Validity of medication-based co-morbidity indices in the Australian elderly population

Objectives: To determine the validity of two medication-based co-morbidity indices, the Medicines Disease Burden Index (MDBI) and Rx-Risk-V in the Australian elderly population. Methods: In Phase I, the sensitivity and specificity of both indices were determined in 767 respondents from wave 6 of the Australian Longitudinal Study of Ageing (ALSA). Medication-defined index disease categories were compared to self-reported medical conditions. Correlation with self-rated health was examined and Cox proportional hazards models were used to assess the predictive validity for mortality. Phase II verified the predictive ability of Rx-Risk-V in a sample of 213,191 veterans from Australian Department of Veterans' Affairs (DVA) database. Results: MDBI and Rx-Risk-V scores could be calculated for 28% and 73% of the ALSA sample respectively. Both indices had high specificities and low to moderate sensitivities compared to self-reported medical conditions. Total weighted scores were significantly related to self-rated health (p<0.001). Both indices were predictive of mortality (Hazard Ratio (HR)=3.690 (95% CI 2.264-6.015) for MDBI and HR 1.079 (95% CI 1.045-1.114) for Rx-Risk-V. The predictive validity for mortality of Rx-Risk-V was confirmed using DVA data (HR=1.090, 95% CI 1.088-1.092). Conclusions: Medication-based co-morbidity indices Rx-Risk-V and MDBI are valid measures of co-morbidity. However, Rx-Risk-V detects more co-morbidity in the Australian elderly population and is likely to be a more suitable index to use in administrative datasets, particularly where studies include large numbers of outpatients.