Red State – by Elizabeth Vaughn
Sean Davis, co-founder of The Federalist, took a serious look at the IHME (Institute for Health Metrics and Evaluation) Coronavirus model and called it “garbage.” Davis makes his case through a series of tweets (below).
Davis found that the actual numbers of hospitalizations on April 1 were a fraction of the numbers projected by the IHME model. The problem with the model, he discovered, is that it uses New York and New Jersey data and applies it to the rest of the states.
The actual numbers range from a low of 9% in Tennessee of the projection to a high of 50% in Virginia. Here are some examples:
The model projected that over 121,000 people would be hospitalized in the country yesterday. The actual number was 31,142. (25.7%)
Texas: Projection – 1,716; Actual – 196 (11.4%)
Georgia: Projection – 2,777; Actual – 952 (34.3%)
Virginia: Projection – 607; Actual – 305 (50%)
Tennessee: Projection – 2,214; Actual number – 200 (9%)
New York: Projection – 50,962; Actual number – 18,368 (36%)
Davis writes that if we’re going to shut down the entire nation’s economy to “flatten the curve” based on the projections of a single model, it shouldn’t be too much to ask that the model approximate reality when it comes to hospitalizations.
Let's look at some state data. The IMHE model predicted that 1,716 people in Texas would have been hospitalized yesterday due to the Wuhan coronavirus. The actual number of hospitalized Texans? 196 people. pic.twitter.com/1YCLKP6TJb
— Sean Davis (@seanmdav) April 2, 2020
In Virginia, the IMHE model predicted that 607 Virginians would have been hospitalized as of yesterday due to the Wuhan coronavirus. The actual number? 305. pic.twitter.com/9iN1TrUSPk
— Sean Davis (@seanmdav) April 2, 2020
In New York, the IMHE model predicted that as of yesterday, 50,962 people would have been hospitalized as of yesterday due to the Wuhan coronavirus. The actual number? 18,368. pic.twitter.com/9aVsVogUQ5
— Sean Davis (@seanmdav) April 2, 2020
All projection data can be found here.
All historical hospitalization data can be found here.