This study develops and validates a model to identify health clinics at high-risk for clinician burnout, which could inform practice improvement efforts to prevent burnout before it occurs.
Study Overview
Problem: Over 500,000 physicians in the United States experience symptoms of burnout at any given time. In addition to the impacts on affected clinicians, burnout also negatively affects quality of care and patient safety. Unfortunately, there is currently no reliable method to proactively identify high-risk clinics and intervene before burnout occurs.
Main Objective: To develop and validate a prediction model that uses existing operational data to identify clinics at high-risk for clinician burnout.
Approach: Researchers are creating a database with practice-specific metrics from Stanford primary care clinics, including electronic health record (EHR) usage measures. Using a machine learning approach, the research team will use the data to develop a model to quantify the risk for clinician burnout and identify clinics at high risk. The research team will then conduct a qualitative assessment to refine the prediction model and then seek to demonstrate the model’s predictive validity at the clinic level based on routinely administered surveys measuring clinician burnout.
Results: To date, researchers have identified two practice-specific operational domains that relate to wellbeing and risk of burnout that could be quantified in a machine learning model alongside EHR measures. The first factor is the effect of people leaders—when leaders are available at expected times, and feedback is effective, predictable, and respectful, employees are more likely to feel a sense of psychological safety and are less likely to feel emotional exhaustion and burnout.1 Second, healthcare workers who are more likely to be at high-risk for burnout include women physicians, midcareer physicians, healthcare workers with adult children, and healthcare workers working less than full-time.2 Additional study findings are forthcoming. Current and future publications can be found here.
Primary Care Relevance
This project has the potential to develop and implement a model that can predict which primary care clinics are at high-risk of burnout before it negatively affects clinicians and patients. This will allow burnout interventions to be tailored to the specific needs of the clinics. Additionally, with a better understanding of which clinic factors impact burnout, leaders can work to proactively make system-level changes to support clinician well-being.
AHRQ Primary Care Priority Area
Research to improve primary care, including regarding quality, access and affordability, the workforce, care delivery models, financing, digital healthcare, person-centeredness, and health equity.
Notes
1. Tawfik, D. S., Adair, K. C., Palassof, S., Sexton, J. B., Levoy, E., Frankel, A., Leonard, M., Proulx, J., & Profit, J. (2023). Leadership behavior associations with domains of safety culture, engagement, and health care worker well-being. Joint Commission Journal on Quality and Patient Safety 49(3), 156–165. https://doi.org/10.1016/j.jcjq.2022.12.006.
2. Tawfik, D. S., Shanafelt, T. D., Dyrbye, L. N., Sinsky, C. A., West, C. P., Davis, A. S., Su, F., Adair, K. C., Trockel, M. T., Profit, J., & Sexton, J. B. (2021). Personal and professional factors associated with work-life integration among US physicians. JAMA Network Open 4(5), e2111575. https://doi.org/10.1001/jamanetworkopen.2021.11575.
