Hawaii was poised to hit its peak coronavirus medical need as soon as this past weekend, according to one widely cited coronavirus research institute. Yet the local numbers continue to climb and are likely to exceed 500 this week, begging the question: why doesn't the state have a solid idea of when the surge will come and whether it is prepared?
U.S. universities using statistical modeling to project the course of the coronavirus through individual states have sought to answer just when cases would peak and whether there are enough health care resources to avoid overwhelming hospitals as seen in New York and Italy.
Across the country, states have scrambled to pull information from the statistical models to help them determine when to impose measures like stay-at-home orders. But the models can be notoriously inaccurate.
Some local researchers think the projections may not apply well to Hawaii and urge developing better information to help local health care and government officials make their life-and-death decisions.
One of the main models used by the White House comes out of the University of Washington's Institute for Health Metrics and Evaluation. IHME projects if there are sufficient health care system needs like hospital beds, and how many deaths states might expect.
In March, the institute said Hawaii's draw on medical resources would peak in early May. According to those early estimates, Hawaii would fall short by 154 hospital beds and 123 intensive care unit beds. Most concerning, it predicted Hawaii would see 374 COVID-19 deaths by early August.
Then earlier this month, the institute's researchers sharply changed those projections. They said Hawaii’s peak would instead occur as early as this past weekend. And, it concluded, the state would see little to no hospital bed shortages and only about 100 deaths by August.
Nick Redding, the executive director of the Hawaii Data Collaborative, a local data analysis project, says the University of Washington model is the best he’s seen. But he isn’t comfortable with the latest changes.
"Basically, because our hospitalization rate and death rate is so low, they had to make some assumptions in their model," he said. When the researchers applied their recent update across all 50 states, Redding believes it made the model less valid for Hawaii, although perhaps more so for New York and other states.
The University of Washington's original, dire projections for Hawaii may be more reliable than the latest optimistic numbers, Redding said.
Regardless, he thinks there’s room for improvement and the Hawaii Data Collaborative is working on a local statistical model of its own.
"We've got to do better. We've got to get a localized model informed by local experts, and daily updated data from the hospitals that we can start to track the model," Redding said.
"We want to put a call out for, 'Hey, let's as a community come together and get better models.'"
Redding said more information about day-to-day hospital capacity would give the state a better understanding of whether it can ride out the pandemic.
Lt. Gov. Josh Green, an emergency room physician who has been critical about aspects of the state administration's response to the pandemic, recently began posting on Instagram information on hospital bed counts and available ventilators, among other data.
Hawaii’s available hospital bed capacity is currently at about 45 percent capacity and ICU beds at 28 percent. Before the COVID-19 outbreak cancelled elective surgeries, hospital occupancy was at about 65 percent.
Perhaps most importantly, experts say, what is needed is more information about who is infected.
Dr. Thomas Lee, an epidemiologist and a professor at the University of Hawaii’s public health program, said Hawaii has challenges.
"We're different from some other states. Unfortunately, we do have a homeless situation. But knowing that that is a potential hot spot, the other hot spots are marginalized communities where English is a second language and they may not be getting the information, the health education that everyone else is getting," he said.
If researchers could see the status of COVID-19 in terms of its potential spread in marginalized, underserved communities, he said it would improve some of the modeling.
Redding said even counting the number of people at hospitals suspected of having the virus, but who haven’t yet tested positive, could go a long way.
According to Lee, the state should be looking at places like South Korea or Taiwan that use extensive contact tracing to notify people who are around those with the illness.
"They send out messages to people who live within those vicinities and say, 'Hey, be aware of the fact that you do have people who have tested positive in your community," Lee said. "If done in the right manner, it's not going to incite fear or aggressive behavior."
The state Department of Health posts maps that generally show where the positive COVID-19 cases are across Hawaii.
But because many of the cases are travel-related, the maps don’t reflect if the virus is circulating in communities with high numbers of cases.
The department says that separately mapping the community-spread cases that number about 92 are not important when it comes to determining the risk of acquiring the coronavirus.
And right now, it says, it’s fine with the 30 or so people tracing and notifying contacts of those who test positive.
