Abstract
The evaluation of the relationship between health related quality of life (HRQoL) and the occurrence of negative events such as deaths and hospitalizations in the elderly is well established for short-term periods, but underexplored in the longterm. We aim to assess the capacity of HRQoL to predict mortality and hospitalization in an elderly population over an 8-year period. Data were collected by means of the “Pianoro Study”, through a questionnaire, which contained the EQ-5D and EQ-VAS instruments to measure the HRQoL. Deaths and hospitalizations occurring over the following 8 years were collected from the registers of the Health Authority. The Cox proportional hazard model and the zero-inflated negative binomial regression model were used to verify the capacity of the EQ-5D to predict deaths and first hospitalizations and the number of hospitalizations respectively, controlling for socio-demographic and clinical variables. High values on the EQ-VAS were significantly associated with a lower risk of death and first hospitalization, whilst the EQ-5D classes of problems and the EQ-5D dimensions were associated with a greater risk (less consistent in predicting mortality). Moreover, the EQ-VAS was found to be significantly associated with both the probability of a zero hospitalization rate and the number of hospitalizations, whereas the EQ-5D classes of problems and the EQ-5D single dimensions were significantly associated with the number of hospitalizations but not with a zero hospitalization probability rate. HRQoL, measured by the EQ-5D questionnaire, is an important prognostic factor in predicting long-term mortality and hospitalization.