Is there a doctor in the house? Comparing attendings, residents, and midlevel providers to increase clinician staffing in an academic emergency department: A cost-efficiency analysis

Abstract

Study objectives: We developed a software-based model to determine the most cost-efficient way to provide incremental clinician staffing in an academic emergency department. Given the government's limits on graduate medical education funding, expansion of residency programs may need to be funded by the sponsoring institution. This study was designed to determine which combination of attending physicians working with residents or midlevel providers was most economical. Methods: A decision tree model using the Tree Age (Williamstown, MA) software package was created to compare the cost-effectiveness of different staffing configurations. It allows an emergency department (ED) administrator to vary the parameters in accordance to their specific productivities and costs and to determine the robustness of the conclusions. The productivity (patients/hour), salary, and working-hour data of different staffing configurations was determined using data from our ED, reported national attending and resident productivity data, and assumptions based on clinician experience. Salary plus benefits for attending physicians is $205,000 a year, for midlevel providers $95,000 a year, and for residents $54,000 a year. Attending physician productivity alone was assumed to be 2 patients per hour; each additional resident or midlevel providers was assumed to add smaller net productivity gains (first adds 0.75 patient/hour, second adds 0.5 patient/hour, third adds 0.25 patient/hour). Resident and midlevel providers productivity were assumed to be equivalent. Job satisfaction and quality of care were not factored into the model. Results: The base cost of an attending physician alone is approximately $68.15 per patient. The cost decreases when they are paired with 1 or 2 residents or 1 midlevel provider. When they are supervising 3 residents or 2 or more midlevel providers, the costs per patient increase (Table). The results are robust as long as attending physician productivity alone is less than 2.5 patients per hour at any physician salary range and, for midlevel provider productivities, up to 20% greater than that of residents. Resident productivity must increase by 30% to change the best model. Conclusion: According to our model, the most cost-efficient staffing configurations in an academic ED are those pairing each attending physician with residents. Substituting midlevel providers for residents in any staffing configuration increased the overall cost per patient. The model allows for the outcomes to vary according to parameters specific to other institutions.Table, abstract 56Cost efficiency results.Base Case Cost/pt, $MD Salary $180,000, $MLP Productivity+20%, $Resident Productivity+30%, $MD alone68.1559.8468.1568.15MD + 1 R64.3358.2864.3358.97MD + 2 R66.9261.8166.9254.38MD + 3 R68.8369.0073.7551.62MD + 1 MLP67.9460.3666.2567.94MD + 2 MLP73.0467.9268.8073.04MD + 3 MLP76.7777.5175.7676.77MD + 1 R + 1 MLP69.9864.8767.8964.98MD + 2 R + 1 MLP71.4871.8374.4667.01MD + 1 R + 2 MLPs74.1274.6775.1374.12 pt, Patient; MD, physician; MLP, midlevel provider; R, resident.