Maccabi, Medial EarlySign develop algorithm to identify high-risk COVID-19 cases
System from leading Israeli HMO and machine learning startup assesses patients based on medical history to guide treatment decisions
Luke Tress is The Times of Israel's New York correspondent.

One of Israel’s leading health care providers on Wednesday announced an artificial intelligence algorithm that it says identifies individuals most at risk for severe COVID-19 complications.
Maccabi Healthcare Services, Israel’s largest HMO with 2.4 million members, said the AI-powered system assesses people based on preexisting conditions and other health variables.
The program analyzes Maccabi members’ anonymized electronic health records, the company said. It was developed by Israeli tech firm Medial EarlySign, which specializes in machine learning analysis of medical data, and the Kahn-Sagol-Maccabi Research and Innovation Institute, which makes the HMO’s medical data accessible to researchers.
The algorithm identified some two percent of Maccabi’s members as potentially high-risk patients.
The algorithm will help Maccabi battle the virus more effectively and allow for fast-tracked testing procedures to cut down on the number of severe COVID-19 cases, Ran Sa’ar, Maccabi’s CEO, said in a statement.
When an individual marked by the algorithm as high-risk for complications contacts medical staff to report possible COVID-19 symptoms, the system will notify the nurse or doctor who receives the notice. They will then forward the individual for immediate testing for the virus.
At-risk individuals will thus receive medical treatment more quickly if they test positive.
Some of the variables considered by the algorithm include: age, respiratory disease history, hospital admission history, weight and body mass index, prescribed medications, tobacco use, and conditions related to immunosuppression.
It classifies high-risk patients into three levels of risk. Maccabi personnel will factor the rankings into their treatment decisions, including whether they are hospitalized, and how often to conduct follow-ups.
Medial EarlySign is in talks with US medical providers that are interested in the algorithm for their COVID-19 response.