patient characteristics, treatment involvement, and dropout. Addictive Behaviors, 20, 279-297.
Ryan, R. M., & Solky, J. A. (1996). What is supportive about social support? On the psychological needs for autonomy and relatedness. In G. R. Peirce, B. R. Sarason & I. G. Sarason (Eds.), Handbook of social support and the family (pp. 249-267). New York: Plenum.
Sallmen, B., Berglund, M., & Bokander, B. (1998). Perceived coercion related to psychiatric comorbidity and locus of control in institutionalized alcoholics. Medicine and Law, 17, 381-391.
Sanford, J. S., & Arrigo, B. A. (2005). Lifting the cover on drug courts: Evaluation findings and policy concerns. International Journal of Offender Therapy and Comparative Criminololgy, 49, 239-259.
Schacht Reisinger, H., Bush, T., Colom, M., Agar, M., & Battjes, R. (2003). Navigation and engagement: How does one measure success? Journal of Drug Issues, 33, 777-800.
Schmidt, L., Dohan, D., Wiley, J., & Zabkiewicz, D. (2002). Addiction and welfare dependency: Interpreting the connection. Social Problems, 49, 221-241.
Schmidt, L., & Weisner, C. (1993). Developments in alcoholism treatment. In M. Galanter (Ed.), Recent developments in Alcoholism. (Vol. 11, pp. 369-396). New York: Plenum Press.
Schmidt, L., Weisner, C., & Wiley, J. (1998). Substance abuse and the course of welfare dependency. American Journal of Public Health, 88, 1616-1622.
Schmidt, L., Zabkiewicz, D., Jacobs, L., & Wiley, J. (2007). Substance abuse and employment among welfare mothers: From welfare to work and back again? Substance Use and Misuse, 42, 1069-1087.
Shane, P., Jasiukaitis, P., & Green, R. S. (2003). Treatment outcomes among adolescents with substance abuse problems: The relationship between comorbidities and post-treatment substance involvement. Evaluation and Program Planning, 26, 393-402.
Sharpley, J., Flam Zalcman, R., Mann, R. E., Brands, B., & Stoduto, G. (2007). Outcome evaluation of the Back on Track program: Program impact on alcohol and drug use and related problems. (Back on Track Report 7). Toronto, ON: Centre for Addiction and Mental Health.
Sheldon, K. M., & Bettencourt, B. A. (2002). Psychological need-satisfaction and subjective well-being within social groups. British Journal of Social Psychology, 41, 25-38.
136
Sia, T. L., Dansereau, D. F., & Czuchry, M. L. (2000). Treatment readiness training and probationers' evaluation of substance abuse treatment in a criminal justice setting. Journal of Substance Abuse Treatment, 19, 459-467.
Simoneau, H., & Bergeron, J. (2003). Factors affecting motivation during the first six weeks of treatment. Addictive Behaviors, 28, 1219-1241.
Simpson, D. (1993). Drug treatment evaluation research in the United States. Psychology of Addictive Behaviors, 7, 120-128.
Simpson, D. D. (2004). A conceptual framework for drug treatment process and outcomes. Journal of Substance Abuse Treatment, 27, 99-121.
Simpson, D. D., & Joe, G. W. (1993). Motivation as a predictor of early dropout from drug abuse treatment. Psychotherapy, 30, 357-368.
Simpson, D. D., & Joe, G. W. (2004). A longitudinal evaluation of treatment engagement and recovery stages. Journal of Substance Abuse Treatment, 27, 89-97.
Simpson, D. D., Joe, G. W., Broome, K. M., Hiller, M. L., Knight, D. K., & Rowan-Szal, G. (1997). Program diversity and treatment retention rates in the Drug Abuse Treatment Outcome Study (DATOS). Psychology of Addictive Behaviors, 11, 279-293.
Simpson, D. D., Joe, G. W., & Brown, B. S. (1997). Treatment retention and follow-up outcomes in the Drug Abuse Treatment Outcome Study (DATOS). Psychology of Addictive Behaviors, 11, 294-307.
Simpson, D. D., Joe, G. W., Dansereau, D. F., & Chatham, L. R. (1997). Strategies for improving methadone treatment process and outcomes. Journal of Drug Issues, 27, 239-260.
Simpson, D. D., Joe, G. W., Greener, J. M., & Rowan-Szal, G. A. (2000). Modeling year 1 outcomes with treatment process and post-treatment social influences. Substance Use and Misuse, 35, 1911-1930.
Simpson, D. D., Joe, G. W., Rowan-Szal, G., & Greener, J. (1995). Client engagement and change during drug abuse treatment. Journal of Substance Abuse, 7, 117-134.
Simpson, D. D., Joe, G. W., & Rowan-Szal, G. A. (1997). Drug abuse treatment retention and process effects on follow-up outcomes. Drug and Alcohol Dependence, 47, 227-235.
Simpson, D. D., Joe, G. W., Rowan-Szal, G. A., & Greener, J. M. (1997). Drug abuse treatment process components that improve retention. Journal of Substance Abuse Treatment, 14, 565-572.
Slaymaker, V. J., & Owen, P. L. (2006). Employed men and women substance abusers: job troubles and treatment outcomes. Journal of Substance Abuse Treatment, 31, 347-354.
Smart, R. G., Mann, R. E., Stoduto, G., & Thomas, R. (2008). Do impaired driving rehabilitation programs need improvement? Paper presented at the Kettil Bruun Society Annual Meeting.
Snyder, L. (2006). Workfare: ten years of pickin' on the poor. In A. Westhues (Ed.), Canadian social policy: issues and perspectives. (4th ed.). Waterloo, ON: Wilfrid Laurier University Press.
137
Sobell, L. C., Cunningham, J. A., & Sobell, M. B. (1996). Recovery from alcohol problems with and without treatment: Prevalence in two population surveys. American Journal of Public Health, 86, 966-972.
Sobell, L. C., Sobell, M. B., & Toneatto, T. (1992). Recovery from alcohol problems without treatment. In N. Heather, W. R. Miller & J. R. Greenley (Eds.), Self control and the addictive behaviours. New York: Maxwell Macmillan.
Sowers, W. E., & Daley, D. C. (1993). Compulsory treatment of substance use disorders. Criminal Behavior and Mental Health, 3, 403-415.
Stahler, G. J., Cohen, E., Greene, M. A., Shipley, T. E., & Bartelt, D. (1995). A qualitative study of treatment success among homeless crack-addicted men: Definitions and attributions. Contemporary Drug Problems, 22, 237-264.
Stahler, G. J., Cohen, E., Shipley, T. E., & Bartelt, D. (1993). Why clients drop out of treatment: Ethnographic perspectives on treatment attrition among homeless male "crack" cocaine users. Contemporary Drug Problems, 20, 651-680.
Staines, G., Magura, S., Rosenblum, A., Fong, C., Kosanke, N., Foote, J., et al. (2003). Predictors of drinking outcomes among alcoholics. American Journal of Drug and Alcohol Abuse, 29, 203-218.
Stanhope, V., Marcus, S., & Solomon, P. (2009). The impact of coercion on services from the perspective of mental health care consumers with co-occurring disorders. Psychiatric Services, 60, 183-188.
Stark, M. (1992). Dropping out of substance abuse treatment: A clinically oriented review. Clinical Psychology Review, 12, 93-116.
Steadman, H. J., Gounis, K., Dennis, D., Hopper, K., Roche, B., Swartz, M., et al. (2001). Assessing the New York City involuntary outpatient commitment pilot program. Psychiatric Services, 52, 330-336.
Stevens, A., Berto, D., Frick, U., Hunt, N., Kerschl, V., McSweeney, T., et al. (2006). The relationship between legal status, perceived pressure and motivation in treatment for drug dependence: Results from a European study of quasi-compulsory treatment. European Addiction Research, 12, 197-209.
Stevens, A., Berto, D., Heckman, W., Kerschl, V., Oeuvray, K., Van Ooyen, M., et al. (2005). Quasi-compulsory treatment of drug dependent offenders: An international literature review. Substance Use and Misuse, 40, 269-283.
Substance Abuse and Mental Health Services Administration. (2008). Treatment Episode Data Set (TEDS). Highlights 2006. National Admissions to Substance Abuse Treatment Services. DASIS Series: S-40, DHHS Publication No. (SMA) 08-4313. Rockville, MD: US Department of Health and Human Services.
Suchman, E. A. (1970). Health attitudes and behavior. Archives of Environmental Health, 20, 105.
Sullivan, M. A., Birkmayer, F., Boyarsky, B. K., Frances, R. J., Fromson, J. A., Galanter, M., et al. (2008). Uses of coercion in addiction treatment: Clinical aspects. American Journal on Addictions, 17, 36-47.
138
Sung, H., Belenko, S., & Feng, L. (2001). Treatment compliance in the trajectory of treatment progress among offenders. Journal of Substance Abuse Treatment, 20, 153-162.
Sung, H. E., Belenko, S., Feng, L., & Tabachnick, C. (2004). Predicting treatment noncompliance among criminal justice-mandated clients: A theoretical and empirical exploration. Journal of Substance Abuse Treatment, 26, 315-328.
Sutton, S. (1996). Can 'stages of change' provide guidance in the treatment of addictions? A critical examination of Prochaska and Diclemente's model. In G. Edwards & C. Dare (Eds.), Psychotherapy, psychological treatments and the addictions (pp. 189-205). Cambridge: Cambridge University Press.
Swanson, J. W., Swartz, M. S., Elbogen, E. B., Wagner, H. R., & Burns, B. J. (2003). Effects of involuntary outpatient commitment on subjective quality of life in persons with severe mental illness. Behavioral Sciences and the Law, 21, 473-491.
Swartz, M., Wagner, J. H., Swanson, J., Hiday, V., & Burns, B. (2002). The perceived coerciveness of involuntary outpatient commitment: Findings from an experimental study. Journal of the American Academy of Psychiatry and Law, 30, 207-217.
Taborda, J. G. V., Bapista, J. P., Gomes, D. A. R., Nogueira, L., & Chaves, M. L. F. (2004). Perception of coercion in psychiatric and nonpsychiatric (medical and surgical) inpatients. International Journal of Law and Psychiatry, 27, 179-192.
Taxman, F. S., & Bouffard, J. (2002). Treatment inside the drug treatment court: The who, what, where, and how of treatment services. Substance Use and Misuse, 37, 1665-1688.
Trice, H., & Sonnenstuhl, W. (1988). Constructive confrontation and other referral processes. In M. Galanter (Ed.), Recent developments in alcoholism (Vol. 6, pp. 159-170). New York: Plenum Press.
Trice, H. M., & Beyer, J. M. (1984). Work-related outcomes of the constructive-confrontation strategy in a job-based alcoholism program. Journal of Studies on Alcohol, 45, 393-404.
Tucker, J., & Anders, S. (2001). Social control of health behaviors in marriage. Journal of Applied Social Psychology, 31, 467-485.
Tucker, J. A., Vuchinich, R. E., & Rippens, P. D. (2004). A factor analytic study of influences on patterns of help-seeking among treated and untreated alcohol dependent persons. Journal of Substance Abuse Treatment, 26, 237-242.
Upshur, R. E. G. (2002). Principles for the justification of public health intervention. Canadian Journal of Public Health, 93, 101-103.
Urbanoski, K., & Wild, T. (2009). Assessing self-determined motivation for addiction treatment: construct validation of the Treatment Entry Questionnaire. Manuscript in preparation.
Vallerand, R. J., Pelletier, L. G., & Koestner, R. (2008). Reflections on Self-Determination Theory. Canadian Psychology, 49, 257-262.
Vansteenkiste, M., & Sheldon, K. M. (2006). There's nothing more practical than a good theory: Integrating motivational interviewing and self-determination theory. British Journal of Clinical Psychology, 45, 63-82.
139
Vasilaki, E. I., Hosier, S. G., & Cox, W. M. (2006). The efficacy of motivational interviewing as a brief intervention for excessive drinking: A meta-analytic review. Alcohol and Alcoholism, 41, 328-335.
Vickers-Lahti, M., Garfield, F., McCusker, J., Hindin, R., Bigelow, C., Love, C., et al. (1995). The relationship between legal factors and attrition from a residential drug abuse treatment program. Journal of Psychoactive Drugs, 27, 17-25.
Walker, M. A. (2008). Program characteristics and the length of time clients are in substance abuse treatment. Journal of Behavioral Health Services and Research, 36, 330-343.
Walker, R. (2009). Retention in treatment--indicator or illusion: An essay. Substance Use and Misuse, 44, 18-27.
Walsh, D. C., Hingson, R. W., Merrigan, D. M., Levenson, S. M., Cupples, L. A., Heeren, T., et al. (1991). A randomized trial of treatment options for alcohol-abusing workers. New England Journal of Medicine, 325, 775-782.
Weisner, C. (1987). The social ecology of alcohol treatment in the United States. In M. Galanter (Ed.), Recent Developments in Alcohol (Vol. 5, pp. 203-243). New York: Plenum Press.
Weisner, C. (1990). Coercion in alcohol treatment. In Institute of Medicine (Ed.), Broadening the base of treatment for alcohol problems. Washington, DC: National Academy Press.
Weisner, C., & Matzger, H. (2002). A prospective study of the factors influencing entry to alcohol and drug treatment. Journal of Behavioral Health Services and Research, 29, 126-137.
Weisner, C., Mertens, J., Tam, T., & Moore, C. (2001). Factors affecting the initiation of substance abuse treatment in managed care. Addiction, 96, 705-716.
Wells-Parker, E., Bangert-Drowns, R., McMillen, R., & Williams, M. (1995). Final results from a meta-analysis of remedial interventions with drink/drive offenders. Addiction, 90, 907-926.
Werb, D., Elliott, R., Fischer, B., Wood, E., Montaner, J., & Kerr, T. (2007). Drug treatment courts in Canada: An evidence-based review. HIV/AIDS Policy and Law Review, 12, 12-17.
West, R. (2005). Time for a change: Putting the transtheoretical (stages of change) model to rest. Addiction, 100, 1036-1039.
White, R. K., McDuff, D. R., Schwartz, R. P., Tiegel, S. A., & Judge, C. P. (1996). New developments in employee assistance programs. Psychiatric Services, 47, 387-391.
Wickizer, T. M., Campbell, K., Krupski, A., & Stark, K. (2000). Employment outcomes among AFDC recipients treated for substance abuse in Washington State. Milbank Quarterly, 78, 585-608.
Wild, T. C. (1999a). Compulsory substance-user treatment and harm reduction: A critical analysis. Substance Use and Misuse, 34, 83-102.
Wild, T. C. (1999b). Treatment Entry Questionnaire (Version 1): User's Guide. Toronto: Addiction Research Foundation.
Wild, T. C. (2006). Social control and coercion in addiction treatment: Towards evidence-based policy and practice. Addiction, 101, 40-49.
140
Wild, T. C., Cunningham, J., & Ryan, R. M. (2006). Social pressure, coercion, and client engagement at treatment entry: A self-determination theory perspective. Addictive Behaviors, 31, 1858-1872.
Wild, T. C., Newton-Taylor, B., & Alletto, R. (1998). Perceived coercion among clients entering substance abuse treatment: Structural and psychological determinants. Addictive Behaviors, 23, 81-95.
Wild, T. C., Roberts, A. B., & Cooper, E. L. (2002). Compulsory substance abuse treatment: An overview of recent findings and issues. European Addiction Research, 8, 84-93.
Williams, G. C. (2002). Improving patients' health through supporting the autonomy of patients and providers. In E. L. Deci & R. M. Ryan (Eds.), Handbook of Self-Determination Research (pp. 233-254). Rochester, NY: University of Rochester Press.
Williams, G. C., Cox, E. M., Kouides, R. W., & Deci, E. L. (1999). Presenting the facts about smoking to adolescents: The effects of an autonomy supportive style. Archives of Pediatrics and Adolescent Medicine, 153, 959-964.
Williams, G. C., Freedman, Z., & Deci, E. L. (1998). Supporting autonomy to motivate patients with diabetes for glucose control. Diabetes Care, 21, 1644-1651.
Williams, G. C., Gagne, M., Ryan, R. M., & Deci, E. L. (2002). Facilitating autonomous motivation for smoking cessation. Health Psychology, 21, 40-50.
Williams, G. C., Grow, V. M., Freedman, Z., Ryan, R. M., & Deci, E. L. (1996). Motivational predictors of weight loss and weight-loss maintenance. Journal of Personality and Social Psychology, 70, 115-126.
Williams, G. C., Lynch, M. F., McGregor, H. A., Ryan, R. M., Sharp, D., & Deci, E. L. (2006). Validation of the "Important Other" Climate Questionnaire: Assessing autonomy support for health-related change. Families, Systems & Health, 24, 179-194.
Williams, G. C., McGregor, H. A., Sharp, D., Levesque, C. S., Kouides, R. W., Ryan, R. M., et al. (2006). Testing a self-determination theory intervention for motivating tobacco cessation: Supporting autonomy and competence in a clinical trial. Health Psychology, 25, 91-101.
Williams, G. C., McGregor, H. A., Zeldman, A., Freedman, Z., & Deci, E. L. (2004). Testing a self-determination theory process model for promoting glycemic control through diabetes self-management. Health Psychology, 23, 58-66.
Williams, G. C., Rodin, G. C., Ryan, R. M., Grolnick, W. S., & Deci, E. L. (1998). Autonomous regulation and long-term medication adherence in adult outpatients. Health Psychology, 17, 269-276.
Young, D. (2002). Impacts of perceived legal pressure on retention in drug treatment. Criminal Justice and Behavior, 29, 27-55.
Young, D., & Belenko, S. (2002). Program retention and perceived coercion in three models of mandatory drug treatment. Journal of Drug Issues, 32, 297-328.
Zeldman, A., Ryan, R. M., & Fiscella, K. (2004). Motivation, autonomy support, and entity beliefs: Their role in methadone maintenance treatment. Journal of Social and Clinical Psychology, 23, 675-696.
141
Zhang, Z., Friedmann, P. D., & Gerstein, D. R. (2003). Does retention matter? Treatment duration and improvement in drug use. Addiction, 98, 673-684.
Zuroff, D. C., Koestner, R., Moskowitz, D. S., McBride, C., Marshall, M., & Bagby, R. M. (2007). Autonomous motivation for therapy: A new common factor in brief treatments for depression. Psychotherapy Research, 17, 137-147.
142
Appendices Appendix 1. Measures Social Pressures Index For each of the following, please indicate the extent to which the following people pressured you in the past 2 months to enter treatment for your drug and/or alcohol use. (For instance, if you do not feel that you were pressured by your spouse to enter treatment, circle “1”):e
No
pressure Extreme
pressure Spouse, girlfriend, or boyfriend ………………………..
1 2 3 4 5
Family member (i.e., mother, father, brother, sister, aunt, uncle, etc.) ………………………………………..
1 2 3 4 5
Friend …………………………………………………..
1 2 3 4 5
Employer or colleague …………………………………
1 2 3 4 5
Legal authority (i.e., police officer, parole officer, judge, lawyer) ………………………………………….
1 2 3 4 5
Children’s aid worker ………………………………….
1 2 3 4 5
Ontario Works or ODSP worker ………………………
1 2 3 4 5
Physician or health care worker ………………………..
1 2 3 4 5
Other: _____________________ ………………...........
1 2 3 4 5
MacArthur Perceived Coercion Scale (MPCS) Please indicate if the following statements are true or false: I had more influence than anyone else about whether I went into treatment……… True False
I had a lot of control over whether I went into treatment…………………………. True False
I chose to come here for treatment………………………………………………… True False
I felt free to do what I wanted about coming here for treatment…………………… True False
It was my idea to come into treatment……………………………………………… True False
e Scores were rescaled to 0 to 4 for analysis
143
Treatment Entry Questionnaire (TEQ) Please indicate whether you agree or disagree with each of the following statements by placing the number that best reflects your own personal opinion in the blank provided. Remember, there are no right or wrong answers, and your responses are completely confidential.
Use the following scale to make your ratings 1 2 3 4 5 6 7
Strongly Disagree Strongly Agree
1. I decided to enter a program because I was interested in getting help. ___ 2. I decided to enter a program because I won’t like myself very much unless my substance abuse problem is under control. ___ 3. If I remain in treatment it will probably be because others will be angry with me if I don’t. ___ 4. I decided to enter a program because I really want to make some changes in my life. ___ 5. I plan to go through with treatment because I’ll be ashamed of myself if I don’t. ___ 6. The reason I am in treatment is because other people have pressured me into being here. ___ 7. If I remain in treatment it will probably be because I’ll feel like a failure if I don’t. ___ 8. I decided to enter a program because it feels important for me personally to deal with my substance abuse problem. ___ 9. I have agreed to follow a treatment program because I will get in trouble with my friends and family if I don't follow all the guidelines. ___ 10. I plan to go through with a treatment program because not abusing alcohol and drugs is a choice I really want to make. ___ 11. If I remain in treatment it will probably be because I’ll feel very bad about myself if I don’t. ___ 12. I have agreed to follow a treatment program because I was pressured to come. ___
144
Substance Problem Scale (SPS) The following is a list of common problems related to alcohol and drug use. Please indicate when was the last time that you had each problem (For example, if you had the problem in the past month, circle “1”. If you had the problem over 1 year ago, or if you never had that kind of problem, circle “3”). When was the LAST time that:
In the past
month
2-12 months ago
> 12 months ago or never
a) you tried to hide that you were using alcohol or drugs? …………………………………………………….
1 2 3
b) your parents, family, partner, co-workers, classmates or friends complained about your alcohol or drug use? ………………………...
1 2 3
c) you used alcohol or drugs weekly? ……………………………….
1 2 3
d) your alcohol or drug use caused you to feel depressed, nervous, suspicious, uninterested in things, reduced your sexual drive? ……...
1
2
3
e) your alcohol or drug use caused you to have numbness, tingling, shakes, blackouts, hepatitis, TB, sexually transmitted disease or any other health problems? ……………………….....................................
1
2
3
f) you kept using alcohol or drugs even though you knew it was keeping you from meeting your responsibilities at work, school, or home? ………………………………………………………………...
1
2
3
g) you used alcohol and drugs where it made the situation unsafe or dangerous for you, such as when you were driving a car, using a machine, or where you might have been forced into sex or hurt? …...
1
2
3
h) your alcohol or drug use caused you to have repeated problems with the law? …………………………………………………………
1
2
3
i) you kept using alcohol or drugs even after you knew it could get you into fights or other kinds of legal trouble? ……………………
1
2
3
j) you needed more alcohol or drugs to get the same high or found that the same amount did not get you as high as it used to? …………
1
2
3
145
When was the last time that:
In the past
month
2-12 months ago
> 12 months ago or never
k) you had withdrawal problems from alcohol or drugs like shaking hands, throwing up, having trouble sitting still or sleeping, or that you used any alcohol or drugs to stop being sick or avoid withdrawal problems? ………………………….....................................................
1
2
3
l) you used alcohol or drugs in larger amounts, more often or for a longer time than you meant to? ……………………………………
1
2
3
m) you were unable to cut down or stop using alcohol or drugs? .......
1 2 3
n) you spend a lot of time either getting alcohol and drugs, using alcohol or drugs, or feeling the effects of alcohol or drugs (high or sick)? ………………………………………………............................
1
2
3
o) your use of alcohol and drugs caused you to give up, reduce or have problems at important activities at work, school, home or social events? ……………………………………………………………..
1
2
3
p) you kept using alcohol or drugs even after you knew it was causing or adding to medical, psychological or emotional problems you were having? ….………………………………............................
1
2
3
146
Treatment Engagement I’m going to ask you some questions about your experiences at PAARC. Please answer the questions as truthfully as you can. Your answers will not be shared with your counsellor or any of the other staff at PAARC. Subscales: [Con] Confidence in treatment [Rap] Counsellor rapport [Com] Commitment to treatment * 1. Please think about all the treatment you have received in this program since your admission.
Overall, how helpful has this treatment been? Not at all Somewhat Very helpful
[Con] 2. Has this treatment helped you stop or cut down on your drug use? Yes No [Con] 3. Would you say it has helped … ? A little
A lot [Con] 4. How much would you say this program helped you with other problems (besides drug
problems) since your admission? Not at all
A little A lot [Con] 5. How likely are you to stay in this program until you complete your treatment? Not at all Somewhat Very likely [Rap] 6. How much do you feel your current counsellor agrees with you about what would be useful
goals for your treatment? Not at all
A little bit Very much
[Rap] 7. How much does your counsellor show a sincere desire to understand you and your problems? Not at all A little bit Very much
147
[Rap] 8. How much do you feel that you are working together with your counsellor that the two of you are joined in a struggle to overcome your problems?
Not at all A little bit Very much [Rap] 9. How satisfied do you feel with treatment so far? Not at all A little bit Very much [Rap] 10. How much has the treatment you have received in this program so far matched with your
ideas about what helps people in treatment? Not at all
A little bit Very much
I’m going to read a list of feelings that some people have about themselves and their drug use, or problems related to their drug use. Please tell me how much you agree or disagree with each statement as it applies to you (strongly disagree, somewhat disagree, somewhat agree, strongly agree) [Com] 11. I am doing something about the problems that have been bothering me
Strongly disagree Somewhat disagree Somewhat agree Strongly agree
[Com] 12. At times my problems are difficult, but I’m working on them Strongly disagree Somewhat disagree Somewhat agree Strongly agree
[Com] 13. Being here is a waste of time because I am not the cause of my problems
Strongly disagree Somewhat disagree Somewhat agree Strongly agree
[Com] 14. Even though I’m not always successful in changing, I am at least working on my problems
Strongly disagree Somewhat disagree Somewhat agree Strongly agree
148
* 15. Sometimes I still find myself struggling with problems that I thought I had taken care of once and for all Strongly disagree Somewhat disagree Somewhat agree Strongly agree
[Con] 16. Maybe this place will be able to help me.
Strongly disagree Somewhat disagree Somewhat agree Strongly agree
[Com] 17. Anyone can talk about changing, but I’m actually doing something about it.
Strongly disagree Somewhat disagree Somewhat agree Strongly agree
* Item 1 was originally used in error instead of item 16, and item 15 was original used in error instead of item 17.
Appendix 2.
149
Consent letter
Social control and coerced addiction treatment: impact on treatment participation and progress
Researchers:
Karen Urbanoski, PhD (Candidate) 416-535-8501 x6121
Edward Adlaf, PhD
Brian Rush, PhD Jurgen Rehm, PhD
Letter of Information:
Dear client, We are researchers interested in conducting a study that examines the relationships between the reasons for seeking substance abuse treatment, motivation, and participation in treatment. We are asking you to participate in our study because you are a client of the Peel Addiction Assessment and Referral Centre (PAARC). This project is being done for a PhD thesis at the University of Toronto, and about 250 people are being asked to participate. The purpose of the study is to measure the relationships between pressures to enter treatment and motivation, and how this affects your progress in treatment. We believe that the results will give us important information about the effectiveness of treatment. We ask your permission to use the information that is being collected from you today and over the next months while you are a client at PAARC, by the therapist or assessment worker. We are also going to ask you to complete a pencil and paper questionnaire asking you about why you are here today. Other questions will ask about problems you may have because of your use of alcohol or drugs. This will take approximately 15-20 minutes of your time today. If you would like to participate in our study but can not do it today, we can set up another time that is more convenient for you. We will contact you again in about one month to ask you some additional questions about your satisfaction with your treatment at PAARC, and about your use of drugs and alcohol. This interview will be conducted by telephone or at PAARC, whichever is most convenient for you. This interview will take about 20-25 minutes. Today we will ask you for your phone number and a second phone number, such as a family member or a friend, where we might be able to reach you in one month. We will not use the second phone number unless we have to. If we do use it, we will not disclose the reason for the phone call in order to protect your privacy. If you feel strongly that you do not wish to provide us with a second number, you do not have to. You will receive $5 in cash for your participation following the first interview today, and $10 following the interview in one month. Your participation is completely voluntary. If you decide
150
that you do not want to allow us to use your information, either now or at any point during your treatment at PAARC, your treatment will not be affected in any way. Non-participation will in no way impact on the level of care or services that you will receive at PAARC. All the information that we collect will be kept strictly confidential to the extent permitted by law. Only members of the research team will have access to the information that you give us in the additional questionnaires. We will not report any information that can allow you to be identified, and will take every possible precaution to ensure that no one is able to identify you from the information that you provide us. Although you may not benefit directly from participating in this study, your involvement may help us help people with addiction problems. If you have questions about your rights as a research participant, please contact Jill Parsons, Health Sciences Ethics Review Office, University of Toronto, at telephone 416-946-5806 or by email: [email protected]. If you have any questions about the study or would like to be informed of the results of the study, please contact Karen Urbanoski at telephone 416-535-8501 x6121 or by email: [email protected]. You are being given a copy of this informed consent to keep for your own records. Thank you for taking the time to read this letter, Karen Urbanoski, PhD (Candidate) Department of Public Health Sciences, University of Toronto and the Centre for Addiction and Mental Health Phone: 416-535-8501 x 6121 [email protected]
151
Consent to Participate: I have read the letter of information and understand that:
• All the information I provide will be kept strictly confidential • I can withdraw my participation at any time without any negative consequences to
myself or my treatment I consent to completing a questionnaire for this study and to allow the Researchers named above to access the information collected as part of my treatment at PAARC for the purposes of this study. I also consent to having one of the Researchers contact me by telephone in one month for a follow-up interview. Name (print): Signature: Date: Main phone number: _____________________________________ Alternate phone number: _____________________________________
Appendix 3.
152
Attrition analysis
Table A3-1. Client characteristics and treatment-related factors by follow-up status (n=276)
Completed second survey: Client Characteristics No
(n=71) Yes
(n=205) Statistic
Gender: Male 27.3% (59) 72.7% (157) χ2
1=1.32, p=.252 Female 20.0% (12) 80.0% (48)
Mean age (±SD) 33.49 ± 9.62 37.25 ± 10.73 t274=-2.61, p=.009 Marital status:
Married 24.4% (30) 75.6% (93) χ22=1.22, p=.542
Single 29.0% (31) 71.0% (76) Widowed, separated, divorced 20.9% (9) 79.1% (34)
Education: Less than secondary school 45.6% (31) 54.4% (37) χ2
3=17.78, p<.001 Secondary school completed 21.6% (16) 78.4% (58) Some post-secondary 15.8% (6) 84.2% (32) Post-secondary completed 19.8% (16) 80.2% (65)
Employment: Employed 26.3% (47) 73.7% (132) χ2
2=0.90, p=.636 Not employed 28.8% (17) 71.2% (42) Not in the labour force or on leave 20.2% (7) 80.0% (28)
Previous treatment: No 27.8% (49) 72.2% (127) χ2
1=1.73, p=.189 Yes 20.6% (20) 79.4% (77)
Treatment mandate: No 19.4% (33) 80.6% (137) χ2
1=9.23, p=.002 Yes 35.8% (38) 64.2% (68)
Legal problems: No 19.2% (25) 80.8% (105) χ2
1=5.59, p=.018 Yes 31.7% (46) 68.3% (99)
Referral source: Addiction service 13.8% (4) 86.2% (25) χ2
5=12.54, p=.028 Family or friends 40.6% (13) 59.4% (19) Legal system 31.5% (29) 68.5% (63) Mental or general health services 16.4% (11) 83.6% (56) Workplace, school, or social services 11.8% (2) 88.2% (15) Self 29.0% (9) 71.0% (22)
Table A3-2. Admission scale scores and retention by follow-up status (n=276)
Completed second survey: Measure No
(n=71) Yes
(n=204) Statistic
Treatment motivation (TEQ-RAI) 76.42 ± 22.48 81.20 ± 20.73 t274=-1.64, p=.103 Perceived coercion (MPCS) 1.66 ± 1.65 1.31 ± 1.61 t272=1.61, p=.108 Social pressures:
Legal 1.77 ± 1.81 1.31 ± 1.68 t274=1.98, p=.048 Other formal 1.03 ± 1.49 0.95 ± 1.37 t274=0.43, p=.671 Informal 1.69 ± 1.64 1.72 ± 1.51 t274=-0.13, p=.900
Addiction problem severity (SPS) 8.76 ± 5.12 10.15 ± 4.90 t274=-2.03, p=.043 2-month retention:
No 38.0% (57) 62.0% (93) χ21=25.91, p<.001
Yes 11.1% (14) 88.9% (112)
Appendix 4.
153
Parametric tests for non-normally distributed variables
Table A4-1. Pearson correlations (r) between admission process (n=276)
TEQ-RAI MPCS SP-L SP-OF SP-I SPS TEQ-RAI 1.000 MPCS -.551** 1.000 SP-L -.391** .519** 1.000 SP-OF .085 -.103 -.053 1.000 SP-I .055 -.069 -.111 † .322** 1.000 SPS .357** -.268** -.222** .254** .464** 1.000 TEQ-RAI: Treatment Entry Questionnaire – Relative Autonomy Index MPCS: MacArthur Perceived Coercion Scale SP-L: Social Pressures – Legal SP-OF: Social Pressures – Other Formal SP-I: Social Pressures – Informal SPS: Substance Problem Severity † p<.10; * p<.05; ** p<.01
Table A4-2. Client characteristics by perceived coercion at admission (n=268)
Client characteristics Mean ± SD Median Statistic Gender:
Male 1.55 ± 1.68 1.0 t128=3.93, p<.001 Female 0.78 ± 1.20 0.0
Marital status: Married 1.42 ± 1.62 1.0 F2,262=0.06, p=.947 Single 1.34 ± 1.62 1.0 Widowed, separated, divorced 1.38 ± 1.65 1.0
Education: Less than secondary school 1.27 ± 1.51 1.0 F3,249=0.22, p=.881 Secondary school completed 1.48 ± 1.58 1.0 Some post-secondary 1.35 ± 1.74 1.0 Post-secondary completed 1.43 ± 1.69 1.0
Employment: Employed 1.42 ± 1.66 1.0 F2,262=0.37, p=.690 Not employed 1.21 ± 1.50 1.0 Not in the labour force or on leave 1.30 ± 1.57 1.0
Previous treatment: No 1.63 ± 1.77 1.0 t255=4.08, p<.001 Yes 0.89 ± 1.16 0.0
Treatment mandate: No 0.74 ± 1.12 0.0 t153=-8.72, p<.001 Yes 2.42 ± 1.74 2.0
Legal problems: No 0.65 ± 0.96 0.0 t217=-7.98, p<.001 Yes 2.04 ± 1.79 2.0
Referral source: Addiction service 0.52 ± 0.83 0.0 F5,255=19.30, p<.001 Family or friends 1.25 ± 1.34 1.0 Legal system 2.55 ± 1.72 2.0 Mental or general health services 0.76 ± 1.31 0.0 Workplace, school, or social services 1.24 ± 1.52 1.0 Self 0.53 ± 0.86 0.0
Problem substances: None 2.89 ± 1.66 3.0 F3,264=33.53, p<.001 Alcohol only 1.03 ± 1.37 0.0 Drugs only 0.86 ± 1.14 0.0 Both alcohol and drugs 0.76 ± 1.28 0.0
154
Table A4-3. Admission process variables by 2-month retention (n=276)
2-month retention Admission process measures No
(n=120) Yes
(n=107) Statistic
Treatment motivation (TEQ-RAI) 74.99 ± 23.39 83.73 ± 21.23 t225=-2.93, p=.004 Perceived coercion (MPCS) 1.71 ± 1.72 0.98 ± 1.37 t266=3.87, p<.001 Social pressures:
Legal 1.74 ± 1.78 1.06 ± 1.57 t273=3.39, p=.001 Other formal 0.77 ± 1.28 1.21 ± 1.49 t248=-2.60, p=.010 Informal 1.55 ± 1.50 1.90 ± 1.58 t274=-1.93, p=.055
Substance problem severity (SPS-12m) 8.91 ± 5.34 10.84 ± 4.31 t274=-3.33, p=.001
Table A4-4. Admission process variables by 2-month engagement (n=112)
Confidence in treatment
Counsellor rapport
Commitment to treatment
Admission scales
r p r p r p Treatment motivation (TEQ-RAI) .299 .004 .103 .333 .221 .034 Perceived coercion (MPCS) -.125 .207 .017 .867 -.153 .118 Social pressures:
Legal .047 .631 .020 .834 -.109 .258 Other formal .058 .551 .164 .091 .011 .908 Informal -.070 .474 .039 .687 -.007 .938
Substance problem severity (SPS-12m) .045 .641 -.072 .459 -.133 .166
155
Table A4-5. Client characteristics by 2-month engagement in treatment (n=112)
Confidence in treatment
Counsellor rapport
Commitment to treatment
Client characteristics
Mean±SD Statistic Mean±SD Statistic Mean±SD Statistic Gender:
Male 10.3 ± 1.5 t106=-0.04 14.0 ± 1.8 t105=0.57 18.3 ± 1.5 t108=-0.62 Female 10.3 ± 1.8 p=.966 14.2 ± 1.0 p=.571 18.1 ± 1.5 p=.540
Marital status: Married 10.5 ± 1.5 F2,105=1.14 14.3 ± 1.5 F2,104=0.98 18.2 ± 1.6 F2,107=0.24 Single 10.1 ± 1.6 p=.323 13.9 ± 1.6 p=.380 18.4 ± 1.3 p=.789 Widowed/separated/ divorced 10.0 ± 2.0 13.7 ± 1.8 18.4 ± 1.6
Education: Less than secondary school 10.5 ± 1.5 F3,97=0.23 13.7 ± 2.3 F3,96=1.20 18.5 ± 1.3 F3,99=0.79 Secondary school completed 10.2 ± 1.6 p=.873 13.8 ± 1.8 p=.313 18.4 ± 1.3 p=.505 Some post-secondary 10.3 ± 1.7 14.3 ± 1.2 17.9 ± 1.6 Post-secondary completed 10.1 ± 1.6 14.4 ± 0.9 18.3 ± 1.5
Employment: Employed 10.2 ± 1.6 F2,105=0.80 14.2 ± 1.2 F2,104=0.45 18.1 ± 1.5 F2,107=1.77 Not employed 10.7 ± 1.5 p=.451 13.8 ± 2.7 p=.641 18.5 ± 1.3 p=.176 Not in labour force, on leave 10.4 ± 1.5 14.1 ± 1.0 18.7 ± 1.4
Previous treatment: No 10.3 ± 1.5 t106=0.11 13.9 ± 2.0 t84=-1.51 18.3 ± 1.4 t108=-0.14 Yes 10.3 ± 1.7 p=.910 14.3 ± 0.9 p=.135 18.3 ± 1.5 p=.887
Treatment mandate: No 10.3 ± 1.6 t106=0.20 14.1 ± 1.7 t105=-0.08 18.3 ± 1.4 t108=0.16 Yes 10.3 ± 1.5 p=.842 14.1 ± 1.3 p=.934 18.2 ± 1.6 p=.870
Legal problems: No 10.2 ± 1.6 t105=-1.13 14.1 ± 1.6 t104=-0.07 18.3 ± 1.4 t107=0.003 Yes 10.5 ± 1.5 p=.261 14.1 ± 1.7 p=.945 18.3 ± 1.5 p=.998
Referral source: Addiction service 10.3 ± 1.5 F5,101=2.43 14.1 ± 1.3 F5,100=1.11 18.6 ± 1.1 F5,103=0.34 Family or friends 9.8 ± 2.0 p=.040 13.8 ± 0.9 p=.362 17.9 ± 1.4 p=.891 Legal system 10.3 ± 1.6 14.1 ± 1.4 18.4 ± 1.5 Mental/general health service 9.9 ± 1.7 13.7 ± 2.3 18.3 ± 1.4 Workplace/school/CAS 10.9 ± 1.2 14.6 ± 0.7 18.4 ± 2.0 Self 11.3 ± 1.0 14.7 ± 0.7 18.1 ± 1.7
Problem substances: None 9.7 ± 1.6 F3,104=0.36 14.1 ± 1.5 F3,103=0.59 18.3 ± 1.1 F3,106=1.17 Alcohol only 10.4 ± 1.6 p=.784 14.2 ± 1.3 p=.624 18.1 ± 1.6 p=.326 Drugs only 10.3 ± 1.6 13.7 ± 2.4 18.1 ± 1.4 Both alcohol and drugs 10.3 ± 1.5 14.2 ± 0.9 18.7 ± 1.3
Table A4-6. Admission scale scores by perceived need for treatment at 2-months among former clients (n=93)
Reason for leaving treatment = lack of perceived need
Admission scales
No (n=52) Yes (n=41) Statistic Treatment motivation (TEQ-RAI) 85.89 ± 18.85 61.95 ± 19.88 t76=5.46, p<.001 Perceived coercion (MPCS) 1.00 ± 1.30 2.44 ± 1.82 t70=-4.27, p<.001 Social pressures:
Legal 0.94 ± 1.55 2.56 ± 1.57 t91=-4.97, p<.001 Other formal 0.85 ± 1.23 0.32 ± 0.88 t90=2.42, p=.017 Informal 2.04 ± 1.46 0.80 ± 1.15 t91=4.45, p<.001
Substance problem severity (SPS-12m) 11.37 ± 4.21 5.66 ± 5.44 t74=5.54, p<.001
156
Table A4-7. Admission process measures by 2-month substance problem resolution (n=194)
Change in past-month substance problems a
Admission scales
ρ p Treatment motivation (TEQ-RAI) .169 .030 Perceived coercion (MPCS) -.258 <.001 Social pressures:
Legal -.144 .045 Other formal .117 .105 Informal .199 .006
a count of substance-related problems at admission – discharge
157
Table A4-8: Client characteristics by 2-month substance problem resolution (n=194)
Client characteristics Mean ± SD Median Statistic Gender:
Male 1.62 ± 4.68 0.0 t192=0.29, p=.774 Female 1.85 ± 5.21 0.0
Marital status: Married 1.91 ± 4.38 0.0 F2,189=0.81, p=.447 Single 1.06 ± 5.25 0.0 Widowed, separated, divorced 2.06 ± 4.52 0.0
Education: Less than secondary school 0.45 ± 4.15 0.0 F3,177=1.12, p=.343 Secondary school completed 2.30 ± 5.34 0.0 Some post-secondary 1.10 ± 3.72 0.0 Post-secondary completed 1.69 ± 5.04 0.0
Employment: Employed 1.70 ± 4.99 0.0 F2,188=0.03, p=.969 Not employed 1.83 ± 5.12 0.0 Not in the labour force or on leave 1.52 ± 3.44 1.0
Previous treatment: No 1.08 ± 4.14 0.0 t119=-2.13, p=.035 Yes 2.70 ± 5.62 0.0
Treatment mandate: No 2.32 ± 5.33 0.0 t187=3.02, p=.003 Yes 0.45 ± 3.29 0.0
Legal problems: No 2.18 ± 5.47 0.0 t173=1.60, p=.112 Yes 1.08 ± 3.89 0.0
Referral source: Addiction service 4.84 ± 6.27 2.0 F5,184=3.35, p=.006 Family or friends 0.94 ± 3.09 0.0 Legal system 0.53 ± 3.49 0.0 Mental or general health services 1.60 ± 4.70 0.0 Workplace, school, or social services 0.85 ± 6.73 0.0 Self 2.67 ± 5.22 0.0
Problem substances: None -0.66 ± 2.69 0.0 F3,190=5.00, p=.002 Alcohol only 1.94 ± 4.36 0.0 Drugs only 2.07 ± 5.99 0.0 Both alcohol and drugs 3.18 ± 5.04 1.5
Appendix 5.
158
Regression diagnostics and alternative models MODEL 1: Regression of Perceived Coercion Table A5-1. Linear regression diagnostics
Independent variables
VIF Maximum |DfBeta|
Social pressures: Legal 1.11 0.25 Other formal 1.14 0.57 Informal 1.35 0.22
Substance problem severity (SPS-12m) 1.40 0.39 Previous treatment experience (vs. no) 1.11 0.25 Gender (vs. female) 1.06 0.23 Maximum |DfFit| = 0.84 Standardized residuals:
Mean = -.0001 |Maximum| = 3.50 Shapiro-Wilks test for normality of distribution: W=0.944, p<.001
Histogram depicting distribution of standardized residuals:
0.2
.4.6
.8D
ensi
ty
-2 -1 0 1 2 3Standardized residuals
159
MODEL 3B: Regression of Client Confidence in Treatment Table A5-2. Linear regression diagnostics
Independent variables
VIF Maximum |DfBeta|
Treatment motivation (TEQ-RAI) 1.67 0.38 Perceived coercion (MPCS) 1.76 0.42 Social pressures:
Legal 1.41 0.31 Other formal 1.14 0.29 Informal 1.41 0.34
Substance problem severity (SPS-12m) 1.53 0.33 Maximum |DfFit| = 0.62 Standardized residuals:
Mean = .001 |Maximum| = 2.58 Shapiro-Wilks test for normality of distribution: W=0.960, p=.003
Histogram depicting distribution of standardized residuals:
0.1
.2.3
.4.5
Den
sity
-3 -2 -1 0 1 2Standardized residuals
160
MODEL 3C: Regression of Counsellor Rapport Table A5-3. Linear regression diagnostics
Independent variables
VIF Maximum |DfBeta|
Treatment motivation (TEQ-RAI) 1.67 0.40 Perceived coercion (MPCS) 1.76 0.56 Social pressures:
Legal 1.41 0.29 Other formal 1.14 0.43 Informal 1.41 0.73
Substance problem severity (SPS-12m) 1.53 0.70 Maximum |DfFit| = 1.24 Standardized residuals:
Mean = .002 |Maximum| = 5.47 Shapiro-Wilks test for normality of distribution: W=0.700, p<.001
Histogram depicting distribution of standardized residuals:
0.2
.4.6
.8D
ensi
ty
-6 -4 -2 0 2Standardized residuals
161
MODEL 3D: Regression of Client Commitment to Treatment Table A5-4. Linear regression diagnostics
Independent variables
VIF Maximum |DfBeta|
Treatment motivation (TEQ-RAI) 1.67 0.58 Perceived coercion (MPCS) 1.76 0.44 Social pressures:
Legal 1.41 0.56 Other formal 1.14 0.34 Informal 1.41 0.33
Substance problem severity (SPS-12m) 1.53 0.28 Maximum |DfFit| = 0.86 Standardized residuals:
Mean = .001 |Maximum| = 2.74 Shapiro-Wilks test for normality of distribution: W=0.969, p=.014
Histogram depicting distribution of standardized residuals:
0.1
.2.3
.4D
ensi
ty
-3 -2 -1 0 1 2Standardized residuals
162
MODEL 6: Regression of Change in Substance Problems Table A5-5. Linear regression diagnostics
Independent variables
VIF Maximum |DfBeta|
Treatment motivation (TEQ-RAI) 1.54 0.30 Perceived coercion (MPCS) 1.77 0.33 Social pressures:
Legal 1.42 0.41 Other formal 1.15 0.60 Informal 1.13 0.38
2-m retention 1.09 0.38 Maximum |DfFit| = 0.80 Standardized residuals:
Mean = .0003 |Maximum| = 2.86 Shapiro-Wilks test for normality of distribution: W=0.950, p<.001
Histogram depicting distribution of standardized residuals:
0.2
.4.6
Den
sity
-4 -2 0 2 4Standardized residuals