Validation of Four Clinical Indicators of Preventable Drug-Related Morbidity

Abstract

BACKGROUND: Clinical indicators are tools that assess quality issues related to the use of medicines. At this time, validated clinical indicators for preventable drug-related morbidity (PDRM) are lacking. OBJECTIVE: To assess the validity and reliability of using population administrative claims data to identify the extent of PDRM in older adults in Canada. METHODS: Four indicators of PDRM related to cerebrovascular and cardiovascular care were chosen for validation. A random sample of cases that represented the indicators and fit the criteria (hits) for PDRM from the retrospective operationalization of the study database and those that did not fit the criteria (near hits) were selected for chart review. One-page abstracts of the cases were prepared for review by a panel of 5 clinical pharmacists. Validity was assessed by calculating sensitivity, specificity, and positive and negative predictive value. Reliability was assessed using reviewers' agreement scores (κ statistics). RESULTS: Overall, 119 case abstracts were reviewed by each panelist. The sensitivity ranged from 33% to 100% and the specificity from 51% to 71%. Predictive values ranged from 5.3% to 43% (positive) and 90% to 100% (negative). The overall κ statistic was fair (0.21). CONCLUSIONS: The validity of the 4 assessed PDRM indicators varied. The reliability was fair; however, these indicators may be useful to screen older adults for PDRM.