Medicare beneficiaries with depression: Comparing diagnoses in claims data with the results of screening

Objective: The study evaluated the accuracy of Medicare claims in identifying elderly patients with depression compared with diagnoses of depression made with validated self-report depression scales. Methods: The study included 1,551 participants in the Medicare Primary and Consumer-Directed Care Demonstration. They resided in 19 counties in three states. Depression diagnoses made by two validated self-rated scales - the Mini-International Neuropsychiatric Interview-Major Depressive Episode Module (MINI-MDE) and the 15-item Geriatric Depression Scale (GDS) were compared with depression identified diagnoses listed in Medicare claims. The main outcome measures were the sensitivity, specificity, and positive and negative predictive values for ICD-9-CM depression codes included in Medicare claims. For validation, two-year periods and additional diagnostic codes were also considered. Results: Compared with the MINI-MDE or GDS, the sensitivity and positive predictive values for Medicare claims were below 50%, and specificity and negative predictive values were over 70%. Conclusions: The study demonstrated the predictive power and limitations of using administrative claims data for identifying patients with depression in the Medicare population. Using Medicare claims to identify patients with depression may lead to underestimation of depression prevalence and may compromise researchers' ability to examine issues related to quality, costs, and utilization.