Impact of an automated report comparison tool on trainee report modification rate at a tertiary hospital

This study aims to compare trainee-modified report percentage rate and trainee/consultant satisfaction regarding the feedback process before and after implementation of an automated report comparison tool.

An automated report comparison tool utilising natural language processing, presenting the trainee's preliminary report beside the final consultant report with changes highlighted, was used in a prospective interventional study. Modification rates, including character counts, of co-authored computed tomography (CT) studies were recorded before and after tool implementation over two 6-month periods and compared with Student's t-test. Trainees and consultants were surveyed before and after the interventional period for time spent and feedback satisfaction.

In total, 3851 (81.7%) of 4175 reports were modified in the baseline preimplementation phase, and 5215 (69.6%) of 7489 reports were modified during the postimplementation phase (p < .001). The average character count change preimplementation was 132, corresponding to 9.0% of the original preliminary report, compared with 91 characters and 7.1% postimplementation, respectively (p < .001). This statistically significant difference generally applied regardless of the level of trainee experience. Prospective data collected in the preimplementation period revealed that for more than two-thirds of after-hours shifts, trainees spent fewer than 5 minutes receiving feedback on their after-hours work. At the conclusion of the implementation phase, 92.3% of trainees and 70% of consultants agreed that the report comparison tool improved feedback.

Following the implementation of an automated report comparison tool, there was a reduction in trainee report modification rates and subjectively improved trainee feedback. This adjunct to existing feedback mechanisms presents a relatively simple intervention to facilitate efficient case review and feedback.