New algorithm helps predict the chance of patient death within a year

Researchers at Stanford University developed a neural network that can examine patient records and estimate a patient's chance of mortality in the next three to 12 months, according to a paper published in arXiv.

Physicians can use the algorithm to better identify patients who have the greatest need for palliative care, the authors found. The algorithm trains on EHR data from previous years and then analyzes a patient's records to generate a prediction about the patient's chance of mortality.

The network also creates a report for physicians to understand how it reached its prediction, which includes details on how factors such as number of days in the hospital, medications and severity of the diagnosis played into its prediction.

"In our preliminary analysis, we find that it is possible to create a model for all-cause mortality prediction and use that outcome as a proxy for the need of a palliative care consultation," the authors concluded. "The resulting model is currently being piloted for daily, proactive outreach to newly admitted patients." The authors did not disclose the pilot program's location. 

More articles on data analytics & precision medicine:
NIH awards National Library of Medicine $4.5M to engage public libraries in All of Us Research Program
DNA sequencing company offers customers $1M guarantee on ransomware protection
U of Delaware, George Washington U team up with NASA to advance precision oncology research

© Copyright ASC COMMUNICATIONS 2017. Interested in LINKING to or REPRINTING this content? View our policies by clicking here.

 

Top 40 Articles from the Past 6 Months