Association of acute-course electroconvulsive therapy pathway markers with same-hospital psychiatric readmission in patients with major depressive disorder: A retrospective cohort study.

BACKGROUND

Electroconvulsive therapy (ECT) is an effective acute treatment for severe major depressive disorder (MDD), but whether two routinely captured acute-course ECT pathway markers-time from admission to first ECT and total acute-course session count-are associated with longer-term readmission is unclear. We examined their associations with same-hospital psychiatric readmission, overall and across baseline suicidality.

METHODS

This retrospective cohort study included 4004 inpatients with MDD who received modified ECT during the index hospitalization at a tertiary psychiatric hospital in China (2013-2020); each patient contributed only the earliest qualifying hospitalization. Baseline suicidality was classified as none documented, ideation, or behavior. Separate marker-specific Cox models estimated baseline-adjusted associations with same-hospital psychiatric readmission within 180 and 365 days.

RESULTS

Readmission occurred in 449 patients (11.2%) by 180 days and 593 (14.8%) by 365 days. Neither time from admission to first ECT (hazard ratio [HR], 1.002 per day; 95% confidence interval [CI], 0.989-1.016) nor session count (HR, 1.003 per session; 95% CI, 0.970-1.037) was associated with 180-day readmission; findings were similar at 365 days. There was no evidence of effect modification by baseline suicidality. Total psychiatric hospitalizations including the index hospitalization showed a graded association with readmission. Sensitivity analyses were consistent.

LIMITATIONS: This single-center electronic health record study did not capture other-institution readmission, mortality, or continuation ECT, and could not distinguish planned from unplanned readmissions.

CONCLUSIONS

Acute-course ECT pathway markers were not detectably associated with same-hospital psychiatric readmission in baseline-adjusted marker-specific models. Findings should not be generalized to symptomatic relapse, other-institution readmission, or all-cause outcomes.