Lectures and seminars Improved measures of comorbidity for registry-based studies

Abstract

Accurate measures of comorbidity are often needed for control of confounding in epidemiological studies involving an elderly population. The impact of baseline comorbidity on predicted survival probability is also a key factor for critical treatment decisions in many clinical situations. The Charlson comorbidity index (CCI) is widely used as a measure of baseline mortality risk from comorbidity. Often the patient stratum with very low or zero CCI is large and there is a concern that CCI may not be sufficiently discriminative. We will describe a new approach, using the history of filled prescriptions, that improved prediction of life expectancy beyond what age and the CCI provided. We will also show a development of Elixhauser categories.