After-hours on-site radiologist availability and emergency CT reporting: a roster-based cohort with staffing-boundary analysis

To compare after-hours emergency CT report turnaround during periods with versus without roster-recorded on-site CT radiologist availability.

This retrospective roster-based cohort included emergency-department CT examinations from 1 July 2025 to 30 June 2026. The common-support cohort comprised non-holiday weekdays from 15:00 to 06:59 and all weekends and public holidays. Components within 15 min were grouped into packages. The exposure was on-site CT radiologist availability recorded in the final as-worked duty roster; actual reporting route was secondary. The primary outcome was scan-to-complete-report turnaround time (TAT). Models used duty-date-clustered standard errors and adjusted for patient, temporal, workload, and examination characteristics.

The cohort contained 13,533 CT packages; 13,270 had usable TAT and 12,333 index encounters were clinically linked. Median TAT was 81 min without and 58 min with an on-site radiologist. Absence of an on-site radiologist was associated with 33.9% longer adjusted TAT (95% CI 25.4-43.0). The association was 28.4% for routine non-contrast head CT and 45.5% for abdominopelvic CT (interaction + 15.1%; p = 0.001). A ± 120-minute staffing-boundary analysis estimated 45.7% longer TAT (16.3-82.5), with positive estimates across ± 60- to ± 180-minute bandwidths and no detected covariate discontinuities. Actual-route TAT was 47.7% longer amid 34.5% route discordance.

During comparable after-hours periods, roster-recorded absence of an on-site CT radiologist was associated with longer complete-report TAT, particularly for abdominopelvic CT. Downstream stay differences were exploratory and may reflect broader hospital processes.