Older adults who are discharged from the hospital to the community face a higher risk of readmission than do individuals of the same age and sex who have not been hospitalized recently (Fethke, Smith, & Johnson, 1986). Early recurrent readmission can be defined as readmission to an acute care hospital within 3 to 4 months following discharge (Berkman, Dumas, Gastfriend, Poplawski, & Southworthe, 1987). Readmission to the hospital is usually for the same, or related, health problems as the original admission. This is a crisis for the family and patient, and costly to society (Berkman et al., 1987; Bull, Jervis, & Her, 1995; Clark, 1996). Hospitalization, discharge, and readmission increase the risk for the patient to become confused and less able to handle post-hospital adjustment. Hospital readmission may also cause sleep deprivation and chronic fatigue resulting in depressive symptoms for the caregiver (George & Gwyther, 1986). Because caregivers of older adults are generally spouses older than age 65 (Schwarz, 1996), hospital readmission compounds the situational causes of depression resulting from multiple losses, poverty, and social isolation (Lockery, Dunkle, Kart, & Coulton, 1993).
Rates of readmission can range from 17% to 38% depending on the time period analyzed, stressors to the family, and primary diagnosis of the patient (Fethke, et al., 1986; Gooding & Jette, 1985; Hennen, Krumholz, Radford, & Meehan, 1995; Naylor, Brooten, Jones, Lavizzo-Mourey, Mezey, & Pauly, 1994; Rieh, Beckham, Wittenberg, Leven, Freedland, & Carney, 1995; Wagnild & Grupp, 1991). Wagnild and Grupp (1991) found that stressors related to hospital readmission included acute illness, caregiver support, and an unsuitable home environment. Although Rich et al. (1995) found that patients with congestive heart failure (CHF) had a very high readmission rate because of their high risk for further cardiac involvement, they can be affected by behavioral and social factors such as social isolation.
Problems related to physical, psychological, and social difficulties are encountered by many hospitalized patients upon discharge (Graham & Livesley, 1983; Helberg, 1993). Graham and Livesley (1983) found many disabled older adults were cared for in the home by relatives who were unable to cope with considerable stress and were not supported by community services. According to Neuman's Systems Model (Neuman, 1995) that organizes the study, stress, and the individual's reaction to stress, vary. Intrapersonal stressors such as increasing age, depression, and declining health, occur within the boundary of the client/client system. Interpersonal stressors such as inadequate social support, occur outside the boundaries of the client/client system at proximal range. Extrapersonal stressors such as lack of adequate health insurance, occur outside the boundaries of the client/client system at distal range. Each stressor is different in its potential to disturb an individual's equilibrium. Individuals either adjust themselves to the environment or adjust the environment to themselves (Neuman, 1995).
Depressive symptomatology, or feelings of hopelessness and helplessness, may result when family caregivers experience increased stress, poor health, and have a negative appraisal of the situation (George & Gwyther, 1986; Schwarz, 1996). Furthermore, lack of satisfaction with social support is a significant predictor of depressive symptomatology (Krause, 1995). These depressive symptoms may have negative consequences for both the caregiver and care recipient (Schwarz & Blixen, 1997; Tennstedt, Caff erata, & Sullivan, 1992). In addition to affecting the health of the caregiver, these feelings of hopelessness may affect the caregiver's willingness or ability to care for the older adult. Thus, quality of life of the caregiver and care recipient are in jeopardy. Because depressive symptoms among caregivers of frail older adults can be as high as 43%, this is a factor worthy of further study (Tennstedt et al., 1992).
Current reductions in hospital lengths of stay are compounding the psychological and physical difficulties for older adult patients (Helberg, 1993). Between 1981 to 1991, the average hospital stay was reduced by 18%, contributing to the extra burden on families and the health care system. In the past, early hospital discharge was associated with growth in services available after discharge for Medicare beneficiaries, including home visits by RNs. Researchers such as Dennis, Blue, Stahl, Benge, and Shaw (1996), Martens and Mellor (1997), and Williams, Blue, and Langlois (1995) found home health care reduced hospital réadmissions within 90 days of hospital discharge. BuIs (1995) found that direct care and teaching from home health care nurses decreased families' anxiety by allowing them to use effective coping strategies for self-care, thus enhancing the patient's health and prognosis.
Nursing attempts to keep the client system stable by assessing the current and potential effects of environmental stressors and in assisting client adjustments for optimal wellness (Neuman, 1995). According to Neuman (1995), the primary goal of nursing is to reduce stressor impact and to increase client resistance. Continuity of care can be provided in an organized and systematic manner by home health care. The goal of tertiary levels of prevention for home health clients is to promote an optimal level of functioning, allowing one to remain in the home.
The type of fit between the caregiver's capacity to meet the needs of the care recipient is an important consideration in whether the care recipient can continue to remain at home (Fink & Picot, 1995). Although home care is the preferred alternative that permits continuity of care, payment rates and visit limits have become quite stringent (Schlenker, 1996). The payment approach to home health care agencies may adversely affect access to quality of care and may increase hospital readmissions. Because home health services will continue to be minimized with managed care, detecting and preventing physical and mental health problems that relate to rehospitalization is even more vital in the present economy.
While predictors of hospital readmission have been examined, gaps exist in the literature. Although family involvement in care may influence patient outcomes after discharge (Helberg, 1993), the role of informal social support as a predictor of hospital readmissions is very limited. Furthermore, home health care nurses are assumed to support caregivers and assist with arduous tasks, but their assistance has only recently been evaluated. The purpose of the study was to determine whether low informal social support, low satisfaction with informal social support, high depressive symptomology of the caregiver, and minimal use of home health care predicted hospital réadmissions of older adults.
A descriptive correlational design was used to assess potential predictors of early readmission of older adults to the hospital. Caregivers of older adults were interviewed in their homes soon after hospital discharge and 3 months later.
Setting and Sample
The convenience sample from two hospitals included 60 family members providing care to patients age 65 years and older who had been hospitalized for a chronic illness and were functionally impaired in two activities of daily living (ADLs). Each patient's functional abilities were identified with a checklist in the chart and were verified by staff nurses. Hospice clients who frequently are hospitalized were excluded. The caregivers were able to speak and understand English and resided with the older adult. Older adults were studied because they occupy more than 34% of hospital beds, and rehospitalizations of Medicare patients account for 25% of all hospital admissions (Naylor et al., 1994).
The sample size was based on power analysis for regression (Cohen, 1988) and a test of the hypothesis: low social support, low satisfaction with social support, high depressive symptomatology of the caregiver, and minimal use of home health care services will predict hospital readmissions of the older adult. Alpha was set at .05, power at .8., and the degrees of freedom for error were set at 20. Because Lockery et al. (1993) found functional capacity, socioeconomic status, and marital status had a combined effect on discharge placement (R2 = .23) and indirect effect on rehospitalization, the effect size expressed as a function of R2 (Cohen, 1988) was .23. Although the number of study participants required for regression was 50, attrition was estimated to be 30% over a 3-month period (Schwarz, 1996), and a total of 71 participants were needed.
After research proposals were submitted to institutional review boards at two large metropolitan hospitals in northeastern Ohio, permission was granted to conduct the study. Clinical nurse specialists and staff nurses of the hospitals identified functionally impaired individuals and their caregivers. They provided each participant with a letter of explanation and gained verbal consent for the principal investigator (PI) to contact him or her by phone to further explain the study. Within a week of hospital discharge, the PI and graduate nursing students trained as research assistants, interviewed the caregivers in their homes. At the first interview, the caregivers signed a consent form, and they were told participation was voluntary. The caregivers were interviewed about informal social support, satisfaction with informal social support, and depressive symptomatology. Three months later, the caregivers were re- interviewed in their homes about social support, satisfaction with social support, and depressive symptomatology. They were questioned about hospital readmission of the care recipient and use of home care. The PI contacted seven home health care agencies to document services provided to the families. The PI documented the number of hospital readmissions for each older adult at the participating hospitals. The caregivers were offered $10 for their participation in the study.
Instrumentation and Measures
Perceived Physical Health. Perceived physical health of the caregiver and care recipient was rated by the caregiver with a one-item measure. The respondents rated health as 4 (excellent), 3 (good), 2 (fair), or 1 (poor). Cadogan (1994) found selfrated health as a strong correlate of objective measures of health.
Hospital Readmissions. Hospital readmissions were documented by the PI through medical record chart audit.
Number of Home Health Care Services Used. This number was determined by reviewing computer printouts documenting services the families received from participating home health agencies.
Informal Social Support. Informal social support was measured with ratings on the Modified Version of the Inventory of Socially Supportive Behaviors Scale (MISSB) (Krause & Markides, 1990). The MISSB measures the amount of social support actually received rather than merely counting supportive individuals, and considers the need for support. Participants rate 40 items on a 4point Likert scale from 1 (never) to 4 (very often). The summative subscales tap dimensions of informal social support: informational (range = 7 to 28), tangible (range = 9 to 36), emotional (range = 11 to 44), and integration (range, 13 to 42). With a random sample of 351 noninstitutionalized adults age 65 and older (Krause & Markides, 1990), Cronbach's alpha for each subscale was as follows: informational support, .81; tangible support, .66; emotional support, .82; and integration, .81. Informational support helps people help themselves by providing them with information such as whom to see for assistance. Emotional support is based on behavioral supportive actions such as physical affection. Tangible support is instrumental behavior such as providing transportation or a loan of money. Finally, integration reflects a reciprocal network of shared obligation by helping others or giving others some form of informal support. Evidence of predictive validity was supported with finding that social support buffered the impact of bereavement on depressive symptoms. A theoretically meaningful, three-factor orthogonal solution supported construct validity (Krause & Markides, 1990). For the present sample, Cronbach's alpha for the total scale was .85. Cronbach's alpha for each subscale was as follows: informational support, .63; tangible support, .71; emotional support, .80; and integration, .83.
Satisfaction with Informal Social Support. Satisfaction with informal social support or needs being adequately met by family and friends was measured by the Satisfaction with Informal Social Support Scale composed of four items (Krause & Markides, 1990). Participants rate satisfaction with informational, emotional, tangible, and integrational support on a 3-point Likert scale (1 [more often] to 3 [less often]; range, 4 to 12). Higher scores indicated greater satisfaction with informal social support. Predictive validity was supported because satisfaction with support buffered the impact of bereavement on depressive symptoms (Krause & Markides, 1990). Cronbach's alpha was .47 in a study of 100 caregivers (Schwarz, 1996) and was .33 for the present sample.
Depressive Symptomatology. Radloff (1977) measured depressive symptomatology with ratings on the Center for Epidemiological Studies Depression Scale (CES-D). The CES-D focuses on distress symptoms prevalent among nonpsychiatric populations during the previous week. Participants rate 20 items on a 4-point scale with a range of 0 to 60 (0 [rarely] to 3 [most or all of the time]). Ratings were summed for a total score with some items reverse scored - higher scores indicated more depressive symptoms. Radloff (1977) reported a high Cronbach's alpha of .85 in a study of nonpsychiatric adults, and construct validity using the contrasted group approach was supported. The Cronbach's alpha for the present sample was .94.
Description of Sample
Initially, 85 family caregivers were contacted soon after hospital discharge, but only 60 caregivers completed the second interview 3 months later. Attrition was a result of either nursing home placement, death of the care recipient, or refusal to participate in the second interview. Those caregivers who finished the study did not differ significantly in age, number of years helping, depressive symptomatology, and satisfaction with informal social support from those who did not finish. Only the informational subscale of the informal social support scale differed significantly between groups based on finishing the study.
The caregivers ranged in age from 27 to 92 years with a mean of 64.4 (SD = 14.29). Of these, 73% were women and 27% were men. Although the population sampled was diverse in race and socioeconomic status, the majority was White (92%), high school graduates (75%), and not presently employed (78%). The sample of caregivers included 36 spouses, 18 children, and 6 other relatives. Eighty-two percent of the caregivers were married. Hours per day helping the care recipient ranged from 2 to 24 with 30% reporting they felt obligated to provide assistance 24 hours per day. The number of years providing care ranged from 1 month to 40 years. The majority (67%) of the caregivers had been providing care for less than 3 years.
The care recipients ranged in age from 66 years to 93 years with a mean of 78.9 (SD = 6.44). The sample of care recipients were almost equally divided between men (n = 29) and women (n = 31). Ninety-two percent of the care recipients were White (97%) and were either retired or had not been employed. The caregivers described the care recipients' health as good (11%), fair (47%), or poor (42%). Most had heart disease (25%), stroke (26%), cancer (15%), or skeletal problems (15%).
Description of Services
Seventy-three percent of the carerecipients (? = 44) received home health care during the 3-month study interval. Seven home health care agencies provided an average of 40.5 (SD = 34.3) visits to each family. The home health aides provided the most visits (mean = 17.5, SD = 26.2) followed by RNs (mean = 13.4, SD - 1 1 .45) and physical therapy (mean = 5.07, SD = 7.66). Eleven families received extra assistance by paying for yard work, household chores, or personal assistance for the carerecipient. Thirteen of the caregivers reported they would like more help with chore services, 6 indicated a desire for more help with personal assistance, and 6 needed more respite or time to oneself. The majority of the caregivers (n = 37) who received home health care were satisfied with the service. However, some of the caregivers thought communication could improve between the nurse and caregiver, and a few requested more counseling, occupational therapy, and physical therapy.
Description of Variables
Scores on the study variables are summarized in Table 1. Twenty of the care recipients (33%) who completed the study were readmitted during the 3-month study period. The number of hospital readmissions for the participants who completed the study ranged from 0 to 8 during a 3-month period (mean = .57, SD = 1.18). When examining readmissions and home health care, a chi-square analysis found no significant between-group differences (?2 = .04, df=ì,p = .84) and the association was weak (f = .03, p = .84).
The total number of hospital readmissions was not significantly related to the total number of home health care services, diagnosis or health of the care recipient, or marital status and physical health of the caregiver. At the second interview, a greater number of hospital readmissions was significantly related to greater depressive symptomatology (r = .24), lower tangible support (r = -.24), and lower emotional support (r = -.25). Greater depressive symptomatology at the second interview was significantly related to lower emotional support (r = -.28) and lower satisfaction with informal social support (r = -.56).
SUMMARY OF MEANS1 STANDARD DEVIATIONS, AND RANGES OF PREDICTOR VARIABLES (N = 60)
SUMMARY OF HIERARCHICAL REGRESSION ANALYSIS FOR VARIABLES PREDICTING EARLY HOSPITAL READMISSIONS
Test of the Hypothesis
Multiple regression was used to evaluate whether low social support, low satisfaction with social support, high depressive symptomatology of the caregiver, and minimal use of home health care would predict hospital readmissions of the older adult. Multiple regression was used because the dependent variable (i.e., hospital readmissions) was measured as interval data rather than as a dichotomy. Subscale scores for informal social support (i.e., informational, tangible, emotional, integration) were entered on the first step to evaluate their individual importance. The hypothesis only approached significance with an R2 = .35, F = 2.14, ? = .07. Tangible support and integration were significant predictors. The adjusted R2 was .19, and with the addition of home health care, the change in R2 was only .01. Table 2 describes the regression analysis with all of the variables entered.
Findings from this investigation concur with others which found that approximately one third of clients discharged were rehospitalized within a 3 -month time period (Fethke et al., 1986; Naylor et al., 1994; Rich et al., 1995). Although these researchers did not indicate why 3 months rather than another time interval was studied, the time frame to study change is similar to that suggested by Krause (1989). According to Krause (1989), selecting the appropriate interval to measure change is difficult in longitudinal studies, and data can be biased when using inappropriate between-round intervals.
According to Neuman (1995), interpersonal stressors such as lack of social support determine how one is affected by chronic illness and adjusts to the environment. Consistent with Neuman 's conceptualization of social support, increased tangible forms of social support, such as transportation, watching possessions, and integration or reaching out to help others, were significant predictors of a lower number of early hospital readmissions. These findings suggested social support is not only that provided by others but the feelings one receives by helping others. According to Krause and Markides (1990) and Krause (1997), this reciprocal network with others promotes a sense of belonging and a decreased sense of guilt from asking for help. Furthermore, unequal intergenerational exchanges can be associated with increased psychological distress (Yamagishi & Cook, 1993).
Contrary to Neuman's Model (1995), emotional support was not a significant predictor of hospital readmissions. Lack of significant findings may be related to the costs involved in a social relationship (Tilden & Gaylen, 1987). Not all ties are supportive and may involve decreased privacy and envy. Hence, social relationships can be a source of stress as well as support. Lack of significant findings for informational support may be related to families receiving most information from a formal support network of health care professionals rather than the informal network of family and friends. Lack of significant findings for satisfaction with informal social support may be related to the low Cronbach's alpha of .33 for this sample of caregivers. According to Kane (1997), methodological issues repress the measurement of satisfaction. Social desirability affects internal validity when respondents answer questions according to what they believe are the values of the interviewer.
The total number of home health care services used was not a significant predictor of hospital readmissions. When assessing readmissions and use of home health care, there were no significant between-group differences. These findings are not consistent with Martens and Mellor (1997) and Williams et al. (1995) who found patients with home health care were readmitted significantly less often after hospital discharge. A possible reason for lack of significance in this study is that the older adults in this study had multiple chronic health problems, while those in Martens and Mellor's (1997) study were diagnosed with CHF. Patients with CHF are at high risk for readmissions and are followed closely after hospital discharge.
According to Neuman (1995), intrapersonal stressors such as depressive symptoms are disturbing to one's equilibrium. Depressive symptomatology of caregivers has not been studied by researchers as a predictor of hospital readmissions and was not a significant predictor in regression. Possibly, because depressive symptoms were low for this sample, less variance accounted for the results. Yet, for this sample, less depressive symptoms and emotional support were significantly related to a lower number of hospital readmissions. When comparing bivariate correlations to squared correlations in regression, the correlation model does not distinguish between the independent and dependent variable (Pedhazur, 1982), and only the association between the variables is noted. Therefore, one does not know if increased readmissions affected depressive symptomatology or if depressive symptomatology affected readmissions.
Greater emotional informal social support was significantly related to lower depressive symptomatology. Questions about emotional support addressed the uplift individuals gained from others listening to problems and expressing interest and concern. With this emotional support, feelings of hopelessness and helplessness were diminished. Greater satisfaction with tangible and emotional support was significantly related to lower symptoms of depression. However, because the reliability of this scale was so low, these results must be treated with caution.
LIMITATIONS OF THE STUDY
While addressing important issues in examining hospital readmissions of functionally impaired older adults, a major limitation is use of a small nonprobability sample that limits the generalizability of findings. The sample was further self-selected because of the high attrition rate and lack of diversity in race and educational status. Future research should extend these findings to include a more heterogeneous sample. Another limitation was not distinguishing between cognitive status of the care recipients. Dellasega, Dansky, King, and Stricklin (1994) found patients with cognitive impairment were readmitted to the hospital more frequently than those who were cognitively intact. This finding indicated the patients' needs were not being met. Another limitation resulted from not measuring the number of days the care recipient was hospitalized or the number of readmissions prior to hospitalization. These factors could have been addressed in studying the relationship of hospital readmissions to social support.
Findings in this study have implications for nursing practice, education, and research. For those families receiving increased tangible support such as assistance with chores, transportation, and shopping, hospital readmissions were decreased. Nurses, especially when involved in discharge planning, should assess the needs of the patients and their families for this material support. Nurses should continue to promote the idea that accepting help from others is not a sign of inadequacy in caring for a family member. Because reaching out to others also predicted less hospital readmissions, caregivers should be praised for gaining support from helping others in addition to the patient. By encouraging one to reach out to others, guilt from receiving support may be reduced.
For nursing education, it is necessary for undergraduate- and graduate-level faculty to teach about the needs of older clients and their caregivers. The results of this study indicated the professional nurse spent less time with the care recipient than did the non-professional. With advanced education in gerontological nursing, the professional nurse will be more prepared and able to document the need for increased time with families caring for older adults. Furthermore, because families are becoming more involved in complex technical skills, nursing education needs to stress the importance of patient-family teaching.
Future nursing research examining predictors of hospital readmissions of older adults should address the patients' physical, mental, and functional health, and their caregivers' social support, attitudes, and psychological status. Studies should limit the population to a single diagnosis such as CHF to develop instruments to facilitate comprehensive assessments. Because the number of home health visits was not a predictor of early hospital readmissions, the hours providing home health visits may be more worthy of study.
Future nursing research needs to refine the measurement of satisfaction with social support. According to Kane (1997), a challenge in the development of satisfaction measures is capturing the respondent's perspective in terms that are meaningful to them. What is important to one person may not pertain to another. Furthermore, a lack of an accepted theoretical model for measuring the determinants of satisfaction is problematic.
- Berkman, B., Dumas, S., Gastfriend, M., Poplawski, J.L., & Southworthe, M. (1987). Predicting hospital readmission of elderly cardiac patients. Health and Social Work, Summerß), 221-228.
- Bull, M.J., Jervis, L.L., & Her, M.A. (1995). Hospitalized elders: The difficulties families encounter. Journal of Gerontological Nursing, 2/(6),19-23.
- Buls, P. (1995). The effect of home visits on anxiety levels of the client with a coronary artery bypass graft and of the family. Home Healthcare Nurse, 13(1), 22-29.
- Cadogan, M.P. (1994). Dimensions of selfperceived health status among older adults. Paper presented at The Fourth Annual Research Conference on Aging and Health, Charlottesville, VA.
- Clark, MJ. (1996). Nursing in the community (2nd ed.). Stamford, CT: Appleton-Lange.
- Cohen, J. (1988). Statistical power analysis for the behavioral sciences. Hillsdale, NJ: Erlbaum.
- Dellasega, C, Dansky, K., King, L., & Stricklin, M.L. (1994). Use of home health services by elderly persons with cognitive impairment. Journal of Nursing Administration, 24(6), 20-25.
- Dennis, L.I., Blue, CL-, Stahl, S.M., Benge, M.E., & Shaw, CJ. (1996). The relationship between hospital readmissions of Medicare beneficiaries with chronic illness and home care nursing interventions. Home Healthcare Nurse, 14(4), 303-309.
- Fethke, CC, Smith, I.M., & Johnson, N. (1986). "Risk" factors affecting readmission of the elderly into the health care system. Medical Care, 24, 429-437.
- Fink, S., & Picot, S.F. (1995). Nursing home placement decisions and post-placement experiences. Journal of Gerontological Nursing, 21(3), 35-42.
- George, L.K., & Gwyther, L.P. (1986). Caregiver well-being: A multidimensional examination of family caregivers of demented adults. The Gerontologist, 26, 253-259.
- Gooding, J., & Jette, A.M. (1985). Hospital readmissions among the elderly. Journal of the American Geriatrics Society, 33, 595-601.
- Graham, H., & Livesley, B. (1983). Can readmissions to a geriatric medical unit be prevented? Lancet, 1(1), 404-406.
- Helberg, J.L. (1993). Patients' status at home care discharge. Image: The Journal of Nursing Scholarship, 25(2), 93-99.
- Hennen, J., Krumholz, H. M., Radford, M.J., & Meehan, TP. (1995). Readmission rates, 30 days and 365 days postdischarge, among the 20 most frequent DRG groups, Medicare inpatients age 65 or older in Connecticut hospitals, fiscal years 1991, 1992, and 1993. Connecticut Medicine, 59(5), 263-270.
- Kane, R.L. (1997). Understanding health care outcomes research. Gaithersburg, MD: Aspen.
- Krause, N. (1989). Issues of measurement and analysis in studies of social support, aging, and health. In K.S. Markides & CL. Cooper (Eds.), Aging, stress, and health (pp. 43-65). New York: John Wiley & Sons.
- Krause, N. (1995). Negative interaction and satisfaction with social support among older adults. Journal of Gerontology. Series B, Psychological Sciences and Social Sciences, 50B(6), 59-73.
- Krause, N. (1997). Anticipated support, received support, and economic stress among older adults. Journal of Gerontology. Series B, Psychological Sciences and Social Sciences, 52(6), 284293.
- Krause, N., & Markides, K. (1990). Measuring social support among older adults. International Journal of Aging and Human Development, 39(1), 37-53.
- Lockery, S.A., Dunkle, RE., Kart, CS., & Coulton, CJ. (1993). Factors contributing to the early rehospitalization of elderly people Health and Social Work, 19(5), 182-191.
- Martens, K.H., & Mellor, S.D. (1997). A study of the relationship between home care services and hospital readmission of patients with congestive heart failure. Home Healthcare Nurse, 15(2), 123-129.
- Naylor, M., Brooten, D., Jones, R., LavizzoMourey, R., Mezey, M., & Pauly, M. (1994). Comprehensive discharge planning for the hospitalized elderly. Annals of Internal Mediane, 120(12), 999-1006.
- Neuman, B. (1995). The Neuman systems model (3rd ed.). Norwalk, CT: Appleton & Lange.
- Pedhazur, E.J. (1982). Multiple regression in behavioral research (2nd ed.). Orlando, FL: Holt, Rinehart, & Winston.
- Radloff, L-S. (1977). The CES-D Scale: A selfreport depression scale for research in the general population. Applied Psychological Measurement, 7(1), 385-401.
- Rich, M.W., Beckham, V., Wittenberg, C, Leven, C, Freedland, K.E., & Carney, R. M. (1995). A multidisciplinary intervention to prevent the readmission of elderly patients with congestive heart failure. The New England Journal of Medicine, 333,1190-1195.
- Schlenker, R.E. (1996). Home health payment legilation: Review and recommendations No. 9611. Washington, DC: American Association of Retired Persons.
- Schwarz, K. (1996). Home health care: Formal social support to family caregivers of older adults. Family Perspective, 30(1), 47-61.
- Schwarz, K., & Blixen, C (1997). Does home health care affect strain and depressive symptomatology for caregivers of functionally impaired older adults? Journal of Community Health Nursing, 14(1), 39-48.
- Tennstedt, S., Cafferata, G.L., Sullivan, L. (1992). Depression among caregivers of impaired elders. Journal of Aging and Health, 4(1), 58-75.
- Tilden, V.P., & Gaylen, R.D. (1987). Cost and conflict: The darker side of social support. Western Journal of Nursing Research, 9(1), 9-18.
- Wagnild, G, & Grupp, K. (1991). Major stressors among elderly home care clients. Home Healthcare Nurse, 9(1), 15-21.
- Williams, H., Blue, B., & Langlois, P.F. (1995). Do follow-up home visits by military nurses of chronically ill medical patients reduce readmissions? Journal of Military Nursing Research, Spring(2), 30-33.
- Yamagishi, T, & Cook, K.S. (1993). Generalized exchange and social dilemmas. Social Psychology Quarterly, 56(2), 235-248.
SUMMARY OF MEANS1 STANDARD DEVIATIONS, AND RANGES OF PREDICTOR VARIABLES (N = 60)
SUMMARY OF HIERARCHICAL REGRESSION ANALYSIS FOR VARIABLES PREDICTING EARLY HOSPITAL READMISSIONS