Sheltered Care Facility Size and the Social Integration of Mentally Ill Adults

Abstract

This paper examines the effects of facility size on the social integration of mentally ill residents in community-based sheltered care homes. Big facilities have come to be associated with many problems (documented and otherwise) related to residential housing for mentally ill adults. However, this image of large, impersonal, and overly restrictive living environments is often not substantiated by the empirical research on sheltered care. Indeed, the present study of sheltered care homes in California finds a much more complex and varied situation. It suggests that homes of all sizes may offer many of the same opportunities for social interaction and that larger homes in some instances may possess certain advantages over smaller facilities.

Sheltered care homes for adults with mental illness have a long-standing image problem in this country. Such facilities include board and care, halfway houses, family care, and all other facilities except non-hospital/nursing homes settings that offer 24-hour supervised residential care. Despite repeated efforts to improve operating standards and practices (Kusserow, 1982; General Accounting Office, 1989), the popular conception of sheltered care homes remains one of therapeutic neglect, marginal living conditions, and inept if not unscrupulous business practices on the part of poorly trained and financially squeezed facility operators. The “excessive size” of many sheltered-care facilities has received its share of the blame for these problems. Indeed, in the eyes of their critics, large residential facilities are virtually synonymous with the oft-decried trend toward reinstitutionalization (Isaac & Armat, 1990; Torrey, 1988), with its evocative imagery of excess restrictiveness and bureaucratic indifference traditionally associated with state mental hospitals (Goffman, 1961).

Against such a backdrop, it may come as a surprise to learn that this one-sided view of large facilities is largely unsubstantiated by the empirical research on the subject which has appeared in the sheltered care literature. To be sure, many boarding homes and single room occupancy hotels bear their share of the blame for the poor living conditions exposed in repeated surveys of the industry. But like most things, size can have benefits as well as liabilities (Blau & Scott, 1962; Segal & Aviram, 1978). It is curious in this respect how few of the many oft-cited advantages of organizational size (e.g., economies of scale, work specialization, financial stability) are ever alluded to in the sheltered care literature.

It seems time, therefore, to take a closer look at facility size—its advantages as well as its drawbacks—as a potential predictor of residential life in sheltered care homes for adults with mental illness. The present paper focuses on a specific side of this broader issue (i.e., whether, and if so, how the size of a sheltered care home affects the “social integration” of its residents). We begin by reviewing past studies which have looked at the empirical effects of facility size on the living environment and social activities of sheltered care residents. Next, we examine these same effects using a representative sample of sheltered care homes in California. Those effects found to be statistically significant in our analysis are then re-examined to determine what it is about facility size that explains them. Finally, we consider the implications of our findings for future research and policy in the area.

Background

A small number of research studies in the sheltered-care literature have examined the effects of facility size on residential life in sheltered care homes. In all of these studies, facility size is conceptualized as the number of residents living in the home. Although some researchers have examined size differences in the psychiatric status of sheltered-care residents, most studies have focused on the way facility size affects the living environment of sheltered-care homes and the involvement of residents in facility-based and community-based activities.

In a five-state study of 210 Veterans Administration patients discharged to community foster care, Linn, Klett, and Caffey (1980) found that the more occupants in the home, the more likely the patient’s social functioning was to deteriorate. This finding was reached by comparing residential size (as measured by the average “number of residents, children and other [than the respondent] patients” in each home) across three “outcome groups”—“improved” (average size = 7.35 residents), “same” (average size = 9.10 residents) and “deteriorated” (average size = 11.63 residents), as measured by the “social dysfunction rating scale. Without offering empirical evidence for their conclusion and despite the narrow size range of their facility sample, the authors (p. 131) interpreted these differences to mean that larger homes foster an institutional-like “atmosphere” in contrast to the “home-like” ambiance characteristic of smaller foster care homes. (Linn et al., 1980).

These methodological weaknesses notwithstanding, Linn et al.’s (1980) conclusion finds support in a subsequent study by Hull and Thompson (1981) of 157 sheltered care facilities in Manitoba, housing from 1 to 30 residents. Among the 296 mentally ill residents living in these homes, the authors found a negative (r = −.364) association between facility size and the “normalization” of the residential environment. Indeed, Hull and Thompson found facility size to be their “most potent predictor” of “normalization,” as measured by the 30-item PASS normalization scale and defined by the authors as “the utilization of means which are as culturally normative as possible, in order to establish and/or maintain personal behaviors and characteristics which are as culturally normative as possible (p. 107).”

The harmful effects of facility size is also a recurring theme in recent studies of the social re-integration of formerly institutionalized patients. For example, in a study of 87 previously hospitalized chronically mentally ill residents in Washington State, Kruzich (1985) found a negative association (r = −.28) between “community integration” and facility size which she interpreted to mean that small facilities encourage higher levels of resident involvement. This relationship, however, must be viewed with some skepticism since no attempt was made to control for other influences (e.g., funding levels, restrictiveness, management style) on resident integration. Kruzich (1985) also ignored important differences in the type of facility being analyzed. By failing to distinguish intermediate care facilities from skilled nursing homes and exclusively residential settings, Kruzich’s study risks confounding possible size effects with other facility characteristics such as case mix, treatment intensity, and service orientation.

Indeed, in a separate study of “internal integration” among the same 87 residents, Kruzich and Kruzich (1985, p. 381) found that size differences in integration, while originally quite pronounced (r = .35), essentially disappeared (beta = .08) once the authors controlled for type of facility and other factors, including number of skills programmed, the rigidity of routine, and social distance.

Work by Segal and Aviram (1978) casts further doubt on the idea that facility size necessarily worsens living conditions or other facility characteristics (e.g., program activities) that potentially benefit sheltered care residents. In their statewide survey of 214 sheltered-care facilities in California, the authors found little evidence of size differences in illness severity. Indeed, using Overall and Gorham’s (1962) “Brief Psychiatric Rating Scale,” the residents in small (1–6 beds) and mid-sized (7–50 beds) facilities were actually found to be more severely disturbed than those in large (over 50 beds) facilities. And in a separate analysis using Langner’s (1962) 22-item psychological distress scale, facility size was found to be unassociated with psychiatric impairment (Segal & Aviram, 1978).

These contrary findings notwithstanding, work by Nagy, Fisher, and Tessler (1988) continues to echo Kruzich’s (1985) earlier conclusions about facility size. In a broadly based study of 851 sheltered-care residents and 210 homes spanning seven different states (CA, CO, FL, MA, MN, TX, and WA), the authors (Nagy, Michael, Fisher, & Tessler, 1988) examined the effects of facility size on resident participation in both community-based and facility-based activities. Although facility size was unrelated to the productive activity of residents, it was found to have a negative effect on both activity measures, even after controlling for other facility characteristics (e.g., ownership, profit status, resident turnover) as well as a number of additional factors. These negative effects were found to be even more pronounced among “socially impaired” residents which the authors took to mean that “residents in larger facilities receive less of the help they need” (Nagy et al., 1988, p. 1284).

In summary, the evidence on facility size is spotty and inconsistent. In fact, how size affects sheltered care seems to depend upon the geographic locale, the outcome being studied, and other characteristics of the facility, including its community setting and residential composition. Another way of interpreting these apparent inconsistencies is to recognize that size may have advantages as well as liabilities. To the extent that “largeness” has both positive and negative influences, the effects of facility size may cancel out one another, the potential result being no noticeable effect whatsoever. This possibility raises the questions of (a) how to separate the good from the bad effects of facility size, and (b) how to structure living environments so as to maximize the benefits of “bigness” while minimizing its negative effects.

Methods

Study Groups

The data for our analysis are taken from the second wave of a statewide longitudinal survey of 214 sheltered-care facilities in California (Segal & Aviram, 1978). The operators of these facilities were originally interviewed in 1973. Follow-up interviews were conducted between 1983 and 1985 with 151 of the original 214 operators (55 from the original sample and 96 new operators) interviewed in 1973 (Segal, Hazan, & Kotler, 1990; Segal & Kotler, 1993; Segal & Aviram, 1978). The analyses in the present paper are based upon these 151 facilities.

Facility Size

Facility size, the primary independent variable in our study, is operationally defined as the number of residents living in each facility as of the interview date. For purposes of this analysis, we recoded facility size into three categories so as to correspond to official zoning and licensing classifications in California (California State Department of Social Services, 1979):

1. Small facilities: Single family units with 1–6 residents.

2. Mid-sized facilities: Halfway houses and other group living facilities with 7–50 residents.

3. Large facilities: hotels, unlicensed nursing homes, and other sheltered-care facilities with more than 50 residents.

All facilities presented themselves as providing twenty-four hour supervised residential care to people with severe mental illness.

Intervening and Dependent Variables

Our study looks at four dimensions of the living environment of sheltered care, the (a) financial resources available for program activities, (b) rules and procedures governing what is considered permissible behavior by residents, (c) style of management employed by facility operators, and (d) extent to which residents interact with each other and participate in social activities within the broader community. The first three factors (i.e., activity programming, restrictiveness, and management style) are viewed as both “dependent” on facility size and as mediating at least some of the effects of facility size on the social integration of residents.

Social Integration

We measure social integration through two sets of survey questions about resident participation in various work, recreational, and learning activities. Since these questions were directed to the manager of the facility, the effective sample size for the analysis is the 151 facility operators rather than the number of residents residing in these homes.

The first set of survey questions covers thirteen social activities taking place within the facility, including discussion groups, classes or lectures and parties. For each activity, the facility operator was asked to indicate how many residents participate in each activity at varying levels, such as “very rarely or never,” “a few times a year.” A single level of participation measure was computed for each activity by taking the proportion of all residents who reportedly participated at least “once or twice a month.” The resulting score for each activity is interpreted as a measure of “internal integration.” An overall “index of internal integration” was subsequently constructed by summing the thirteen separate scores for each facility.

The second set of survey questions cover fifteen activities that take place outside the facility, including “visiting friends or relatives,” “going to a sports event” or “going on a picnic.” The responses of facility operators were coded the same as before. A separate “level of participation” measure was likewise computed for each activity and a summated index was constructed by adding together the respective scores on all items. The sixteen resulting scores are interpreted as measures of “external integration.”

Explanatory Factors

From a public policy perspective, knowing how size affects social integration is as important as knowing the direction (positive or negative) and the magnitude of its effects. Segal & Aviram’s (1978) sheltered care database permit us to examine three policy relevant ways in which facility size may affect the social integration of sheltered care residents.

1. Program Expenditures. Larger homes enjoy certain economies of scale, not the least of which being proportionately smaller fixed costs, such as mortgage payments and utilities. Other things being equal, therefore, the more (paying) residents in a home, the more revenues potentially available to invest in program activities and other areas of group life (e.g., facility maintenance and upkeep) which directly benefit the residents. We hypothesize, therefore, that larger homes will spend proportionately more per resident on such activities with a net gain in social integration being the likely result. We measure program costs by the actual dollar estimates given by facility operators when asked to itemize their monthly operating costs.

2. Restrictiveness. Restrictiveness, in the context of a group living facility, typically refers to the extent to which residents’ actions and behaviors are regulated by policies and procedures imposed from without, usually by the facility manager. In the classical Weberian view of bureaucratic organizations, increased size is associated with greater reliance upon formal rules and procedures (Weber, 1947; Blau & Scott, 1962; Mouzelis, 1967), largely because of the increased efficiency thought to result. By this rationale, we would expect the level of restrictiveness to rise with the number of residents living in a home. If so, facility size should have a dampening effect on social integration, since larger homes presumably impose more rules and procedures on their residents and since more restrictive settings presumably reduce a resident’s opportunity to interact with others. For the purposes of this analysis, we measure rules and procedures by how many of nineteen actions and behaviors (e.g., “keeping a pet,” “leaving the building”) facility operators say they “encourage,” “allow,” “discourage” or “do not tolerate.”

3. Management Style. The “management style” of facility operators can obviously vary along many different dimensions—the extent to which a manager shares power and responsibility, the manager’s level of involvement in the day-to-day operations of the facility, the relative importance the manager places on therapeutic goals as opposed, say, to financial matters, etc. Because of our interest in the “institutional aura” of sheltered care, we constructed a revised version of Segal & Moyles (1979) “client/management scale” which measures the relative “openness” of the living environment in sheltered care homes (Segal & Moyles, 1979). The original Segal-Moyles scale consists of nine survey items addressed to facility operators. Eight of the nine are taken from the well-known Community Oriented Programs Environment Scale (Moos, 1974) and focused on the extent to which residents’ activities were planned or regulated by facility staff. In line with the other eight, the ninth item asked whether residents must obey a “curfew?” The second wave of the Segal and Aviram (1978) survey contained only seven of the original eight items included in the first wave. As a result, our “revised” scale includes eight rather than nine items overall. Otherwise, we followed the same scaling procedures used by Segal and Moyles (1979) to construct the Client/Management Scale.

Results

Size Effects on Program Finances

According to our results, facility size has no apparent effect on profitability, whether measured by earnings margins (p = .730) or by profit per resident (p = .440). Likewise, large and small facilities spend virtually the same (p = .971) amount on their residents—$381.08/month per resident for large homes, $393.72 for mid-sized facilities and $392.57 for small homes. These basic similarities notwithstanding, the way expenditures are distributed varies markedly by facility size. While small homes spend more on “nondiscretionary” items like utilities ($31.80 for large facilities, $37.09 for mid-sized homes, and $59.55 for small homes; where p =.001) and food ($73.53, $86.90, $135.51, respectively; where p = .000), larger homes invest more in “discretionary” items like program activities ($83.78/month, $79.33/month and only $9.47/month, respectively; where p =.021).

Size Effects on Restrictiveness

Contrary to our expectations, large facilities tend to be less restrictive than mid-sized or small homes. Of nineteen facility rules and regulations (e.g., “smoking in bed,” “keeping a pet”) covered by the Segal and Aviram (1978) survey, restrictiveness varies significantly (at the .05 level) on seven. Of these seven significant items, the level of restrictiveness decreases monotonically with size in every instance but one. While smaller facilities are more lenient about “alcohol at meals,” larger facilities are more likely to allow residents to (a) lock their doors, (b) have sexual relations with a friend, (c) keep a hotplate or coffeemaker in their room, (d) leave the building unannounced, (e) use their own furniture, and (f) keep their door closed. Larger facilities are also more tolerant in the aggregate, as measured by our summated restrictiveness scale which was found to be statistically significant at the .001 level.

Size Effects on Management Style

To judge by our revised index of “management style,” larger facilities are no more “management-centered” than smaller facilities. In fact, of the eight items included in our revised “management-client” scale, only one was found to vary significantly by size (p = .053)—larger homes being more likely to require residents to “make detailed plans before leaving the house (i.e., going out for an activity).”

Size Effects on Social Integration

Internal Integration

As shown in Table 1, in most cases, residents of larger facilities are no more nor less likely to participate in social activities (discussion groups, social groups, religious services, etc.) that take place within the home. Of the thirteen facility-based activities covered by Segal and Aviram’s (1978) survey, only four showed statistically significant (at the .05 level) size differences in resident participation. Moreover, for three (parties, films and movies, and reality groups) of these four activities, participation levels rise monotonically with facility size. The sole exception is “outside entertainment” where participation levels actually decrease with size of home. As with most of its constituent items, our summated index of “internal integration” was not found to yield statistically significant differences by facility size (p = .349).

Table 1.

“Internal Integration” of Board and Care Residents by Size of Home, 1985

	Percent of residents engaging in activity
Activity	D*	n	All homes (mean)	Small (1–6)	Mid-size (7–50)	Large (51–400)	p-value
Parties	+	124	40.0%	22.4%	42.4%	58.9%	.005
Films/movies	+	123	36.8%	27.7%	35.9%	53.4%	.048
Reality groups	+	117	30.3%	14.2%	34.1%	44.9%	.014
Outside entertainment	-	123	36.3%	50.9%	26.2%	42.0%	.014
Board games	U	119	50.5%	51.4%	44.1%	66.4%	.073
Social groups	-	123	29.2%	40.1%	27.2%	18.9%	.128
Religious services	U	118	40.9%	48.9%	36.6%	40.5%	.265
Self-help group	+	122	31.7%	24.8%	33.9%	35.8%	.512
Social hour	U	122	48.5%	52.8%	44.0%	54.9%	.515
Discussion groups	+	121	53.0%	46.6%	54.5%	57.9%	.568
Exercise	-	122	59.2%	64.7%	58.0%	54.9%	.628
Arts & crafts	U	121	35.7%	38.8%	34.3%	34.9%	.853
SUMMATED SCALE	U	128	41.0%	40.9%	38.9%	46.7%	.349

SOURCE: 1985 Facility Survey (unweighted facility sample).

^*Direction of relationship, where “+” = positive, “−” = negative, and U = curvilinear.

External Integration

As with “internal integration,” in most cases, residents of larger homes are just as likely to participate in social activities that take place outside the home (see Table 2). Of the fifteen activities covered in the Segal and Aviram (1978) survey, only two—’eating out’ and ‘attending church’—showed statistically significant (.05-level) size differences in resident participation. In contrast to the four significant “internal integration” measures, however, participation levels decreased monotonically with facility size in both cases. Our summated index of external integration was also found to be negatively associated with facility size, although the relationship was not statistically significant at the .05 level (p = .145).

Table 2:

“External Integration” of Board and Care Residents by Size of Home, 1985

	Percent of residents engaging in activity
Activity	D*	n	All homes (mean)	Small (1–6)	Mid-size (7–50)	Large (51–400)	p-value
Eat out	-	116	51.1%	70.6%	44.9%	37.8%	.003
Attend church	-	114	34.8%	45.6%	33.0%	23.1%	.050
Attend concert/play	+	118	7.8%	2.9%	8.1%	14.9%	.083
Attend funeral	U	117	2.0%	5.9%		1.6%	.117
Go to recreat’n center	-	118	29.9%	40.3%	26.7%	22.9%	.158
Go shopping	-	119	73.7%	81.1%	72.7%	62.7%	.176
Go to parties	-	118	18.0%	24.0%	18.0%	8.4%	.239
Go to seniors’ center	-	119	8.3%	12.8%	7.7%	4.2%	.291
Volunteer/paid work	U	120	12.9%	7.9%	15.7%	12.9%	.429
Attend movie	U	119	26.0%	29.3%	22.2%	31.3%	.469
Visit friends/relative	U	120	40.0%	48.5%	47.2%	39.1%	.602
Go to sports event	+	115	12.9%	10.3%	13.7%	14.6%	.808
Go on overnight trip	U	120	7.7%	6.4%	8.5%	7.4%	.850
Go on picnic	+	119	23.7%	22.9%	22.9%	27.4%	.883
Go on ride/tour	-	115	52.3%	55.1%	51.8%	49.8%	.892
SUMMATED SCALE	-	122	27.7%	31.7%	26.3%	24.1%	.145

SOURCE: 1985 Facility Survey (unweighted facility sample).

^*Direction of relationship, where “+” = positive, “−” = negative, and U = curvilinear.

Explaining Size Effects

Table 3 shows the significant size effects from Tables 1 and 2 after controlling, first separately and then jointly, for the three facility characteristics—program spending, restrictiveness, and management style—described above. The first column of the table reports the zero-order association (i.e., “eta”) between facility size and each of the seven (four “internal” and three “external”) activity items, as estimated from the “multiple classification analysis” procedure (Nie, Hull, Jenkins, Steinbrenner, & Bent, 1975, p. 416). Columns two through four show the change in eta after controlling for each explanatory factor. A decrease in the zero-order association measured by eta indicates that part of the size effect in question is due to the explanatory factor being controlled. In such cases, the explanatory factor can be said to “mediate” the effect of facility size on social integration. An increase, on the other hand, indicates that the covariate being controlled actually serves to “suppress” the full magnitude of the relationship between size and integration. The final column of the table shows the combined effects of controlling for all three explanatory factors simultaneously.

Table 3:

Size Effects on Social Integration Controlling for Three Covariates, 1985

Activity	Change in eta/beta, controlling for
	Unadjusted ETA*	Program spending	Restrictiveness	Management Style	All three Cov’s
	Internal Integration
SUMMATED SCALE	.11	(.02)	.06	.01	(.05)
Parties	.26	(.05)	.01	.01	(.05)
Films/movies	.29	(.04)	.03	.01	(.07)
Reality groups	.32	(.02)	.00	.01	(.02)
Outside entertainment	−.33**	.00	.02	(.01)	.00
	External Integration
SUMMATED SCALE	−.21**	.02	(.01)	(.01)	.02
Eat out	−.39**	.02	.03	.01	.06
Attend church	− 29**	(.01)	.02	.00	.02

SOURCE: 1985 Facility Survey (unweighted facility sample).

^*Eta values based upon unadjusted effect in ANOVA models including all three covariates.

^**Negative signs (“−”) added to eta statistics to highlight direction of relationship, as noted in Table 1 or Table 2.

Internal Integration

We will begin with the four measures of internal integration which were found to differ significantly by facility size. As shown by the statistics in the second column of Table 3, at least part of the effect of size on internal integration is due to differences in program spending among large, mid-sized, and small residential facilities. Except for “outside entertainment,” controlling for program spending reduces the original zero-order correlation between facility size and social interaction within sheltered care homes.

By contrast, neither “restrictiveness” nor “management style” help explain the significant size differences in facility-based participation levels which we have found. Indeed, restrictiveness operates more as a suppressor variable than a mediating factor, causing three of the four zero-order size effects to appear smaller than they actually are. Management style has similar but somewhat smaller suppressor effects, except for “outside entertainment” where management-centeredness operates as a minor mediating influence. Finally, controlling for all three factors simultaneously has a substantial effect on two (i.e., “parties” and “films/movies”) of the four activities, reducing the zero-order size effect on “parties” by (.05 / .26 =) 19.2% and on “films and movies” by (.07 / .29 =) 24.1%. It should also be noted that the same three covariates produce a (.05/.11 =) 45.5% reduction in the zero-order correlation with our summated measure of “internal integration,” although, as already noted, this zero-order relationship was not found to be statistically significant at the .05 level.

External Integration

We turn next to the two measures of external integration which were found to be statistically significant. In contrast to the size effects we have found within facilities, program spending has no consistent effect on the size differences we have found in external participation levels. On the other hand, restrictiveness and management style both act predominantly as suppressor variables, attenuating the full effect of facility size on social interaction outside sheltered care homes. Not surprisingly, the size of these suppressor effects is particularly pronounced when simultaneously controlling for all three factors. In the case of “eating out,” for example, failing to take account of these three factors tends to understate the effect of facility size by (.06/(.39 + .06) =) 13.3%. Thus, facility size seems to be an important factor in inhibiting activities outside the facility.

Discussion

The size of residential homes has been widely blamed for contributing to the poor living environments repeatedly found to characterize many sheltered care facilities for the mentally ill in this country. In the eyes of many, “bigness” per se has come to be synonymous with much that is wrong with our sheltered care system as well as a popular point of departure for reforming the current system. Indeed, in a related context, some even see the passing of large residential facilities as a natural extension of current trends away from “mass congregate care to smaller forms of group care” (Dore, Young, & Pappenfort, 1984).

Our review of research on the topic suggests that the actual situation may be considerably more complex. While large facilities have been found deficient in a number of areas, much of the relevant sheltered care literature is based upon disparate samples of facility types and locales as well as widely varying measures of resident outcomes. Still other studies (e.g., see Segal & Aviram, 1978) have reported offsetting results which further cloud our empirical picture of the way facility size affects residential life in sheltered care homes. There is also a troubling tendency throughout this entire literature to ignore the possible benefits of facility size, notwithstanding a sizable body of sociological and economic research on the advantages enjoyed by large organizations. Finally, and perhaps most importantly, the sheltered care literature has been silent on what it is about facility size that accounts for the poor outcomes supposedly associated with large facilities.

The present paper has sought to offer a more balanced interpretation of size differences in sheltered care homes, examining various ways in which facility size might affect resident outcomes, while highlighting possible benefits as well as problems of “bigness” in sheltered care.

To summarize we have found that facility size involves significant tradeoffs with respect to the social integration of sheltered care residents. For those activities where size differences do exist, larger homes tend to promote greater resident participation within the facility, while simultaneously inhibiting participation in activities taking place in the broader community. Such tradeoffs, however, appear to be lifestyle choices rather than indicators of the type of association between size and institutionalization commonly associated with mental hospitals (Linn et al., 1980).

We actually found, however, that large facilities are no more “management-centered” than smaller homes and even less “restrictive” in their rules and practices regarding permissible resident behavior. We also found that large facilities spend more on program activities for their residents (i.e., a by-product, presumably, of the superior scale economies of large homes). In fact, these added expenditures may help explain why residents of larger homes tend to be somewhat more involved in facility-based activities while at the same time participating less in certain community activities.