Development of a "COVIDVax" Model to Estimate Risk of COVID-19 Related Deaths among Veterans for Use in Prioritizing Vaccination
There are no evidence-based strategies that offer guidance on how to best prioritize persons for COVID-19 vaccination. Explicitly prioritizing persons’ vaccination according to their risk of COVID-19 related death would minimize the number of deaths that would occur in the time it takes to vaccinate a large enough proportion of the population to achieve herd immunity. Therefore, we sought to develop a model to estimate the risk of COVID-19 related death among VA enrollees to aid vaccination prioritization. We used electronic health record data from the Corporate Data Warehouse (CDW) supplemented by the COVID-19 Shared Data Resource (CSDR) developed by VINCI to identify Veterans in VA care (7.6 million) during the study observation period (May 21, 2020 to November 2, 2020). We then developed and internally validated in different time periods a logistic regression model – called COVIDVax – to predict the risk of COVID-19 related death. In estimating the risk, COVIDVax used the following 10 patient characteristics: sex, age, race, ethnicity, body mass index (BMI), Charlson Comorbidity Index (CCI), diabetes, chronic kidney disease, congestive heart failure, and the Care Assessment Need (CAN) score. We found that:
Implications and Implementation
We demonstrated that we could ingest all necessary data streams and execute the model for all VA enrollees in real time to generate a “Dashboard” with risk scores. We also developed a web-based calculator that executes the COVIDVax model (https://COVIDVax.xyz). Although VA has already vaccinated many VA enrollees (2.26 million have received at least one shot as of April 6, 2021), our model may still be helpful to identify high-risk Veterans who remain unvaccinated and for if or when booster vaccinations are deemed necessary. Models such as ours would also be very helpful in countries or systems that are still lagging behind in vaccination.