Validity of the Medication-Based Disease Burden Index Compared with the Charlson Comorbidity Index and the Cumulative Illness Rating Scale for Geriatrics

Background: Co-morbidity is common in older people. A co-morbidity index reduces coexisting illnesses and their severity to a single numerical score, allowing comparison with scores from other patients. Recently, the Medication-Based Disease Burden Index (MDBI) was developed.

Objective: The aim of the study was to assess the MDBI’s validity in hospitalized elderly patients.

Methods: Clinical and demographic data and data on patients’ medications on admission were obtained prospectively. Retrospectively, we applied the MDBI to the patients’ medication regimens, determining their co-morbidity using the Charlson Comorbidity Index and Cumulative Illness Rating Scale for Geriatrics (CIRS-G). The MDBI’s criterion validity was assessed against the Charlson and CIRS-G indices. Convergent and discriminant validities were also assessed. The MDBI’s predictive validity was assessed by its ability to predict 3-month post-discharge readmissions or mortality compared with the Charlson and CIRS-G indices.

Results: MDBI scores were correlated with the Charlson and CIRS-G indices’ scores (r = 0.44 and r = 0.37, respectively [p<0.001]). MDBI, Charlson Co-morbidity Index and CIRS-G scores were correlated with the number of drugs (r = 0.52, r = 0.34 and r = 0.40, respectively [p<0.001]) and were the same in both sexes. No significant differences in MDBI scores were found between cognitively normal and impaired mental status (IMS) patients or between the functionally independent and partially/fully dependent patients. Charlson Comorbidity Index and CIRS-G scores were significantly lower in IMS patients and in dependent patients. The MDBI had no predictive ability for 3-month mortality but had good predictive power for a composite of 3-month mortality or readmissions (odds ratio [OR] 2.99 [95% CI 0.99, 9.03; p = 0.051]). However, CIRS-G and Charlson indices had good predictive ability for mortality (OR 1.50 [95% CI 1.22,1.84; p<0.001] and OR 2.06 [95% CI 1.40, 3.02; p<0.001], respectively) and for a composite of 3-month mortality or readmissions (OR 1.24 [95% CI 1.11, 1.34; p<0.001] and OR 1.39 [95% CI 1.12, 1.72; p = 0.003], respectively).

Conclusions: The MDBI showed satisfactory criterion, convergent and discriminant validities and good predictive validity for mortality or readmission, but failed to differentiate between cognitive and functional patient groups. The MDBI should be investigated in larger studies to determine its validity in settings where medication data rather than diagnostic data are more readily available. In clinical practice with elderly patients, we recommend employing co-morbidity indices that are based on medical records, such as the Charlson Comorbidity Index and CIRS-G.