Introduction: The Canadian CT Head Rule (‘the rule’) is widely used across the country and its use is specifically recommended by Choosing Wisely Canada. Studies in Canadian hospitals have shown appropriate declines in CT scans when decision tools have been made readily available and useable at the point of care. Research into the implementation of the Canadian CT Head Rule in particular has shown that barriers to its use include an inability to accurately recall each criteria and forgetting to attempt to apply the rule altogether. In an attempt to provide our clinicians with effective access to the rule, we modified CT requisitions and order procedures to facilitate the use of the rule for every head CT in our emergency department (ED). Methods: A quality improvement (QI) approach was used to pilot, implement, and evaluate the modified CT requisition at our hospital. Several Plan-Do-Study-Act cycles involving stakeholders in the hospital resulted in iterative changes to the requisition leading to the implemented version. The new requisition required physicians to indicate which rules or exclusion criteria were met and this was made mandatory for all head CTs ordered. Demographic data was collected on all patients presenting to the ED on age, gender, CTAS level, disposition, and length of stay. Data on which exclusion criteria were appropriate, the rules met leading to CT scans, whether each requisition was used appropriately, and whether there was a significant injury found was collected for each patient receiving a head CT after implementation. Results: In our primary outcome (% of ED visits receiving a head CT), preliminary results have demonstrated a relative reduction in head CT ordering of 10.9%. Our study at completion is powered to detect a ~10% relative change in ordering behaviour, and a Chi square of the data to date yields a P-value of 0.0147. There are no significant differences in visit volume or any of the demographics collected to date. Final results including analysis are anticipated in March, 2016. Conclusion: Preliminary results on this simple, no-cost intervention are very promising. The reduction in head CTs ordered suggests that with mandated access to an easy-to-use, well validated decision tool, ED physicians have been able to confidently defer scans that have a very low risk of having any significant injury present, reducing cost, radiation exposure, and perhaps time in department.
A Deep Learning Algorithm for Automatic Detection and Classification of Acute Intracranial Hemorrhages in Head CT Scans