State, local experts push back on university’s bleaker coronavirus forecast

Massachusetts officials say that the stay-at-home advisory has been very effective.Erin Clark/Globe Staff

An influential University of Washington group forecasting the spread of coronavirus worldwide is now anticipating a much gloomier outcome for Massachusetts, with far more deaths than its original models, based in part on how residents may interpret state guidance on social distancing to slow the spread of the virus.

But local officials and experts on Wednesday pushed back against the group’s model, saying it conflicts with in-state forecasts and does not accurately account for the social distancing taking place in Massachusetts.

Two weeks ago, the University of Washington’s Institute for Health Metrics and Evaluation predicted that demand on Massachusetts hospitals would peak around April 14, and that by summer the virus would have killed nearly 1,800 people in the state. It updates its forecasts frequently as new data is available.

As of Wednesday afternoon, the study group was predicting an April 28 peak and had upped its forecast for COVID-19 deaths in Massachusetts to a staggering 8,219, a more than fourfold increase from earlier projections. Local state modeling as of last week, however, forecasts no more than about 4,300 deaths.

That massive jump in projected mortality in the university’s model is being propelled by the number of deaths already recorded in Massachusetts, as well as the weight the model gives to a statewide stay-at-home order to promote social distancing, said Ali H. Mokdad, a professor of health metrics sciences at the Institute for Health Metrics and Evaluation and chief strategy officer for population health at the University of Washington.

Increases in the number of fatalities is troubling, Mokdad said. As of Wednesday, the state reported 29,918 cases of COVID-19 and 1,108 deaths. Among all infected people reported, roughly 3.7 percent have died, a percentage that has crept up since the beginning of the month.

Mokdad and his colleagues also take issue with the social-distancing messaging in the state. Governor Charlie Baker announced last month that nonessential businesses must close by March 24 and directed the Department of Public Health to issue a statewide stay-at-home advisory, urging residents to “avoid unnecessary travel and other unnecessary activities.”

Other states, such as New Hampshire, have issued what they call stay-at-home orders, though they come with many exceptions for the necessities of daily life, such as going out for groceries or gas, visiting a relative, picking up takeout food, or just getting fresh air and exercise.

It may seem largely like a semantic difference, but the power of the “order” counts significantly in the University of Washington model, and Massachusetts’ long-term outlook appears far worse without it, Mokdad said.

Mokdad defended the model, saying an advisory is a weaker message and does less to promote social distancing. “When the government comes out and says, 'Hey, we have a serious problem, stay at home,’ people are more likely to stay at home,” he said. “When you say it’s an advisory and leave it up to the people, I mean — it is sending mixed messages at this time. The language is very important.”

The Baker administration disagrees, saying the advisory has been extremely effective in promoting social distancing, the primary practice to prevent infections.

“All available data show Massachusetts residents are staying home," said Tim Buckley, a senior adviser to the governor, in an e-mail, "as mobility trends demonstrate the Commonwealth outperforms the national average and tracks closely to states that use different terminology for their guidance. Massachusetts’ social distancing protocols are in line with national standards and balance social isolation with mental health needs.”

Samuel Scarpino, a professor in the Network Science Institute at Northeastern University, also takes issue with the University of Washington model, which he said has some built-in weaknesses.

“The fact they have a variable that says there’s this huge difference between an order and a strong recommendation — when we can see in the data that that is not bearing out — that is clearly a flaw,” he said.

Scarpino, who has studied local social distancing data, said that commuting patterns in Massachusetts show a “precipitous drop,” and that the movement of people within their own census tracts “in and around their homes, is down to levels that have never been recorded.”

The death toll in Massachusetts, which topped 1,000 on Wednesday, “is horrifying and unimaginable,” Scarpino said, “but realistically are we going to see eight times as many deaths, given where we are in the curve? No.”