A recent study found clinical notes could be used to identify fatigued physicians.
The study, published July 1 in Nature, used data from 129,228 ED visits to a single academic medical center over 2010 to 2012 to train a language learning model. This dataset included patient demographics, the patients’ chief complaint and key outcomes related to an important physician decision: whether or not the patient is tested for heart attack (via stress testing or catheterization), and the outcome of testing. The model was then asked to identify physicians who were showing signs of fatigue.
The model was able to flag high-workload physicians both in the ED and in other settings with overnight shifts and high patient volumes. In cases where fatigue was flagged in notes, physician decision-making for that patient appeared to be worse. For example, the yield of testing for heart attack was 19% lower with each standard deviation increase in model-predicted fatigue.
A key identifying factor of notes written by fatigued physicians was the predictability of the next word.
“We demonstrated the potential of detecting physician fatigue using the notes that they wrote,” the study authors wrote. “Our predicted fatigue allowed us to reveal connections with decision quality. Our finding highlights the role of clinical notes not only as a medium of storing medical information but also as a window into physician decision-making.”
At the Becker's 11th Annual IT + Revenue Cycle Conference: The Future of AI & Digital Health, taking place September 14–17 in Chicago, healthcare executives and digital leaders from across the country will come together to explore how AI, interoperability, cybersecurity, and revenue cycle innovation are transforming care delivery, strengthening financial performance, and driving the next era of digital health. Apply for complimentary registration now.