Predictors of rehabilitation outcomes: a comparison of Israeli and Italian geriatric post-acute care (PAC) facilities using the minimum data set (MDS)

OBJECTIVES: To understand the relative contribution of sociodemographic, clinical, and health care features to rehabilitation outcomes in Israel and in Italy in post-acute care (PAC) facilities. DESIGN: Prospective cross-national study SETTING: Two hospital geriatric PAC departments: Harzfeld Geriatric Hospital, Gedera, Israel, and Catholic University of Sacred Heart Geriatric Hospital, Rome, Italy. PARTICIPANTS: Post-acute care patients aged 65 and older admitted consecutively for stabilization, improvement, or rehabilitation to 3 departments in Harzfeld Geriatric Hospital, Gedera, Israel from April, 1999 through February, 2002 (N = 364), and to the post-acute Geriatric Rehabilitation Unit of the "A. Gemelli" Hospital, Catholic University of Sacred Heart, Rome, Italy, between February, 1999, and April, 2002 (N = 351), for whom there were complete assessments at admission and discharge (the total number admitted in Israel was 505, and in Italy, 409). MEASUREMENTS: Minimum Data Set for Post-Acute Care (MDS-PAC) assessments conducted within 4 days of admission and at discharge; data collected identically in both sites. Predictors of functional recovery> were identified using multivariate binary logistic regression. The dependent variable: improvement of 1 or more points in the ADL scale. RESULTS: The staffing pattern of the PAC department in Italy had about double the physicians and physio- and occupational therapists than in Israel, but about the same number of nurses and somewhat fewer aides than in Israel. Multivariate binary logistic regression that includes country, age, sex, and marital status, found that the patients in Italy had about triple the probability of improvement in ADL function (OR 3.3, CI 2.4-4.6) (P < .001) than PAC patients in Israel. Even after health system characteristics were added to the model, ADL improvement was most significantly associated with higher cognitive ability and a diagnosis of hip fracture, as well as longer length of stay and being admitted to PAC directly from an acute hospital. For each additional point (worse cognition) in a cognitive scale, there was a 30% decrease in the probability of ADL improvement (OR 0.7, CI 0.6-0.8, P < .001). Those who had a stroke were about half as likely to show ADL improvement (OR 0.5, CI 0.3-0.7) than those without stroke, but those with a hip fracture had more than double the probability of ADL improvement (OR 2.7, CI 1.7-4.2) than those without hip fracture. Those who stayed in the PAC ward an additional block of time had a 30% higher probability of ADL improvement (P < .1), and those who were admitted directly to PAC from an acute hospital had more than 4 times the probability of ADL improvement (OR 4.1, CI 2.3-7.0, P < .001) than those who were admitted from a private home. CONCLUSIONS: We found support for the hypothesis that differences in sociodemographic and clinical factors cannot account for all differences in ADL improvement, and that the organization of care and constraints of the health system also influence functional outcomes. Policymakers should examine the policy-amenable features of the Italian and Israeli systems so that optimal ADL recovery can be encouraged. Any reduction in disability will help both patients and the health care system; slightly higher short-term PAC treatment costs may have large long-term future benefits, if they result in the reduction of ADL disability. This study is one of the first to examine outcomes of PAC in 2 countries, and can provide an initial assessment of how rehabilitation can be enhanced or limited by health policies and staffing patterns.