MP38: Barriers and facilitators to physician use of computerized clinical decision support for mild traumatic brain injury and suspected pulmonary embolism

Introduction: As utilization of CT imaging has risen dramatically, evidence-based decision rules and clinical decision support (CDS) tools have been developed to avoid unnecessary CT use in low risk patients. However, their ability to change physician practice has been limited to date, with a number of barriers cited. The purpose of this study was to identify the barriers and facilitators to CDS adoption following a local CDS implementation. Methods: All emergency physicians at 4 urban EDs and 1 urgent care center were randomized to voluntary evidence-based CT imaging CDS for patients with either mild traumatic brain injury (MTBI) or suspected pulmonary embolism (PE). CDS was integrated into the computerized physician order entry (CPOE) software and triggered whenever a CT scan for an eligible patient was ordered. Physicians in both the MTBI and PE arms were ranked according to their CDS use, and a stratified sampling strategy was used to randomly select 5 physicians from each of the low, medium and high CDS use tertiles in each study arm. Each physician was invited to participate in a 30-minute semi-structured interview to assess the barriers and facilitators to CDS use. Physician responses were reported using a thematic analysis. Results: A total of 202 emergency physicians were randomized to receive CDS for either MTBI or PE, triggering CDS 4561 times, and interacting with the CDS software 1936 times (42.4%). Variation in CDS use ranged from 0% to 88.9% of eligible encounters by physician. Fourteen physicians have participated in interviews to date, and data collection is ongoing. Physicians reported that CDS use was facilitated by their confidence in the evidence supporting the CDS algorithms and that it provided documentation to reduce medico-legal risk. CDS use was not impeded by concerns over missed diagnoses or patient expectations. Reported barriers to CDS use included suboptimal integration into the CPOE such as the inability to auto-populate test results, it disrupted the ordering process and was time consuming. A common concern was that CDS was implemented too late in workflow as most decision making takes place at the bedside. Physicians did not view CDS as infringing on physician autonomy, however they advised that CDS should be a passive educational option and should not automatically trigger for all physicians and eligible encounters. Conclusion: Physicians were generally supportive of CDS integration into practice, and were confident that CDS is an evidence-based way to reduce unnecessary CT studies. However, concerns were raised about the optimal integration of CDS into CPOE and workflow. Physicians also stated a preference to a passive educational approach to CDS rather than an automatic triggering mechanism requiring clinical documentation.