Enhancing Depression Identification and Stratification with a Claims-Based Analytical Framework

Depression, a prevalent health condition, substantially impacts both socioeconomic outcomes and individual wellbeing. Despite the availability of diagnostic tools, existing approaches for identifying depression severity often rely on single-indicator approaches, limiting accuracy. This retrospective study evaluates a multi-parameter analytics-enabled Identification and Stratification (IDS) framework designed to improve depression identification and severity stratification by leveraging health insurance claims and electronic health record data. For the evaluation, Highmark Health dataset was used, consisting of records for members aged 18 + with at least one healthcare encounter. The IDS framework identified 720,882 members with depression (16.6% of the population). The framework identified 258,206 more members (5.9% of the population) compared to using diagnoses alone. The stratification rules revealed variability in prevalence, with 5.0% mild, 8.5% moderate, 2.2% severe, with the remaining 0.9% in unknown, remission, or minimal. The IDS rules escalated 46% of mild and 19% of moderate cases to higher severity compared to single indicator assessments. Expenses for severe depression were, on average, 2.5 times higher than for minimal. The IDS framework demonstrated utility in identifying members with depression by linking fragmented data sources. Aligning multiple indicators provided a more comprehensive identification and a more nuanced severity evaluation compared to individual data elements. This enables targeting of cost-effective digital self-care tools to milder cases while reserving higher cost interventions for the most severely ill, potentially reducing costs while maintaining health outcomes. Implementation of this integrative platform can help focus efforts on those with the highest need and bridge the gap in treating depression.