portion of survivors appeared in both samples, the t test for uncorrelated means was used.

Despite the variation across sites, the pattern clearly shows that the tenant populations in the aggregate had declined in competence and well-being as the housing environments aged. Interestingly enough, the decline was smallest in the two measures of functional health. No overall change was seen in the Rosow-Breslau Scale (Rosow & Breslau, 1966) and the change in Instrumental Activities of Daily Living difficulty (Lawton & Brody, 1969) arose from significant declines in only 2 of the 4 non-congregate housing sites. Overall, the incidence of hospitalization did increase, however, as did restrictions in mobility represented by leaving the neighborhood, shopping frequency, public transportation frequency, and frequency of going out after dark.

A general decline in quality of life is indicated by the increase in passive activities (television, looking out window, sitting and thinking) and the decrease in 11 active activities. Diminished contact with children is seen as well as a decrease in availability of children. A decrease in the frequency of seeing the most frequently seen relative was observed in every site, and in two sites frequency of interaction with children decreased among tenants with living children. An overall decrease in contact with in-building friends also occurred, especially in the most active social behavior, that of visiting in one another's apartments. The sites differed in the extent of change in contact with friends outside the building, with no overall change being shown. On a subjective level, although no change appeared in general level of satisfaction with housing, a decline was observed in satisfaction with the number of activities in the housing complex. In a 15-item index of perceived change over the last year in many aspects of life quality, the net perceived change was generally on the negative side, although the modal response for 14 of the items was "no change." Finally, in a shortened version of the Philadelphia Geriatric Center Morale Scale (Lawton, 1975), a decline was noted in three of the five sites.

and less than 2% for bathing); impairments in instrumental activities ranged from 4% who could not use any form of transportation to 21% who either could not shop or would have difficulty shopping if they had no help. It is of interest to note that, aside from shopping, the highest rates of impairment were for housework and cooking in Sholom Aleichem, the one congregate setting that provided these services. Impairment in the first five activities (housework, laundry, cooking, within-unit ambulation, and bathing) was considered a major impairment on the assumption that impairments in the other two (shopping and transportation) could be tolerated or compensated for more easily. Thus the basic overall prevalence of impairment was the existence of an impairment in one or more of the major activities; this rate was 21%. Seven percent displayed two or more such impairments, while only 2% had three or four impairments.

The most direct indicator of service need is the number of tenants who, by their own statement, that of the manager, or the judgment of the interviewer, require additional assistance. Table 3 shows estimates of this type. Unmet need prevalence was relatively low for the major needs (in-home nursing care, housework, laundry, cooking, and personal care) ex

Tenants' current service needs

The last entries in Table 1 show small increases in the wish for three types of services. The most important data for service-planning purposes are estimates of the functional capacity and needs of tenants in these "aging" projects- that is, the number who are impaired in their ability to perform activities of daily living and the number who need assistance or increased amounts of assistance in these tasks, determined after the projects had aged (1980). Table 2 shows the percentages of people who were classified as "'impaired" in each of the activities measured (tenants were considered impaired if they reported either that they could not perform an activity even with help or, if they performed the activity with help, they indicated that they would have difficulty doing it without help). Impairment in basic self-care activities was very low (none for getting about the apartment

cept for housework. It is of interest to note that the administrators considered fewer tenants to have unmet needs than did tenants themselves or the interviewer. Overall unmet need estimates indicate that 17% of all tenants perceived that they had one or more major unmet needs.

Environmental Responses to Changing Needs

The degree of accommodation made by the housing facility is indicated by (a) physical changes made to allow an on-site service to be delivered, (b) the addition of services sponsored by the facility, and (c) the provision of services through means other than by the sponsor of the housing. The discussion to follow will show that major accommodation to service needs was accomplished with minimum change in either physical arrangements or housing sponsorship of services.

None of the five sites had made a physical addition designed to accommodate supportive services or done any substantial remodeling for this purpose. Changes in room use were the major type of change made. Usually these changes were minor; for example, offices were assigned new uses, as in Blueberry Acres, where an administrative office became the preventive health office used by a nurse or physician. The only major change in room use was the location of a weekday hot lunch program (OAA Title III) in the single major social space of President House. This change entailed the addition of a small amount of fixed equipment, an upgrading of the existing kitchenette, and the preempting of the community room for this purpose for about 2.5 hours in the middle of the day. The only additional sponsor-provided onsite service was Sholom Aleichem's implementing a service included in the original plan but not in place for a few years, in which a medical suite was staffed by a full-time nurse and was used by a private practitioner from the community to see tenants at their own initiative and expense.

Thus, accommodation in a major form was virtually nonexistent. It should be noted that housing designed for upper-income individuals, to whom the

costs of additional care could be charged, might possibly have made more major accommodations. In our five sites accommodation was, however, major in the third type of accommodation, that is, services not sponsored by the site. Table 4 shows the presence of congregate meals, health services, and other supportive services for the five sites during their first year as compared to 1980. Sholom Aleichem, the only congregate housing site among the five, provided all of these services; all were located on-site and most of them were sponsored by the housing facility. The other sites show a very similar pattern: a virtual absence of all services when the housing was new and an almost-complete package of services in 1980, all being sponsored by community agencies. In addition to the services shown in Table 4, all sites had activity programs in both 1966 and 1980. By 1980 most also had some ability to connect tenants to counseling, mental health services, or service management; these resources were too diverse and difficult to characterize and accordingly they and a great variety of other services are not included in the table.

The service patchwork

What Table 4 cannot portray is the immense diversity of the agencies, types of services, and manners of delivery of those services. For the reason to be stated below, we refer to the cluster of services available in non-congregate housing as a "patchwork of services." This service patchwork has five characteristics. First, different services are delivered to tenants by two or more community agencies. Second, the locus of the service may be the housing site or a nearby site. Third, the initiative for establishing the service may come from the housing facility, the community agency, or both. Fourth, the services delivered by each agency are dispensed relatively autonomously by each, sometimes with little or no involvement of the housing site. Finally, services are targeted to a small percentage of all tenants.

The term "patchwork” thus refers to the relatively unplanned and uncoordinated nature of the total service package. For example, the executive director of one housing authority felt left out because he was

not involved in the early planning discussions between the area agency for the aging and a provider. The manager of another site, by contrast, indicated that communication with local agencies was not her job: "If anything needs to be said, that's my supervisor's job." Yet, this person's supervisor was in the housing authority office and had no direct tenant contact. In another city planning engaged in by housing management and a senior center director resulted in mutual agreement that a lunch program would serve everyone's needs (those of both tenants and non-tenants) better if it were located in the senior center, two blocks from the housing site. Interviews with managers, agency personnel, and local coordinating groups made clear that a great deal was being lost by having each agency pursue its own practices so independently. It is thus easy to recommend that every housing sponsor (whether private or public) instruct its administrator to take the initiative, beginning with the pre-occupancy period, in establishing a service planning and monitoring structure that includes both the housing and the community service sector. Before community-based services are in place, this structure might mean only that the housing administrator, a AAA representative, and one or two agency staff members meet periodically (preferably with regular participation by tenant representatives). As an agency decides to direct attention to tenants in the housing, it should add a representative to the group, whose purpose then is enlarged to include service monitoring and outreach. In addition to the rational planning, coordination, and efficient deployment of resources this process would afford, it would reduce greatly the present unfortunately frequent feeling of uninvolvement and disinterest among some of the people best qualified to know tenants and their needs.

To say that the patchwork needs improvement does not mean that it represents a failure. Two strengths of the patchwork far outweigh its weaknesses. First and foremost, at least in our sample of four sites with no planned services, every site had managed to obtain the basic package of required services. Second, since the overall supply of service resources is limited, an incentive is provided to target a service to the person with a real need for it. This practice not only encourages giving priority to the needy but discourages overserving capable people in a way that jeopardizes their independence. Does this kind of targeting reduce the costs of serving tenants' needs over that accomplished by traditional congregate housing? Although this question was not addressed in the present research, the answer is almost certain to be "yes." It may even be more costeffective than the supportive services being offered in the current "congregate services" demonstration (U.S. Department of Housing and Urban Development, 1983), since the demonstrations carry a substantial amount of administrative overhead as well as some compulsion to fill a quota of tenants in order to reduce the unit cost of services. However, this question requires hard research findings for a definitive

answer.

Vol. 25, No. 3, 1985

263

In addition to the coordination problem, the major risk in the service patchwork is that, unlike the services delivered in purposive congregate housing, a continuing financial commitment to providing the services is not guaranteed. Our suggestion is that the existence of the planning and monitoring group and the links that this structure forges with the larger community constitute the best possible insurance for continuity of funding.

How Has the Housing Performed Over 12 Years?

In conclusion, the research has shown clearly that, when a change in tenant competence occurred, that change was in a downhill direction, as is inevitable over a 12- to 14-year period during which both original and replacement tenants experienced the process of biological aging. The change was most marked where physical motility was concerned; people moved about into the wider community less, managed their own transportation less frequently, and altered their habits so as to shop less frequently. The richness of their everyday lives changed for the worse as well, in both objective (social activity and participation in active solitary or organized activities) and subjective senses (perceived quality of life and morale). Number of hospital days also increased. With all the foregoing declines having been demonstrated, it is surprising that decline was absent or only mild in the indicators of functional health. Two possible explanations may apply. There is general concern among tenants about being forced to move to a nursing home should health decline become obvious. It is conceivable that this anxiety, as well as the well-known tendency to emphasize one's own intactness, masked some actual decline in functional health. It is also possible that the relative lack of decline in functional capacity demonstrates the resourcefulness of aging tenants in finding alternative ways of satisfying instrumental needs, so that limitations are not perceived as impairments. For example, a person whose shopping needs have increasingly been satisfied through the help of others may not call this change "difficulty." Thus, the wish to remain in an independent setting may express itself through either of the mechanisms mentioned in a laudable determination to persevere in the independent role.

Despite such determination, people who cannot perform many activities on their own cause problems within the community, both because they appear physically disabled to other tenants and because their impairments may lead to actual deprivation. Our attempts to assess the degree of true deprivation suffered from the lack of an absolute criterion for deprivation. Nonetheless, we suggest as a rule of thumb that having either two or more major impairments or self-assessed unmet needs represents a significant level of present deprivation. By each of these estimates the percentage of deprivation was 7% of these tenant populations.

Whether a 7% "deprivation rate" is high or low is, of course, the major question and one that is not easily answered. Were there little hope that these

people's needs could be served, even this small percentage of unmet needs would be intolerable from a humanitarian point of view. Our research has shown, however, that a substantial degree of accommodation to changing needs did, in fact, occur. None of the sites attempted a total redesign of the environment so as to provide service-rich congregate housing, but all of them in their own ways augmented services in the patchwork fashion described above. Without the patchwork the deprivation rate would have been much higher. We still must ask whether a further "patchwork increment" designed to decrease the 7% deprivation rate is feasible.

Current community resources often seem strained to their maximum. How can we stretch them even more thinly? Our research conclusions suggest two related approaches to closing the unmet need gap. First, some slippage at the administrative level is indicated by the results shown in Table 3, where administrators consistently judged fewer tenants to need services than did tenants themselves or interviewers. We do not, of course, know that managers are underestimating. However, it does seem that, at the very least, these results indicate the need for management to recognize that there is a difference between their estimates and those of tenants. It is likely that further managerial training that includes humanservice as well as economic and administrative content would be beneficial. With greater professiona! expertise, managers might recognize tenant need more quickly and better matches between need and services received might be attained. It is, of course, probable that such an increase in expertise might result in the earlier separation of a few more deeply impaired tenants. This result would also reduce the deprivation rate.

The second approach to attaining greater needservice congruence also requires sensitized and informal management. A more carefully planned approach to the service patchwork by local service planners, service-delivery agencies, and housing personnel could undoubtedly succeed in adding enough efficiency and accuracy of service targeting to pay the additional costs associated with serving the deprived 7%.

A last point suggests that there is a real place in the continuum of care for housing for persons who are no longer independent. If the present five case studies are any guide, planned housing for the elderly is very likely to evolve into environments with a strong

protective function. Protection is required for those who survive as tenants in such housing environments but become more impaired. Protection of a different kind is also given to the socially isolated (those who are not married and those without children), who appear to be increasingly served by this type of housing. The costs of continuing to address these problems are justified by the critical ingredient of protected housing, the relative autonomy atforded the individual.

Although the service patchwork has clearly served a real need, our confidence in its long-term dependability must be tempered by the fact that most of the 12 or 14 years during which the patchwork developed were characterized by a steady expansion of community-based services to the aged. That period is clearly concluded. Is the observed accommodation to tenant need also concluded? The patchwork for tenants and services of all kinds for non-tenants are now in competition for the same service dollars. The present findings suggest that the patchwork will be efficient enough to warrant only slightly larger total financial allocations over and above those required for services to the non-tenant community. Continued parallel development of the full congregate model is required to simply keep up with the increasing number of persons of marginal independence.

References

Ehrlich, P., Ehrlich, I., & Woehlke, P. (1982). Congregate housing thirteen years later. The Cerontologist, 22, 399–403.

Lawton, M. P. (1972). The dimensions of morale. In D. P. Kent. R. Kastenbaum, & S Sherwood (Eds.). Research, planning, and action for the elderly. (pp 114-165). New York Behavioral Publications. Lawton, M. P. (1975). The Philadelphia Genatric Center Morale Scale: A revision, Journal of Gerontology, 30, 85-89.

Lawton, M. P. (1980). Social and medical services in housing for the aged Rockville, MD: National Institute of Mental Health.

Lawton, M. P., & Brody, E. M. (1969). Assessment of older people: Selfmaintaining and instrumental activities of daily living. The Gerontolo gist, 9, 179-186.

Lawton, M. P. & Cohen, J. (1974). Environment and the well-being of elderly inner city residents. Environment and Behavior, 6, 194–211. Lawton, M. P., Greenbaum, M., & Liebowitz, B. (1980). The lifespan of housing environments for the aging. The Gerontologist, 20, 56–64. Regnier, V. (1975, June). Assessment of preferred supportive services for luxury retirement housing. Paper presented at the annual meeting of the Environmental Design Research Association, Lawrence KS. Rosow, I., & Breslau, N. (1966). A Guttman health scale for the aged. Journal of Gerontology, 21, 556-559.

Urban Systems Research and Engineering Inc. (1976). Evaluation of the effectiveness of congregate housing for the elderly. Final report. Con tract No. H-2255R. Washington, DC: U.S. Department of Housing and Urban Development.

U.S. Department of Housing and Urban Development (1983). Third annual report to Congress on the Congregate Housing Services Program, 1982 Washington, DC: Author.