P75. Spine call coverage and seasonality are key variables involved in the workload placed on orthopaedic residents

Abstract

BACKGROUND CONTEXT Excessive resident duty hours (RDH) is a recognized issue with implications in both physician well-being and patient safety. A major component of the RDH concern is on-call duty. While considerable research has been put forth to address this problem of reducing resident call workload, there is a paucity of research in optimizing resident call scheduling. Currently, call coverage is scheduled arbitrarily rather than demand-based and may lead to over-scheduling coverage in order to prevent a service gap. Machine learning and forecasting has been widely applied in other industries to prevent such issues of a service supply-demand mismatch. However, it has gained little traction in health care. We believe that forecasting principles can be used to efficiently schedule residents to optimize their time while ensuring high quality patient care. PURPOSE We test a mathematical model to determine key variables driving resident scheduling and identify potential scheduling optimization. METHODS Daily handover emails of the orthopaedics department at a major academic hospital over a 12-month period were collected. The following data was extracted: 1) spine call coverage, 2) date sent, 3) number of operating rooms (ORs) completed, 4) number of traumas, 5) number of admissions, and 6) number of consults. Combined with observation of resident responsibilities over a 12-month period, the above variables were used to calculate the demand per day in hours (demand function) required of orthopaedic residents. A linear regression model was used to determine variables that significantly contribute to the demand function. Reduction capacity was measured by determining the number of “second-call” resident shifts that were not necessary, defined by days when the daily demand was less than 20 hours. RESULTS A linear regression model determined three statistically significant variables (P CONCLUSIONS Key drivers of resident demand have been identified using a linear regression model, demonstrating potential for significant optimization of resident scheduling. Further work would involve creating a forecasting model that accurately predicts future demand based on past and current demand patterns. This work has the potential to significantly impact resident well-being while maintaining high quality patient care. FDA DEVICE/DRUG STATUS This abstract does not discuss or include any applicable devices or drugs.