Survival prediction in nursing home residents using the Minimum Data Set subscales: ADL Self-Performance Hierarchy, Cognitive Performance and the Changes in Health, End-stage disease and Symptoms and Signs scales

Abstract

Background: With the intention to aid planning for elderly focused public health and residential care needs in rapidly aging societies, a simple model using only age, gender and three Minimum Data Set (MDS) subscales (MDS-ADL Self-Performance Hierarchy, MDS-Cognitive Performance and the MDS-Changes in Health, End-stage disease and Symptoms and Signs scales) was used to estimate long-term survival of older people moving into nursing homes. Methods: A total of 1820 nursing home residents were assessed by the MDS 2.0 and their mortality status 5 years later was used to develop a survival prediction model. Result: In December 2006, 54.2 of subjects were dead. Older age at nursing home admission (HR 1.036 per 1-year increment, 95 CI 1.0281.045), men (HR 1.895, 95 CI 1.6512.175), higher impairment level according to the MDS-ADL (HR 1.135 per 1-unit increment, 95 CI 1.0991.173) and MDS-CPS (HR 1.077 per 1-unit increment, 95 CI 1.0331.123), and more frail on the MDS-CHESS (HR 1.150 per 1-unit increment, 95 CI 1.0421.268), were all independent predictors of shorter survival after nursing home admission in multivariate analysis. Survival function was derived from the fitted Cox regression model. Survival time of nursing home residents with different combinations of risk factors were estimated through the survival function. Conclusion: The MDS-ADL, MDS-CPS and MDS-CHESS scales, in addition to age and gender, provide prognostic information in terms of survival time after institutionalization. The model may be useful for health care and residential care planning in an ageing community.