Coronavirus predictive modeling more dire than actual results

N.C. Department of Health and Human Services Mandy Cohen briefs media at the Emergency Operations Center in Raleigh. Photo via Robert Clark, North State Journal

RALEIGH – On Monday’s COVID-19 media availability call-in, N.C. Department of Health and Human Services Secretary Mandy Cohen said the team that put together the April 6 preliminary estimate of North Carolina’s hospital capacity predictive model has not provided a second update.

When asked on the call if the group had shared an updated policy brief, Cohen said the group had not done so.

The team, at the request of Cohen, included experts from Blue Cross and Blue Shield of NC, Duke University, NoviSci, RTI International and the University of North Carolina at Chapel Hill. Their stated goal was to develop North Carolina-specific data to help guide the state’s response to COVID-19.

“I know they have continued to do their work, obviously things are changing so rapidly, as are those models, so I think they have been fine-tuning their work,” said Cohen.

In the group’s policy brief, the models’ composite estimates indicate that approximately 750,000 people in North Carolina may be infected with the virus by the end of our forecast period, June 1, 2020. Their model assumes social distancing policies are fully lifted at the end of April and are not replaced by other policies with equal assumed effectiveness to reduce transmission. The group adds that if those policies (or some other policies with similar effect) remain in place, the estimates indicate that nearly 250,000 North Carolinians may be infected by June 1.

NCDHHS currently shows that with 79,484 lab tests completed, the number of confirmed cases of COVID-19 in the state are 6,764. Estimates of those who carried the virus are higher, which some are hopeful that antibody tests will show the true number of how prevalent the virus has been in the general population.

The model used by many sources across the country has been the University of Washington’s Institute for Health Metrics and Evaluation (IHME) COVID-19 projections.

The IHME has updated its model six times in the month of April and covers all 50 states. One of the most-used data items from the IHME is the estimated peak resource use, which encompasses hospital beds, ICU beds, and invasive ventilators needed. The April 17 model, which is the most recently released, showed that April 15 was the peak of North Carolina’s expected resource use.

The daily update on April 20 provided by NCDHHS shows that with 77% of hospitals reporting, 28% of ICU beds are empty and 43% of hospital beds are empty. This data, though, is incomplete because it is unclear which hospitals have not reported and if the same hospitals are not reporting daily figures.

In hard-hit Mecklenburg County, where COVID-19 cases have been twice as high as any other in the state, Atrium and Novant health systems said the original need for a 3,000-bed field hospital were no longer needed in a letter to county health officials.

“As a result of our combined efforts, we believe we are now in position to meet the 600 medical beds needed that were previously requested in a field hospital, assuming the effects of social distancing trend continues the current trajectory,” the letter says.

April Umminger, the senior manager of external and corporate communications at RTI International, said the group which provided the predictive model was continuing to work ad would release new briefs periodically.

“We do not expect to have a standard release schedule as the work is guided by key policy questions, which change over time” said Umminger in an email to NSJ.