Patterns and Impact of Electronic Health Records-Defined Depression Phenotypes in Spine Surgery

BACKGROUND Preoperative depression is a risk factor for poor outcomes after spine surgery. OBJECTIVE To understand effects of depression on spine surgery outcomes and healthcare resource utilization. METHODS Using IBM's MarketScan Database, we identified 52 480 patients who underwent spinal fusion. Retained patients were classified into 6 depression phenotype groups based on International Classification of Disease, 9th/10th Revision (ICD-9/10) codes and use/nonuse of antidepressant medications: major depressive disorder (MDD), other depression (OthDep), antidepressants for other psychiatric condition (PsychRx), antidepressants for physical (nonpsychiatric) condition (NoPsychRx), psychiatric condition only (PsychOnly), and no depression (NoDep). We analyzed baseline demographics, comorbidities, healthcare utilization/payments, and chronic opioid use. RESULTS Breakdown of groups in our cohort: MDD (15%), OthDep (12%), PsychRx (13%), NonPsychRx (15%), PsychOnly (12%), and NoDep (33%). Postsurgery: increased outpatient resource utilization, admissions, and medication refills at 1, 2, and 5 yr in the NoDep, PsychOnly, NonPsychRx, PsychRx, and OthDep groups, and highest in MDD. Postoperative opioid usage rates remained unchanged in MDD (44%) and OthDep (36%), and reduced in PsychRx (40%), NonPsychRx (31%), and PsychOnly (20%), with greatest reduction in NoDep (13%). Reoperation rates: 1 yr after index procedure, MDD, OthDep, PsychRx, NonPsychRx, and PsychOnly had more reoperations compared to NoDep, and same at 2 and 5 yr. In NoDep patients, 45% developed new depressive phenotype postsurgery. CONCLUSION EHR-defined classification allowed us to study in depth the effects of depression in spine surgery. This increased understanding of the interplay of mental health will help providers identify cohorts at risk for high complication rates, and health care utilization.