“It’s not a model that most of us in the infectious disease epidemiology field think is well suited” to projecting Covid-19 deaths, epidemiologist Marc Lipsitch of the Harvard T.H. Chan School of Public Health told reporters this week, referring to projections by the Institute for Health Metrics and Evaluation (IHME) at the University of Washington.
Other experts are even harsher in their criticism of the IHME model, in particular, warning against relying on it as the basis for government decision-making, including “re-opening America.”
“That the IHME model keeps changing is evidence of its lack of reliability as a predictive tool,” said epidemiologist Ruth Etzioni of the Fred Hutchinson Cancer Center. “That it is being used for policy decisions and its results interpreted wrongly is a travesty unfolding before our eyes.”
The fact that the White House is basing public-policy on the IHME model is worrisome – especially now that President Donald Trump on Thursday announced a blueprint for how and when to “re-open” the country.
The IHME model comes from the University of Washington’s Institute for Health Metrics and Evaluation (IHME), whose director is Christopher Murray. And the IHME model differs from those used by almost all other epidemiologists worldwide.
According to Forbes, the IHME model forecasts that by August 4 the outbreak will be fully contained. By then, daily death rates will have dropped to zero, and total deaths will have reached 68,000 in the United States.
WATCH: Trump on reopening the country as #coronavirus ravages the U.S. #economy:
“We are not opening all at once, but one careful step at a time…Now that we have passed the peak in new cases, we’re starting our life again, we’re starting rejuvenation of our economy again” pic.twitter.com/MXFYM3X5QP
— Face The Nation (@FaceTheNation) April 17, 2020
This same forecast implies that one month from now, the pandemic will be 96 percent contained, meaning that of the total eventual U.S. deaths from COVID-19, only 4 percent will occur after five weeks from now. So what is the confidence level in these projections?
IHME is now projecting that the eventual number of COVID-19 deaths in the U.S. will be 68,841, with a confidence interval between 30,188 and 175,963. Most of us remember the initial fatality projections – up to 240,000 U.S. deaths, compared with fewer than 70,000 now.
Is there a problem with the model?
For those of us who want to reopen the economy – the projections are supportive for doing so. The same goes for those of us who argue that the stay-at-home measures are too restrictive.
But for those of us who believe the country needs to avoid taking measures that would weaken our containment of the virus – The IHME’s forecasts are problematic and appear unrealistically optimistic.
Ali Mokdar, Chief Strategy Officer and Professor of Global Health at IHME, in a C-Span interview said their model projecting deaths is not the same as those used by most other research groups. IHME’s emphasis is on fitting the patterns of daily mortality observed in the experiences of other geographic areas such as Wuhan, Italy, and Spain.
Last Sunday, on CBS News’ “Face the Nation,” IHME director Murray, stated that the U.S. would “very clearly have a rebound” in COVID-19 cases were social distancing guidelines to be eased on May 1.
MODEL DISCREPANCY: @IHME_UW‘s Dr. Christopher Murray says the swings in death projections – from just over 90k down to 60k – over the past two weeks were a result of data reporting. WATCH –> pic.twitter.com/VweZj3GW73
— Face The Nation (@FaceTheNation) April 12, 2020
This brings to mind the State of Florida. Even as Governor DeSantis announced the reopening of some beaches on Friday, the state saw its highest one-day increase in COVID-19 cases, with 700 reported. There were 58 new deaths since Thursday evening.
Also on Friday, the IHME revised its coronavirus model for Florida significantly, forecasting that the state saw its peak of deaths two weeks ago will see 1,363 deaths from COVID-19 by Aug. 4, far lower than the 4,748 projected earlier this week. The IHME model shows the most deaths per day in the state was 77 on April 2 even though FDOH data showed only 34 people died that day.
When can social distancing be relaxed? Dates range between May 4 and early July, assuming the current level of social distancing is in place until infections reach below 1 per million, and containment strategies are implemented. pic.twitter.com/eP8Sh8Ulvd
— IHME_UW (@IHME_UW) April 19, 2020
What does IHME have to say?
Even though the IHME model may be different from other mortality models, especially when emphasizing and comparing caseloads in the U.S. to what is occurring in Wuhan, China or Italy, the IHME website explains that “As data continue to come in, our estimates may change. Specifically, new death data … have changed our projections.”
But there is one thing that is clearly written out on the IHME’s website – “It may be possible to relax social distancing, assuming strategies for testing, contact tracing, isolation, and limitations on mass gathering are in place. The timeline could change based on what data show about the trajectory of the pandemic.”
All models are wrong, but some are useful. Our director, Dr. Christopher Murray, shares on @AC360 with @drsanjaygupta that we hope ours is useful for hospitals and state authorities trying to plan for the next phase of #COVID19. pic.twitter.com/g2eLI02kIF
— IHME_UW (@IHME_UW) April 18, 2020