PDFlib PLOP: PDF Linearization, Optimization, Protection

Page inserted by evaluation versionwww.pdflib.com – sales@pdflib.com

Outcome Measurement in Clinical Genetics Services:A Systematic Review of Validated Measures

Katherine Payne, PhD, MSc, BPharm, MRPharmS,1,3 Stuart Nicholls, MSc, BSc,1,3

Marion McAllister, PhD, MSc, BSc,1,2 Rhona MacLeod, PhD, MSc, BSc,2,3 Dian Donnai, MD, FRCP, OBE,1,2,3

Linda M. Davies, MSc, BSc3

1Nowgen,The North West Genetics Knowledge Park, Manchester, UK; 2Central Manchester and Manchester Children’s University HospitalsNHS Trust, Manchester, UK; 3The University of Manchester, Manchester, UK

ABSTRACT

Objective: This systematic review aimed to inform research-ers and policymakers about what validated outcome mea-sures are available to evaluate clinical genetics services (CGS)and the need for new measures.Methods: Validated outcome measures used to evaluate CGSwere identified from a systematic literature review. Subjectiveoutcome measures were assumed to have been validated onlyif some form of psychometric assessment was reported.Results: A total of 1688 titles and abstracts were identified,and 61 articles met the inclusion criteria for the final review,which covered 67 validated outcome measures. There were37 nongenetics-specific and 30 genetics-specific measuresidentified. No single validated outcome measure encom-passed all potential patient benefits from using a CGS. Avariety of different domains were identified, includinganxiety and depression, coping, decision-making, distress,family environment, health status, knowledge, mood, per-

ception of risk, perceived personal control, psychologicalimpact, quality of life, satisfaction and expectations, self-esteem, spiritual well-being, and worry. Some importantaspects of patient benefit from CGS are not covered by exist-ing outcome measures.Conclusions: New research is necessary to develop the arrayof outcome measures required to quantify the benefits CGSoffer patients living with the effects of genetic conditions.These need to be suitable for use in prospective evaluationstudies to provide robust evidence for decision-makersto inform service development and commissioning. Thisincludes prioritization of the existing validated outcome mea-sures in terms of their usefulness and relevance to the mea-surement and valuation of patient benefits from a CGS.Keywords: clinical genetics services, decision-making, mea-surement, outcomes research, systematic review.

Introduction

Clinical genetics services (CGS) have evolved sincethey first emerged as a medical specialty in the 1970s[1]. Clinics initially offered clinical diagnosis and coun-seling services to guide informed decision-making. Ser-vices have since developed, along with the improvedknowledge and technological advances, to offer genetictests for more conditions. With improved understand-ing of molecular pathways, CGS will be able to offerimproved care pathways with options for managingand treating genetic conditions. Genetic counseling isan important part of the overall genetic consultationand the aspect of the service most apparent to patients.Genetic counseling involves providing information andsupport to patients. Fraser [2] offered a definition ofgenetic counseling that included a number of aspects,

but the key factors were: a communication process sothat the patient can understand the medical facts;understand the hereditary nature of the condition;know the options available; and make the best possibleadjustment to the disorder. Genetic testing can formpart of CGS as an aid to genetic diagnosis or to predictthe risk of developing a genetic condition in the future.The value of the genetic test and associated CGS maybe the individual’s and his or her family’s valuations ofthe inherent knowledge of the diagnosis, which canthen act as an aid to decision-making both medically(e.g., determining future treatment pathways) and per-sonally (e.g., reproductive decision-making). Ability tomake informed decisions and knowledge of the geneticcondition can be assumed to be key domains tomeasure given the role of genetic counseling, specifi-cally, and CGS, generally.

There has been, and will continue to be, a rapidexpansion in genetics knowledge about the basis ofdisease and associated developments in new genetictests and CGS [3]. CGS compete with other health-care services for scarce resources. In this climate of

Address correspondence to: Katherine Payne, North WestGenetics Knowledge Park (Nowgen), The Nowgen Centre, 29Grafton Street, Manchester M13 9WU, UK. E-mail: Katherine.payne@manchester.ac.uk

10.1111/j.1524-4733.2007.00259.x

Volume 11 • Number 3 • 2008V A L U E I N H E A L T H

© 2007, International Society for Pharmacoeconomics and Outcomes Research (ISPOR) 1098-3015/08/497 497–508 497

expansion, with the potential to impact on NationalHealth Service (NHS) resources, it is essential to evalu-ate the impact of new CGS on patients and their fami-lies. Two main types of outcome measures are used.Objective clinical measures simply count the numberof times an event occurs and by definition do not takeaccount of the patient’s perspective. Subjective mea-sures quantify an outcome using patient (or clinician)assessment of an event. Patient-reported outcomes areinherently subjective, and examples include patient’sreport of a health condition and its treatment, health-related quality of life, satisfaction with treatment, ortreatment preferences [4].

Clinical genetics services, and associated genetictesting, are offered as part of health-care services, butthere is a paucity of robust economic evidence thatuses appropriate measures of outcome to support theiruse [5,6]. A variety of approaches to outcome mea-surement in clinical genetics have been attempted overthe past 20 years, none of which have proved adequateto take account of the full range of the service offeredto patients and their families [7]. It is difficult toachieve satisfactory measures of outcome and effec-tiveness in clinical genetics [8]. In view of this, qualityindicators as proxies for outcome measures were advo-cated until there is more robust evidence [9]. Wanget al. [7] suggest that outcome measurement is difficultbecause the definition of “success” for genetics servicesis rarely stated explicitly. The clinical genetics profes-sion generally argues that traditional approaches tooutcome measurement in health care will not be rel-evant or appropriate [10,11]. This is because of thenature of the problems with which families come tothe genetics clinics and the lack of agreement about thegoals of CGS. This lack of agreement about the goalsof CGS may stem from the fact that there are differenttypes of services that have developed with specificgoals in mind. CGS for familial cancers offer patientsinformation about the risk of developing cancer in thefuture and subsequent management and treatmentstrategies and as such have different goals from ser-vices for noncancer-related genetic conditions forwhich no treatments are often available.

A number of reviews have called for the develop-ment of new measures of outcome, and the associatedimportance of process measures, for CGS [7,12,13].The literature to date, however, has not systematicallyidentified the breadth of outcome measures applied tothe evaluation of CGS. There has been no structuredapproach to identify validated outcome measures thatcould be viewed as realistic options for robust,evidence-based approaches to identify patient benefitsof CGS [7]. This systematic review aimed to informresearchers and policymakers about what validatedoutcome measures are available to evaluate CGS andthe need for new measures. Specific objectives were toidentify existing validated outcome measures used in

evaluations of CGS and the key domains captured bythese measures.

Methods

Validated outcome measures used to evaluate CGSwere identified from a systematic literature review. Anoutcome measure was defined as “any measured con-sequence or impact for the patient of using geneticsservices.” This systematic review focused on validatedoutcome measures, which for the purpose of thisarticle was defined in terms of whether any form ofpsychometric assessment was reported. A CGS wasdefined as “a specialist service that offers diagnosis ofgenetic conditions; information about genetic condi-tions including information about inheritance andrisks to family members; genetic testing; and support-ive counseling to help the family make decisions andcope better with the genetic condition in their family.The service is offered to all members of a family inwhich a genetic condition may be present, not justthose who have the condition.”

Search StrategyElectronic searches of MEDLINE (Ovid, 1966 todate), Embase (Ovid, 1980 to date), PsychInfo (Ovid,1806 to date), HAPI (Ovid, 1985 to date) and theCochrane Library (1900–2005), which includes NHSEconomic Evaluation Database and Database ofAbstracts of Reviews of Effectiveness, were conducted(February 2006). Figure 1 shows a summary of the keysearch terms, which were based on text terms relevantto outcomes from genetics services. The search strategywas identified from consultation with three geneticshealth-care professionals, a narrative outcomes review[7], and a PhD thesis evaluating the genetic counselingprocess [10]. Bibliographies of included studies werechecked for relevant articles.

SelectionThree independent reviewers (KP/SN/LMD) screenedidentified studies. Articles were rejected on an initialscreen of titles and abstracts only if the reviewers coulddetermine that the articles did not meet the inclusioncriteria. If an article could not be rejected with cer-tainty, the full text of the article was obtained forfurther evaluation. The following interventions wereexcluded because these interventions are not generallypart of routine CGS: population screening, genetherapy, and preimplantation genetic diagnosis. Inaddition, reviews discussing the implications of theHuman Genome Project, gene patenting, genomics andcommercialization, licensing of genetic testing, storageof tissue or blood samples, and ethical and legal impactof genetic testing/services were excluded. This system-atic review included empirical studies and systematicreviews of empirical studies that either developed or

498 Payne et al.

applied outcome measures to CGS. Commentary-typereview studies and studies that did not involve humansubjects or were non-English language articles wereexcluded from the review.

Differences in opinion about inclusion wereresolved by discussion within the group and withothers. Articles were included only if they involved anapplication of a validated outcome measure to evalu-ate CGS, using the definitions assumed by this study.

Assessment and Data ExtractionAssessment of the outcome measure, not the studydesign, was the focus of this review. Two main types ofoutcomes were defined: objective and subjective mea-sures. The concept of an objective outcome measurerefers to a measure that quantifies the fact of an eventoccurring, such as giving a clinical test, attending acounseling session, or giving a positive or negative testresult to a patient. Subjective outcome measures referto measures that quantify an outcome using patient (orclinician) assessment of an event. This review onlyincluded subjective outcome measures that had under-gone some degree of psychometric assessment aimed atmeasuring the extent to which the outcome measuredwhat it claimed to measure. For the purpose of thisarticle, we refer to these measures as validated mea-sures if any form of psychometric assessment wasreported. The validation was confirmed only if theappropriate published primary source, reporting thepsychometric assessments, could be located. Dataextraction forms were used to extract informationabout the validated outcome measure and its applica-tion to CGS. Nonvalidated outcome measures usedwere reported but no further details on these measureswere summarized.

Data SynthesisThe primary output of this review was a qualitativedescription of validated outcome measures used toevaluate CGS. Because of the nature of the data col-lected, they were not synthesized using quantitativetechniques, but extracted data were summarized intables and in narrative form.

Results

The search strategy identified 1668 titles and abstracts.Of these, 121 full papers were retrieved for furtherassessment and 61 papers selected for inclusion in thefinal review (Fig. 2).

Study CharacteristicsThe study characteristics for 61 included papers andthe outcome measures used, both validated and non-validated, are described in the supplementary materialfor this article. Three systematic reviews were includedin this review and a fourth systematic review by Bekkeret al. [14] was excluded because the seven studiesreferred failed to meet the review inclusion criteria.Meiser and Halliday [15] and Vadaparampil et al. [16]clearly reported a list of validated outcome measuresused by the studies identified in the systematic reviews.Katapodi et al. [17] reported a set of instrumentsfor measuring perceived risk to describe emotionalresponse to breast cancer. The article did not reporteach instrument and its validation; therefore, the mea-sures of perceived risk, and extent of validation foreach measure, could not be reported.

The study designs used in the 61 studies were notalways reported clearly, and in some instances thestudy design was defined by the reviewers (KP/SN).

Terms relating to clinical

genetics services

Examples include:genetic assessmentgenetic evaluation

Terms combined using OR

Terms relating to

outcome measurement

Examples include:patient outcomequality of lifesubjective healt effectdecision-making

adjustmentpatient expectation

psychosocial

Terms combined using OR

Examples include:psychometric

reliabilityreproducibilityappropriateinterpretability

Terms combined using OR

Terms relating to

validation

Terms combined using

AND

Terms combined using

AND

:

outcome measure

psychological assessment

health statusdisability scale

patient preferences

clinical geneticsnon-directive counseling

validity

responsivenessacceptabilityprecisionfeasibility

Figure 1 Summary of key search terms.

Outcomes in Clinical Genetics Services 499

The majority of the studies (37) used only surveydesigns, but five studies used randomized controlledtrials [18–22]; six were before and after studies[23–28]; and seven used prospective follow-up studydesigns [29–35]. Two intervention studies [36,37],one case-control [38], and three systematic reviews[15–17] were also reported. The stated aims of the 61studies varied and included not only studies aimed atdeveloping a new outcome measure applicable to clini-cal genetics but also studies using existing validatedoutcome measures to evaluate the psychologicalimpact and psychosocial issues associated with avariety of CGS.

Validated Outcome MeasuresA total of 70 validated outcome measures werereferred to in these 61 studies. Three outcome mea-sures, Life Coping Skills, the Giessener ComplaintsInventory, and perceived barriers to mammography,were reported by their authors to have been validated,but this could not be verified from a published sourceand were therefore excluded from this review. The 67validated outcome measures were summarized interms of the key items or domains assessed, purpose ofthe measure, and the extent of validation reported.(See supplementary material.)

The 67 identified outcome measures can be broadlyclassified into: 37 nongenetics-specific (Table 1) suit-able for the evaluation of any health-care service and30 genetics-specific (Table 2). Tables 1 and 2 also showthe number of times each outcome measure was usedin the 61 identified studies. The majority (n = 46) ofthe validated outcome measures were used, andreported, in just one article, but 21 of the outcome

measures were used on more than one occasion byvarious authors. Most of the outcome measures wereused in studies (n = 41) that evaluated CGS for inher-ited cancers.

The framing of the outcome measures varied, butgenerally they presented respondents with a series ofstatements that required a rating (63 measures), a true/false answer (3 measures), or a combination of ratingand a multiple-choice of responses (1 measure). TheHealth Orientation Scale is unusual in that is uses aspecific form of rating scale, the differential scale. Adifferential scale differs from a rating scale in that itdoes not have a middle or a “neutral” option. Some ofthe measures gave results in the form of a single scoreor index (28 measures), and some measures gave theresult as a profile that described more than one dimen-sion (34 measures). Five measures presented the resultas both a profile and an index. None of the measureswere preference-based measures. Preference-basedoutcome measures show the weight, or importance,that individuals attach to changes in health state orwell-being as a result of a health-care intervention.

Critical Assessment of the Outcome MeasuresThe quality of the reporting and extent of psychomet-ric assessment of the outcome measures varied. Thisreview focused on identifying the extent of psychomet-ric assessment that followed the initial design of anoutcome measure. Table 3 summarizes the extent ofpsychometric validation for the 67 outcome measures.Approximately one-quarter of the outcome measures(n = 19) were assessed in terms of internal reliabilityonly. Internal reliability measures the internal consis-tency of a measure and assesses the extent to which the

1668 potentially relevant articles identified via electronic searching

121 papers on outcome measures

retrieved for more detailed evaluation

61 papers included

(including 3 systematic reviews)

9236 patients (current or future service users)

1266 health-care staff

3 systematic reviews: 66793 participants

(1) 64,276 (2) not reported (3) 2517

Excluded—reasons:669 not clinical genetics service

485 not validated outcome245 not empirical study

10 nonhuman138 non-English

Excluded—reasons:4 not clinical genetics service

53 not validated outcome

3 not empirical study

Figure 2 Flow of studies through the system-atic review.

500 Payne et al.

items (questions) that relate to a particular dimensionin a scale (e.g., perceived empathy toward patientstress in the Genetic Counseling Satisfaction scale)assess only this dimension and no other. The remainingmeasures underwent more extensive psychometricassessment that involved test–retest reliability and

validity but there was limited assessment of sensitivityto change or interpretability.

Objective Outcome MeasuresThe review identified 11 articles that described threeobjective outcome measures: accuracy of diagnosis,accuracy of genetic test results, and rate of terminationof pregnancies. Eleven articles used objective measures(Table 4) [36,39–48].

Discussion

This systematic review identified a large body ofevidence describing validated outcome measures, either

Table 1 Nongenetics-specific outcome measures

Outcome measure Times used

Anxiety and depressionBeck Depression Inventory 5(Breast) Cancer Attitude Inventory and Anxiety

subscale1

Cancer Anxiety and Helplessness Scale 1Center for Epidemiologic Studies Depression-Scale and

brief form6

Hopkins Symptom Checklist 4Hospital Anxiety and Depression Scale 8Self-Rating Depression Scale 1Spielberger State Trait Anxiety Inventory and state scale

of the Spielberger State Trait Anxiety Index20

CopingUtrechtse Coping List 1

Decision-makingDecision Evaluation scales 1Decisional Conflict Scale 1

DistressImpact of Event Scale 20

Family environmentFamily Environment Scale 1Openness to Discuss Cancer in the Family Scale 1

Health statusMedical Outcomes Short-Form 2

KnowledgeKnowledge about Breast Cancer 1Knowledge Scale about Breast (and Ovarian) Cancer

and Hereditary2

MoodProfile of Mood State 5

Perceived riskPerceived Risk of Breast Cancer 2

Personality profilesHealth Beliefs Model (screening and breast cancer) 1Life Orientation Test 4Monitoring Blunting Style Scale, see also Miller

Behavioral Style Scale and Threatening MedicalSituation Inventory

4

Minnesota Multiphasic Personality Inventory plussupplementary ego strength scale

2

Modified Tolerance for Ambiguity Scale 2Psychological impact

Brief Symptom Inventory 4General Health Questionnaire 6Global Severity Index of the Symptom Check List-90 3Psychological Consequences Questionnaire 1

Quality of lifeFunctional Assessment of Cancer Therapy-General 2Subjective Quality of Life Profile 1

SatisfactionMedical Interview Satisfaction Scale—modified 2Satisfaction with Decision Scale 1

Self-esteemRosenberg Self-Esteem Scale 1Tennessee Self-Concept Scale 1

Social supportMedical Outcomes Study Social Support Scale 1

Spiritual well-beingSpiritual Well-Being Scale 1

WorryBreast Cancer Worry 9

Table 2 Genetics-specific validated outcome measures

Outcome measure Times used

CopingPsychological Adaptation to Genetic Information Scale 1

Decision-makingDecision-making process 1Intention to act on shared decision-making program 1

ExpectationsBeliefs about Breast Cancer Genetic Testing 1Prostate cancer genetic screening survey 1Quality of Care through the Patients’ Eyes 1

KnowledgeBreast Cancer Genetic Counseling Knowledge

Questionnaire1

Genetic Knowledge Index 1Knowledge about genetic testing for inherited cancer

(HNPCC and breast cancer)3

Knowledge about genetic risk for breast cancer 1Measure of Counselees’ Knowledge of Down

Syndrome1

Modified Maternal Serum Screening KnowledgeQuestionnaire

1

Risk comprehension and subjective knowledge ofwomen in the shared decision-making program

1

Outcomes of genetics serviceAudit Tool for Genetic Services 1

Perception of risk (benefit)Assessment of benefits and risk of breast cancer testing 1Perceptions of the benefits, limitations, and risks of

genetic testing1

Perceived personal controlPerceived personal control 1

Personality profilesMedical Communication Behavior System 1Desire to participate in the shared decision-making

program1

Psychological impactAnticipated impact of results 1Emotional reaction to the program information 1Health Orientation Scale 1

SatisfactionGenetic Counseling Satisfaction Scale 1Patient Satisfaction with Genetic Counseling 1Satisfaction with shared decision-making program 1Shared decision-making program rationale acceptability 1

Self-esteemBody Image/Sexuality Scale 1

WorryBreast cancer (hereditary) concern 1Multidimensional Impact of Cancer Risk Assessment 1Worry Interference Scale 1

HNPCC, hereditary nonpolyposis colorectal cancer.

Outcomes in Clinical Genetics Services 501

developed or used to evaluate CGS. Previous systematicreviews focused on cancer genetics services andexplored different aspects of outcome measures: mea-sures of perceived risk [17]; measures of psychologicaldistress; accuracy of perceived risk of developingcancer; knowledge and screening uptake associatedwith cancer [15]; and instruments used to measuredepression, anxiety, and distress in individuals atincreased risk for hereditary breast, ovarian, or coloncancer [16]. Wang et al. [7] gave an overview of thepotential outcomes from genetics services, includinggenetic counseling and genetic testing, and used thepublished literature as an illustration. Jarrett andMugford [6] critically reviewed the available literatureon published economic evaluations of genetic healthtechnologies and concluded that methods need to bedeveloped to consider the value of benefits other thansimple health gain. This article presents the results fromthe first systematic review to identify and explore therange of existing outcome measures used across thebroad range of services offered in clinical genetics.

Objective measures quantified three domains: accu-racy of diagnosis, accuracy of tests, and rate of termi-nated pregnancies. Objectives measures, althoughrelatively easy to apply in practice, may need to beused in conjunction with other measures that takemore account of the patient’s perspective. It is ap-parent that a large number of validated subjectiveoutcome measures have been used to evaluate CGS but

each measure focuses on a restricted number ofdomains. The majority of studies have evaluated genet-ics services for familial cancers, rather than othergenetic conditions, which may reflect the type of mea-sures used. The outcome measures used most often haveincluded domains that measure psychological concepts,such as anxiety, depression, worry, and mood. Gener-ally, the aim of using such measures was to explore thepsychological impact of receiving the result of a genetictest, such as breast cancer genes 1 and 2 (BRCA1/2).Perception of risk was a measure used in a number ofstudies [19,20,49]. The systematic review by Katapodiet al. [17] focused on the relationship between per-ceived breast cancer risk and breast cancer preventionand early detection but generally did not report theextent of psychometric assessment of the measures.

Nearly half of the outcome measures identified werebespoke measures designed specifically for the purposeof valuing some aspect of the benefits from CGS. Withthe exception of the measure of knowledge aboutgenetic testing for inherited cancers, each of thesebespoke measures was used on one occasion. Impor-tantly, the extent of psychometric assessment for thesebespoke measures was generally limited and the major-ity of such measures, approximately 80%, were vali-dated only in terms of internal reliability measuringinternal consistency using Cronbach’s alpha. Thisfinding adds further support to the suggestion thatresearchers should look for existing validated mea-sures first before designing their own and not useunpublished and nonvalidated measures because theycan produce biased results [50]. Approximately one-quarter of both generic and genetics-specific measureswere assessed in terms of internal reliability (consis-tency) only. For the purpose of this article, we includedmeasures that had any form of psychometric testing.Nevertheless, it could be argued that using internalconsistency alone is not a sufficient test for completevalidation of a measure to assess the extent to whichthe outcome measures the underlying attribute it pur-ports to measure. The extent of validation of themajority of the genetics specific measures and some of

Table 3 Reporting of outcome measure validation

Validation Number of studies

ValidityFace 9Content 23Construct 26Criterion 15

ReliabilityInternal 63Test–retest 25

Sensitivity to change 5Interpretability 2

Table 4 Studies reporting objective outcome measures

Objective outcome measure Author (Year) Condition

Testing accuracy Ambros et al. (2003) [39] NeuroblastomaAndersson et al. (1995) [40] Acute intermittent porphyriaChamberlain et al. (1992) [41] Duchenne and Becker muscular dystrophyTaylor et al. (2003) [42] Hereditary nonpolyposis colorectal cancer

Diagnosis accuracy Cabana et al. (1998) [43] Ataxia-telangiectasiaChristianson et al. (1995) [44] Termination of pregnancy for abnormalityPark et al. (2002) [45] Hereditary nonpolyposis colorectal cancerWard and Jamison (1991) [46] Craniofacial anthropometry

Termination of pregnancy Kim et al. (2002) [47] Sex chromosome abnormalityKromberg et al. (1999) [36] Huntington diseaseModel et al. (2001) [48] Beta thalassemia

502 Payne et al.

the generic measures identified in this review wastherefore limited.

There are no examples of measures designed spe-cifically to quantify the ability to make informed deci-sions, but some proxy measures have been used. Brainet al. used a measure of the decision-making process toquantify the extent to which women thought or “ago-nized” about the decision to have a genetic test forhereditary nonpolyposis colorectal cancer (HNPCC)[21,51]. Stalmeier et al. [27] devised two measurescentered around decision-making—intention to act onshared decision-making program and risk comprehen-sion and subjective knowledge—to evaluate a shareddecision-making program for women with familialbreast cancer. In a more recent study, Stalmeier et al.used the Decisional Conflict Scale to measure howpatients at high risk for breast and ovarian cancer,awaiting a genetic test result, and facing the choicebetween prophylactic surgery or screening, evaluatedtheir medical treatment choices [52–54].

A number of instruments have been developed tomeasure knowledge, which is a necessary but not suf-ficient component of the ability to make informeddecisions. Measures that assess an individual’s level ofknowledge after a consultation assume that improvingknowledge will facilitate informed decision-making,but adequate knowledge is only one aspect associatedwith making a good decision. The measures of knowl-edge used to date have mainly been cancer-specific, forexample, the Knowledge Scale about Breast (andOvarian) Cancer and Hereditary [55] and knowledgeabout genetic testing for inherited cancer (HNPCC andInherited Breast Cancer) [56]. Measures of knowledgespecific to genetics, rather than cancer genetics, havealso been used in evaluations, such as the GeneticKnowledge Index [57], Measure of Counselees’Knowledge of Down Syndrome [58], and the ModifiedMaternal Serum Screening Knowledge Questionnaire[59]. No existing measures are capable of making theconnection between level of knowledge and the abilityto make informed decisions.

Perceived personal control (PPC) is a concept thatincludes domains relating to knowledge, decision-making, and behavior [60] Berkenstadt et al. designed ameasure of PPC specifically for use in the evaluation ofgenetic counseling [23]. The domain coping has beenincluded in genetic-specific measures such as the Psy-chological Adaptation to Genetic Information Scale[61] to identify people at risk for coping difficulties.Generic measures, such as the Utrechtse Coping List, aDutch adaptation of the Westbrook Coping Scale [62],were used to evaluate coping strategies to describeindividuals seeking a predictive test for HuntingtonDisease [30]. The Openness to Discuss Cancer in theFamily Scale contains domains that are related to atheoretical model of coping with cancer stress. VanOostrum et al. used the Openness to Discuss Cancer

in the Family Scale to explore the long-term psychoso-cial consequences of carrying a BRCA1/2 mutation[63,64].

Family issues may be more significant than health-related outcomes when evaluating services aimed tobenefit people with conditions that are inherited andaffect the whole family. As such, family environmentcould be a key domain to evaluate CGS, consideringwhether using the service had changed the family envi-ronment of the patient, such as closeness, communica-tion, and relationships within the family. Bieseckeret al. used the Family Environment Scale, among othermeasures, to identify the family variables that charac-terize members of hereditary breast and ovarian cancerfamilies who are more likely to choose to undergopredictive testing after pretest education and counsel-ing [29,65]. The Family Environment Scale [65] is notdesigned to measure patient benefits from a service butcharacterize a family environment into one of threetypologies. Van Oostrum et al. used the Openness toDiscuss Cancer in the Family Scale to explore theimpact of carrying a BRCA1/2 mutation in the family[63]. Other measures that mention the possible impacton the family include the Beliefs about Breast CancerGenetic Testing, measure, which asks about beliefs onunrestricted flow of information about test resultsamong family members and physicians [18]. TheBreast Cancer Genetic Counseling Knowledge Ques-tionnaire assesses knowledge of information providedduring breast cancer genetic counseling and specificallyasks about knowledge about the implications ofBRCA1/2 status for family members’ risks [66]. Nostudies have formally evaluated the impact of CGS onfamily environment explicitly.

The interventions offered by CGS often cannotprovide the clear health benefits offered by othermedical services involving pharmacological or surgicalinterventions. It is therefore not surprising that genericmeasures of health status and quality of life are notgenerally used in evaluations of CGS. Cancer genetics isan exception to this and studies evaluating the impact ofcancer risk evaluation clinics have used the SF-36 tomeasure health status [28,67]. Two quality of life mea-sures have been used in evaluations of CGS. The Func-tional Assessment of Cancer Therapy-General (FACT)was designed specifically for cancer-related quality oflife [68,69]. Schwartz et al. used FACT, along withother measures, to evaluate the impact of pretreatmentgenetic counseling on BRCA1/2 testing and on surgicaldecision-making among breast cancer patients at highrisk for carrying a mutation [70,71]. Dazord developedthe Subjective Quality of Life Profile (SQLP) to assesssubjective quality of life in patients or healthy peopleand explore the various dimensions of quality of life[72]. The SQLP was used by Freyer et al. to measurequality of life of individuals belonging to medullary-thyroid carcinoma families [73].

Outcomes in Clinical Genetics Services 503

Satisfaction with service is an outcome measureinextricably linked with the process of how a service isprovided. Measures of satisfaction generally look atsatisfaction with the process of service delivery as awhole, such as the Medical Interview Satisfaction Scale[74]. Some measures have been developed specificallyto evaluate satisfaction with the decision rather thanthe process [75]. Quality of Care Through the Patients’Eyes (QUOTE)-geneCA was designed to measure theneeds and preferences in genetic counseling for heredi-tary cancer [76]. This measure also incorporatesanother key domain related to satisfaction, which iswhether patients’ expectations from using a servicehave been met. Satisfaction is a difficult concept tomeasure or interpret because it is often not clear whataspects of the service are driving the observed levels ofsatisfaction. It is questionable whether satisfaction isan outcome measure and may be considered a proxyoutcome measure in that a change in satisfaction mayhave an impact on the patient’s health and social well-being, but this causal link is based on a substantialassumption.

A number of the measures were designed as predic-tors of outcome and used to categorize the individualsusing a CGS into a typology of, for example, differentpersonality traits, rather than using the measure tovalue the outcome in terms of the possible patientbenefit from using the service. Biesecker et al. used twooutcome measures that included domains of subjectivewell-being, self esteem, and spiritual well-being, intheir description of family variables to characterizehereditary breast and ovarian cancer family memberswho are more likely to choose to undergo predictivetesting [29]. Self-esteem was measured using theRosenberg Self-Esteem Scale, and this measure wasoriginally developed to measure adolescents’ globalfeelings of self-worth or self-acceptance [77]. TheSpiritual Well-Being Scale was designed to assess per-sonal spiritual meaning and satisfaction and definedspiritual well-being as “the affirmation of life in arelationship with God, self, community and environ-ment that nurtures and celebrates wholeness” [78].These measures were used to describe a person’s char-acteristics rather than evaluate changes in their char-acteristics. A number of other studies used measuresthat are more accurately described as measuring per-sonality profiles of individuals rather than patient ben-efits from using a service. Three examples of measuresthat can be described as personality profiles are theMonitoring Blunting Style Scale [79], the MinnesotaMultiphasic Personality Inventory [80], and the LifeOrientation Test [81]. Personality profiles do notmeasure the outcome of the service per se but ratherassess the predisposition of the person using a CGS toexhibit negative feelings as a result of hearing badnews and display symptoms of distress, anxiety, ordepression.

Outcome measures for CGS are needed for: 1)prospective evaluation studies to provide robust evi-dence for commissioning and service developmentdecisions; and 2) audits, to identify whether local ser-vices are effective and of value to patients and 3) toensure that service quality is maintained or improved.The choice of outcome measure depends on thequestion being addressed, so a standardized set ofoutcome measures is required to meet each of thesepossible needs. National decision-making bodies in-volved in the appraisal of the benefits and cost-effectiveness of health-care interventions, such as theNational Institute for Health and Clinical Excellence(NICE), have called for outcome measures that aregeneric and preference-based [82]. A generic measureof patient benefit is necessary to allow interventionswith different effects to be compared directly. Exist-ing generic health status measures, however, may notbe appropriate to value the potential patient benefitsfrom using a CGS. Preference-based outcome mea-sures are necessary to show the weight, or impor-tance, that individuals attach to changes in healthstate or well-being as a result of a health-care inter-vention. This allows interpretation of the significance,to the patient, of the magnitude of change in ameasure. Outcome measures, such as satisfaction orknowledge gain, described in this review do not meetthe definition of a preference-based measure. None ofthe outcome measures used to evaluate CGS waspreference-based. If an intervention that forms partof a CGS were to be appraised by NICE, no outcomemeasure is available that meets their suggestedrequirements.

Ideally, to be useful in a prospective economicevaluation that collects cost and outcome data along-side a randomized controlled trial, as a primaryoutcome, the measure should be reported as an indexrather than as a profile. A profile presents an array ofscores for each domain of the outcome measures. Halfof the identified outcome measures used a profile only.If an outcome measure is reported as a profile, adecision-maker must use some subjective judgmentabout which domains of the profile are relevant as theprimary outcome. It is methodologically feasible toconvert measures that report profiles into indexes bypreference-weighting the dimensions in the measure.With the exception of the SF-36, this has not been donefor the outcome measures reporting profiles identifiedin this review, and further research will be necessary toproduce indexes for use in economic evaluations. Incontrast, a profile measure may be useful as a measurefor use in clinical practice. Each dimension may berelevant to a different aspect of the service. It will allowa decision-maker to identify which aspects of the goalof the service are proving useful.

No single validated outcome measure encompassesall aspects of the potential patient benefits from using

504 Payne et al.

a CGS. Individual outcome measures cover specificdomains in isolation, such as the level of knowledgeabout genetic risk for breast cancer or a measure ofPPC. Particular aspects of patient benefits that aremissing from existing measures evaluating CGSinclude assessment of the benefits to future generationsthat includes some measure of hope in the patient andother family members [83].

LimitationsDifferent degrees of psychometric validation were per-formed and reported, but each measure had some formof validation to be included in this review. One limi-tation of this review is the focus on validated outcomemeasures, which was a practical decision driven by thedifficulties in identifying nonvalidated outcome mea-sures that may use different nomenclature. Further-more, excluding the “validation” component of thesearch strategy resulted in an unmanageable number ofstudies, approximately 29,000, being identified. Vali-dated outcome measures are appropriate for robustevaluations but may exclude potentially useful mea-sures that contain certain domains relevant to the goalsof a CGS. A further limitation is that only studieswritten in the English language were included.

Conclusion

This systematic review has identified a number of vali-dated outcome measures and the domains in thesemeasures. It is clear that no single validated outcomemeasure encompasses all aspects of the potentialpatient benefits from using a CGS. The focus to datehas been on mainly using measures of psychologicalimpact, which represents a limited perspective on whata CGS can offer patients. A variety of differentdomains were identified, but aspects of the potentialpatient benefits from using CGS are not being capturedby existing measures. This systematic review did notfind sufficient evidence to recommend which of theexisting outcome measures are appropriate to valuethe patient benefits from using CGS. Further researchis needed to develop and validate an array of outcomemeasures that are capable of quantifying the benefitsCGS offer patients living with the effects of geneticconditions. It will be necessary to incorporate a varietyof perspectives when designing new measures. Pa-tients’ views should be used to target the measure tothe aspects of the CGS that benefits them directly.Health-care professionals’ views may be used to facili-tate the design of measures that are user-friendly in theNHS not only for the purpose of evaluation studies butalso for audit of clinical practice. One possible nextstep is to prioritize which of the existing validatedoutcome measures may be usefully incorporated intoevaluation studies. This review concludes that a coreset of outcome measures may be necessary when evalu-

ating genetics services. This core set of measures,however, cannot yet be assembled: Gaps identified inthe range of existing measures suggest possibilities forfuture research.

The authors would like to express their sincere thanks toHelen McEvoy, NHS Liaison and Human Services FacultyLibrarian, University of Manchester, for guidance with thesearch strategy; Dr. Lee Hooper, Lecturer in Research Syn-thesis, University of East Anglia; and Dr. Fay Crawford,Senior Research Fellow, University of Dundee, for reviewingthe study protocol. We would also like to thank Dr. HelenMiddleton-Price, Director of The North West GeneticsKnowledge Park, who commented on drafts of this article.We are also grateful to the three anonymous reviewers whoprovided thoughtful comments on this article.

Source of financial support: Nowgen, The North West Genet-ics Knowledge Park, is funded by a grant from the Depart-ment of Health and the Department of Trade and Industry.Funding is also acknowledged from Central Manchester andManchester Children’s Hospitals NHS Trust and the Univer-sities of Manchester, Liverpool, and Lancaster. The viewsexpressed in this article are those of the authors and not ofthe funding bodies.

Supplementary Material

Supplementary material for this article can be found at:http://www.ispor.org/valueinhealth_index.asp.

References

1 Coventry PA, Pickstone JV. From what and why didgenetics emerge as a medical specialism in the 1970sin the UK? A case-history of research, policy andservices in the Manchester region of the NHS. Soc SciMed 1999;49:1227–38.

2 Fraser FC. Genetic counselling. Am J Hum Genet1975;27:240–2.

3 Donnai D. Genetic services. Clin Genet 2002;61:1–6.4 Acquadro C, Berzon R, Dubois D, et al. Incorporating

the patient’s perspective into drug development andcommunication: an ad hoc task force report of thePatient-Reported Outcomes (PRO) HarmonizationGroup meeting at the Food and Drug Administration.Value Health 2001;6:522–31.

5 Carlson JJ, Henriskon NB, Veenstra DL, et al. Eco-nomic analysis of human genetics services: a system-atic review. Genet Med 2005;7:519–23.

6 Jarrett J, Mugford M. Genetic health technology andeconomic evaluation. A critical review. Appl HealthEcon Health Policy 2006;5:27–35.

7 Wang C, Gonzalez R, Merajver SD. Assessment ofgenetic testing and related counselling services:current research and future directions. Soc Sci Med2004;58:1427–42.

8 Royal College of Physicians. Clinical Genetics Ser-vices: Activity, Outcome, Effectiveness and Quality.London: RCP, 1998.

Outcomes in Clinical Genetics Services 505

9 Royal College of Physicians. Commissioning ClinicalGenetics Services: Activity, Outcome, Effectivenessand Quality. London: RCP Clinical Genetics Commit-tee, 1998.

10 Clarke AJ. Outcomes and process in genetic counsel-ling. Clin Genet 1997;50:462–9.

11 MacLeod R. The genetic counselling process: aninterpretative phenomenological analysis of patient–counsellor interactions. PhD Thesis, The University ofManchester, 2003.

12 Bernhardt BA, Biesecker BB, Mastromarino CL.Goals, benefits and outcomes of genetic counselling:client and genetic counsellor assessment. Am J MedGenet 2000;94:189–97.

13 Biesecker BB, Peters K. Process studies in geneticcounselling: peering into the black box. Am J MedGenet 2001;106:191–8.

14 Bekker H, Thornton JG, Airey CM, et al. Informeddecision making: an annotated bibliography and sys-tematic review. Health Tech Assess 1999;3:1–199.

15 Meiser B, Halliday JL. What is the impact of geneticcounselling in women at increased risk of developinghereditary breast cancer? A meta-analytic review. SocSci Med 2002;54:1463–70.

16 Vadaparampil ST, Ropka M, Stefanek ME. Measure-ment of psychological factors associated with genetictesting for hereditary breast, ovarian and coloncancers. Fam Cancer 2005;4:195–206.

17 Katapodi MC, Lee KA, Facione NC, Dodd MJ. Pre-dictors of perceived breast cancer risk and the rela-tion between perceived risk and breast cancerscreening: a meta-analytic review. Prev Med 2004;3:388–402.

18 Bowen DJ, Burke W, Yasui Y, et al. Effects of riskcounselling on interest in breast cancer genetic testingfor lower risk women. Genet Med 2002;4:359–65.

19 Brain K, Norman P, Gray J, Mansel R. Anxiety andadherence to breast self-examination in women with afamily history of breast cancer. Psychosom Med1999;61:181–7.

20 Brain K, Gray J, Norman P, et al. Randomized trial ofa specialist genetic assessment service for familialbreast cancer. J Natl Cancer Inst 2000;92:1345–51.

21 Brain K, Sivell S, Bennert K, et al. An exploratorycomparison of genetic counselling protocols forHNPCC predictive testing. Clin Genet 2005;68:255–61.

22 Cederholm M, Axelsson O, Sjoden PO. Women’sknowledge, concerns and psychological reactionsbefore undergoing an invasive procedure for prenatalkaryotyping. Ultrasound Obstet Gynecol 1999;14:267–72.

23 Berkenstadt M, Shiloh S, Barkai G, et al. Perceivedpersonal control (PPC): a new concept in measuringoutcome of genetic counselling. Am J Med Genet1999;82:53–9.

24 Collins V, Meiser B, Gaff C, et al. Screening and pre-ventive behaviors one year after predictive genetictesting for hereditary nonpolyposis colorectal carci-noma. Cancer 2005;104:273–81.

25 Dorval M, Patenaude AF, Schneider KA, et al. Antici-pated versus actual emotional reactions to disclosure

of results of genetic tests for cancer susceptibility:findings from p53 and BRCA1 testing programs. JClin Oncol 2000;18:2135–42.

26 Drake ER, Engler-Todd L, O’Connor AM, et al.Development and evaluation of a decision aid aboutprenatal testing for women of advanced maternal age.J Genet Couns 1999;8:217–33.

27 Stalmeier PF, Unic IJ, Verhoef LC, van Daal WA.Evaluation of a shared decision making program forwomen suspected to have a genetic predisposition tobreast cancer: preliminary results. Med Decis Making1999;19:230–41.

28 Trask P, Paterson AG, Wang C, et al. Cancer-specificworry interference in women attending a breast andovarian cancer risk evaluation program: impact onemotional distress and health functioning. Psy-chooncology 2001;10:349–60.

29 Biesecker BB, Ishibe N, Hadley DW, et al. Psychoso-cial factors predicting BRCA1/BRCA2 testing deci-sions in members of hereditary breast and ovariancancer families. Am J Med Genet 2000;93:257–63.

30 Decruyenaere M, Evers-Kiebooms G, Boogaerts A,et al. Predictive testing for Huntington’s disease: riskperception, reasons for testing and psychologicalprofile of test applicants. Genet Couns 1995;6:1–13.

31 Decruyenaere M, Evers-Kiebooms G, Boogaerts A,et al. Psychological functioning before predictivetesting for Huntington’s disease: the role of the paren-tal disease, risk perception, and subjective proximityof the disease. J Med Genet 1999;36:897–905.

32 Fries MH, Hailey BJ, Flanagan J, Licklider D.Outcome of five years of accelerated surveillance inpatients at high risk for inherited breast/ovariancancer: report of a phase II trial. Mil Med 2004;169:411–16.

33 Grosfeld FJ, Beemer FA, Lips CJ, et al. Parents’responses to disclosure of genetic test results of theirchildren. Am J Med Genet 2000;94:316–23.

34 Michie S, French S, Allanson A, et al. Informationrecall in genetic counselling: a pilot study of its assess-ment. Patient Educ Couns 1997;32:93–100.

35 Schwartz MD, Kaufman E, Peshkin BN, et al. Bilat-eral prophylactic oophorectomy and ovarian cancerscreening following BRCA1/BRCA2 mutation testing.J Clin Oncol 2003;21:4034–41.

36 Kromberg JG, Krause A, Spurdle AB, et al. Utilisationof predictive, prenatal and diagnostic testing for Hun-tington’s disease in Johannesburg. S Afr Med J1999;89:774–8.

37 Miller SM, Fleisher L, Roussi P, et al. Facilitatinginformed decision making about breast cancer riskand genetic counselling among women calling theNCI’s Cancer Information Service. J Health Commun2005;10:119–36.

38 Wolraich ML, Albanese M, Stone G, et al. MedicalCommunication Behaviour System. An interactionalanalysis system for medical interactions. Med Care1985;24:891–903.

39 Ambros IM, Benard J, Boavida M, et al. Qualityassessment of genetic markers used for therapy strati-fication. J Clin Oncol 2003;21:2077–84.

506 Payne et al.

40 Andersson C, Thunell S, Floderus Y, et al. Diagnosisof acute intermittent porphyria in northern Sweden:an evaluation of mutation analysis and biochemicalmethods. J Intern Med 1995;237:301–8.

41 Chamberlain JS, Chamberlain JR, Fenwick RG, et al.Diagnosis of Duchenne and Becker muscular dystro-phies by polymerase chain reaction: a multicenterstudy. J Am Med Assoc 1992;267:2609–15.

42 Taylor CF, Charlton RS, Burn J, et al. Genomic dele-tions in MSH2 or MLH1 are a frequent cause ofhereditary non-polyposis colorectal cancer: identifica-tion of novel and recurrent deletions by MLPA. HumMutat 2003;22:428–33.

43 Cabana MD, Crawford TO, Winkelstein JA, et al.Consequences of the delayed diagnosis of ataxia-telangiectasia. Pediatrics 1998;102:98–100.

44 Christianson AL, van den Berg HJ, van RensburgP, et al. Mid-pregnancy genetic terminations ofpregnancy–postnatal assessment and management. SAfr Med J 1995;85(10 Suppl.):S1084–7.

45 Park JG, Vasen HFA, Park YJ, et al. SuspectedHNPCC and Amsterdam criteria II: evaluation ofmutation detection rate, an international collabora-tive study. Int J Colorectal Dis 2002;17:109–14.

46 Ward RE, Jamison PL. Measurement precision andreliability in craniofacial anthropometry: implicationsand suggestions for clinical applications. J CraniofacGenet Dev Biol 1991;11:156–64.

47 Kim YJ, Park SY, Han JH, et al. Parental decisions ofpre-natally detected sex chromosome abnormality. JKorean Med Sci 2002;17:53–7.

48 Modell B, Khan M, Darlison M, et al. A nationalregister for surveillance of inherited disorders: betathalassaemia in the United Kingdom. Bull WorldHealth Organ 2001;79:1006–13.

49 Donovan KA, Tucker DC. Knowledge about geneticrisk for breast cancer and perceptions of genetictesting in a socio-demographically diverse sample. JBehav Med 2000;23:15–36.

50 Marshall M, Lockwood A, Bradley C, et al. Unpub-lished rating scales: a major source of bias inrandomised controlled trials of treatments for schizo-phrenia. Br J Psychiatry 2000;176:249–52.

51 Michie S, Smith D, McClennan A, Marteau TM.Patient decision making: an evaluation of twodifferent methods of presenting information about ascreening test. Br J Health Psychol 1997;2:317–26.

52 Stalmeier PF, Roosmalen MS, Verhoef LC, et al. Thedecision evaluation scales. Patient Educ Couns2005;57:286–93.

53 O’Connor AM. Validation of a decisional conflictscale. Med Decis Mak 1995;15:25–30.

54 Bunn H, O’Connor AM. Validation of client decision-making instruments in the context of psychiatry. CanJ Nurs Res 1996;28:13–27.

55 Ondrusek N, Warner E, Goel V. Development of aknowledge scale about breast cancer and heredity(BCHK). Breast Cancer Res Treat 1999;53:69–75.

56 Lerman C, Narod S, Schulman K, et al. BRCA1testing in families with hereditary breast-ovarian

cancer. A prospective study of patient decision makingand outcomes. JAMA 1996;275:1885–92.

57 Furr LA, Kelly SE. The genetic knowledge index:developing a standard measure of genetic knowledge.Genet Test 1999;3:193–9.

58 Braitman A, Antley RM. Development of instrumentsto measure counselees’ knowledge of Down syn-drome. Clin Genet 1978;13:25–36.

59 Goel V, Glazier R, Holzapfel S, et al. Evaluatingpatient’s knowledge of maternal serum screening.Prenat Diagn 1996;16:425–30.

60 Averill JR. Perceived control over aversive stimuli andits relationship to stress. Psychol Bull 1973;80:286–303.

61 Read CY, Perry DJ, Duffy ME. Design and psycho-metric evaluation of the Psychological Adaptation toGenetic Information Scale. J Nurs Scholarsh 2005;37:203–8.

62 Westbrook MT. A classification of coping behaviourbased on multidimensional scaling of similarityratings. J Clin Psychol 1979;35:407–9.

63 Van Oostrum I, Meijers-Heijboer H, Lodder LN,et al. Long-term psychological impact of carrying aBRCA1/2 mutation and prophylactic surgery: a 5-yearfollow-up study. J Clin Oncol 2003;21:3867–74.

64 Mesters I, van den Borne H, McCormick L, et al.Openness to discuss cancer in the nuclear family:scale, development and validation. Pychosom Med1997;59:269–79.

65 Moos RH, Moos BS. Family Environment ScaleManual. Mountain View, CA: Consulting Psycholo-gists Press 1994.

66 Erblich J, Brown K, Kim Y, et al. Development andvalidation of a Breast Cancer Genetic CounsellingKnowledge Questionnaire. Patient Educ Couns2005;56:182–91.

67 Keller M, Jost R, Haunstetter CM, et al. Comprehen-sive genetic counselling for families at risk forHNPCC: impact on distress and perceptions. GenetTest 2002;6:291–302.

68 Cella DF, Tulsky DS, Gray G, et al. The functionalassessment of cancer therapy scale: development andvalidation of the general measure. J Clin Oncol1993;11:570–9.

69 Brady M, Cella DF, Bonomi AE, et al. Reliabilityand validity of the functional assessment of cancertherapy-breast quality of life instrument. J Clin Oncol1997;15:974–86.

70 Schwartz MD, Lerman C, Brogan B, et al. Impact ofBRCA1/BRCA2 counselling and testing on newlydiagnosed breast cancer patients. J Clin Oncol2004;22:1823–9.

71 Schwartz MD, Lerman C, Brogan B, et al. Utilizationof BRCA1/BRCA2 mutation testing in newly diag-nosed breast cancer patients. Cancer Epidemiol Biom-arkers Prev 2005;14:1003–7.

72 Dazord A. Quality of life assessment in medicine:presentation of a new instrument (Subjective Qualityof Life Profile (SQLP)) and first results. Eur RespirRev 1995;5:66–71.

73 Freyer G, Dazord A, Schlumberger M, et al. Psycho-social impact of genetic testing in familial medullary-

Outcomes in Clinical Genetics Services 507

thyroid carcinoma: a multicentric pilot-evaluation.Ann Oncol 1999;10:87–95.

74 Wolf MH, Putnam SM, James SA, Stiles WB. MedicalInterview Satisfaction Scale: development of a scale tomeasure patient perceptions of physician behaviour. JBehav Med 1978;1:391–401.

75 Holmes-Rovner M, Kroll J, Scmitt N, et al. Patientsatisfaction with healthcare decisions: the satisfactionwith decision scale. Med Decis Making 1996;15:58–64.

76 Pieterse A, van Dulmen S, Ausems M, et al. QUOTE-gene (ca): development of a counselee-centred instru-ment to measure needs and preferences in geneticcounselling for hereditary cancer. Psychooncology2005;14:361–75.

77 Rosenberg M. Society and the Adolescent Self-Image.Princeton, NJ: Princeton University Press, 1965.

78 Ellison CW, Smith J. Towards an integrative measureof health and well-being. J Psychol Theol 1991;19:35–48.

79 Miller SM. Monitoring and blunting: validation of aquestionnaire to assess styles of information seekingunder threat. J Pers Soc Psychol 1987;52:345–53.

80 Barron F. An ego-strength scale which predictsresponse to psychotherapy. J Consult Psychol 1953;17:327–33.

81 Scheier MF, Carver CS, Bridges MW. Distinguishingoptimism from neuroticism (and Trait anxiety, self-mastery and self-esteem): a re-evaluation of the LifeOrientation Test. J Pers Soc Psychol 1994;67:1063–78.

82 National Institute for Health and Clinical Excellence.A Guide to Technology Appraisals. London: NICE,2004.

83 McAllister M, Nicholls S, Payne K, et al. Process andoutcome in clinical genetics services: a new modeldeveloped from qualitative research. J Med Genet2006;43(Suppl. 1):S62.

508 Payne et al.