EQ-5D as a predictor of mortality and hospitalization in elderly people.

Purpose: The objective of this study was to assess whether the EQ-5D independently predicts all-cause mortality and first hospitalization in a generic elderly population and to evaluate the relationship between the EQ-5D "Class of problems," the EQ-5D index, the EQ-Visual Analogue Scale (EQ-VAS), and the number of hospitalizations occurring over a 2-year period.

Methods: The data were collected in a longitudinal study (over 5 years) conducted on 9,644 elderly people in northern Italy with 5,256 completed questionnaires returned. Multivariate Cox's proportional hazards models were used to assess the effect of the EQ-5D questionnaire on 12 and 24 month mortality and first hospitalization risk, controlling for sociodemographic and medical covariates. Multivariate Zero-Inflated Poisson models were performed to test the association between the EQ-5D "Class of problems," the EQ-5D index and EQ-VAS and the number of hospitalizations over 2 years.

Results: The EQ-5D "Class of problems," EQ-5D index, and EQ-VAS were consistently associated with mortality and first hospitalization rates. The EQ-VAS was associated with the number of hospitalizations over 2 years, though the EQ-5D "Class of problems" and EQ-5D index were not.

Conclusions: The EQ-5D questionnaire is a useful predictor of mortality and first hospitalization in a generic elderly population in Italy. This suggests that it would be a good instrument for evaluating the effect of social/medical actions in a population-based study.