State leaders are relying on a hodgepodge of statistical models with wide-ranging numbers to guide their paths through the deadly coronavirus emergency and make critical decisions, such as shutting down businesses and filling their inventory of medical supplies.
During hurricane season, coastal states can trust the same set of computer models to warn of a storm’s track. During this pandemic, there is no uniform consensus to predict the toll and direction of the virus that is tearing through communities around the country.
With little agreed-upon information, governors and local officials are basically creating do-it-yourself sources of information with their own officials and universities.
The models have resulted in conflict in several locations.
About 75 protesters on Thursday called on Ohio Gov. Mike DeWine to reopen businesses and questioned the models used by his health director to continue the state’s shelter-at-home order. Critics have denounced an Iowa Health Department a matrix as arbitrarily devised and being used by the governor to rationalize her decision to not issue a stay-at-home order.
The federal government and many states rely on a University of Washington model that’s the closest thing to a benchmark but it is so imprecise that the latest projection for the death toll had a range of more than 100,000. In Washington D.C., health officials took the unusual step of publicly announcing that they didn’t trust the University of Washington’s updated model and embraced far more pessimistic predictions from a model created by researchers at the University of Pennsylvania.
Fine-tuning models
Some states, including Alaska, Ohio, Pennsylvania, Minnesota, Louisiana, are incorporating the work of local researchers and other experts to fine-tune their models.
Some elected officials have cited the most dire forecasts in issuing stay-at-home orders. Others have seized on more optimistic figures from their models to justify their calls to loosen restrictions.
“We know now that a lot of the models out there are not accurate,” Missouri Gov. Mike Parso said in describing why he waited until Monday to issue his stay-at-home order.
President Donald Trump’s handling of the pandemic could well decide whether he is re-elected, and the models the White House relies on for forecasting will be an issue if they miss the mark.
Unlike with the National Hurricane Center, the federal government doesn’t have a national clearinghouse for virus models. The Centers for Disease Control and Prevention hasn’t publicly released any coronavirus models of its own. The Trump administration favors the University of Washington model, but the CDC hasn’t identified a modeling consensus for states to follow.
Dr. Georges Benjamin, executive director of the American Public Health Association, said some public officials tend to act according to what “politically plays the best” instead of “following the science.”
