Radiation exposure from CT imaging accumulates silently over a patient's lifetime, and emergency departments have long struggled with over-imaging in lower-risk cases. A structured campaign showing measurable, statistically significant reductions in unnecessary head and cervical spine CT use challenges the assumption that cultural and clinical habits in emergency settings are too entrenched to change — and has direct implications for patient safety and healthcare efficiency.
Across two emergency departments — an academic trauma center and a community hospital — a multimodal quality improvement intervention spanning 23 months produced consistent reductions in CT utilization among low-risk trauma patients. Head CT rates fell from 5.00 to 4.01 per 100 patients at the academic site (rate ratio 0.80; p < .001) and from 6.64 to 5.12 per 100 patients at the community site (rate ratio 0.77; p < .001), representing roughly 20–23% reductions. Cervical spine CT showed comparable directional improvements. Critically, these reductions occurred without a corresponding signal of missed injuries, suggesting the eliminated scans were genuinely low-yield.
This work sits within a well-established but persistently underacted-upon evidence base. Decision instruments like the Canadian CT Head Rule and NEXUS criteria have existed for decades, yet emergency physician adherence remains inconsistent — driven by medicolegal pressure, time constraints, and patient expectation. The novelty here lies not in the clinical criteria themselves but in the implementation architecture: a sustained, multimodal campaign that achieved durable behavior change across two institution types simultaneously. That the community hospital showed a slightly larger proportional reduction hints that baseline over-imaging may be more pronounced in non-academic settings, where specialist backup is less available and defensive imaging more common. Limitations include the absence of a randomized control group, single-network design, and the prospective-observational nature of the study, which limits causal inference. Still, for health systems seeking scalable approaches to radiation stewardship, this represents a pragmatically replicable model with meaningful patient safety dividends.