Purpose: To develop a software-based model to determine which combination of attendings working with/without residents and/or midlevel providers (MLP) was most cost-efficient for incremental staffing in an academic emergency department (ED).
Method: A decision tree model using standard decision analysis software was created to compare different staffing configurations for the Rhode Island Hospital ED. The productivity, salary, and working hour data of different staffing configurations were determined using data from the ED, reported productivity data, and assumptions based on the authors’ experience. Attending physician productivity alone was assumed to be 2.1 patients per hour, and each additional resident and/or MLP was assumed to add smaller net productivity gains (the first one adds 0.75 patients/hour; the second, 0.5 patients/hour; the third, 0.33 patients/hour). Resident and MLP productivity were assumed to be equivalent in the base case and varied during subsequent analysis. Noneconomic variables were not included in the model.
Results: The lowest base case cost option is to pair one attending with one resident; all other approaches are more expensive. The difference between most options is less than $5/patient. Only at extremes of variables do overall choices differ.
Conclusions: Incrementally staffing an academic ED with a ratio of one attending per resident achieves the lowest cost, but other models are minimally more expensive. The model allows an ED administrator to determine the costs of different models.