The objective of this study is to test the feasibility of using deliberate practice - goal-oriented training in the presence of a coach who can provide personalized, immediate feedback - to increase engagement. The research design involves recruitment of a national convenience sample of board-certified emergency physicians who will serve as trainees (n=30), pairing of the trainees with a coach, delivery of three 30-minute coaching sessions using the existing games as the training task, and assessment of the effect of the combined intervention on performance in the laboratory. The specific aims are:

1. To assess the fidelity of intervention delivery by measuring coaching skill acquisition, coaching skill drift and protocol adherence.
2. To assess the potential effect size of the intervention by comparing trainee performance on a validated virtual simulation with a control group of physicians (n=30).
3. To assess the acceptability of the intervention by using a mixture of validated instruments and semi-structured debriefing interviews with trainees to assess their engagement with the intervention.

Deliberate practice - goal-oriented training in the presence of a coach who can provide personalized, immediate feedback - has successfully improved performance across multiple domains, including sports, music, and combat. When used in conjunction with simulation to improve surgical skill, it has a large effect on educational outcomes. It has characteristics that make its application in this context potentially powerful (e.g. personalized feedback/relationship with coach increase engagement) but also potentially challenging (e.g. the diagnostic process does not lend itself easily to assessment). The objective of this study is to test the feasibility of using deliberate practice to amplify the effect of a video game intervention. The team will recruit a national sample of board-certified emergency physicians (n=30) to serve as trainees, with members of the team (n=3) serving as coaches. Trainee-coach dyads will meet for 30 minutes/week for 3 weeks, by video-conferencing, to play one of the existing video games and to use it to practice pattern recognition. Aims are:

1. To assess the fidelity of intervention delivery. Approach: the team will standardize coaching skill during an 'on-boarding session,' measure skill drift over the course of training sessions, and measure protocol adherence (primary outcome). Hypothesis: >90% of dyads will complete three training sessions.
2. To assess the potential effect size of the intervention. Approach: the team will compare performance of trainees (n=30) with a control group of physicians (n=30) on a validated virtual simulation. Hypothesis: Trainees will make ≥25% fewer diagnostic errors than control physicians (large effect size).
3. To assess the acceptability of the intervention. Approach: the team will conduct semi-structured debriefing interviews with trainees, assessing elements of the intervention that promote engagement.

This proposal will inform a future Stage III trial to compare the effect of different interventions on diagnostic error in trauma triage. If successful, this program of research will have an impact on patients by reducing the burden imposed by injury and by addressing the refractory problem of diagnostic error. It is novel conceptually in its effort to make heuristics a source of power, methodologically in its use of deliberate practice to improve diagnosis, and translationally in its use of video game technology. It is feasible because the investigative team has clinical and behavioral science expertise, experience developing deliberate practice interventions, and a track record of successfully building video games that can transform physician behavior. It responds to two national research priorities: 1) improving the diagnostic process; 2) maintaining health and independent living among the aging.

Inclusion Criteria:

Board certified physicians who work at non-trauma centers in the United States

Exclusion Criteria:

Physicians without board certification (i.e., residents). Physicians who work only at trauma centers in the US.

De-identified participant data will be shared with researchers conditional on approval from the University of Pittsburgh Office of Research

Data will be made available at the conclusion of the analytic phase of the trial, and for 3 years thereafter.

Access will be provided to researchers and interested parties.

3 months

First, we used a passive rather than an active control as the comparator, which may have magnified the effect of the intervention. Second, the differential rate of attrition for completing the virtual simulation in the control and intervention groups could introduce bias, if missingness was not random. Third, we used a simulation to assess behavior, albeit one with evidence of criterion validity, which limited inference about the efficacy of the intervention.

Allocation to treatment v. control group will be masked to the person doing the analysis until the data is locked.