However, the administration’s dependence on its projections has nevertheless frustrated a large part of the public health community, which warns that the IHME has not respected traditional disease modeling procedures or incorporated crucial variables. The result is a more rosy picture of the crisis than that presented by much of the rest of the modeling world.
“The IHME model is a strange duck in the math model pool,” said Gregg Gonsalves, epidemiologist at the Yale School of Medicine. “I’m afraid the White House is looking for data that tells them a story they want to hear, and so they turn to the model with the lowest mortality projection.”
At the center of these concerns is a key element, say critics of the IHME model. The projection does not try to take into account the determining characteristics of the virus, such as its ease of spread or the time during which a person can be infected before presenting symptoms.
Instead, it draws on data from cities already affected by the coronavirus, including Italy and China, and maps the United States to a similar curve. The result is a projection that is easily digestible and more accurate in forecasting than most models of infectious disease, but much more volatile as the situation unfolds on the ground.
The IHME has frequently revised its model in the past month. Since April 9, for example, his forecast for the death toll in the country has risen from around 61,000 to almost 70,000, before returning to around 67,000.
Each of its projections also includes an upper limit and a lower limit, bounded by a shaded area, which ranges from 48,000 – a figure the U.S. has already exceeded – to 123,000 deaths.
“It is a statistical model adapting the curves of the epidemics in China and elsewhere to what they think might happen in the United States,” said Gonsalves, “then continually rearranged based on news. data.”
IHME Director Christopher Murray Defended also rigorous work and among the best models available, arguing that forecasts are simply aimed at achieving objectives different from more traditional projections. The model was originally intended to help hospitals forecast their supply needs as suppliers around the world were preparing for a wave of coronavirus patients.
“We are ready to forecast. Most academics want to cover their bets and never go wrong,” said Murray. “It is not useful for a planner – you cannot go to the hospital and say that you may need 1,000 ventilators, or that you may need 5,000.”
He added that the IHME model is much more optimistic than the others, largely because it strongly explained the impact of social distancing – a decision that Murray credits for helping to locate the national peak of pandemic even as others warned of massive and continuous growth of cases.
“We are more optimistic orders of magnitude. On the other hand, we have also called the peak correctly,” he said. “We believe in fitting models to data, without making assumptions and then say how my hypothesis would materialize in a hypothetical world. ”
This caught the attention of the 2-4 million people – including many public health officials and hospital administrators – who visit the site daily. He also won the trust of the Trump administration, which first contacted IHME in late March while fighting to allocate limited supplies and prevent an overwhelming health care system, and continued to trade data and observations with the group since.
In early April, IHME became more confident that the country’s brutal foreclosure had started to work, as did senior public health officials.
Shortly after IHME released its forecast of 60,000 deaths, Fauci echoed on April 8, saying that the administration now thought the toll would be “closer to 60,000 than 100,000 to 200,000” deaths of health officials previously estimated.
But the White House coronavirus task force has, in recent conversations with the group, focused its attention on a new challenge: how to handle a gradual reopening of the country, representing a new phase that will be much more difficult to model.
“This opens up a whole new set of challenges,” said Murray, noting that Georgia – whose governor Brian Kemp called for business reopening much earlier than its counterparts in other states – has not even not reached its peak coronavirus under the IHME model.
“If Georgia is going to experience a resurgence, what about neighboring states?” Said Murray.
These concerns are compounded by what he has called a “worrying” trend of slower dropouts in new cases in some countries like Italy, a sign that the crisis may persist longer than expected.