Predicting Incident Treatment-Resistant Depression: A Model Designed for Health Care Systems
Background: Treatment resistant depression (TRD) is burdensome to health systems. Earlier identification of at-risk patients may reduce clinical inertia and improve outcomes.
Methods: Integrated claims and electronic health records across health systems identified adults with major depressive disorder initiating antidepressant (AD) treatment (7/1/2014 - 1/1/2017 who completed 1+ course at adequate dose and duration (qualifying). TRD was defined by initiating a 3rd course following two failed qualifying courses. Regression identified significant predictors of TRD; model fit was assessed using receiver operating characteristics determined by area under the curve (AUC).
Results: A total of 35,246 people met eligibility; 7,098 (20.1%) met TRD criteria after an average of 402 days. Predictors included age, insomnia, hypertension, psychiatric medications attempted, psychiatric office visits, nurse telephonic encounters, index AD class, qualifying initial AD episode, non-psychiatric prescriptions (count), suicidality, and physician specialty. Psychiatric office visits, nurse encounters, and suicidal ideation or attempt were protective. ADs prescribed by a psychiatrist were protective (OR: 0.89); ADs prescribed by a pain specialist were predictive (OR: 1.70). Other significant positive predictors included: lag from initial depression diagnosis to medication initiation; unique ADs attempted, claims for 1+ unique anticonvulsants (OR: 1.4) and 2+ unique anxiolytics (OR: 1.34). Compared to SNRIs, initiating on SSRIs or on multiple classes was significant (ORs=1.3 and 0.72). The final model achieved an AUC=0.83.
Conclusion (Implications for Practice): TRD transition occurs after an extended treatment period, suggesting clinical inertia. Monitoring risk factors may allow health systems to identify patients at risk for TRD earlier, potentially improving outcomes.
This poster was presented at the 32nd annual Psych Congress, held Oct. 3-6, 2019, in San Diego, California.
