As the death toll climbs in the Bay Area’s coronavirus battle, an alarming projection that the pandemic would kill at least 2,000 in Santa Clara County has exposed a rift among local officials and a debate among scientists: What is the best — and most responsible — way to estimate how many might die?

On Friday, Santa Clara County Executive Jeff Smith, who has a medical degree and whose job includes overseeing the county’s public health department, blasted San Jose officials for unveiling a controversial projection that the virus would kill 2,000, even with the most stringent social distancing. Up to 16,000 would die, the city reported, if nothing was done to stop the spread.

“It misrepresents the reality,” Smith said in his first public comment about the report. “It’s a projection based on data from other parts of the country and the world, and it’s a worst-case scenario. This was really irresponsible for them to put this out because there’s no real basis in fact for that information.”

San Jose Mayor Sam Liccardo said the city’s modeling was aimed at getting a handle on how the disease might affect city services and to motivate residents to observe social distancing mandates aimed at “bending the curve” of new infections downward by showing the dire consequences if they aren’t followed.

“They are not, in any way, intended to scare,” Liccardo said. “The model merely illustrates why we all need to do our part to halt virus transmission. … The city will continue to be transparent, because our residents deserve to know what we know.”

It’s not the first time a government leader’s alarming projection raised eyebrows. Gov. Gavin Newsom last week, in a letter to the Trump administration seeking federal aid in bolstering hospital capacity, said 56 percent of the state’s nearly 40 million residents would be infected over the next eight weeks, which he said would mean 25.5 million people. But 56 percent actually would be 22.4 million people, and his office later clarified that the figure did not take into account the statewide stay-home order he had imposed.

A number of studies projecting the crest of infections, hospitalizations and fatalities have surfaced recently as public officials desperately try to asses the social distancing needed to avoid a surge that overwhelms hospitals, typically with a range of outcomes based on the stringency of those measures.

Smith said Santa Clara County is working with Stanford University on projections that will be more accurate and meaningful based on local data that takes into account age and illness distribution in the county as well as local social distancing efforts, but they aren’t quite ready for prime time.

“We don’t have enough data to know in an extremely accurate way what presumptions need to be made,” Smith said.

SAN JOSE, CA – FEBRUARY 28: Santa Clara County Executive Dr. Jeff Smith speaks during a news conference in San Jose, Calif., on Friday, Feb. 28, 2020. Santa Clara County health officials confirmed a third case of the novel coronavirus. (Anda Chu/Bay Area News Group) 

A study this week by Imperial College London assessing the global impact said that without mitigation measures, 7 billion of the world’s 7.8 billion people would be infected and 40 million would die, including nearly 3 million in North America. With the most successful mitigation, global infections would fall below 500 million and deaths below 2 million, 92,000 of them in North America.

A University of Washington study this week on U.S. impacts projected more than 81,000 U.S. deaths from COVID-19 over the next four months, including more than 6,100 in California and more than 10,200 in New York.

Both of those models suggest just how aggressive the San Jose estimates are for Santa Clara County, which has less than one percent of the U.S. population and about five percent of California’s population, though it is one of the nation’s coronavirus hot spots.

Like Smith, the authors of the University of Washington study noted that modeling is only as good as the quality of data plugged into it, which they said in many cases has been flawed. They said that with U.S. testing limited and largely restricted to those who are seriously ill, it skews the picture of infections and fatalities among the population.

“Right now we don’t have enough testing and don’t know how many have COVID-19,” said University of Washington health metrics sciences professor Ali H. Mokdad. He said the modeling takes into account social distancing measures in California and other states, and the projections will be updated as those change.

The San Jose projections, first reported Thursday, raised alarm as Santa Clara County recorded its 574th case of coronavirus Friday — second only to Los Angeles County, with five times its population. For weeks, Santa Clara has been the center of the storm for the deadly virus in California, which has confirmed 4,040 cases and 83 deaths as of Friday.

SAN JOSE, CALIFORNIA – MARCH 18: San Jose Mayor Sam Liccardo, with a group of elected officials socially distanced behind him, announces the creation of Silicon Valley Strong, a regional effort forming to help fight the coronavirus pandemic, Wednesday, March 18, 2020, in San Jose, Calif. (Karl Mondon/Bay Area News Group) 

On March 16, Santa Clara County health officials orchestrated an unprecedented stay-at-home order for six counties in the Bay Area in an effort to slow the virus’ spread so hospitals could prepare for a patient surge.

The tension between city and county officials revealed the pressure local leaders are under to push for public cooperation of the nation’s first coronavirus shelter-in-place order while not inducing panic.

Liccardo noted that county executive Smith had told the city in an email that their projection model “was not wrong in any way.”

But Smith said that while the model wasn’t wrong, it was based on national and international numbers, not exclusively on county numbers, resulting in a skewed picture of the local impact.

Though Smith said he raised serious concerns about the data used in San Jose’s modeling, Liccardo insisted Smith never told the city not to publish their own projections. Email exchanges Wednesday and Thursday however show a dispute between the mayor and county executive over such projections, with the city pushing the county to provide them.

“I strongly suggest that we not focus upon predictions more than preparation,” Smith wrote in a March 25 email to Liccardo and City Manager Dave Sykes. “Modelling can only give us clues about the future, not predictions. The completely honest answer is that we can’t predict a specific time when resources will have to be rationed. We can’t even be certain that that will happen. We do know that it is certainly a risk and that we must prepare for it.”

“A message that creates more fear and anxiety,” Smith wrote, “is just as bad as a message that does not motivate action.”

But Liccardo responded the next day in an email copied to Sykes and several county officials that “obviously, I disagree with your conclusion, as you undoubtedly expect.” The mayor said he was concerned that people will stop observing social distancing mandates if they see hospitals overwhelmed and conclude they were ineffective.

“We could (and probably should) use assumptions more likely to reveal a more severe outcome,” Liccardo continued. “I could be wrong, of course — we need you to communicate with us about that, or else we only have our own model, and our own assumptions.”