How hospitals use algorithms to prioritize COVID-19 vaccine distribution

Hospitals and health systems across the U.S. developed plans to prioritize COVID-19 vaccine distribution based on the CDC's recommendations.

Some organizations used artificial intelligence to develop an algorithm that helped to decide who would receive the first doses. Renton, Wash.-based Providence used an algorithm to decide how to distribute the vaccine.

"We used a simple form to allow our caregivers to self-attest to their risk and role," Executive Vice President and CIO B.J. Moore told Becker's. "Based on a series of questions, we had a simple scoring system that was then used to prioritize caregivers into cohorts."

Washington, D.C.-based George Washington University Hospital also used an algorithm that took age, medical conditions and infection risk into consideration to decide when employees would receive the COVID-19 vaccine, according to The New York Times. The hospital used an employee survey to gather information for the algorithm.

However, the method used by some other organizations has come under scrutiny. Stanford Medicine in Palo Alto, Calif., used an algorithm to select the first 5,000 employees to receive the vaccine based on several factors to prioritize those at high risk. According to a Dec. 18 article in the Times, the algorithm Stanford used assigned individuals a risk score based on age, job description and number of COVID-19 cases within their department.

As a result, the initial vaccine administration left out more than a thousand medical residents as well as fellows and nurses who care for COVID-19 patients while food service and environmental service workers as well as older employees were prioritized. The residents' age coupled with the inability to assign a location to them in the algorithm put them at a disadvantage, according to ProPublica.

Hospital administrators did not review the list generated by the algorithm before beginning to administer the shots, according to the report. On Dec. 20, Stanford Medicine told Becker's it is revising the plan to better sequence vaccine distribution going forward.

North Country Healthcare in Lancaster, N.H., is not using artificial intelligence algorithms to prioritize COVID-19 vaccine distribution. Instead, the health system is relying on CDC and HHS guidance to rank employees and clinical providers by risk on a 1 to 100 scale. In Phase 1a, the health system is focused on critical workers and distributing the vaccine to those willing to receive it.

"There is a significant percentage that will not get the vaccine," said CIO Darrell Bodnar. "It is not mandatory for anyone in our organization. As we move forward, non-patient facing staff will be included in Phase 1b. There is also a priority list ranked in similar fashion. A lot of this has been managed through HR and employee health."

Charleston (W.Va.) Area Medical Center established the prioritization algorithm without artificial intelligence.

"We had a group of leaders who established tiers of associates to receive the vaccine," said CIO Daniel Stross. "The tiers were based on exposure risk. For example, an ED or ICU clinician would be head of an [information services] associate who has been working from home."

At Deborah Heart and Lung Center in Philadelphia, CIO Richard Temple said the hospital set up a vaccine schedule emphasizing front-line patient-facing workers first but staggered vaccine administration to make sure there weren't any large-scale adverse events.

"We didn't want to be in a position where we would have a large number of clinicians unavailable simultaneously due to adverse reactions," he said. "As a small hospital, we are hoping that we can stretch our dose allocation to reach as much of our hospital staff as possible, thereby protecting our employees and our patients, and [we] hope to accomplish this by the end of the year."


Copyright © 2024 Becker's Healthcare. All Rights Reserved. Privacy Policy. Cookie Policy. Linking and Reprinting Policy.


Featured Whitepapers

Featured Webinars