Deep learning algorithm predicts patient mortality

Twitter icon
Facebook icon
LinkedIn icon
e-mail icon
Google icon
 -  Algorithms show link between rads’ gaze, diagnostic decision and image content

Researchers from Stanford University have developed an algorithm capable of predicting patients’ three to 12-month mortality rate, according to a study published in arXiv.

Oftentimes, patients needing palliative or end-of-life care miss their chance to be home during their final days due to the overestimated prognoses of physicians. In this study, researchers evaluate the feasibility of using a deep learning algorithm to predict patient mortality in order to improve the quality of end-of-life care.

Researchers developed the Deep Neural Network (DNN) using deep learning and data from electronic health records. The DNN algorithm was trained using EHR data from previous years to predict all-cause, three to 12-month mortality of patients.

Patients admitted to the hospital were automatically evaluated by the algorithm, which was then able to identify patients who were more likely to require palliative care services.

“We demonstrate that routinely collected EHR data can be used to create a system that prioritizes patients for follow up for palliative care,” concluded first author Anand Avati and colleagues. “Our predictions enable the Palliative Care team to take a proactive approach in reaching out to such patients, rather than relying on referrals from treating physicians, or conduct time consuming chart reviews of all patients.”