Theories about inflated and inaccurate death counts have become common as the number of people who succumb to the coronavirus continues to mount.
But health experts say COVID-19 fatality totals are undercounted across the nation — even in Texas, where a change in reporting methods last month led to an increase of hundreds of deaths.
So how do health experts really know how many people have died from COVID-19? Here’s what you need to know.
Deaths are not easy to count during a pandemic
Many factors make COVID-19 fatality calculations complex.
Spencer Fox, associate director of the COVID-19 Modeling Consortium at the University of Texas at Austin, which models deaths from the virus nationally and by state, said the calculations are similar to those used for the flu.
“Seasonal flu is something that has been investigated for decades, and still every year, there are refinements to how we determine the true mortality costs of a seasonal flu epidemic,” he said. “We often look at influenza mortality, that is a case confirmed by influenza, but then we also look at pneumonia mortality. There’s metrics that have been developed over the past few decades to try and encapsulate all the ways in which the flu can cause mortality.”
That’s why health experts say COVID-19 deaths are underestimated. There are many ways COVID-19 can kill because of the complications it causes. And many of the people killed by the virus are already medically fragile because of other conditions, making cause of death determinations difficult.
There are many ways people could be left out of the totals. Health experts say people who died during the early months of the pandemic and weren’t tested probably were excluded.
People can be tested for the virus after they die, but places with limited testing capacity may reserve kits for people who are sick. Homeless and low-income individuals who don’t have access to testing also may be left out of the totals for that reason, health experts say.
People who die in their homes, which has become more common as area hospitals face surges in capacity, also could be left out, Fox said.
Recent changes in Texas may cause discrepancies with local data
The Texas Department of State Health Services announced July 27 that it would use death certificate data to report COVID-19 deaths, instead of relying on reports from local health departments.
The state said the change would allow deaths to be reported faster, make more demographic information available and allow deaths to be reported based on the day people died, instead of by the day deaths were reported to the state.
The new method initially added more than 600 deaths to the total, but 225 were later removed because of an automation error.
Chris Van Deusen, a spokesman for the department, said the state’s old process was unlikely to have caused deaths to be overreported, because fatalities were added to the total only after local health departments investigated and verified each death.
“We don’t have any reason to believe that deaths not related to COVID were included in the counts before the change in methodology, and the fact that the number of deaths went up after the change would seem to reinforce that,” he said in an email.
But the new method means there probably will be differences in the death totals reported by the state and local health authorities.
In Dallas County, health officials have been following guidance issued by the Centers for Disease Control and Prevention in April to report deaths, said Lauren Trimble, Dallas County Judge Clay Jenkins’ chief of staff.
Deaths are included if a person has a positive COVID-19 test and a death certificate denotation, a record of symptoms consistent with COVID-19 that progress from illness to death, or a determination by the medical examiner’s office that there is no other cause of death, Trimble said.
Even though the county uses death certificate data to confirm deaths, there could still be differences in county and state reports.
A Texas death certificate contains two parts under the cause of death section: the first being the cause of death, and the second covering significant conditions that contributed to the death but didn’t directly cause it.
Although Dallas County includes deaths in which COVID-19 is listed in both parts, the state is only counting deaths that list COVID-19 in part one.
Lags in the time it takes the state to obtain and process death certificates also could account for differences.
As of Wednesday, the county reported 50 more deaths than the state did.
Difference in reporting methodologies between local and state authorities account for differences in case and death totals nationwide, health experts say. Although the CDC offers guidance on how authorities should count cases, it’s up to each department what they choose to include.
A review conducted by the Washington Post in June found that most states weren’t reporting probable cases or deaths — instances in which COVID-19 is suspected but isn’t confirmed with a positive test — despite CDC recommendations to do so.
Texas doesn’t include probable deaths in its total, nor does Dallas County. Probable deaths are included as a separate category in Dallas County’s demographic reports, which are released Tuesdays and Fridays. There have been two probable deaths so far.
It’s not likely death counts are being significantly inflated
Health experts say false positives and negatives are possible with any reporting method. But even with limitations in data and differing methodologies, it’s still unlikely that fatalities are significantly overreported.
One common theory has been that people may die in the hospital of an unrelated cause but get added to death totals because they also test positive for the virus.
However, Fox said the researchers in the modeling consortium looked into that possibility and found that such scenarios would “absolutely not” meaningfully inflate the numbers.
The reason comes down to probabilities, he said. The chance of dying in general is typically very low, while the chance of dying from COVID-19 only grows as the disease spreads in a community.
“The risk of winding up being hospitalized or dying from COVID is so much higher than the general probability that someone winds up in a hospital,” he said. “If that’s the case, then there never could be a dominating effect of random people showing up in the hospital and winding up in the data. And the same would be said for mortality.”
Will we ever know how many people COVID-19 really killed?
Health experts say the most accurate death counts will become clear years after the pandemic is over. Even then, there probably will be only estimates.
“It will always be an estimate. … The estimates will get more and more accurate through time,” Fox said. “I don’t think any changes that happen will make mortality estimates more inaccurate.”