Randomised controlled trial of audit-and-feedback strategies to reduce imaging overutilisation in the emergency department

BACKGROUND: Evaluation of neck trauma is a common reason for emergency department (ED) visits. There are several validated clinical decision rules, such as the National Emergency X-Radiography Utilization Study (NEXUS) Cervical Spine (C-spine) Rule, that can be used to risk stratify these patients and identify low-risk patients who do not require CT imaging. Overutilisation of CT imaging exposes patients to unnecessary radiation, impairs hospital throughput and increases healthcare costs. Various audit-and-feedback strategies have been described in other settings, but it is not known whether these strategies are effective for reducing imaging overutilisation in the ED. Additionally, the effectiveness of face-to-face feedback strategies as compared with digital feedback strategies for addressing this problem has not been previously evaluated. The aim of this study was to compare audit-and-feedback strategies to reduce CT overutilisation in the ED. METHODS: This was a prospective randomised controlled trial, in which emergency medicine clinicians were randomised into three arms to receive digital feedback, hybrid face-to-face/digital feedback or no feedback. Each clinician received three rounds of feedback on patient encounters in which they ordered a CT of the C-spine. Patient encounters were retrospectively reviewed to determine each clinician's overutilisation rate, defined as the percentage of patients who underwent CT of the C-spine despite being classified as low risk by NEXUS criteria. RESULTS: A total of 78 emergency medicine clinicians were randomised into three arms. Baseline overutilisation rates for each group were 46%-47% of CT of the C-spine studies. After three rounds of audit-and-feedback strategy, the clinicians in the digital feedback group had an overutilisation rate of 33%, compared with 44% in the control group (p=0.020). The hybrid feedback group had an overutilisation rate of 36% (p=0.055 vs control; p=0.577 vs digital feedback). Over the study period, the digital group saw a reduction of 1.26 CT of the C-spine studies per provider per month (p=0.049), and the hybrid feedback group saw a reduction of 1.43 CTs per provider per month (p=0.044). CONCLUSION: A digital audit-and-feedback strategy is effective for reducing overutilisation of CT imaging of the C-spine in the ED, while the effectiveness of a hybrid strategy requires further investigation.