P4.012: Who's Working Christmas: Using Principles of Game Theory for Holiday Scheduling

Background: We will share a novel auction-based system incorporating principles of economic game theory to allocate holiday shifts for a pediatric hospitalist group that provides 24/7 in-house coverage at two freestanding children’s hospitals. The system fits within one year and avoids the need for year-to-year promises, it is fully transparent and algorithmic, and it provides a near-perfect indication of each person’s specific preferences weighted against the preferences of the rest of the group. This system has resulted in a substantial improvement in physician satisfaction regarding holiday coverage.

Methods: The surge census during the pandemic put significant pressures on our staffing which led to physician burnout, and we found that fair and equitable allocation of holidays was a crucial component in ensuring wellness of our physicians. We developed a live-auction-based holiday system to allow for holiday allocation based on an algorithmic approach that would indicate the group’s weighted preferences. Each participant was given a number of points to allocate across the holidays as they wanted, and we developed a shared spreadsheet that would allow adjustments to point allocations to determine each person’s rank for each holiday. This rank would then determine one’s likelihood of being off for each holiday, and the spreadsheet would display updated ranks in real-time as participants changed their points. By running this system for five to seven days, we were able to allow for nearly infinite adjustments. This would allow the group to achieve a steady state at which point very few further adjustments were requested, therefore perfectly reflecting each person’s exact preferences and under their full control. Physicians would be scheduled on each holiday lowest rank to highest rank, and we established a color-coding system to determine likelihood of being off on a holiday. We separated our ranks into four color categories. Our “red” ranks would work the holidays. “Orange” ranks would have a fair chance of being off, “green” ranks had a high chance of being off, and “blue” ranks were essentially guaranteed to be off.

Results: Our first iteration running this system, we had a 97.2% approval by participants. We were able to successfully grant the entire holiday stretch off to 100% of “blue” ranks, 91.5% of “green” ranks, and 35.1% of “orange” ranks.

Conclusion: Our system resulted in an almost perfectly fair and equitable allocation of holidays that was both intuitive to understand and fully transparent. We believe this could serve as a model for shift allocation for other groups that have to work holidays, and we believe this system could result in a notable improvement in physician wellness and would be of particular interest to leaders of groups that provide around-the-clock holiday coverage who are struggling with similar issues.