Physician-level variation exceeds time-of-day variation in diagnostic resource utilization in low-acuity emergency department care.

Abstract

Background: Emergency physicians working prolonged shifts may show temporal variation in clinical practice. We examined whether diagnostic resource utilization differed across scheduled shift periods in a structured low-acuity emergency department setting.

Methods: This retrospective observational study included 11 032 visits for upper respiratory tract infection, myalgia, abdominal pain, or headache across 12 nonconsecutive 24-hour shifts (5 December 2025-19 January 2026) in a high-volume tertiary emergency department in Turkey. Visits were classified as early (0-5 h; 08:30-13:30), mid (5-10 h; 13:30-18:30), or late (10-16 h; 18:30-00:30) based on elapsed time since the scheduled shift start at 08:30. Multivariable logistic regression with physician-level clustered standard errors (19 clusters) adjusted for age, sex, and diagnostic category.

Results: Laboratory testing rates were 9.5% across all periods (P = .997); imaging rates were 13.5%, 12.8%, and 12.6% (P = .485). The composite utilization score did not differ significantly (P = .682). Late-period consultation was lower than in the early period (odds ratio = 0.541, 95% confidence interval: 0.306-0.957), although this finding was limited by the low event count (n = 82). Inter-physician variation substantially exceeded shift-period variation, with laboratory rates ranging from 3.5% to 21.5% and imaging rates from 6.3% to 28.7%.

Conclusion: In this low-acuity emergency department setting, scheduled shift period was not measurably associated with increased laboratory or imaging utilization. Physician-level variation appeared more prominent than shift-period variation, supporting prospective studies incorporating fatigue, workload, and decision-quality measures.