Purpose or Case Report: In our pediatric radiology department, radiographs (XR) are the shared responsibility of body radiologists, expected to be read in addition to daily modality-based or site-specific assignments. Due to concerns that the increasing XR volume was unevenly shared amongst colleagues, a software intervention was developed to improve weekday 7am-5pm workload balance by auto-distributing exams at 10-minute intervals during peak hours to rotation worklists within PACS; a cap of 20 distributed exams to each target worklist was set. Metrics to evaluate the intervention’s effectiveness were assessed. Additionally, as there was concern that assigning exams may result in slower turnaround times (TATs) and increased errors, TATs and error rates were evaluated.
Methods & Materials: This study was HIPAA compliant and exempt from IRB approval. Data were retrieved from the electronic health record, scheduling software, and peer learning database. Daily mean XR volume was determined. Data from 12 months pre and 6 months post intervention on weekdays were compared, metrics included: number of studies read by each radiologist on target rotations, variance of daily XR count per target rotation, frequencies of reading 5 or fewer XR exams on target rotations, daily median TATs and variance, and frequency of errors. Levene’s test for equality of variance was used to determine significance of variance differences. Fisher’s exact test was used for frequency differences.
Results: Post intervention, the variance of the daily XR counts read on target rotations decreased by 33% (p<0.0001). Mean daily XR volume on target rotations increased by 9% (p=0.0063). Mean daily institutional XR volume increased by 6%. Following intervention, the percentage of days where a body radiologist read fewer than 5 XR decreased from 18.5% to 0.9% (p<0.0001). Mean of daily median XR TAT decreased by 11% from 31.1 min pre to 27.7 min post, with decreased variance of median daily TATs from 418.1 to 114.3 (p<0.0022). There was no statistically significant difference in error frequency reported in the peer learning database (2.65% pre, 2.62% post, p=1).
Conclusions: Utilizing a novel software intervention, there was improved workload balance with more equitable radiologist contribution and decreased variability. Despite increased volumes, there was an improvement in median daily TAT, with decreased variance. There was no effect on error frequency reported in the peer learning database.