Self-reported and recent claims-based major depressive disorder identifies distinct risk profiles in a nationwide population study.

Background

Major depressive disorder (MDD) in population studies is commonly identified from self-reported history or healthcare records, but agreement is often modest and the implications of discordance remain unclear.

Methods

We included 135,287 Taiwan Biobank participants enrolled from 2008 to 2020 whose baseline self-reported depression status was linked to nationwide health insurance claims. Claims-based MDD required at least two outpatient claims or one inpatient claim in the preceding year. Participants were classified as meeting neither definition, self-reported depression only, claims-based MDD only, or both. Healthcare utilization was assessed during the year before and after baseline; all-cause mortality was followed through 2022 using Cox proportional hazards models.

Results

Most participants met neither definition (94.9%); 2.4% had self-reported depression only, 1.5% had claims-based MDD only, and 1.2% met both definitions. Claims-positive participants had more recorded physical comorbidities and greater healthcare utilization than the self-reported-only group. During a median 6.41-year follow-up, fully adjusted mortality hazard ratios versus the neither-definition group were 1.34 (95% CI, 1.00-1.79) for self-reported depression only, 1.76 (1.31-2.36) for claims-based MDD only, and 1.87 (1.36-2.58) for both definitions.

Limitations

Self-report reflected lifetime history, whereas claims captured diagnoses from the preceding year. Claims-based identification also depended on healthcare use and diagnostic coding. Findings may not generalize beyond Taiwan's healthcare system.

Conclusions

Self-reported and recent claims-based depression identified partially overlapping populations with different recorded health profiles and mortality estimates. Combining survey and claims data may improve depression phenotyping in population research.