Nursing home placement is often an undesirable outcome of functional disability among elderly persons. Patient preference for independent living arrangements in the community and societal pressure to contain medical costs have provided the impetus to identify risk factors for institutionalization.
Previous cross-sectional research suggests that nursing home residents, when compared to community residents, are more likely to be old; poor; Caucasian; female; living alone; chronically ill; functionally impaired with respect to activities of daily living; mentally impaired either because of dementia or other mental illness; and selectively burdened by such medical conditions as cerebrovascular disease, cancer, or incontinence.1"8 Prospective identification of placement predictors from the community into an institution (over a period of several years) confirms many of these correlates. Specifically, living alone, being Caucasian, and being cognitively impaired and dysfunctional in terms of activities of daily living predict future nursing home placement.1·6 Identified predictors of nursing home placement after hospitalization include preadmission residence in a nursing home, physician anticipation of discharge to a nursing home, extensive use of medications, age, mental illness, and certain chronic or unstable medical conditions such as heart failure and severe pulmonary disease.9"13
The objectives of the present study were to identify predictors of nursing home placement among a vulnerable elderly population and to identify "change variables" predictive of subsequent nursing home placement. These predictors would describe those most at risk in an already high-risk group and contribute to the gerontology literature by identifying changes in functional abilities that precede and are associated with institutionalization.
In the present study, data from the Washington State longitudinal study, entitled Community-Based Care Systems for the Functionally Disabled, were reviewed retrospectively. I4 Predictors of placement into a nursing home within a sixmonth, follow-up period were identified. Furthermore, variables that changed in status over a six-month period were identified when they predicted subsequent nursing home placement.
The Community-Based Care Systems for the Functionally Disabled was a research and demonstration project started in 1975 and conducted by the Washington State Department of Social and Health Services.14 Its major purpose was to determine whether a program of providing long-term social and health care to lòw-income aged and functionally disabled adults in a community would affect community resource and nursing home utilization. The Older Americans Resources and Services (OARS) questionnaire was used to assess patient function and service utilization.
Cowlitz and Wahkiakum Counties served as a demonstration community, and Whatcom County served as the control community. In this study only the longitudinal data from Cowlitz- Wahkiakum Community were retrievable.
The study population consisted of low-income individuals who were either eligible for or receiving supplemental income or social services. One group of subjects was a random sample of those listed on the State Register of Eligible Recipients. The Register of Eligible Recipients is a listing of Medicaid individuals eligible for benefits and services under Titles XVI and XTX of the Social Security Act. A second sample of subjects consisted of functionally disabled persons eligible for Title XIX or Title XX social services through the Department of Social and Health Services. These individuals were considered at high risk for nursing home placement because they:
1 . Were recently discharged from an acute care facility and, except for the availability of coordinated services, would be placed in a long-term care facility; or
2. Resided in the community with physical disabilities that limited their ability to manage the tasks of daily living.
The median age of the total sample of high-risk patients in Cowlitz- Wahkiakum Community was approximately 78 years of age. More than three quarters of the patients were female, 60% were widowed, and a majority (68%) lived alone. Approximately one half (48%) of the study population considered their physical health to be fair.
STATIC PREDICTORS OF NURSING HOME PLACEMENT
The OARS instrument provided functional assessment information in several areas: physical health, mental health, social and economic resources, and activities of daily living (ADLs). Investigators at Duke University developed the questionnaire, which contains 101 questions and takes approximately 1 hour to administer.15·16
Data from Cowlitz-Wahkiakum patients were retrieved, and questions from the OARS were analyzed. Questions that predicted nursing home placement for individuals within six months (146 cases, 143 missing observations, 92 variables) and that indicated a change in status over a six-month period predictive of nursing home placement within a subsequent sixmonth period (109 cases, 34 missing observations, 157 variables) were identified.
Chi-square and chi-square for linear trend analyses were conducted because the dependent and independent variables were categorical. In these analyses, the independent variables (question responses) were ordinal and the dependent variables were dichotomous (staying in community vs moving into a nursing home). When cell sizes were small in 2 x 2 analyses, Fisher's exact statistic was employed.17 One-tailed tests were employed because of the obvious hypothesis that poorer functioning is predictive of nursing home entry.
The results of the analyses regarding predictors of nursing home placement demonstrated that five questions were associated with future nursing home placement (Fisher's exact statistic, p<0.01). These were identified from the Cowlitz- Wahkiakum Community's data, in which 21 (14%) community subjects entered a nursing home over a six-month period. The predictors, presented in Table 1, were the use of a walker; the need for assistance with handling one's finances, preparing meals, and walking; and getting to places beyond walking distance. Although these variables were statistically predictive of institutionalization, none demonstrated a predictive value greater than 50%.
Several change variables (ie, questions in which responses changed over six months) also were associated with prediction of future nursing home placement (chi-square for linear trend). These variables also were identified from data obtained from the CowlitzWahkiakum Community, in which ten (9%) entered a nursing home within six months after a six-month observation period, even though a multitude of community support services were available. These variables included the development of gastrointestinal ulcers, an alteration in the use of nitroglycerin for chest pain, deterioration in the patient's ability to get in and out of bed, increased difficulty in walking, initiation of wheelchair use, diminution of the frequency with which the patient engaged in telephone conversations with others, and the cessation of "retired on disability" status. These statistically significant variables are presented in Table 2. Table 2 also indicates which of these changes in functional status were reliably associated with nursing home placement both within two months and six months of the second administration of the OARS.
Although all of these change variables were predictive of nursing home placement with statistical significance, none of them demonstrated a positive predictive value greater than 50%. When the patient's baseline functional status was taken into consideration (controlled for statistically), the only change variable predictive of nursing home placement (within two and six months) was deterioration in the patient's ability to get in and out of bed (p<.05).
Nursing home placement has become an expensive consequence of functional impairment associated with aging. To identify predictors of placement into nursing homes from the community, data from a Washington State study conducted in the late 1970s was reviewed and analyzed. The analyses revealed that among poor, elderly, widowed females, dependency in several activities of daily living predicted nursing home placement within a relatively short period. In addition, several changes in patient function occurred that predicted subsequent nursing home placement. These predictive changes reflected alterations in several functional domains: socioeconomic, medical, and activities of daily living.
Data from this project and other studies support the predisposing nature of several areas of dysfunction for nursing home placement. Specifically, these predisposing factors include using an ambulatory aid and requiring help with certain activities of daily living, such as handling money, preparing meals, walking, and dealing with transportation. Other research studies have identified several predictors of nursing home placement that were not identified in this study. These included living alone; being cognitively impaired, female, financially secure, and white; and having limited family support. The most likely explanation for not identifying these variables in this study population is that the subjects in this study were generally white, widowed, and poor. The homogeneity of these risk factors in the sample frame could have prevented their identification in this study.
Change in status over time can help clinicians appreciate trends and facilitate clinical evaluation. Nurses and physicians have learned the value of temperature flow diagrams as aids in distinguishing fevers from infectious causes from those secondary to drug reactions. Similarly, monitoring changes in serum electrolytes has often pointed to medication side effects. Heretofore, change variables predictive of nursing home placement have not been identified.
In this study, loss of income due to cessation of retirement pensions appeared to be a predictor of nursing home placement. This might reflect a loss of the ability to pay for community support services, or could reflect the result of other financially dependent issues, such as nutritional status and access to medical care. The decrease in the number of telephone conversations and the association of this variable with subsequent nursing home placement probably reflects the previously demonstrated relationship that increased social isolation predisposes an individual to nursing home placement. Two changes in medical status were associated with subsequent nursing home placement; both changes may reflect reduction in health status. Finally, the increased disability in bed transferring and walking, and using a wheelchair approximate the static predictors, which reflect dysfunction in activities of daily living.
CHANGE PREDICTORS OF NURSING HOME PLACEMENT WITHIN SIX MONTHS*
Institutionalization still occurred for some individuals in the demonstration site despite attempted interventions, such as chore, home nursing, nutrition programs, and transportation services, to prevent nursing home placement. Although change variables predictive of subsequent nursing home placement were identified, this evidence alone is not sufficient to indicate that these predictors necessarily lead to nursing home placement regardless of interventions. This is true, in part, because we do not know the quality, consistency, and appropriateness of the services provided in the study. However, the predictors in Table 2 may reflect an abbreviated listing that has been reduced by some successful community services.
In addition to concerns about self-reported data and the small numbers of subjects, there are several other limitations to the study that suggest the need for further validation. The amount of missing data in the CowlitzWahkiakum Community may have introduced a bias. Another limitation may be that several systematic differences exist between the communities described herein and other communities. For example, the CowlitzWahkiakum Community has fewer family providers and an increase in selfrating of health by the geriatric population, when compared to another area in Washington State.
Surprisingly, cognitive dysfunction and incontinence were not demonstrated to be predictors of nursing home placement. Although this might suggest a study limitation, the reason why these apparent omissions occurred was that the mental status score and incontinence were not coded in the data.
The identified predictors in this study can facilitate targeting of medical and social attention in an attempt to prevent institutionalization for the vulnerable elderly. For example, a precipitous decrease in the financial status of an elderly person might alert clinicians to request consultation from a social worker. Similarly, a decrease in social contacts might suggest the need to encourage greater family involvement with the patient or patient referral to either a senior citizens' center or adult, day healthcare program. The development of functional disabilities could also be used as a rationale for specific chore or attendant assistance. Of course, these interventions might not be effective, but that should not prevent clinical attempts to promote patient function and well-being. While clinicians continue to provide care based on need, research can further validate predictors of institutionalization and determine the effectiveness of interventions aimed at preventing nursing home placement.
Even if targeting services fails to prevent institutionalization, clinician knowledge of predictors of nursing home placement allows for better clinician-patient communication. Nurseinitiated discussions about the possibility of future healthcare needs offer older patients the opportunity to plan accordingly and exercise some degree of autonomy. This would also be an important outcome, as it reflects a basic goal of geriatric care: maximum independence for the patient.
- 1 . Palmore E: Total chance of institutionalization among the aged. Gerontologist 1976; 16(6):504-507.
- 2. Kraus AS, Spasoff RA, Beattie EJ, et al: Elderly applicants to long-term care institutions, I: Their characteristics, health problems and state of mind. / Am Geriatr Soc 1976; 24(3):117-125.
- 3. Brody SJ, Poulshock SW, Masciocchi CF: The family caring unit: A major consideration in the long-term support system. Gerontologist 1978; 18(6):556-561.
- 4. Weissen W, Scanlon W: Determinants of institutionalization of the aged. Working paper. The Urban Institute #1466-21, 1983.
- 5. Medicaid and Nursing Home Care: Cost Increases and the Need for Services are Creating Problems for the States and the Elderly, US General Accounting Office, IPE-84-1.13, 1983.
- 6. Branch LG, Jette AM: A prospective study of long-term care institutionalization among the aged. Am J Public Health 1982; 72(12):1373-1379.
- 7. Kane RL, Kane RA: Care of the aged: Old problems in need of new solutions. Science 1978; 200:913-919.
- 8. Greenberg J, Ginn A: A multivariate analysis of the predictors of long-term care placement. Home Health Care Services Quarterly 1979; l(l):75-99.
- 9. Kane RL, Matthias R, Sampson S: The risk of placement in a nursing home after acute hospitalization. Med Care 1983; 21(11):1055-1061.
- 10. Davis JW, Shapiro MF, Kane RL: Level of care and complications among geriatric patients discharged from the medical service of a teaching hospital. J Am Geriatr Soc 1984; 32(6):427-430.
- 1 1 . Lamont CT, Sampson S, Matthias R, et al: The outcome of hospitalization for acute illness in the elderly. J Am Geriatr Soc 1983; 31(5):282-288.
- 12. Runenstein LZ, Wieland D, English P, et al: Data on four-year outcomes and predictors of improved patient outcomes. J Am Geriatr Soc 1984; 32(7):503-512.
- 13. Wachtel TJ, Derby C, Fulton JP: Predicting the outcome of hospitalization for elderly persons: Home versus nursing home. South Med J 19M; 77(10):1283-1285.
- 14. Solem RA, Garrick MA, Nelson H: Community-Based Care Systems for the Functionally Disabled: A Project in Independent Living. Olympia, WA, Department of Social and Health Services, 1979.
- 15. Kane RA , Kane RL: Assessing the Elderly: A Practical Guide to Measurement. Lexington, MA, Rand Corporation, 1981.
- 16. Fillenbaum GG, Smyer MA: The development, validity, and reliability of the OARS multidimensional functional assessment questionnaire. J Gerontol 1981; 36(4): 428-434.
- 17. Nie NH, Hull CH, Jenkins JG, et al: Statistical Package for the Social Sciences. New York, McGraw-Hill Book Co, 1975.
- This project was supported by the Seattle Veterans Administration Geriatric Research, Education and Clinical Center, and the Pacific Northwest Long-Term Care Gerontology Center (Administration on Aging 90-AT-2152/04), Seattle, Washington.
STATIC PREDICTORS OF NURSING HOME PLACEMENT
CHANGE PREDICTORS OF NURSING HOME PLACEMENT WITHIN SIX MONTHS*