Trends in Mortality, Morbidity and Healthcare Utilization : Does It Make Sense to Use Healthcare As a Proxy for Health?

Medical utilization, such as a count of hospitalizations, is routinely used as a health proxy for both policy and research purposes. Over time, trends in how medicine is practiced has impacted this relationship, as technological improvements or policy changes have altered hospitalizations at the same level of health. In this work, we document how hospitalizations have evolved as a marker for subsequent health (mortality and morbidity) and subsequent utilization (hospitalizations, emergency department visits, physician visits, and others). We assess these trends in the DI population (both fee-for-service and managed care), the aged Medicare fee-for-service population, and a commercial managed care population. We find that hospitalizations continue to predict subsequent mortality well, but that the relationship is more nuanced for morbidity. We find that hospitalizations are declining over time, but that they decline less for those with a prior hospitalization. Other types of utilization are increasing with time, with mixed evidence of the relationship with prior hospitalizations. We conclude that the relationship between hospitalizations and health continues to evolve and that the utility of this proxy depends on what measures of health one is assessing