Why we can’t really calculate the true coronavirus (COVID-19) ‘death rate’ in Michigan

Number of uncounted coronavirus cases skews data

DETROIT – There’s a lot of confusion surrounding the term “death rate” when it comes to the coronavirus (COVID-19). What does that really mean, and how high is the virus’ death rate in Michigan and beyond?

At the start of the coronavirus outbreak, infection rates appeared to be in the 1-2% range in China, but that’s not what we’ve seen as COVID-19 moves around the globe and in the United States.

Dr. Frank McGeorge wanted to clear up why there has been such a large discrepancy in death rates.

Many viewers have asked Local 4 why the death rates from COVID-19 are so high in some states and not in others. More specifically, why are they so high in Southeast Michigan?

The answer is that many are using the term “death rate” imprecisely without an apples to apples comparison.

UPDATE -- May 26, 2020: Michigan coronavirus cases up to 55,104; Death toll now at 5,266

“Death rate” implies the percentage of people who die from a given condition. In order to know that number, you have to know exactly how many people die of the infection and divide that by the total number of people infected.

The problem is that neither of those numbers are exactly known in this pandemic.

When you divide the recorded number of deaths by the number of confirmed cases, you’re actually calculating a case fatality rate.

Looking at the date from the state on May 19, the number of COVID-19 deaths was 5,017. Dive that by the number of confirmed cases at the time -- 52,350 -- and you get a case fatality rate of 9.5%. That’s a concerning number, but there’s a problem.

“Unfortunately, the confirmed cases are from a population of very sick people,” said Jim Collins, the director of the Communicable Disease Division at the Michigan Department of Health and Human Services. "Up until recently, our testing strategy was to look at the sickest people to try to help out in terms of treatment, right? So those people were more likely to have serious consequences.

"A lot of our tests were limited to those with severe illness, the hospitalized cases, because we didn’t have resources and we wanted to focus on people that we could help -- life and death decisions. So we’re seeing the impact of that. There are more deaths in that population.

“A testing strategy when resources are limited targets the most vulnerable people. It’s going to artificially elevate the case fatality rate."

Basically, in the beginning of the pandemic, officials were testing more of the people most likely to die, which skewed the numbers.

But there’s an even more fundamental flaw with case fatality rate.

“In epidemiology, we always talk about the cases we have in our surveillance system representing the tip of the iceberg,” Collins said. “Those are cases that sought medical care, were tested and then showed up in our surveillance system as a case. That’s the tip of the iceberg because there’s a whole bunch of cases out there that were not ill enough to seek care, a bunch were probably asymptomatic altogether. They’re still infected.”

READ: Retail reopens in Metro Detroit with uncertainty of coronavirus changes

Since those infections go unidentified and uncounted for, the number of cases is actually much higher. If you were to divide the number of deaths by the much higher actual number of total cases, the infection fatality rate would be much lower.

There are some challenges to counting deaths. It’s likely deaths have been under counted because people died at home without a formal diagnosis of COVID-19.

The case fatality rate for older people is higher. In fact, the number increases consistently for every additional decade after 40, so that’s a real issue, not an artificial one from a testing bias.

View: COVID-19 data section

About the Authors: