The College Fix

Number has gone down by over a third

The university-produced coronavirus model cited repeatedly by White House officials, one which has consistently and significantly overestimated the number of hospitalizations that would result from the disease, is still wildly over-projecting the number of Americans currently hospitalized with COVID-19 even after being substantially revised over the weekend.

The “Murray model,” as it’s popularly known—a project headed by Christopher Murray out of the University of Washington—has been touted by White House medical officials such as Deborah Birx as a significant contributor to the federal government’s ongoing “social distancing” guidelines. The model initially predicted a grimly dire forecast for the United States, in which peak resource use around mid-April would see over 220,000 Americans hospitalized with the disease and a shortage of over 54,000 hospital beds.

The efficacy of that model was called into serious question over the past week. A review of reported COVID-19 statistics in America compared with the model’s predictions showed that it was actively overestimating the number of hospitalized Americans by as much as a factor of four. On Apr. 2, for instance, it predicted around 123,000 Americans in the hospital due to the disease. The reported number on that day was about 31,000.

Over the weekend the researchers issued a significant downward revision of their estimates, now claiming that at its peak the U.S. will need around 141,000 hospital beds with a shortage of 36,654. Yet even within 24 hours the model was already significantly diverging from state-level coronavirus reporting.

The new model estimated on Apr. 5, for instance, that there would be over 90,000 Americans hospitalized due to that disease, yet the COVID Tracking Project, which updates nationwide coronavirus infection, hospitalization and death totals on a daily basis using state-reported data, lists only 23,000 currently hospitalized nationwide.

The flawed projections were notable at the state level, too. The model projected around 24,000 hospitalizations in New York yesterday, while that number was actually closer to 16,500, according to the New York governor’s office.

Here is the newly updated mIHME_UW model with predictions for NY state today (which doesn’t need to be predicted, as it’s already happened). And here’s the actual data from today. As you can see, the NEW model can’t even get today right. pic.twitter.com/4y2rat5jcz

— Alex Berenson (@AlexBerenson) April 6, 2020

The model’s revised death projections have been roughly unchanged: It estimates around 82,000 coronavirus deaths in the country through July, where before it had predicted 84,000. It is unclear why the modelers’ death estimates remained relatively static even as their hospitalization estimates dropped by well over a third.

MORE: University projections of COVID-19 hospitalizations turn out to be four times too high

IMAGE: Velli Sky / Shutterstock.com

Like The College Fix on Facebook / Follow us on Twitter