Variables correlated with elderly referral from nursing homes to general hospitals

BACKGROUND Referring patients from nursing homes to general hospitals exposes them to nosocomial diseases, and may result in the development of a broad spectrum of physical, mental and social damages. Therefore, minimizing the referring of nursing home patients to hospitals is an important factor for keeping the elderly healthy and minimizing health expenditures. In this study we examined the variables related to the referral rates from nursing homes to general hospitals and the relationship between the referral and the mortality rates among the elderly who live in nursing homes in the Haifa Sub-district. METHODS Thirty-two nursing homes were included in a cross-sectional study. All medical directors and head nurses were interviewed using a structured questionnaire between November 2006 and October 2007. Statistical analysis, including the ANOVA and the nonparametric Spearman tests, were employed to determine the factors that influence referral rates and the correlation between referral rates and mortality rates. RESULTS The referral rate ranged between 18 and 222 per 100 financed elderly in a single year. In the multivariate analysis, the absence of a physician from the nursing home at the time of the referral to general hospitals was the only significant variable related to referral rates. No significant relationships were found between referral rates and mortality rates. CONCLUSIONS Absence of a significant relationship between referral rates and mortality rates may indicate that high referral rates do not necessarily protect the elderly from death. Therefore, any recommendations issued by the Ministry of Health (MOH) should emphasize in-house treatment rather than hospitalization. Clear instructions on referral from nursing homes to general hospitals need to be constructed by the MOH. The MOH should increase the presence of physicians in the nursing homes, especially, when the need to refer a patient arises. Further quantitative and epidemiologic studies should be conducted in order to, more fully and reliably, create guidelines for policy recommendations.