June 21, 2021

Prediction Tool Shortcomings Highlighted in JAMA Internal Medicine Attracts Media Coverage

Singh along with other DLHS faculty co-author study on accuracy in sepsis prediction tool by Epic

In a new paper published in JAMA Internal Medicine, Karandeep Singh along with colleagues, including DLHSers John Donnelly and Andrew Krumm, describe a study they conducted to evaluate a tool developed by Epic Systems to predict sepsis.  Sepsis is a dangerous, and too often deadly, response to infection in the body. An article in Michigan Medicine's Health Lab Blog Popular sepsis prediction tool less accurate than claimed interviews Singh and describes the findings reported in JAMA which show Epic's method for identifying sepsis is less effective than the company claims.

Because sepsis is responsible for one in three hospital deaths the study has drawn the attention of WIRED magazine with an article entitled An Algorithm That Predicts Deadly Infections Is Often Flawed and health and medicine outlet STAT in their article A popular algorithm to predict sepsis misses most cases and sends frequent false alarms, study finds.

In Health Lab article Singh describes Epic's flawed methodology in their algorithm design, “In essence, they developed the model to predict sepsis that was recognized by clinicians at the time it was recognized by clinicians. However, we know that clinicians miss sepsis.”

The following is a complete List of the Authors:  Andrew Wong, MD; Erkin Otles, MEng; John P. Donnelly, PhD; Andrew Krumm, PhD; Jeffrey McCullough, PhD; Olivia DeTroyer-Cooley, BSE; Justin Pestrue, MEcon; Marie Phillips, BA; Judy Konye, MSN, RN; Carleen Penoza, MHSA, RN; Muhammad Ghous, MBBS; Karandeep Singh, MD, MMSc1,4

Karandeep Singh picture

Karandeep Singh, MD, MMSc

Assistant Professor of Learning Health Sciences
Assistant Professor of Internal Medicine
Assistant Professor of Urology
Assistant Professor of Information