Implementation of a Software Distribution Intervention to Improve Workload Balance in an Academic Pediatric Radiology Department

Abstract

In our pediatric radiology department, radiographs (XR) are the shared responsibility of the body section and interpreted in addition to modality or site-specific assignments. Given an unequal contribution amongst radiologists to the XR workload, a software solution was developed to distribute radiographs and improve workload balance. Metrics to evaluate the intervention's effectiveness were compared before and after the intervention. Data was retrieved from the radiology analytics platform, scheduling software, and the peer learning database. Metrics were compared 12 months pre (March 2018-February 2019) and 6 months post (March 2019-August 2019) intervention on non-holiday weekdays, 7 am-5 pm. To evaluate the intervention's effectiveness, variance between radiologists' contributions to XR volume was assessed using Levene's and Fisher's tests. Changes in turnaround times (TATs) and error rates pre- and post-intervention were evaluated as secondary metrics. Following the intervention, the average number of XR interpreted on target rotations increased by 8.9% (p = 0.011) while the departmental volume of radiographs increased only 4.5%. The variance between radiologists' daily XR contribution was 21.3% (p < 0.0001) higher prior to the intervention. Days where target rotations read fewer than 5 XR decreased from 17.8 to 1.1% (p < 0.0001) after the intervention. Days in which more than 75% of all XR had a TAT less than 60 min improved from 26.8 to 39.7% (p = 0.017) after the intervention. There was no statistically significant difference in error frequency (error rate 2.49% pre and 2.72% post, p = 0.636). In conclusion, the software intervention improved XR workload contribution with decreased variability. Despite increased volumes, there was an improvement in turnaround times with no effect on error rates.

Keywords: Burnout; Motivation; Radiologists; Software; Workflow; Workload.

Fig. 1

Screenshot of the typical daily worklists using the worklist manager. Radiologists select their primary assignment (*) and also are expected to opt-in (denoted by checkmark) to all XR (arrow). Subspecialty rotations which were not targeted for the distribution intervention are marked by the orange box

Fig. 2

Within the existing software manager, a new filter is created to include all exams that should be distributed. At our facility, we excluded stat, urgent, and inpatient exams, as well as specialized exams such as genetic bone surveys and leg lengths. The software rule is configured to take exams from the source filter and distribute them to existing target worklists

Fig. 3

Worklist screenshot showing the arrow icon that was added to distributed exams to allow radiologist to know a distributed exam was their responsibility. For example, in this case, the exam with the arrow would only be visible to the radiologist interpreting off the “Woodlands Body” primary assignment, while the other XR on the list are on the “ALL XR” list

Fig. 4

Prior to the intervention, there were many days where the radiologist interpreted fewer than five exams. After the intervention, the number of exams read by radiologists per shift were more centrally concentrated with a significantly lower number of shifts with fewer than 5 exam read (see Fig. 5 for a breakdown of this last change by radiologist)

Fig. 5

Prior to the intervention, a few radiologists frequently interpreted fewer than 5 XR (despite expectation of interpreting 20). Following the intervention, the frequency of interpreting fewer than 5 XR improved for all radiologists