Coronavirus: Decreased death of Big Covid-19 in new US model

At a sober press conference at the White House last week, members of President Donald Trump’s Corona Virus Task Force released data supporting the need to continue national efforts to limit the spread of the virus.

While maintaining policies aimed at limiting human-to-human contact, the government said that 100,000 to 240,000 Americans would die as a result of Covid-19, a disease caused by the virus. I predicted.

One slide using data from the Health Metrics and Evaluation Institute at the University of Washington shows the predicted peak daily mortality from disease arriving in mid-April.

read more:
* Coronavirus: Kiwi’s popular question about lockdown, test answer
* Coronavirus: US “wasteful” months before preparing for pandemic
* Coronavirus: What Scientists Know About How It Spreads
* Coronavirus: young adults are the majority of hospitalizations in the US and Europe, data shows

Anthony Fauci, a key member of the Task Force, made a key point about these predictions yesterday.

“The models are just as good as the assumptions you put into them, and if we get more data, you may put it in, it may change,” he said.

That point was quickly reinforced, and IHME estimates shifted upwards shortly after the White House briefing.

However, late Tuesday night (hours local time), IHME estimates shifted in the opposite direction.

The model predicted about 94,000 deaths by the end of summer last week, but its new estimates estimate that 60,400 people were affected by August, a 26% decrease from previous estimates of the model .

In addition, the days with the highest expected deaths shifted from more than 3,100 deaths on April 16 to 2,200 deaths on April 12.

This is clearly good news, but there are some caveats.

First, the accuracy of the IHME model has already been suspected by virus-fighting authorities.

For example, in Colorado, IHME estimates that the state’s daily mortality peak has already passed. This is a claim that state public health authorities believe they are inaccurate.

“I am convinced that the peak has not been reached,” a Colorado Public Health and Environment Administration director told reporters this week. Denver post.

of Washington Post Washington DC reported last week that it relied on another model that provided a completely different picture of how the virus affected the district.

At that time, IHME had identified the number of dead peaking in mid-April. The DC model developed by Penn Medicine estimates that the peak will come in late June.

This variance is a function of the difficulty we modeled last month, modeling a pandemic for voting site FiveThirtyEight.

However, the model should become more accurate as the actual peaks get closer. However, identifying when a peak has arrived is itself a tricky predictive model.

In part, this is because the United States is likely not to see one peak. The series of peaks are geographically and temporally separated.

In New York, the epicenter of the Covid-19 infection, there are encouraging signs that the number of new cases is falling despite the peak daily deaths. (This is to be expected, experts have repeatedly reminded Americans: The delay between infection and death means that the number of deaths follows infection rates.)

However, state data still provides unclear information.

William Hanage, an epidemiologist at Harvard TH Chang Public Health School, said there was evidence that outbreaks in New York were slowing in response to last month’s aggressive efforts to stop the spread of the virus. Stated.

For example, a drop in new hospitalization rates is evidence for that.

“But this does not mean an” inflection point. ” Because they are still up, “Hanage said in an email. “It’s not entirely clear how long it will take to turn it around.”

More importantly, even after New York’s peak has passed, the worst effects of the virus are beginning to appear elsewhere.

“The pandemic just wears boots elsewhere,” Hanage said. He added that he remains “incredibly insecure” about smaller communities throughout the country, where viruses can be caught and cause new hotspots.

“Because they lack the resources to deal with and are probably not tested until it’s too late, they can be hit as hard as urban areas,” he said.

Marynia Kolak, Assistant Director of Medical Informatics at the Center for Spatial Data Science at the University of Chicago, a member of a team studying county-level data on outbreaks, agreed that local outbreaks remained a problem .

“I hope the situation gets better quickly, but all the data we have suggests that we are starting to approach the peak of some parts of the country,” said Colac.

“It’s all happening regionally,” she added. “Many parts of the country have just begun to expand as hotspots in the past week or two.”

The IHME model attempts to estimate the peaks for each state, but there is another important constraint to keep in mind.

Coronaviruses can retreat in summer and recur in autumn. Same as the pattern seen with seasonal flu.

Also, if federal and local leaders begin to relax regulations without any other intervention, it could lead to a recurrence of the case.

The IHME estimate of 60,000 deaths is through August, and if the virus continues to spread without effective treatments and vaccines, it excludes those that could occur in autumn and winter .

The model also relies on social distance initiatives encouraged by the White House staying in place until May.

This assumes existing efforts to stop the spread of the virus. Relaxing these efforts means greater spread and more deaths.

Perhaps the most important thing to remember about migrating the IHME model is that Fauci warned last week. The model depends on the data, and as the data changes, the model changes.

A downward shift in estimated deaths reflects a positive change in the status of the data reporting on the effects of the virus, while a shift in the opposite direction means that the model will soon disappoint. .

“It’s the worst scenario for people to slow down now,” Colac said of the need to maintain current containment efforts.

“Rather, leaders should assure those who are doing the right thing that they are doing this for the community.”

“In public health, we want everyone to think we’ve overreacted a few weeks from now,” she added. “We don’t want to start preparing a lot of tombs.”