Report incident 🔥 | Improve page 💁 | Access database 🔢
An algorithm to predict whether or not patients with infections have contracted sepsis has been discovered to have missed about two-thirds of actual cases, rarely found cases medical staff did not notice, and frequently issued false alarms.
Electronic health record company Epic Systems' Epic Sepsis Model is used by hundreds of hospitals across the US and is marketed as being 76 percent accurate.
Website: Epic Sepsis Model
Released: 2016
Developer: Epic Systems
Sector: Health
Purpose: Predict sepsis infection
Type: Prediction algorithm
Technique: Machine learning
Epic Systems (2021). For Clinicians, by Clinicians: Our Take on Predictive Models
The Epic Sepsis Model has been criticised for its limited predictive power, with studies indicating it fails to detect sepsis in approximately two-thirds of patients.
August 2023. Atrium Health research study found that Epic’s product was more accurate at the highest scoring thresholds, when it was most confident that a patient had sepsis. However, it also found that those thresholds were only reached after clinicians had already taken steps to treat the condition.
February 2022. Stat published the findings of a research study conducted with the Massachusetts Institute of Technology that small shifts in data fed into well-known health care algorithms, including the Epic Sepsis Model can cause their accuracy to degrade over time. Instead of transforming care, the study found, the algorithms are unable to keep pace with fast-moving clinical conditions, potentially resulting in mis-diagnoses and raising the prospect AI could do more harm than good. Epic Systems later said it had overhauled its sepsis prediction model to improve its accuracy and make its alerts more meaningful to clinicians. It also said it had changed its definition of sepsis to match the international consensus definition.
June 2021. Study: Epic sepsis model misses two-thirds of cases
Schertz A., et al (2023). Sepsis Prediction Model for Determining Sepsis vs SIRS, qSOFA, and SOFA
Yang J., et al (2022). STAT and MIT rooted out the weaknesses in health care algorithms. Here’s how we did it
Wong A., et al (2021). External Validation of a Widely Implemented Proprietary Sepsis Prediction Model in Hospitalized Patients
Habib A.R., et al (2021) The Epic Sepsis Model Falls Short—The Importance of External Validation
https://www.statnews.com/2021/06/21/epic-sepsis-prediction-tool/
https://www.wired.com/story/algorithm-predicts-deadly-infections-often-flawed/
https://www.theverge.com/2021/6/22/22545044/algorithm-hospital-sepsis-epic-prediction
https://www.healthcareitnews.com/news/research-suggests-epic-sepsis-model-lacking-predictive-power
https://www.statnews.com/2022/02/28/sepsis-hospital-algorithms-data-shift/
https://khn.org/morning-breakout/warnings-over-falling-accuracy-of-health-care-algorithms/
https://www.statnews.com/2022/10/03/epic-sepsis-algorithm-revamp-training/
https://www.statnews.com/2022/10/24/epic-overhaul-of-a-flawed-algorithm/
https://www.beckershospitalreview.com/ehrs/epic-overhauls-sepsis-algorithm.html
Page info
Type: System
Published: February 2022
Last updated: December 2024