Critics of deinstitutionalization have maintained that the reduction of stays in state mental hospitals led to increased hospital recidivism and to transinstitutionalization to nursing homes (1,2,3) and jails (4), where mentally ill persons were being warehoused and which provided little treatment for their illness. Nursing homes were seen as the new back wards in the community (5). In Illinois, during the 1970s and early 1980s, the most common nursing home placement for persons with serious mental illness was in intermediate care facilities; few placements were in skilled nursing facilities (6).
Although a large number of mentally ill persons came to reside in nursing homes after deinstitutionalization, relatively few received mental health services in those facilities (3,7,8). A 1981 study of a random sample of eight intermediate care facilities in Illinois found that the facilities provided, on average, only 7.5 hours of psychiatric services each week for an average of 95 residents with mental disorders, which is equivalent to about five minutes of treatment a week for each resident (6). Illinois state inspectors commonly referred to these facilities as "sheets and eats," because little more than a bed and food were offered (9).
Other studies using national samples have reported that less than one-fifth of mentally ill persons in nursing homes received any mental health services in a year, and only 4.5 percent received such services over a month (7,8). Furthermore, only about 7 percent received any psychotherapy (3).
On a national level, these concerns led to the inclusion of provisions in the Omnibus Budget Reconciliation Act of 1987 (OBRA-87) that sought to reduce the number of people with mental illness in nursing homes. The Illinois Department of Human Services developed the community integrated living arrangement program in 1989 in response to this legislation (10,11,12). A second initiative involved converting intermediate care facilities to intermediate care facilities for the mentally ill. The facilities continued to serve large numbers of mentally ill persons, but additional supportive services were mandated as part of the conversion. The state gave all individuals in intermediate care facilities for the mentally ill the option of moving to a community integrated living arrangement facility.
Through the community integrated living arrangement initiative, the Illinois Department of Human Services provided funds to community mental health centers to pay for case management, supportive services, and housing in the community for persons with severe mental illness. The community integrated living arrangement program was designed to offer residential treatment on a small scale and to integrate mentally ill persons into the community. Each facility provides residential treatment for no more than eight persons in a group of subsidized apartments. Although the program was developed to provide an alternative to nursing homes, mentally ill individuals could gain access to a program residence after they were discharged from a state mental hospital or if they were not functioning well in other community residential situations.
Community integrated living arrangements use one of two modes of residential treatment, which vary by level of supervision. Continuous-care programs have staff on-site 24 hours a day. Intermittent-care programs provide staff monitoring periodically during the week. These two approaches allow matching between the level of supervision and the residents' level of need and allow provision of individualized supportive services. The community integrated living arrangements exemplify the "housing-as-treatment" approach, in which clinical considerations govern the type of residence provided for the client (13,14).
The community integrated living arrangement facilities offered an opportunity for a natural experiment examining whether satisfaction with housing differed between the two levels of supervision. The first purpose of the study was thus to compare residents' satisfaction with housing in continuous-care and intermittent-care residences. Previous research has shown that residents favor less restrictive living arrangements and that ratings of satisfaction with housing increase as mentally ill individuals move from more to less restrictive care (15). Residents of state hospitals have reported the lowest level of satisfaction with housing, residents of group homes were more satisfied than those in hospitals, and individuals in supervised apartments were the most satisfied.
These findings suggested that the satisfaction ratings of residents of community integrated living arrangement housing would be similar to those of residents of supervised apartments. Another possibility is that the residents of intermittent-care community integrated living arrangements might be more satisfied with their housing than the residents of continuous-care arrangements. Alternatively, the effort to match the level of residential supervision to individual needs may have lessened potential differences by providing appropriate care.
A second purpose of the study was to examine whether residents' hospital use was reduced after they were placed in community integrated living arrangements housing. The arrangements were expected to function as a substitute for hospitalization because they were developed to enable mentally ill persons to receive residential treatment in the community.
Structured interviews were conducted from 1994 to 1997 with individuals who had at least one year of residence in a community integrated living arrangement. Interviews were conducted at nine randomly selected community agencies that operated community integrated living arrangements. All were within the metropolitan Chicago area. Residents were randomly selected for interviews in each program, except for two very small programs, in which all residents were interviewed. Data on hospital use during the year before placement and the first year of residence in the community integrated living arrangement were collected. Written informed consent was required for participation in the study. The study was approved by the institutional review board of the University of Chicago Division of Biological Science.
Seventy-four mentally ill residents of community integrated facilities were selected for the study. Forty-three lived in intermittent-care residences and 31 in continuous-care residences. Most participants were men (65 percent, N=48). The majority were white (60 percent, N=44), and the remainder were African American (35 percent, N=26), Hispanic (3 percent, N=2), or Asian (3 percent, N=2). Their mean±SD age was 41± 10 years. Most had either schizophrenia (57 percent, N=42) or schizoaffective disorder (23 percent, N=17). The remainder had a diagnosis of biopolar disorder (11 percent, N=8) or other diagnoses (9 percent, N=7).
Satisfaction with housing was assessed with the Quality of Life Interview (16), which has well-established reliability and validity. The interview includes measures of satisfaction with six aspects of housing: living arrangements, food, rules, privacy, freedom, and the prospect of continuing to live in the current location. Respondents are asked for their subjective evaluation of each aspect on a 7-point Likert scale ranging from 1, terrible, to 7, delighted. Responses to these six questions were summed and divided by six to produce a mean housing satisfaction score. The Quality of Life Interview also includes an objective measure of quality of care in which the presence or absence of 12 amenities, including private space, cooking facilities, a bathroom, and a shower, is determined. These responses were summed to produce a total amenities score.
Information about hospitalization during the baseline year (one year before placement in a community integrated living arrangement facility) and during the first year of residence in the program was obtained from chart reviews and from the state database of utilization of state-operated facilities. The charts included information on both state hospital use and the use of private hospitals funded by Medicaid.
Satisfaction scores, data on hospital use, and demographic data were analyzed with descriptive statistics, such as means and frequency distributions. The relationship between subjective residential satisfaction scores and total amenities scores was evaluated with Pearson's product-moment correlation. Comparisons between mean satisfaction scores for intermittent-care and continuous-care residences were made with Student's t test. Differences in residential satisfaction and hospital use by age, gender, and race were examined by using appropriate univariate statistics. For example, the relationship between age and time spent in the hospital was tested with Pearson's product-moment correlation. Demographic variables for which differences were found were included in a multivariate analysis (repeated-measures analysis of variance with covariates). SPSS version 7.5.1 was used for these analyses.
Satisfaction with residences
The mean±SD residential satisfaction score was 5.2±1, or "mostly satisfied." No significant differences in satisfaction scores were found between participants who lived in continuous-care residences (mean±SD=5.1±1) and those who lived in intermittent-care residences (mean±SD=5.3±1). No demographic differences in satisfaction scores were found. The data reflected a nonsignificant tendency for satisfaction scores to be correlated with the total amenities score. However, the presence of amenities did not differ by the level of supervision of the residence. Both continuous-care and intermittent-care residences had almost all of the 12 amenities assessed (mean±SD=11.3±1.1).
The mean±SD number of days in the hospital decreased significantly from 47.7±103 during the baseline year to 5.3±17 during the first year in community living arrangement facilities (repeated-measures t=3.67, df=73, p<.001). No differences in hospital use were noted between participants in intermittent-care residences and those in continuous-care residences. Younger residents tended to have been in the hospital longer during the baseline year (r=-.24, p< .04). No other demographic differences were found. In a second multivariate analysis with age as a covariate, the decrease in number of days hospitalized remained significant (repeated-measures F=8.82, df=1, 72, p<.01).
The type of housing that should be provided for persons with serious mental illness has been the subject of much debate (14). Advocates of supportive housing ("housing as housing") argue that consumers prefer housing that is not linked to services; however, results of research in support of this approach have been equivocal (17,18). Other clinicians and advocates argue that services for persons with mental illness should be integrally related to housing ("housing as treatment") (13). The community integrated living arrangements fit with the housing-as-treatment model and provide an opportunity to better understand consumer satisfaction with this approach. The high satisfaction ratings reported by the program's clients support the view of those who believe that housing should be integrated with treatment.
Our findings suggest that residents of the community integrated living arrangement facilities view the program positively. The results are consistent with those of other studies that have found high quality-of-life ratings in less restrictive community settings (15). The residents' mean rating of satisfaction with their living arrangements—5.2 on a 7-point scale—was comparable to the highest score given to a residential program in studies reviewed by Lehman and colleagues (15). The program with the highest score—a supervised apartment program in Rochester, New York—was the least restrictive program described in that review and was given a mean rating of 5.4 by residents. The next highest rating was 5.2, given to a group home program, also in Rochester. On the other hand, our study found no differences in residents' satisfaction between the less restrictive intermittent-care residences and more heavily supervised continuous-care residences.
It would be useful to take a broader look at the extent to which other factors, such as type of treatment program, affected satisfaction with housing. However, given the small number of subjects in our study, the lack of statistical power limited the issues that could be addressed.
Community integrated arrangements may be successful in maintaining residents in the community. However, our findings of a decrease in the number of days spent in the hospital after enrollment in the program must be considered inconclusive. In the absence of a control group, we cannot be certain that the community integrated living arrangements facilities functioned as a substitute for hospitalization. Selection factors and temporal trends might account for the decrease. Selection at entry to the program and the selection of persons who remained in the program for at least one year could have biased the sample toward residents with higher functioning. Studies that include control groups would provide more persuasive evidence. Unfortunately, such studies are not possible at this time, because the community integrated living arrangements facilities are filled and new slots open infrequently.
Assessment of the extent to which the residential program can be considered a successful response to OBRA-87 is beyond the scope of this paper. However, a related analysis found that only 6 percent of persons with mental illness in Illinois were transferred from nursing homes to community integrated living arrangements residences (19). This finding is consistent with an earlier prediction by one of the authors (20), although higher estimates of up to 20 percent have also been projected (21).
Most of the study population remained in skilled nursing facilities or in intermediate care facilities for the mentally ill. Possible reasons for the low rate of transfer to community integrated living arrangement facilities include nursing home residents' reluctance to relocate, their need for skilled nursing care, the presence of dementia, and subtle pressures exerted by nursing home staff.
Concerns about the quality of intermediate care facilities for the mentally ill have paralleled the earlier dissatisfaction with nursing homes. Not only did most mentally ill persons who were eligible for the community integrated living arrangements program remain in intermediate care facilities for the mentally ill, but also recent changes in the delivery of health services have created incentives to recruit more mentally ill residents for these facilities.
The expansion of home health care has left Illinois nursing homes with many empty beds. Nursing home bed brokers regularly visit homeless shelters and hospital wards in search of persons who have mental illness (22). The inadequacy of mental health services in the nursing homes has caused problems in providing appropriate care for these difficult-to-treat recruits.
The lack of adequate alternative housing for mentally ill persons is a national problem. This issue will likely receive greater attention in the wake of Olmstead v. L.C. and E.W., in which the Supreme Court found that inadequate housing opportunities overly extended hospital stays (23). New initiatives are needed to provide appropriate housing opportunities for mentally ill persons in nursing homes and intermediate care facilities for the mentally ill as well as for those who are homeless or being discharged from hospitals. Ensuring access to appropriate housing should become a cornerstone of community care.
Our study further supports the use of the residential satisfaction measures developed by Lehman and colleagues (16) as a useful component of assessments of the quality of housing initiatives.
The community integrated living arrangements program provided mentally ill residents with an alternative to nursing homes that was quite acceptable to them. It is unfortunate that very few nursing home residents participated in the program.
This study was supported by a postdoctoral fellowship awarded to Dr. Hanrahan from the National Association of State Mental Health Program Directors Research Institute and the National Institute for Mental Health; a Home Health Care Research grant from the University of Chicago awarded to Dr. Luchins; and grants from the Clinical Evaluation Program of the Office of Mental Health of the Illinois Department of Human Services. The authors thank Noel Mazade, Ph.D., and Lawrence Appleby, Ph.D., for consultation during the study.
Dr. Hanrahan, Dr. Luchins, and Ms. Savage are affiliated with the department of psychiatry at the University of Chicago School of Medicine and the Illinois Department of Human Services. Dr. Goldman is with the department of psychiatry at the University of Maryland School of Medicine and the National Association of State Mental Health Program Directors Research Institute. Address correspondence to Dr. Hanrahan at the Department of Psychiatry, University of Chicago, MC3077, 5841 South Maryland Avenue, Chicago, Illinois, 60637 (e-mail, firstname.lastname@example.org).