special care units for people with dementia: findings from

Chapter 4

Special Care Units ForPeople With Dementia:

Findings From Evaluative Studies

ContentsPage

INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111TYPES OF EVALUATIVE STUDIES OF SPECIAL CARE UNITS . . . . . . . . . . . . . . . . 112EVALUATIVE STUDIES WITHOUT A CONTROL GROUP:

EFFECTS ON RESIDENTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112EVALUATIVE STUDIES WITH A CONTROL GROUP:

EFFECTS ON RESIDENTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117STUDIES OF PARTICULAR FEATURES AND INTERVENTIONS IN

SPECIAL CARE UNITS: EFFECTS ON RESIDENTS . . . . . . . . . . . . . . . . . . . . . . . . . . . 121STUDIES THAT EVALUATE THE EFFECTS OF SPECIAL CARE UNITS

ON UNIT STAFF MEMBERS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123STUDIES THAT EVALUATE THE EFFECTS OF SPECIAL CARE UNITS

ON RESIDENTS’ FAMILIES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . 125CONCLUSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127

TablesTable Page4-1. Evaluative Studies Without a Control Group . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1134-2. Evaluative Studies With a Control Group . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1184-3. OTA’S Conclusions From the Evaluative Studies of Special Care Units . . . . . . . . . . 128

Chapter 4

Special Care Units For People With Dementia:Findings From Evaluative Studies

INTRODUCTIONAs noted in chapter 3, much of the literature on

special care units consists of descriptive reportsabout an individual unit. These descriptive reportsoften present anecdotal evidence of the unit’spositive outcomes. Frequently, the reports includecase examples that show how the unit benefited oneor more of its residents. Many of the reports alsodescribe positive outcomes of the unit for residents’families and unit staff members.

Anecdotal evidence of the positive outcomes ofindividual special care units is compelling. The caseexamples are particularly compelling: the individualresidents they describe seem typical of nursing homeresidents with dementia who do not do well innonspecialized units; these individuals often areadmitted to the special care unit in a very agitated orwithdrawn condition; they frequently have beenovermedicated and physically restrained; character-istics of the unit, including its physical designfeatures, patient care philosophy, and activity pro-grams, seem to match their needs exactly; and theyrespond positively and dramatically to the unitenvironment.

Case examples and other anecdotal evidence ofthe positive outcomes of individual units are notadequate, however, to evaluate the effectiveness ofspecial care units. In the past few years, a number ofevaluative studies of special care units have beenconducted. These studies attempt to measure objec-tively the effectiveness of one or more special careunits in terms of changes in aspects of theirresidents’ condition and functioning over time.Several of the evaluative studies also measure theeffects of special care units on residents’ familiesand unit staff members.

This chapter reviews what is known about specialcare units from the available evaluative studies. Itdoes not include information from descriptive re-ports on individual special care units. Findings of theavailable evaluative studies are discussed in somedetail because, like the descriptive studies discussedin chapter 3, they provide a basis for informed policydecisions about the development of special regula-

tions and reimbursement for special care units, aboutthe need for and content of consumer educationabout special care units, and about the futuredirection and level of government support forresearch on special care units.

OTA’s conclusions from the evaluative studiesdiscussed in this chapter are summarized in table 4-3at the end of the chapter. The findings differ,depending on whether the study used a controlgroup. The nine evaluative studies that did not use acontrol group found positive outcomes for specialcare unit residents in a variety of areas. If contradic-tory findings are excluded, the only positive out-comes found in more than one of the nine studies aredecreased nighttime wakefulness, improved hy-giene, and weight gain. A few of the studies foundimprovements over time in the important areas ofresidents’ ability to perform activities of daily livingand residents’ behavioral symptoms, but an equalnumber of studies did not find such improvements.

Only two of the six evaluative studies that used acontrol group found any positive outcomes forspecial care unit residents. One of these studiesfound that over a l-year period, 14 residents of onespecial care unit showed significantly less declinethan 14 residents with dementia in nonspecializedunits of the same nursing home in their ability toperform activities of daily living (392). The secondstudy found that 13 residents of one special care unitexhibited significantly fewer catastrophic reactionsthan 9 residents with dementia in nonspecializedunits of the same facility (265). The 13 special careunit residents also interacted significantly moreoften with staff members.

Only one of the four evaluative studies thatmeasured the impact of a special care unit on unitstaff members found any positive outcomes. Thefindings with respect to outcomes for residents’families are contradictory, as described later in thechapter.

The limited positive findings in many of theseevaluative studies and the complete lack of positivefindings in some of the studies are surprising andappear to contradict the conviction of special care

–111–

112 . Special Care Units for People With Alzheimer’s and Other Dementias

unit operators and others that the units benefitresidents, residents’ families, and unit staff mem-bers. Each of the available studies suffers from oneor more methodological problems that could invali-date its findings, e.g., small sample sizes and use ofnonrandom samples. Citing these problems, somespecial care unit advocates discount the lack ofpositive findings. In contrast, OTA concludes thatsome of the studies-particularly the six studies thatused a control group-constitute credible researchin an area in which good research is difficult todesign and conduct. Despite methodological prob-lems, the studies’ findings are meaningful anddeserve careful consideration by policymakers, spe-cial care unit advocates, and others.

TYPES OF EVALUATIVE STUDIESOF SPECIAL CARE UNITS

Three types of evaluative studies of special careunits have been conducted. In one type, selectedcharacteristics of individuals with dementia, theirfamilies, and/or unit staff members are measured atdesignated intervals before and after the individuals’admission to a special care unit. Changes or lack ofchanges in the measured characteristics over timeare then attributed to the impact of the special careunit. This type of study does not use a separatecontrol group.

The second type of evaluative study does use aseparate control group. In this type of study, selectedcharacteristics of the special care unit residents, theirfamilies, and/or unit staff members and selectedcharacteristics of other individuals with dementia,their families, and/or staff members in nonspecial-ized nursing home units or other settings aremeasured at designated intervals. Changes or lack ofchanges in the measured characteristics of the twogroups of subjects are compared, and any differencesbetween the two groups are attributed to the impactof the special care unit.

A third type of evaluative study measures theeffectiveness of particular features and interventionsin special care units. One example is research on theeffectiveness of various types of devices to deterresidents who wander from leaving the unit.

The findings of these three types of evaluativestudies are discussed in the following sections.Findings with respect to the effects of special care

units on residents, residents’ families, and unit staffmembers are discussed separately.

EVALUATIVE STUDIESWITHOUT A CONTROL GROUP:

EFFECTS ON RESIDENTSOTA is aware of nine evaluative studies of special

care units in which a control group was not used (seetable 4-l). Seven of the nine studies were conductedin a single special care unit. The other two studieswere conducted in two and three special care units,respectively. The samples for 6 of the 9 studies werevery small (under 12 individuals each). One of the 3remaining studies had a sample of 32 subjects, andone had a sample of 53 subjects (24,245). Thesample size for the ninth study is not specified in thestudy report (22).

Table 4-1 lists the physical design and otherchanges made to create the special care units, asdescribed in the study reports. These changesdiffered from one special care unit to another. Somechanges that were made to create one or more of theunits may not have been mentioned in the studyreports.

Each of the nine studies found some positiveoutcomes of the special care units, as summarizedbelow. The study reports emphasize these positiveoutcomes. Negative outcomes are also reported, butthey receive less emphasis in the study reports. Thestatistical significance of the studies’ findings wascomputed in only four of the nine studies. In thefollowing discussion, OTA uses the terms statisti-cally significant and significant for research findingswith a P value of 0.05 or less.

Bell and Smith found statistically significantimprovements in behavior among residents of anewly created 24-bed special care unit (22). Over a3-month period, the residents became significantlymore likely to exhibit three behaviors defined as“positive’ by the researchers-having a clean face,having clean clothes, and walking alone. At the endof the 3-month period, the frequency of thesebehaviors among residents of the newly created unitwas similar to their frequency among residents of a26-bed special care unit that had been operating forover a year. This outcome fit the researchers’hypothesis that behaviors they defined as positivewould increase over time in the new unit andbehaviors they defined as negative would decrease

Table 4-l—Evaluative Studies Without a Control Group

Year of Funding Duration ofCitation the Study Source Subjects study Changes Made to Create the Special Care Unit

Bell and Smith,unpublishedmanuscript

1986 no funding sourcereported

residents of one 24-bedspecial care unit andone 28-bed special careunit

32 residents of a 46-bed special care unit

6 months, from 3months before to 3months after the 24-bed unit opened

locked access doors; secure outdoor area; separate lounge, dining area andnurses’ station; increased staff-to-resident ratio compared to nonspecializedunits in the same two facilities; staff training by the Denver Alzheimer’sAssociation Chapter; efforts to involve families.

unlocked access doors with alarms and double doorknobs; special activityprograms; sensory stimulation; reality orientation; personal markers on resi-dents’ doors; orientation boards; ongoing staff training; family support groups.

Benson et al.,1987 andCameron et al.,1987

1984-1985 no funding sourcereported

one year, from just be-fore to one year afterthe unit opened

11 residents of a 20bed special care unit

8 months, from 4months before to 4months after the unitopened

6 months, from 3months before to 3months after the unitopened

“quiet, predictable environment;” increased staff-to-resident ratio compared tothe rest of the facility.

Bullock et al.,unpublishedmanuscript,1988

1987 no funding sourcereported

no funding sourcereported


closed access doors; separate dining and activity areas; efforts to reducestimulation; consistent dally routine; neutral colors and design; no TV or radio;only one phone; visitor and staff traffic through the unit limited to reducestimulation; training programs for staff and families.

Cleary et al.,1988

notreported


6 residents of a 26-bedspecial care unit

4 months, from beforeadmission to 4 monthsafter admission for 5subjects, and one month,from before admissionto one month after ad-mission for one subject

3 months, from the timethe unit opened to 3months after it opened

locked access doors; separate dining room and day room; calm, reassuringapproach; flexible daily routine; familiar background music; residents encour-aged to bring in personal Items; 40 hours of staff training; efforts to Involvefamilies.

Greene et al.,1985

notreported

Hall et al.,1966


12 residents of a 24-bed unit that alsohoused nondementedchairfast residents

unlocked access door; minimal remodeling; efforts to reduce stimulation; nomirrors; no TV; no public address system; home-like atmosphere; textured wallhangings; chairs placed in the corridor to encourage resting; flexible dallyroutine; residents fed In small groups; visitor and staff traffic through the unitlimited to reduce stimulation; no Increase in staff; ongoing staff training; effortsto involve families; family support groups.

notreported

locked access doors; resident bedrooms situated on three skies of a largecentral space; designated dining and activity areas; open, centrally locatednurses’ station; therapeutic kitchen for residents; lounge for residents and theirfamilies; staff offices located just outside the unit; movable furniture in centralarea; washable, vinyl wall coverings in neutral colors; fabric wall hangings;mirrors in residents’ rooms; large dock; orientation board; color-coded doorjams and bedrooms; residents’ name on bedroom door; toilet In each bedroom.


53 residents of 3 iden-tical 40-bed special careunits in a 120-bed nurs-ing home designed forpersons with dementia

19 months, from oneyear before to 7 monthsafter the units opened

Lawton et al.,1984 andLiebowitz et al.,1979

one year, from before toone year after the unitopened

closed unit; no other features of the unit are described in the study report.McCracken andFitzwater, 1988

notreported


11 residents of a spe-cial care unit; unit sizenot reported

doorways painted in contrasting colors; enclosed outdoor area with nonpoison-ous plants; furniture with rounded edges; medication carts and housekeeping

Mummah-Castillo,1987



one year, from 6months before to 6

carts locked; residents encouraged to bring in personal items; home-likeatmosphere; visual cues; clocks, calendars, and orientation boards; remi-niscence therapy; pet therapy; cooking; encourage resident participation Inactivities; staff training; staff selected specifically for the unit; efforts to involvefamilies.

months after admission

SOURCE: Office of Technology Assessment, 1992.


and eventually reach the same frequency as in the oldunit.

Other findings of Bell and Smith’s study did notfit their hypothesis. Use of physical restraints, whichwas significantly higher in the new unit than the oldunit at the beginning of the study, increased in bothunits over the course of the study (22). In addition,at all times during the study, residents of the old unitwere significantly more likely than residents of thenew unit to exhibit two behaviors defined asnegative by the researchers-being incompletelydressed and talking to oneself. One “positive”behavior—talking with others-was significantlymore common in the new unit than the old unit, butincreased over time in the old unit. Thus somenegative behaviors were more common in the oldthan the new unit, and one positive behaviorincreased in the old unit over time. ‘‘Negative’behaviors, such as shouting, swearing, and hitting,were rare on both units, and their frequency did notchange over time.

Benson et al. found statistically significant im-provements in mental and emotional status, hygiene,and other physical functions among 32 residents ofone 46-bed special care unit (24). Compared withbaseline values at the time of the residents’ admis-sion to the unit, significant improvements werefound at both 4 months and 1 year in the followingaspects of the residents’ mental and emotionalstatus: the residents made more decisions, compre-hended more, were more responsive, exhibitedgreater interest in themselves and others, and werejudged by the researchers to be less lonely, anxious,apathetic, depressed, and self-centered. Improve-ments in hygiene and other physical functionsincluded increased cleanliness and neatness, bettereating habits, normal bowel habits, and normalurinary function. Residents also had less difficultysleeping, took fewer sedatives, had less diarrhea, andwere less malodorous. No statistically significantchanges were noted over the l-year course of thestudy in the proportion of residents who weredependent in activities of daily living (i.e., bathing,dressing, eating, transferring, or walking) or in theproportion of residents who exhibited five behav-ioral symptoms (i.e., regressive behavior, wander-

ing, nighttime agitation, assaultiveness, and abusive-ness) (70).

Bullock et al. found improvements in behavioramong 13 female residents of a 20-bed special careunit (56).1 The researchers compared the frequencyof 11 behavioral symptoms over an 8-month periodfrom 4 months before until 4 months after the unitopened. The 11 behavioral symptoms were agita-tion, anxiety, combativeness, insomnia, resistive-ness, uncooperativeness, restlessness, withdrawal,verbal abusiveness, yelling, and taking off one’sclothes. In the 4 months after the special care unitopened, the frequency of 9 of the 11 behavioralsymptoms was greatly reduced (from 12 to 84percent, depending on the behavior). The frequencyof the other 2 behavioral symptoms—resistivenessand verbal abuse-increased 5 percent and 20percent, respectively. No other negative outcomesare noted in the study report. On the positive side, thereport notes slight reductions in the dosages ofpsychotropic medications received by some of theresidents. The statistical significance of the study’sfindings was not computed.

As part of the study by Bullock et al., briefinterviews were conducted with the unit residents(56). The residents were asked whether they likedthe unit; whether they were “very happy,” “prettyhappy,” or ‘‘not so happy;’ whether they weretreated well; and whether they were worried orrelaxed. In general, the residents expressed positiveattitudes toward the unit. No attempt was made toevaluate the reliability or validity of their responses.Moreover, since the interviews were conducted onlyonce, after the unit opened, it is not clear whetherthere were changes in the residents’ attitudes thatcould be attributed to the impact of the special careunit.

Cleary et al. found statistically significant im-provements in several aspects of the functioning andphysical condition of 9 residents of a 16-bed specialcare unit which is described in the study report as a“reduced stimulation unit” (88). Over a 6-monthperiod from 3 months before to 3 months after theiradmission to the unit, the residents’ average scoresimproved significantly on the Haycox DementiaBehavior Scale (176), an assessment instrument that

1 This study diilers from the other studies discussed in this section because the special care unit was in a mental hospital rather than a nursing home.OTA has included the study in this analysis of evaluative research on special care units because, like the other special care units included in the analysis,this special care unit is intended to serve only individuals with Alzheimer’s disease and related dementias. Other studies that have evaluated specializedunits in mental hospitals have focused on units that serve elderly persons with a variety of psychiatric conditions as well as dementia.

Chapter 4--Findings From Evaluative Studies ● 115

includes measurements in 8 areas (language/conver-sation, social interaction, attention/awareness, spa-tial orientation, motor coordination, bowel andbladder control, eating and nutrition, and dressingand grooming). The special care unit residents alsobecame significantly less agitated; use of physicalrestraints was significantly reduced; and the resi-dents’ weight increased. No changes were noted inresidents’ sleep patterns or use of psychotropicmedications. The researchers observed more interac-tions among residents and between residents andstaff members, but the study design did not includea measure of these interactions.

As part of the study by Cleary et al. interviewswere conducted with the unit residents to assess theirfeelings of security and well-being (88). The resi-dents were asked the same six questions at four timesbefore and four times after the unit opened. Theywere asked whether they felt safe; whether they gotthe help they needed; whether they got enough toeat; whether the unit was “a good place;’ whetherthey had a place to sleep; and whether they wereafraid. Nine of the 11 residents in the study samplecompleted all the interviews. In general, the resi-dents expressed a high level of security. Theirresponses were also highly consistent, suggesting itis possible to obtain consistent responses from somenursing home residents with dementia. Whether theresidents’ responses reflect their true feelings is notknown.

Greene et al. found improvements in behavior andother aspects of functioning among 6 residents of a26-bed special care unit (160). The researcherscompared the frequency of 10 negative indicatorsover a 4-month period for 5 of the residents and overa l-month period for one resident. The 10 negativeindicators were hostility, agitation, decreased appe-tite, failure to feed oneself, combativeness, failure toambulate, incontinence, inability to dress oneself,withdrawal, and hallucinations. The frequency ofeight of these indicators decreased to zero over thecourse of the study, and the frequency of the othertwo indicators-hostility and failure to ambulate—was greatly reduced. An improvement in cognitiveskills was found in two of the three residents inwhom cognitive skills were measured. An improve-ment in mood was found in the three residents inwhom mood was measured. The statistical signifi-cance of the study’s findings was not computed.

Hall et al. found reduced use of psychotropicmedications and desirable weight gain in residentsof a 24-bed special care unit described in the studyreport as a “low stimulus unit” (171). In the3-month period after their admission to the unit,psychotropic medication use was reduced or elimi-nated in 5 of the 12 individuals in the study sample.Prior to their admission to the special care unit, all12 individuals had been losing weight. In the 3months after their admission to the unit, 6 of theresidents gained weight; 5 stopped losing weight,and one continued to lose. The statistical signifi-cance of the study’s findings was not computed.

The study by Hall et al. was intended to evaluatethe effectiveness of the special care unit in reducingcatastrophic reactions, defined by the researchers toinclude outbursts of noisiness, agitation, combative-ness, sudden withdrawal, increased confusion andfear, intensified pacing, and nighttime wakefulness(171). The study did not include quantitative meas-urements of these indicators, however. The research-ers observed a decreased incidence of two of theindicators-agitation and nighttime wakefulness.Other positive outcomes were also observed, includ-ing increased social interaction among the residents,decreased wandering, and reduced incidence ofdelusions. The researchers point out, however, thatthese positive findings are based on subjectiveevaluations and that objective measurements ofvarious outcome indicators are needed.

Lawton et al. found statistically significant in-creases in friendliness and interest among 53 resi-dents who were moved from a 350-bed nursinghome to three 40-bed special care units in a new120-bed nursing home (245). The researchers com-pared the residents’ cognitive and self-care abilities,behavior, and mood at 4 times in the l-year periodbefore the move and 2 times, one month and 7months, after the move. Over the 19-month period ofthe study, the subjects showed a significant decreasein cognitive and self-care abilities. Following themove, the subjects spent less time in their bedroomsand more time in the social spaces, but there were nosignificant changes in social behavior, involvementin planned or staff-supervised activities, ambulation,behavioral symptoms, use of restraints, or time spentsleeping or doing nothing. There was an increase insolitary activities and a decrease in self-maintenanceactivities. Although the residents were judged by thestaff to be significantly more friendly and interestedafter the move, they were also judged to be

116 ● Special Care Units for People With Alzheimer’s and Other Dementias

significantly more depressed. There were no statisti-cally significant changes in any of the other moodstates studied (i.e., anxiety, anger, happiness, amuse-ment, agitation, and tranquility).

Lawton et al. also compared the behavior of 80residents of the 3 special care units and 40 residentsof the old 350-bed nursing home (245).2 Thecomparison showed the special care unit residentswere significantly more likely than the residents ofthe old nursing home to be involved in planned andstaff-supervised activities and significantly lesslikely to exhibit behavioral symptoms. On thenegative side, the special care unit residents weresignificantly less likely to be involved in self-maintenance activities. There were no significantdifferences between the special care units residentsand the residents of the old nursing home in socialbehavior, ambulation, involvement in solitary activ-ities, or time spent sleeping or doing nothing.

McCracken and Fitzwater found improvements inspecial care unit residents’ scores on the HaycoxDementia Behavior Scale (297), (as did Cleary et al.,discussed earlier). Over the l-year period of theMcCracken and Fitzwater study, 8 of the 11individuals in the study sample showed improve-ments in their overall scores on the scale. Improve-ments were noted in all but two of the measuredcharacteristics-motor coordination and dressingand grooming. The three subjects whose overallscores on the scale did not improve showed thegreatest decline in these two areas, as well as boweland bladder control, eating and nutrition, and spatialorientation. The statistical significance of thesefindings was not computed.

Mummah-Castillo found reductions in the dos-ages of psychotropic medications and desirableweight changes in residents of a 22-bed special careunit (312). Over a l-year period from 6 monthsbefore to 6 months after their admission to the unit,9 of the 10 individuals in the study sample showeda weight gain, and the dosages of psychotropicmedications were decreased for 7 of the 10 subjects.The statistical significance of these findings was not

computed. The researchers observed that aggressivebehaviors and catastrophic reactions were rare on theunit, but the incidence of these behaviors was notmeasured.

In Summary, all nine studies found some positiveoutcomes of the special care units they evaluated.The positive outcomes vary from one study toanother, and some of the findings are contradictory.As noted earlier, if the contradictory findings areexcluded, the only positive outcomes found in morethan one of the nine studies are decreased nighttimewakefulness, improved hygiene, and weight gain.

These studies are frequently cited as evidence thatspecial care units are effective. Often the research-ers’ general observations, rather than a study’sspecific findings, are cited. In many instances,findings that are cited from one study are contra-dicted by findings of another study.

All the studies suffer from one or more problemsthat raise questions about the validity of theirfindings-both positive and negative. One of theseproblems is small sample sizes. The second problemis the lack of rigorous research design and imple-mentation. In many of the studies, the outcomes tobe measured are not clearly defined, and themeasurement process is more impressionistic thanobjective or standardized. As noted earlier, thestatistical significance of the findings was computedin only four of the nine studies. Failure of the studiesto include a control group is another problem sincewithout a control group, the impact of the specialcare unit cannot be separated from the impact ofother factors that may affect resident outcomes.Finally, many of the studies were conducted by unitstaff members or other individuals who were in-volved in planning or operating the unit. Theseindividuals have an obvious interest in findingpositive outcomes. The potentially powerful effectof their expectations coupled with small samplesizes, lack of a rigorous research design, and lack ofcontrol groups means the results of the studies mustbe suspect.

Z ‘rh.is component of the study had a pre-post design like the other studies discussed in this SWiOn and an apptUent control group me the studiesin the following section of the chapter. The study is included in this section because the status of the control group is unclear. Some, but not all, of the80 special care unit residents were among the 40 residents of the old nursing home who constituted the control group (245). The study report providesno information about the special care unit residents who were not among the 40 residents of the old nursing home that constituted the control group,


EVALUATIVE STUDIES WITH ACONTROL GROUP: EFFECTS

ON RESIDENTSOTA is aware of six evaluative studies of special

care units in which a control group was used (seetable 4-2). The samples for these six studies are, onaverage, larger than the samples for the studiesdiscussed in the previous section. The six studiesvary in the outcomes they studied and their duration.The control groups they used also vary: four of thestudies used a control group consisting of individu-als with dementia in nonspecialized nursing homeunits that also serve nondemented residents; onestudy used a control group consisting of individualswith dementia in a segregated but nonspecializedunit; and one study used a control group consistingof individuals on the waiting list for admission to aspecial care unit. As described below, only two ofthe six studies found any statistically significantpositive outcomes for the special care unit residents.

Chafetz compared changes in cognitive and be-havioral characteristics over a 15-month period in 12residents of a 30-bed special care unit and 18residents of a 60-bed nursing home unit that servedonly individuals with dementia but provided nospecialized services (80). The study was designed totest the hypothesis that cognitive abilities woulddecline equally over time in residents of the twounits, whereas behavior would decline less inresidents of the special care unit. As shown in table4-2, the staff-to-resident ratios were similar in thetwo units, but the special care unit staff memberswere specifically selected and trained to work on theunit. The special care unit provided family meetingsand a more extensive activity program than thenonspecialized unit, and a few physical designfeatures distinguished the special care unit from thenonspecialized unit. The study found that bothcognitive abilities and behavior worsened over timein residents of the two units. The special care unithad no statistically significant effect on residents’cognitive abilities or their behavior, and there wereno positive outcomes that could be attributed to thespecial care unit.

Coleman et al. compared the rate of hospitaliza-tion over a l-year period for 47 residents of 2 specialcare units and 58 residents of 2 nonspecialized unitsin the same nursing home (99). The 58 residents ofthe nonspecialized units included 36 individuals

who had a diagnosis of dementia and 22 individualswho did not have a diagnosis of dementia. The studywas designed to determine whether special care unitresidents are less likely than residents of nonspecial-ized units to be hospitalized. The staff-to-residentratios were the same for the special care units and thenonspecialized units. The study report does notdescribe the differences in physical design or otherfeatures of the units. The study found no statisticallysignificant difference in the rate of hospitalizationfor the special care unit residents and the residents ofthe nonspecialized units. There was, however, anonsignificant trend for a larger proportion of thespecial care unit residents to be hospitalized over thecourse of the study (21 percent vs. 14 percent,respectively). The higher rate of hospitalization forthe special care unit residents was due primarily toa higher incidence of hip fractures: 9 percent of thespecial care unit residents, compared with only 3percent of the residents of the nonspecialized units,were hospitalized for hip fractures.

Holmes et al. compared changes in cognitive,functional, and behavioral characteristics over a6-month period in 49 residents of 4 special care unitsand 44 individuals with dementia in nonspecializedunits in the same 4 nursing homes (195). The studywas designed to measure the impact of a special careunit vs. a nonspecialized nursing home unit onindividuals with dementia. Table 4-2 lists the manydifferences between the special care units and thenonspecialized nursing homes in terms of staff,activity programs, and physical design features.Baseline measurements indicated there were statisti-cally significant differences between the special careunit residents and the residents of the nonspecializedunits at the start of the study. The special care unitresidents were, for example, more likely thanresidents of the nonspecialized units to be disori-ented and to exhibit behavioral symptoms. Thespecial care unit residents were also more likely tobe able to ambulate independently. After 6 months,the study found little change in any of the measuredresident characteristics, including cognitive abili-ties, mood, ability to perform activities of dailyliving, frequency of behavioral symptoms, sleepproblems, and ability to ambulate independently.Taking into account differences between the specialcare unit residents and residents of the nonspecial-ized units at the beginning of the study, theresearchers found no statistically significant positive

Table 4-2—Evaluative Studies With a Control Group

Year of Funding Duration ofCitation the Study Source subjects study Changes Made to Create the Special Care Unit

_- - -- ----Chafetz, 1981

Coleman et al.,1990

Holmes et al.,1990

Maas andBuckwalter,1990

Rovner et al.,1990

Wells and Jorm,1987

1988-1987

1987-1988

not reported

1986-1988

1885-1886

1986

University of Texassouthwestern Med-ical Center and Itsaffiliated Alzheimer’sDisease ResearchCenter

University of Cal-ifornia, San Fran-cisco, School of Meal-icine, and U.S.Health Resourcesand Services Admin-istration


National Center forNursing Research

Johns Hopkins Uni-versity’s affiliatedAlzheimer'sDiseaseResearch Center


12 residents of a 30-bedspecial care unit and 8 resi-dents of a 60-bed unit Inwhich individuals with de-mentia were segregated butno special services wereprovided

46 residents of two 28-bedspecial care units and 58residents of two 28-bed non-specialized units in the samefacility (of the 58 residentsof the nonspecialized units,36 had dementia, and 22did not)

49 residents of special careunits in 4 nursing homesand 44 residents with de-

13 residents of a 20-bedspecial care unit in a State-owned veterans home and9 residents with dementia innonspecialized units of thesame facility

14 residents of a 22-bedspecial care unit which iSpart of a 31-bed unit and 14residents with dementia innonspecialized units of thesame facility

12 residents of a specialcare unit in Australia and 10individuals with dementiawho were on the waiting listfor the unit and living athome

13 to 15 months

one year

6 months

2 years

one year

3 months, from justbefore admission tothe unit to 3 monthsafter admission

in the special care unit: access door secured with special locks; secure outdoorarea; 34 hours per week of specialized activities; staff selected specifically for theunit; staff training over a 10-week period and ongoing training; efforts to involvefamilies; family meetings every 6 to 8 weeks.

in the comparison unit: no special physical design features; 5 hours per week ofnonspecialized activities; no special staff training or special efforts to involvefamilies.

no physical design or other special features of the special care units are describedin the study report; the report says that the distinguishing features of the specialcare units “are similar to those found in the literature;” the staff-t-resident ratioswere the same on the special care units and the nonspecialized units.

in the special care units: dosed access doors with alarms; furniture with roundededges; special activity rooms; nurses’ station located near the exits to facilitatemonitoring residents; special activity programs; reality orientation; music programs;Increased staff-to-resident ratios; staff training; multidisciplinary team care.

in the comparison units: no special physical design features, activity programs, orstaff training.

in the special care unit: locked access doors; access doors camouflaged withmurals; secure outdoor area; separate day room/dining room; dividers in residentrooms to provide privacy; residents’ beds dose to the floor; curtains and wallhangings with velcro fasteners to prevent damage if residents pull on them; safetymirrors; safety glass; supplies stored out of view; no highly waxed floors; no stairsIn the unit; residents’ lockers and all but one drawer are locked to preventrummaging; flexible daily routine; efforts to reduce stimulation; subdivided diningroom to allow residents to eat In small groups; fabric wall decorations; colors thatare “functionally stimulating and reassuring;” orientation signs; piped-in music; pettherapy; specialized activity programs; activity barrel filled with pliable plastic itemsfor residents; multidisciplinary team; consistent staff; efforts to involve families.

in the comparison units: no special physical design features, activity programs, orstaff training.

in the special care unit: an activity room; staff training; weekly rounds with apsychiatrist and internist; staff efforts to identify residents’ specific cognitiveImpairments, to treat depression, delusions, and hallucinations, to recognizemedication side effects, to maintain residents’ physical health, to reduce use ofphysical restraints, and to increase participation in activities; 40 hours a week ofspecialized activities.

in the comparison units: no special physical design features, activities, or stafftraining; less hours of nursing care per resident (2.1 hours/day in the nonspecializedunits vs. 2.9 hours/day In the special care unit).

In the special care unit: corridors designed for wandering; secure outdoor area;private rooms; several activity rooms; home-like atmosphere; residents encouragedto bring their own furniture; unit philosophy of “normalization.”

for the comparison group: respite care, adult day care, and in-home services asneeded.

SOURCE: Office of Technology Assessment, 1992.


outcomes that could be attributed to the special careunit.

Maas and Buckwalter compared changes in cog-nitive, fictional, behavioral, and other characteris-tics in 13 residents of a 20-bed special care unit and9 individuals with dementia in nonspecialized unitsin the same facility (265). The study was designed tomeasure the effect of a ‘‘low stimulus’ special careunit vs. nonspecialized nursing home units onresidents with dementia, their families, and the unitstaff members. As noted in table 4-2, many physicaldesign and other changes were made to create thespecial care unit. Extensive baseline data werecollected in the year before the unit opened (264).After the unit opened, data were collected for oneyear at 2-month intervals. Due to subject attrition,complete data for the 22 subjects are available foronly a 10-month period, from 4 months before to 6months after the unit opened (265). These data showno statistically significant differences over time inthe cognitive or functional abilities of the specialcare unit residents and the individuals with dementiain the nonspecialized units. The most frequentlyreported behaviors for both groups of residents were‘‘sleeping/resting, ’ ‘ ‘quiet, ’ and ‘‘pleasant/happy.” Catastrophic reactions occurred, but theirfrequency decreased significantly from baselinelevels for both groups of residents.3 Nevertheless,catastrophic reactions were significantly less fre-quent in the special care unit residents than in theindividuals with dementia in the nonspecializedunits. The special care unit residents were alsosignificantly more likely than the individuals withdementia in the nonspecialized units to interact withstaff. There were no significant differences betweenthe two groups in the frequency of their interactionswith other residents or family members. The re-searchers noted a general trend for the subjects tobecome more active after being admitted to thespecial care unit. This increased activity includedboth positive and negative behaviors. The research-ers point out that:

Behaviors such as “screaming/yelling,” “pacing,”‘‘noisy,’ and “restless,” as well as a decrease in‘‘cooperative’ behavior may be seen as non-constructive. Positive behaviors such as “pleasant/happy, “ “talking/visiting,” “a wake,” and “up and

about,’ were all reported more frequently among theexperimental group. . . Viewed singly, no one be-havior (changed) significantly. However, whenviewed (together), it seems that important changes inoverall level of activity were occurring after intro-duction of the special care unit (265).

Other results of the study show that for their first fourmonths in the unit, the special care unit residentswere significantly less likely to be physicallyrestrained than the individuals with dementia on thenonspecialized units, but for the next 2 months, thespecial care unit residents were significantly morelikely to be physically restrained. Use of antipsy-chotic medications was significantly higher for thespecial care unit residents both at baseline andfollowing their admission to the special care unit.There was no significant difference between the twogroups in the total number of medications of allkinds that they were taking. Lastly, the special careunit residents were significantly more likely to fallthan the individuals with dementia on the nonspe-cialized units, but the increased incidence of fallswas not accompanied by an increase in injuries dueto falls.

Rovner et al. compared changes in fictionalability over a l-year period in 14 residents of a22-bed special care unit and 14 individuals withdementia in nonspecialized units in the same nursinghome (392). As shown in table 4-2, the special careunit provided more hours of nursing care and moreactivity programs than the nonspecialized units.Only one physical design change was made to createthe unit. In the view of the researchers, the distin-guishing features of the special care unit were theefforts of its multidisciplinary staff to accomplishsix objectives: 1) to identify residents’ specificcognitive impairments and associated disabilities,2) to treat depression, delusions, and hallucinations,3) to identify medication side effects; 4) to maintainresidents’ physical health; 5) to reduce use ofphysical restraints, and 6) to increase residents’participation in activities. Baseline measurementsindicated that the special care unit residents weresignificantly younger, on average, than the residentsof the nonspecialized units and that the special careunit residents were less likely to be taking medica-tions of all types. The study found that over a l-year

3 c~taS&ophic ~eaction was def~~ in tis study as ‘ ‘a reaction (mood change) of the resident in response to Wtit WY aPPe~ to s~to ~ ~stimuli (bathing, dressing, having to go to the bathmo~ a question asked of the person) which can be characterized by weeping, blushing, anger,agitatio~ or stubbornness. The reaction is not necessarily very dramatic or violen~ but may appear over-emotional or not appropriate for the stimulus’(265).


period, there was much less decline in the fictionalabilities of the special care unit residents than theresidents of the nonspecialized units: 14 percent ofthe special care unit residents and 64 percent of theresidents of the nonspecialized units declined intheir “level of care” as determinedly the number ofactivities of daily living with which they neededassistance. This statistically significant positiveoutcome is attributed by the researchers to theimpact of the special care unit.

Wells and Jorm compared changes in cognitive,functional, and behavioral characteristics over a3-month period in 12 residents of an Australianspecial care unit and 10 individuals who were on thewaiting list for the unit and living in the community(489). The study was designed to compare the effecton individuals with dementia and their families ofbeing in a special care unit vs. being deferred fromadmission. The study findings with respect to theimpact on the subjects’ families are discussed laterin this chapter. The physical changes made to createthe special care unit included an environmentaldesign to allow wandering, a secure outdoor area,and efforts to create a home-like atmosphere. Mostof the individuals on the waiting list received respitecare, and some received adult day care or in-homeservices. The study found that over a 3-monthperiod, the cognitive and functional abilities andbehavior of all the subjects declined. Except for atemporary worsening of behavioral symptoms amongthe special care unit residents in the first month ofthe study, there was little difference in the rate ofdecline in these characteristics between the specialcare unit residents and the individuals on the waitinglist.

In summary, four of the six evaluative studies thatused a control group found no statistically signifi-cant positive resident outcomes that could beattributed to the special care unit. One of the studieswith a positive resident outcome found that over al-year period the special care unit residents showedsignificantly less decline than individuals withdementia in the nonspecialized units in their abilityto perform activities of daily living (392). The threeother studies that used a control group and measuredresidents’ ability to perform activities of daily livingfound no significant effect of the special care units

in this area. The second study with positive residentoutcomes found that special care unit residentsexhibited significantly fewer catastrophic reactionsthan residents with dementia in the nonspecializedunits (265). The special care unit residents alsointeracted significantly more with staff members.

The research design and implementation of thesesix studies are far more rigorous than the design andimplementation of the nine studies discussed earlierthat did not use a control group. The outcomes aremore precisely defined and measured in these sixstudies, and their use of a control group increases thepresumed validity of their findings.

On the other hand, there are one or more problemswith each of the studies that could affect the validityof their findings-both positive and negative. Oneproblem is that several of the studies were conductedby individuals who were involved in planning orworking on the special care unit that was the focusof the study. In one of the two studies that found apositive resident outcome (392), the nurses whoevaluated the residents’ ability to perform activitiesof daily living were unit staff members whosejudgments about the residents could have beenbiased by their expectations about the effectivenessof the special care unit.4

A second problem that could affect the validity ofthe findings of some of the studies discussed in thissection is selection bias. If the special care unitresidents and the control group subjects differed insignificant ways at the start of the studies, thesedifferences, rather than the impact of the special careunit, could account for any observed differences inoutcomes. To address this problem, all six studiesdiscussed in this section compared the characteris-tics of the special care unit residents and controlgroup subjects at the beginning of the study, andseveral of the studies used statistical methods tocorrect for any observed differences in the twogroups.

As discussed in chapter 1, randomization ofsubjects to the special care unit or the control groupwould be the ideal way to address the problem ofentry point differences among subjects. Two of thestudies discussed in this section (265,489) randomlyassigned subjects to the special care unit or the

4111 addition to bias iIIrm&KXXI by StimemberS’ expectations, a more subtle form of bias could arise in this and other St’UdieS thttt rely On stimembers’ evaluations of residents’ ability to perform activities of daily living as a result of differences in the way impairments in activities of daily livingare pereeived on a special care unit vs. a nonspecialized nursing home unit.


control group. Randomization of subjects apparentlyworked well in the 3-month study by Wells and Jorm(489). Randomization also worked well initially inthe longer study by Maas and Buckwalter buteventually broke down, in part because some fami-lies were reluctant to move their relative who wasdoing well in a nonspecialized unit to the specialcare unit to meet the requirements of the studydesign (265).

A third methodological problem—and one thatcould affect the validity of the findings of Rovner etal. (392)—is failure to measure differences in thecognitive abilities of the special care unit residentsand control group subjects at the end of a study. Asnoted earlier, the outcomes measured in the study byRovner et al. were changes in the subjects’ ability toperform activities of daily living (392). In individu-als with dementia, ability to perform activities ofdaily living is related to some degree to cognitiveability (369,508). The special care unit residents andcontrol group subjects in this study did not differsignificantly in their cognitive abilities at the begin-ning of the study, but their cognitive abilities werenot measured at the end of the study, and significantdifferences could have developed. If such differ-ences did develop, they, rather than the impact of thespecial care unit, could account for the observeddifferences in the proportion of special care unitresidents vs. control group subjects that declined intheir ability to perform activities of daily living.

In addition to these methodological problems,there are difficulties in interpreting the findings ofthe six evaluative studies. In all six studies, thespecial care units differ in many ways from thecontrol group settings. It is unclear whether particu-lar features of the special care units or their overallmilieu account for the studies’ findings. A thirdpossibility proposed by Rovner et al. as an explana-tion for the findings of their study is that increasedstaff attention to the unit residents could account forthe positive outcome, irrespective of any specialfeatures of the unit (392). In all these studies, it isalso possible that certain aspects of the special careunits (i.e., particular features, milieu, or staff atten-tion) have a positive impact and other aspects havea negative impact, and that the two types of impactscancel each other out. Still another possibility is thatcertain aspects of the special care units have apositive impact on some residents and a neutral ornegative impact on other residents, and that theseimpacts cancel each other out. Small sample sizes,

lack of a common taxonomy for classifying individ-uals with dementia across studies, and lack of aprecise description of the features of each of thespecial care units make it impossible at present todifferentiate among these various explanations.

The one study that found a significant positiveeffect of a special care unit on the residents’ abilityto perform activities of daily living focused on a unitthat was created with the addition of an activity roombut no other physical design changes (392). Instead,the “special” features of the unit, in the view of theresearchers, were staff efforts to identify residents’specific cognitive impairments, to treat depression,delusions, hallucinations, and medication side ef-fects, to maintain residents’ physical health, and toincrease their involvement in activities. Ongoinginvolvement of a psychiatrist on the staff seems to beunique to this study. Whether any of these featuresare different enough from the features of the specialcare units in the other studies to explain theircontradictory findings cannot be determined fromthe available data.

STUDIES OF PARTICULARFEATURES AND INTERVENTIONS

IN SPECIAL CARE UNITS:EFFECTS ON RESIDENTS

Unlike studies that evaluate the overall impact ofa special care unit, some studies evaluate the effectof particular features and interventions in a specialcare unit. Such studies do not constitute special careunit research in the same sense as the studiesdiscussed earlier in this chapter because the featuresand interventions generally can be used in nonspe-cialized nursing home units and other residential andnonresidential care settings as well as in special careunits. The research to evaluate these features andinterventions can also be conducted in other settings.For these reasons, studies of particular features andinterventions in special care units are not discussedin the same detail in this report as studies thatevaluate the overall impact of the units.

The particular features and interventions that havebeen studied most in special care units are variousdevices and visual barriers to stop individuals withdementia from escaping or wandering away from theunit. To OTA’s knowledge, the frost research oninterventions of this kind was a study conducted inthe geriatric ward of a psychiatric hospital (198).

122 ● Special Care Units for People With Alzheimer’s and Other Dementias

That study found that when strips of tape wereplaced in any of four different grid patterns on thefloor in front of the exit doors, the frequency withwhich demented patients approached and touchedthe doors decreased significantly. Two other studieshave attempted unsuccessfully to replicate theseresults in special care units (75,77,316). Both studiesfound that use of strips of tape in front of the exitdoors resulted in a temporary increase but nosignificant long-term change in the number of timesper day the special care unit residents opened the exitdoors.

Other interventions to stop individuals with de-mentia from escaping or wandering away have alsobeen tested in special care units. Chafetz found thatuse of a second spring-loaded latch on the exit doorsstopped residents of one special care unit fromopening the doors (75,77). Namazi et al. found thatconcealing the exit doors with either a beige cloth ora green patterned cloth stopped residents of anotherspecial care unit from opening the exit doors (316).Two other interventions-painting the door knobthe same color as the door and using a door knobcover that allows the knob to turn only whenpressure is applied-also decreased the frequencywith which special care unit residents opened theexit door (316). The latter two interventions were notas effective as concealing the doorknob with apieceof cloth, however.

Researchers at the Corinne Dolan Alzheimer’sCenter in Chardon, OH, have conducted studies onmany other features and interventions in special careunits. The center was designed to facilitate researchof this kind. It has 2 separate but essentially identicalwings, each housing 12 residents, so alternateinterventions that require physical design or othermodifications to the unit can be tested in the 2 wingssimultaneously and their outcomes compared. Eightinterventions studied recently at the center are:

1.

2.

use of “significant” vs. “nonsignificant”personal belongings in showcases next toresidents’ rooms to help them identify theirrooms;

use of clearly visible toilets in residents’ roomsvs. toilets that are concealed behind a curtainto help them locate the bathroom and remaincontinent;

3.

4.

5.

6.

7.

8.

use of certain types, colors, and placements ofsigns to help residents locate the bathroomsand remain continent;use of partitions of various heights in thedining room and the activity rooms to reducedistractions for residents;use of unlocked vs. locked doors to an en-closed courtyard to enhance residents’ sense ofautonomy;use of special closet doors that allow residentsto see only one set of clothing at a time vs.ordinary closet doors to help residents dressthemselves independently;use of refrigerators with glass doors vs. ordi-nary refrigerators with opaque doors to allowresidents to see food and thereby encouragethem to eat when they are hungry; anduse of familiar tasks (e.g., washing dishes anddusting) vs. unfamiliar tasks (e.g., untanglinga box of hangers) to engage residents’ attentionand sustain their interest (314).

Results of some of these studies were published inlate 1991 (317), and results of the other studies willbe published in 1992.

The Dementia Study Unit in the Geriatric Re-search, Education, Clinical Center (GRECC) at theE.N. Rogers Memorial Veterans Hospital inBedford, MA, has also conducted studies on manyparticular interventions in special care units. TheDementia Study Unit includes three special careunits that serve elderly veterans with dementia. Theinterventions evaluated in the Dementia Study Unitinclude:

●

●

●

●

●

use of a hospice-like approach in the care of 40severely demented special care unit residents(474);substitution of normal feeding for tube feedingin six special care unit residents who werebeing tube fed on admission to the unit (475);use of a few beds on one of the special care unitsto provide respite care for 22 veterans withdementia who were still living in the commu-nity (238,405);use of antibiotics vs. palliative measures to treatfevers in special care unit residents (135); anduse of dietary changes and enforced rest periodsto maintain normal body weight in six specialcare unit residents who paced constantly (376).

Studies to evaluate the impact of other features andinterventions have been conducted or are underway


in special care units at several other VA medicalcenters (159).

An analysis of the results of studies of particularfeatures or interventions in VA and nonVA specialcare units and a comparison of these results with theresults of similar studies conducted in nonspecial-ized nursing home units and other residential andnonresidential care settings is beyond the scope ofthis report. The important point is that the existenceof special care units probably encourages research toevaluate particular features and interventions. It iseasier and more efficient to conduct this type ofresearch in a special care unit, in part because all theresidents have dementia. In addition, as discussed inchapter 1, the existence of special care units focusesattention on the special needs of nursing homeresidents with dementia and thereby encouragesresearch to evaluate particular features and interven-tions to address those needs.

Research on particular features and interventionsmay help to explain the findings of studies thatevaluate the overall impact of special care units. Ifparticular features or interventions are shown to beeffective or ineffective in general or for certain typesof residents, those findings may explain the contra-dictory results of studies that evaluate the overallimpact of the units. More importantly, however, thisresearch may identify features and interventions thatcan be used not only in special care units but also innonspecialized nursing home units and other resi-dential and nonresidential care settings to improvethe care of individuals with dementia.

STUDIES THAT EVALUATE THEEFFECTS OF SPECIAL CARE UNITS

ON UNIT STAFF MEMBERSOTA is aware of four studies that evaluate the

effect of special care units on unit staff membersover time. Two frequently cited reasons for estab-lishing special care units are: 1) a belief that trainingabout dementia can be more easily and effectivelyprovided for the staff of a special care unit than forthe staff of nonspecialized nursing home units andtherefore that special care unit staff members arelikely to be more knowledgeable about dementia,and 2) a belief that it is less stressful for staffmembers to work with residents with dementia on aspecial care unit than on nonspecialized units. Threeof the available studies measured the effect of aspecial care unit on staff members’ knowledge about

dementia; two studies measured the effect of aspecial care unit on staff stress and burnout, and onestudy measured the extent to which special care unitand other staff members were disturbed by thebehavioral symptoms of residents with dementia.

Chafetz and West compared knowledge aboutdementia among 1) 11 staff members of one specialcare unit, 2) 13 staff members of nonspecializedunits in the same nursing home, and 3) 30 staffmembers of nonspecialized units in another nursinghome (81). During the 9- to 12-month period of thestudy, the special care unit staff members partici-pated in 10 weekly training sessions about dementia.The staff of the nonspecialized units did not receivethis training. All staff members’ knowledge aboutdementia was measured at the beginning and end ofthe study using a 20-item true-false quiz. The studyfound that despite the training received by thespecial care unit staff members, there were nosignificant differences among the three groups ofstaff members in the extent to which their test scoreschanged over time. The researchers concluded thatthe training provided for the special care unit staffmembers did not have a significant or lasting effecton their knowledge about dementia.

Maas and Buckwalter compared knowledge aboutdementia among 21 special care unit staff membersand 55 staff members of nonspecialized units in thesame facility (265). During the frost 3 months afterthe special care unit opened, its staff members andthe staff members of the nonspecialized unitsreceived 80 hours of training about dementia. Thestudy found that during the baseline period beforethe unit opened and throughout the course of thestudy, the special care unit staff members scoredslightly higher than the staff members on thenonspecialized units on a 33-item test of knowledgeof dementia, but this difference was not statisticallysignificant. There was also no statistically signifi-cant change in the scores of the special care unit staffmembers over the course of the study. Registerednurses (RNs) scored significantly higher than li-censed practical nurses (LPNs), nurse aides, andnon-nursing staff members, regardless of whetherthey worked on the special care unit or the nonspe-cialized units.

Cleary et al. compared knowledge of dementiaamong the staff of a 16-bed special care unit at onepoint 3 months before the unit opened and again 3months after it opened (88). Despite a staff training

328-405 - 92 - 5 Ql 3


program conducted during this time period, thestudy found no significant change in the staffmembers’ knowledge of dementia. This study didnot have a control group.

With respect to job satisfaction, Cleary et al.compared special care unit staff members’ scores ona questionnaire administered at one point 3 monthsbefore the 16-bed unit opened and again 3 monthsafter it opened (88). The 83-item questionnaireaddressed 6 aspects of job satisfaction (workingconditions, professional considerations, professionalpreparation, emotional climate, supervision, andsocial significance). The study found no significantchange in the staff members’ scores before and afterthe unit opened. On the positive side, the researcherspoint out that the staff members did not seem to reactnegatively to the isolation of the special care unit, asmight have been expected. Moreover, in open-endedinterviews, some staff members reported they werespending much less time retrieving patients whowandered away from the unit and were experiencingfewer interruptions when caring for patients. Nomeasurements were made of the latter two outcomes.

Using the same 83-item questionnaire, Maas andBuckwalter compared job satisfaction among 21special care unit staff members and 55 staff memberson nonspecialized units in the same facility (265).The study found job satisfaction was “moderatelyhigh” for both groups of staff members during thebaseline period before the special care unit openedand throughout the course of the study. There waslittle difference between the scores of the two groupsof staff members on the questionnaire as a whole orany of its six subscales. RNs scored significantlyhigher than LPNs, nurse aides, and non-nursing staffmembers on one of the subscales-satisfaction withprofessional preparation-regardless of whether theyworked on the special care unit or the nonspecializedunits. After the special care unit opened, LPNs,nurse aides, and other non-nursing staff memberswho worked on the special care unit scored signifi-cantly higher on the same subscale than comparablestaff members on the nonspecialized units. Therewere no significant differences for the staff memberson any of the other subscales.

With respect to staff stress, Maas and Buckwalterfound a generally low level of stress among 15special care unit staff members and 49 staff memberson nonspecialized units in the same facility bothbefore and after the special care unit opened (265).

The special care unit staff members consistentlyreported less stress than the staff members on thenonspecialized units. Nevertheless, the study foundthat after the special care unit opened, its staffmembers experienced a statistically significant re-duction in stress, whereas the staff members on thenonspecialized units experienced an increase instress. The special care unit staff members also hadsomewhat lower scores than the other staff memberson a test of three indicators of burnout-emotionalexhaustion, depersonalization, and lack of a feelingof personal accomplishment; this difference inscores was statistically significant for deperso-nalization but not for the other two indicators. Thestudy’s findings with respect to use of sick leave,leave without pay, and overtime are still beinganalyzed (54).

Finally, in their study of special care units andnonspecialized units in the same four nursing homes,Holmes et al. compared staff members’ attitudestoward residents’ behavioral symptoms (195). At thebeginning of the study, although the special care unitresidents had significantly more behavioral symp-toms than the demented residents of the nonspecial-ized units, there was no significant differencebetween the staff members in the two types of unitsin the extent to which they reported being disturbedby the residents’ behavioral symptoms. After 6months, there was still no significant differencebetween the staff members in the two types of unitsin this regard.

In addition to these four longitudinal studies, twodescriptive studies have addressed the issue of staffstress in special care units. One study that comparedstaff stress on two special care units found that stresswas related to the severity of the residents’ impair-ment (506). Staff members on the unit with moreimpaired residents were more likely to report feelinghighly stressed than staff members on the unit withless impaired residents. Interestingly, many of thespecific types of stressors identified by staff mem-bers on both units were unrelated to residentcharacteristics and therefore might be expected tooccur as frequently in work with nondementedresidents and on nonspecialized nursing home unitsas on special care units. In another study of anonrandom sample of special care units, the re-searchers concluded staff stress was related tostaff-to-resident ratios: units with less staff perresident were much more likely than units with more


staff per resident to report problems with staff stress(332).

The University of North Carolina study of 31randomly selected special care units and 32 matchednonspecialized units in 5 States found staff turnoverwas significantly lower for RNs and LPNs on thespecial care units (291). Turnover was also lower fornurse aides on the special care units, but thisdifference was not statistically significant. Accurateinterpretation of these findings is difficult becausethey are based on data collected at one point in time.It is possible that pre-existing differences betweenthe staff members on the two types of units ratherthan differential effects of the units account for thedifferences in staff turnover.

I n summary, the three longitudinal studies thatmeasured staff knowledge of dementia found nostatistically significant effect of the special careunits. One of the two studies that measured jobsatisfaction found a statistically significant improve-ment in the scores of LPNs, nurse aides, and othernon-nursing staff of the special care unit on one ofsix aspects of job satisfaction. There were no othersignificant effects of the special care units on jobsatisfaction. The one longitudinal study that meas-ured staff stress and burnout found a statisticallysignificant reduction in stress among the special careunit staff members and a statistically significantdifference between the special care unit staff mem-bers and other staff members on one of threeindicators of burnout. There were no other signifi-cant effects of the special care unit on staff stress orburnout. Lastly, the study that measured the extentto which staff members were disturbed by residents’behavioral symptoms found no significant differ-ences over time for the special care unit staffmembers and no significant difference between thespecial care unit staff members and other staffmembers in this respect.

STUDIES THAT EVALUATE THEEFFECTS OF SPECIAL CARE UNITS

ON RESIDENTS’ FAMILIESOTA is aware of four studies that evaluate the

effect of a special care unit on residents’ familiesover time. One study conducted in Australia com-pared the psychological status of 12 family membersof individuals with dementia who were admitted toa special care unit and 10 family members ofindividuals with dementia who were placed on the

waiting list and offered in-home services (489). Atthe beginning of the study, the family members inboth groups showed high levels of symptoms onpsychological tests of anxiety, depression, guilt, andgrief. After 3 months, family members of the specialcare unit residents showed a statistically significantreduction in symptoms on all the tests. In contrast,family members of the individuals who had beenplaced on the waiting list showed little change in anyof the symptoms, except guilt, which was slightlyreduced.

Chafetz measured knowledge about dementia andattitudes toward older people among 12 familymembers of residents of a 30-bed special care unit(76). Anxiety and depression were measured among9 of the 12 family members. The study found nostatistically significant changes over a l-year periodin any of these areas, although there were somenonsignificant improvements in each of the areasexcept anxiety. This study had no control group.

Cleary et al. measured family satisfaction withcare among 11 family members of individuals withdementia who were moved from a nonspecializedunit to a new special care unit in the same nursinghome (88). Family satisfaction with the care pro-vided by the nonspecialized unit was quite high, asmeasured by a 38-item satisfaction questionnaire;nevertheless, family satisfaction increased signifi-cantly in the frost 3 months after the special care unitopened. This study had no control group.

In addition to the questionnaire, Cleary et al.conducted open-ended telephone interviews withthe family members (88). According to the studyreport, only 7 of the 11 family members visited theirrelative with dementia frequently enough in thespecial care unit to be able to respond in any detailto the open-ended questions. These seven familymembers reported their relative with dementia wasless agitated in the special care unit than he or shehad been in the nonspecialized unit. Five of theseven family members also reported they were betterable to communicate with their relative in the specialcare unit. None of the seven family membersexpressed concern that the special care unit wasisolated, and none described difficulties in visiting.

Lastly, Maas and Buckwalter compared familysatisfaction with care at 2-month intervals over al-year period among family members of special careunit residents and residents with dementia in non-specialized units of the same facility (265). Due to


subject attrition and replacement, the number offamily members varied over the course of the study,from 17 to 22 family members of special care unitresidents and from 12 to 21 family members ofindividuals with dementia in nonspecialized units.Both groups of family members reported fairly highlevels of satisfaction with the care their relative wasreceiving. They tended to be most satisfied with theirrelative’s overall care and least satisfied with thenursing care he or she was receiving. Familymembers of the special care unit residents hadsomewhat higher satisfaction scores than familymembers of the individuals with dementia in thenonspecialized units, but these differences were notstatistically significant.

In addition to these four longitudinal studies, anumber of cross-sectional studies have comparedvarious characteristics of families of special careunit residents and families of individuals withdementia in nonspecialized nursing home units.Since the findings of these studies are based on datacollected at one point in time, it is unclear whetherthey are attributable to the effect of the special careunits vs. the nonspecialized units or to preexistingdifferences between the two groups of families.

The study by Chafetz discussed above had across-sectional component that compared knowl-edge of dementia, attitudes toward older people,anxiety, depression, and guilt among three groups offamily members: 1) 18 family members of specialcare unit residents, 2)7 family members of residentsof a nonspecialized nursing home unit that servedboth demented and nondemented residents, and 3) 8family members of residents of a unit that servedonly individuals with dementia but provided nospecial services (76). The study found no significantdifferences between family members of the specialcare unit residents and family members of residentsof the two nonspecialized units in any of themeasured characteristics. Interestingly, all threegroups of family members had low levels of anxiety,depression, and grief. Moreover, in comparison withfamily members of the individuals in the segregatedbut nonspecialized unit, family members of thespecial care unit residents were significantly moredepressed and anxious.

A small pilot study done by researchers at theUniversity of North Carolina found that families ofindividuals with dementia in one special care unitwere, on average, more likely than families of

individuals with dementia in two nonspecializedunits to be satisfied with the physical aspects of theunit and the care their relative received and to feeltheir relative with dementia was better off in the unitthan at home (292). The findings differed for the twononspecialized units, however. Compared with fam-ilies of the special care unit residents, families ofindividuals with dementia in one of the nonspecial-ized units were as satisfied with the care theirrelative received, more satisfied with the physicalaspects of the care environment, and more likely tobelieve their relative was better off in the unit thanat home. In contrast, families of the residents in theother nonspecialized unit were less likely thanfamilies of the special care unit residents to besatisfied with the physical aspects of the unit and lesslikely to believe their relative was better off in theunit than at home.

Another small pilot study of two special care unitsand two nonspecialized nursing home units inCalifornia found that families of the special care unitresidents were less likely than families of residentsof the nonspecialized units to be satisfied with thephysical aspects of the unit and less likely to believetheir relative was better off in the unit than at home(256). Families of the special care unit residentswere also less likely to be satisfied with the numberof staff members, the adequacy of the care receivedby their relative, and the willingness of staffmembers to discuss the family members’ concerns.

Finally, the University of North Carolina study of31 randomly selected special care units and 32nonspecialized nursing home units in 5 States foundthat families of the special care unit residents weresignificantly more likely than families of individualswith dementia in the nonspecialized units to visittheir relative regularly (413).

Accurate interpretation of the findings of thesecross-sectional studies is difficult because the find-ings are based on data collected at one point in timeand therefore cannot be attributed with certainty tothe differential impact of the special care units vs.the nonspecialized units. It is possible, for example,that the finding of the University of North Carolinastudy--i.e., that families of special care unit resi-dents were significantly more likely than families ofindividuals in the nonspecialized units to visit theirrelative with dementia—reflects pre-existing differ-ences between the two groups of families rather thanthe impact of programs and policies of the two types


of units that might encourage or discourage familyvisiting.

Insummary, two of the four longitudinal studiesthat evaluate the impact of special care units onresidents’ families had statistically significant posi-tive findings. One of the studies found a significantincrease in family members’ satisfaction with care,and the other study found a significant reduction infamily members’ feelings of anxiety, depression,guilt, and grief. The other two longitudinal studiesfound no significant differences in these areas. Thetwo studies that had statistically significant positivefindings were much shorter than the two studies thatdid not have significant positive findings (3 monthsvs. 1 year, respectively).

One of the four cross-sectional studies had astatistically significant positive finding with respectto the frequency of visiting by families of the specialcare unit residents, but it is unclear whether thisfinding is attributable to the impact of the specialcare units. The findings with respect to familysatisfaction with care are contradictory, perhapsreflecting differences among the particular units inthe study samples.

CONCLUSIONBased on the preceding review of findings from

the available evaluative studies, some conclusionscan be drawn about the effectiveness of special careunits. Table 4-3 lists OTA’s conclusions from thestudies’ findings. In general, these studies show fewpositive outcomes of special care units. With respectto residents’ ability to perform activities of dailyliving, the findings of studies that did not use acontrol group are contradictory. Three of the studiesthat used a control group and measured residents’ability to perform activities of daily living found nosignificant effect of the special care units. Incontrast, one study (392) found less decline in abilityto perform activities of daily living over a l-yearperiod among the special care unit residents thanamong residents of the nonspecialized units. Like-wise, three of the studies that used a control groupand measured residents’ behavioral symptoms foundno significant effect of the special care units. Incontrast, one study (265) found fewer catastrophicreactions among the special care unit residents thanamong residents of the nonspecialized units. Onlyone of the four studies that measured the effect of aspecial care unit on the unit staff members found any

significant positive outcomes. The findings withrespect to family members’ feelings of depression,anxiety, and guilt and their satisfaction with care arecontradictory.

As noted at the beginning of this chapter, the factthat many of the available evaluative studies do notshow significant positive outcomes of special careunits is surprising. The failure of most of the studiesto show the expected positive outcomes is attributedby some commentators to methodological problems.The preceding discussion has noted many methodo-logical problems with the available studies. Asdiscussed in chapter 1, there are also numerousdifficult conceptual and methodological issues in-volved in designing special care unit research. Theseconceptual and methodological issues include un-certainty about which outcomes should be meas-ured; the difficulty of measuring certain outcomes inindividuals with dementia; the lack of validatedinstruments for measuring these outcomes; thedifficulty of identifying and correcting for differ-ences between special care unit residents andresidents of nonspecialized units that could affectthe study outcomes; and attrition in sample sizesover time which means even studies that started witha sample of a respectable size may end up withusable data on so few individuals that only a verystrong effect of the special care unit could bedetected.

Methodological problems and the difficult con-ceptual and methodological issues involved indesigning special care unit research probably ex-plain part of the failure of many of the availablestudies to find positive outcomes. Moreover, it mustbe noted that very few evaluative studies of specialcare units have been conducted thus far. Thepreceding sections discuss a total of only 15 studiesthat have measured impacts on residents and a fewadditional studies that have measured impacts onresidents’ families and/or unit staff members. On theother hand, some of the available studies, particu-larly the studies that used a control group, are welldesigned and carefully conducted, despite methodo-logical difficulties. The special care units theystudied incorporated the patient care philosophies,staff training, programmingg, and physical designfeatures recommended by special care unit advo-cates, and the researchers used accepted statisticalmethods to correct for baseline differences amongthe subjects that could affect the study outcomes.Thus, it is unlikely that the failure of these studies to


Table 4-3-OTA’S Conclusions From the Evaluative Studies of Special Care Units

● Evaluative studies of special care units that did not use a control group have found a variety of positiveoutcomes in special care unit residents. If contradictory findings are excluded, the positive outcomes foundin more than one of these studies are decreased nighttime wakefulness, improved hygiene, and weight gain.

● A few evaluative studies of special care units that did not use a control group have found improvements overtime in the important areas of residents’ ability to perform activities of daily living and residents’ behavioralsymptoms, but an equal number of studies of this type have not found such improvements.

● For of the six evaluative studies of special care units that used a control group have found no statisticallysignificant differences between the special care unit residents and the control group subjects in the followingareas: cognitive abilities, ability to perform activities of daily living, behavioral symptoms, mood, and rateof hospitalization. Two of the six studies of this type found certain statistically significant positive residentoutcomes: one study found that over a l-year period, 14 special care unit residents showed significantly lessdecline than 14 residents with dementia in nonspecialized nursing home units in their ability to performactivities of daily living; the other study found that 13 special care unit residents had significantly fewercatastrophic reactions than 9 residents with dementia in nonspecialized nursing home units; the 13 specialcare unit residents also interacted significantly more with the unit staff members. These two studies had noother statistically significant positive resident outcomes.

● Evaluative studies of particular features and interventions in special care units have focused primarily onmethods to deter individuals with dementia from escaping or wandering away from the unit. The mostsuccessful methods identified thus far are latches and locks the residents cannot open and various methodsof concealing the exit doors.

. Three of the four studies that evaluated the impact of special care units on the unit staff members found nostatistically significant effects. One of the 4 studies of this type found a statistically significant reduction instaff stress among 15 special care unit staff members and a statistically significant difference between the15 special care unit staff members and 49 staff members on nonspecialized nursing home units in one ofthree indicators of burnout. The study also found a statistically significant improvement in the scores of 16special care unit staff mernbers (licensed practical nurses, nurse aides and other non-nursing staff members)on 1 of 6 indicators of job satisfaction. None of the three studies that measured staff knowledge of dementiafound any significant effect of the special care unit.

. Two of the four studies that evaluated the impact of special care units on the residents’ families hadstatistically significant positive findings. One of the studies found a significant increase in the familymembers’ satisfaction with the care provided for their relative with dementia, and the other study found asignificant reduction in the family members’ feelings of anxiety, depression, guilt, and grief. The other twostudies of this type found no significant changes in either of these areas. One cross-sectional study foundthat families of special care unit residents are more likely than families of individuals with dementia innonspecialized units to visit their relative regularly, but it is not clear whether this finding is attributable tothe effect of the special care unit or to preexisting differences between the two groups of families.

SOURCE: Offke of ‘lkhnology Assessment, 19924

show positive outcomes is due entirely to methodo- effect of special care units is on residents’ quality oflogical problems. Alternate explanations include thepossibility that some or many of the featuresrecommended for special care units are not effectiveand the possibility that some of the recommendedfeatures have a positive effect on some or allresidents, families, and staff members, that otherfeatures have a negative effect, and that thesepositive and negative effects cancel each other out.Still another possibility is that the primary positive

life-an outcome that is difficult to define opera-tionally and one that has not been measured directlyin any of the studies conducted thus far. Furtherresearch is needed to differentiate among these andother possible explanations.

Research on specific interventions in special careunits may help to explain the findings of studies thatevaluate the overall effect of the units by showingthat certain interventions have positive outcomes


and others do not. This type of research is alsoimportant because some and perhaps many interven-tions that are shown to be effective in special careunits can also be used in nonspecialized nursinghome units, residential care facilities, and othersettings to improve the care of individuals withdementia in these settings.

Finally, it is important to note certain findings ofseveral of the studies discussed in this chapter thatdo not fit with widely held beliefs about nursinghome residents with dementia, their families, andnursing home staff members who work with resi-dents with dementia:

three studies found that the incidence of behav-ioral symptoms was much lower than expectedamong residents with dementia (22,265,312);one study found that three groups of familymembers—family members of special care unitresidents, family members of residents of anonspecialized nursing home unit, and familymembers of residents of a unit in which

●

●

individuals with dementia were segregated butno special services were provided—had muchlower levels of anxiety, depression, and guiltthan expected (76);two studies found moderately high familysatisfaction with the care provided for individu-als with dementia in nonspecialized nursinghome units (88,265,266); andone study found that staff members in fourspecial care units and four nonspecializednursing home units were not particularly dis-turbed by the residents’ behavioral symptoms(195).

It is unclear whether these findings reflect uniquecharacteristics of particular study samples or aremore generally representative. Certainly, if thebaseline levels of behavioral symptoms amongresidents, negative feelings among family members,and distress among staff members are low in generalor in particular study samples, it is unrealistic toexpect large positive changes in a special care unit.

special care units for people with dementia: findings from

Documents