Weighted multimorbidity indexes predicted mortality, health service use, and health-related quality of life in older women

Abstract

Objective: To develop indexes of multimorbidity, based on self-reported data, to predict mortality, health service use, help with activities of daily living (ADL), and health-related quality of life (HRQOL) in older women.

Study design and setting: Cross-sectional survey of 10,434 women, aged 73-78 years, in the Australian Longitudinal Study of Women's Health in 1999, with mortality follow-up to 2005. For analysis, the sample was equally split into a development and validation sample. Weighted and unweighted multimorbidity indexes were developed and tested.

Results: Outcomes ranged from 14% for mortality to 47% for specialist doctor visits. Mortality was predicted by heart disease, stroke, low iron, diabetes, cancer (nonskin), bronchitis/emphysema, and Alzheimer's disease. Different patterns of morbidities were associated with the other outcomes. Weighted and unweighted multimorbidity index scores were linearly related to increasing risk of each outcome. For each outcome, the weighted scores fitted the data better and had a wider range of possible values.

Conclusion: These multimorbidity indexes predict mortality, health service use, help with ADL, and HRQOL in older women. The indexes could be used as covariates in research with weighted scores having a better chance of discriminating between patient groups than unweighted scores.