Researchers Combine AI with Data to Predict Caregiver Burnout and Improve Patient Care

Combining Artificial intelligence with data may help improve patient care and outcomes and predict burnout for caregivers before it even arises, according to a paper presented at this year’s ACM SIGKDD Conference on Knowledge Discovery and Data Mining.

The research led by Chenyang Lu, the Fullgraf Professor of computer science and engineering at the Washington University in St. Louis McKelvey School of Engineering, presents the first end-to-end deep learning framework for predicting physician burnout based on electronic health record (EHR) activity logs and digital traces of physician work activities that are available in any EHR system. 

The researchers collected data from eighty-eight interns and resident physicians in Internal Medicine, Pediatrics, and Anesthesiology at the Washington University School of Medicine, BJC HealthCare, and St Louis Children’s Hospital from September 2020 through April 2021. 

Unlike other studies that rely exclusively on surveys for burnout measurement, the framework of Lu and colleagues called the Hierarchical burnout Prediction based on Activity Logs (HiPAL) directly learns deep representations of physician behaviors from large-scale clinician activity logs to predict burnout. 

Furthermore, the researchers could predict who would require more time in surgery and who was more likely to experience delirium following surgery using unique algorithms developed by the Lu lab. Compared to existing approaches, the model was able to transform hundreds of clinical variables into just ten, which it then used to predict outcomes with greater accuracy and interpretability.

“The experiment on over 15 million real-world physician activity logs collected from a large academic medical center shows the advantages of our proposed framework in the predictive performance of physician burnout and training efficiency over the state-of-the-art approaches,” the authors concluded.

 

Source: News Medical net

Journal Source: ACM