Clinician clusters significantly influence antidepressant prescribing patterns and patient outcomes for major depressive disorder, according to a recent study published in JAMA Psychiatry.
The cohort study examined electronic health record data from 11,934 clinicians across 2 academic medical centers, 6 community hospitals, and affiliated clinics in eastern Massachusetts from March 1, 2008, to April 27, 2022. Researchers included clinicians who billed at least 10 major depressive disorder (MDD) diagnoses per year and used unsupervised learning techniques to identify clinician clusters based on diagnostic coding.
Ten distinct clinician clusters were identified, revealing considerable variability in antidepressant prescribing patterns. Statistically significant differences were found between clusters in rates of prescribing for major antidepressant classes, including selective serotonin reuptake inhibitors (SSRIs), selective norepinephrine reuptake inhibitors (SNRIs), and tricyclic antidepressants (TCAs). Variability was also observed in prescribing heterogeneity and patient outcomes, with the rate of loss to follow-up ranging from 27% to 69% and the achievement of stable antidepressant treatment from 22% to 42%.
Psychiatric clinicians demonstrated the greatest heterogeneity in antidepressant prescribing. The cancer and kidney disease clusters had the highest rate of antidepressant prescribing, similar to outpatient psychiatry after excluding non-prescribing clinicians. SSRIs were prescribed at the highest rate in the obstetrics/gynecology (OB/GYN) cluster. The study also examined the predictive validity of clinician clusters for patient outcomes, finding that they were modestly predictive of dropout and achieving stability.
The researchers discuss implications for precision medicine approaches in psychiatry, highlighting the importance of considering treatment setting alongside patient-level features in developing strategies for depression treatment.
The analysis was conducted between September 2023 and January 2024. Full disclosures can be found in the published study.