The Register Citizen (Torrington, CT)
5 pieces of data we don’t have
Virus information that has not been collected or shared publicly in Conn.
Since the first case of coronavirus was confirmed in Connecticut, the state has been sharing data on the virus’ victims.
That data has included the number of tests performed, the number of hospitalizations, the number of deaths and a variety of demographic and geographic breakdowns.
It’s that data — shared by Gov. Ned Lamont each day at his pandemic news briefing — that has guided policy as the state first restricted businesses and individual gatherings in an effort to control the virus, and then began to reopen the state in phases.
But some data has been missing — figures that could tell the public and policymakers much more about how to move forward in the crisis.
“The reporting that’s come out is the top layer of basic facts,” said Richard Martinello, medical director of infectious diseases at Yale New Haven Hospital.
For example, while the state had shared the raw number of COVID-19 tests performed each day across Connecticut, it was not known until recently how many
tests were performed townby-town.
The objective, said state Sen. Alex Kasser, D-Greenwich, who had pushed for that town-by-town data to be released, should be “to formulate the best policy for the state … really precise guidelines and really precise, data-driven metrics to guide our policy.”
More robust data collection and sharing would be helpful, Martinello said, for both physicians and politicians as the state begins to open up.
“It would be very helpful on a community level in order to figure out this problem we call COVID,” he said.
Here are five pieces of data about coronavirus in Connecticut that are either not being collected, have been collected but not compiled into a usable form, or have not been shared publicly:
1. Excess mortality
One measure of deaths in the crisis is the direct toll from COVID-19. But how many people are dying over and above the normal rate? That’s called excess mortality and we don’t have it — even though the state Department of Public Health reported it promptly, broken down by age, after the 1918
flu pandemic.
Excess mortality is determined by taking a five-year average of total deaths over a period of time, say January through April, and then comparing it against the same period of time in the target year. That’s another way of gauging the death toll in a crisis.
“What epidemiologists are able to do then is look and see the first wave of COVID and how that compares to historical data,” Martinello said. “Mortality is a very objective outcome.”
This data is shared with the CDC by the state, but The New York Times, in a state-by-state comparison, reported that Connecticut was one of a few states that had not been doing so recently for an analysis.
2. Underlying medical conditions
People often don’t die of just one thing, and sick people often have more than one condition. Knowing those other factors a patient may have, known as comorbidities — diabetes, for example, or chronic lung disease — can point to potential outcomes for patients and help guide policy.
For most COVID-19 cases, we don’t have that data in a usable form. That’s because health professionals reporting deaths and illnesses often don’t record co-morbidities. Until just a few weeks ago, the state Department of Public Health didn’t include relevant co-morbidities in the forms that were used for coronavirus data collection.
“The longer a form is and the more we collect, the less likely people are to fill it out,” said Lynn Sosa, the deputy state epidemiologist. “In the past month we have have expanded it greatly.”
Connecticut’s chief medical examiner, James Gill, called it “a very important public health issue,” adding that some people filling out death certificates “just list the main cause of death.”
“This is unfortunate because it deprives the public health investigators important information about risk factors,” Gill said. “If it is not on the death certificate, it is essentially lost for public health purposes.”
The state does share this data as part of the CDC’s COVID-NET project, but it’s only collected from Middlesex and New Haven counties.
3. Hospital discharges and admissions
The state shares the raw number of current hospitalizations — how many people are in Connecticut hospitals with COVID-19 on any given day. The continued decline in this metric has been one of Lamont’s main justifications for reopening the state.
What we don’t know is how many patients are admitted to the hospital each day, and how many patients have been discharged. A decline of 50 can mean 25 were admitted and 75 discharged, or 350 admitted and 400 discharged.
The numbers are reported by hospitals but the state has not compiled them publicly, except once or twice in April.
“That is data that we follow for our specific institution,” said Kevin Dieckhaus, chief of UConn Health Center’s Division of Infectious Diseases. “How that is reported to the state and how that’s reported out to the community I’m not sure.”
4. Survival rate on a ventilator
Once a patient is put on a ventilator, what are their chances of successfully coming off it and going home?
It’s a slice of data that has been shared publicly in other states — New York Gov. Andrew Cuomo, for example, has shared it several times on his daily briefings.
“That is actually a very complex figure to report accurately,” Martinello said. “There could be other reasons why patients end up on ventilators.”
The typical survival rate is 20 percent to 60 percent, said Dr. Michael Parry, chair of infectious diseases at Stamford Hospital, but it’s hard to compile early in a crisis because people can be on ventilators for a long time and remain hospitalized even longer.
“Six to eight months from now the data will be much more complete,” Parry said, referring to ventilator survival rates but also a lot of other data related to coronavirus.
5. Breakdown of deaths outside of nursing homes
Nursing homes account for 60 percent of all deaths from coronavirus in Connecticut and assisted living facilities add another 10 percent. That leaves 30 percent of all deaths in the general population — and the state has released very little about that group separately from the big picture.
For example, we know the age breakdowns for all COVID-19 patients and deaths in the state, and we know the race of about half the patients. But little is publicly shared about those inside vs. outside of nursing homes.
Those facts, and co-morbidities of patients outside of nursing homes, could help policymakers decide which segments of the population are at less of a risk from the virus.
The information exists; too much of it, in fact, coming from too many directions, requiring the state to meticulously reconcile details that don’t line up. “That is an ongoing process,” said Sosa, at the Department of Public Health, “and is challenged by the fact that we have just mountains of data that we’ve never had before.”
Hearst Connecticut Media columnist Dan Haar contributed to this story.