AI modeling of nursing workload to understand burnout

This K01 award application is for Dr. Victoria L. Tiase, a PhD-trained nurse informaticist with a commitment to improving the systems, structures, and policies that support the nursing workforce. Her overarching career goal is to become an independent nurse researcher focused on obtaining a broad understanding of nursing workload, specifically in relation to reducing nurse burnout, through the application of electronic health record (EHR) data and data science. This K01 will support three key areas of career development: (1) strengthen existing knowledge of EHR audit log data for use in research; (2) acquire advanced knowledge and skills in data science and computational modeling; and (3) transition from a mentored researcher to independent investigator. The training and research will be conducted at an institution with a strong record of providing excellent support and rich training and educational resources. The candidate’s department is committed to the success of this early-stage researcher, providing any additional resources necessary to complete the proposed career development and research aims, and ensuring ongoing protected research time. The coursework and mentoring will be overseen by a complementary team of experienced researchers and experts in these fields. Nurse burnout is persistent in the U.S. with nurses reporting concerns over their work environment, particularly those working in primary care and community clinics. Increased workload is reported as the main contributor. Although nurse burnout has been studied for decades, little has changed in the organization of clinical care, and the measurement of nursing workload is not well understood. Workload measurement has traditionally taken the form of self- report, surveys, and time and motion studies which are time-consuming, expensive, and difficult to scale. Gathering sufficient data that are reliable, reproducible, generalizable, and that represent nursing contributions within the context of work activities remains a complex, unsolved problem. Advances in informatics and electronic health record (EHR) audit log data have shown promise in measuring clinician work activities. State-of-the-art data science paradigms are needed to fully understand the complexity of nursing work activities and their relevance to workload. Thus, this study’s Specific Aims include: (1) Characterize and extract features from EHR audit log data and develop a data representation amenable to state-of-the-art data science techniques; and (2) Develop realistic and reproducible computational models of nursing EHR interactions. The proposed research is innovative because it will extend an existing untapped data source to nursing activities and will create a model to quantitatively measure workload influencers. This research lays the groundwork to test scalable interventions that mitigate nurse burnout, improve nurse wellness, reduce costs, and ultimately, improve health for all.