137

Abstract

Introduction: New resident duty hour restrictions aimed at reducing sleep deprivation have the unintended consequences of limiting exposure to high-risk, low frequency diseases and reducing closed loop learning during each case. Sepsis is a bloodstream infection that presents in relatively subtle ways, is often confused with more common conditions that are less urgent and can be deadly unless detected and treated relatively quickly. Critical decision method (CDM) is an interviewing technique that allows improved recall and description of healthcare provider diagnostic and management thought processes. Simulation-based training offers a potential strategy to increase resident exposure to critical illnesses, such as sepsis, over a range of presentations and in a safe environment. Methods: Five study team members were trained in the CDM approach. Interviewer pairs included one experienced pediatrician [from critical care (CC) or emergency medicine (EM)] and one qualitative researcher. Interviews lasted one hour during which respondents identified challenging cases of sepsis. Interviews were transcribed and then coded for identified cues and strategies surrounding recognition of sepsis and early management. Prior to coding, a card sort technique was employed to identify potential coding categories. Post-coding investigators met to reconcile variances. The coding sheet was iteratively refined and clear definitions for each category and sub-category were developed. Simulation scenarios were developed based on critical incidents, interview findings and coding results, then a seperate cohort of pediatric residents and attendings were enrolled to pilot the scenarios. Measures included an adapted version of the situation awareness global assessment technique (SAGAT) to assess each participant’s understanding of the situation at three points in each scenario and novel PALS-based and early goal-directed therapy (EGDT) checklists to evaluate assessment and management skills. Results: Twenty-three patient-based sepsis incidents were discussed in 14 interviews (8 pediatric resident physicians, 3 pediatric EM attendings, and 3 pediatric CC attendings). Coding categories included classic indicators/consensus criteria, experience-based criteria, medical history and source of information. For example, for pediatric interns in particular concerns voiced by parents were an important cue as to the seriousness of the patient’s condition whereas experienced-based criteria were articulated more commonly by experienced clinicians, including distal perfusion, mental status changes, and ill-appearance. Critical cues identified drove the design of five scenarios presenting sepsis cases (2 compensated, 2 decompensated and 1 garden path) to support the development of perceptual skills and pattern recognition in the context of a challenging incident. Eighteen physicians (6 interns, 6 third-year residents, 5 EM attendings and 1 CC attending) were enrolled in the pilot course and randomized to 3 sepsis scenarios, as well as one non-sepsis scenario. Of the 54 sepsis scenarios, participants correctly identified sepsis in 33% (interns 29%, 3rd-year residents 38%, attendings 39%) as shown by the SAGAT, including only recognizing sepsis in 8 of 18 hypotensive (decompensated) patients. The novel PALS and EGDT checklists showed moderate levels of reliability (0.534, 0.446 respectively) when applied by video-based reviewers. Conclusions: CDM interviews revealed a set of critical cues and cue clusters that were integrated into simulation-based scenarios. Piloting of scenarios to distinguish novice and expert differences in recognition of sepsis prompted further revision of scenarios to realign cues misinterpreted in these pilot simulations. Additionally, novel assessment checklists were derived, validated and shown to be reliable. Currently, a full simulation-based course has been implemented for all first-year pediatric residents to accelerate their recognition and early management of sepsis.