Michigan's coronavirus death toll much higher than reported as state readies release of new data
Michigan’s death toll from the coronavirus is expected to grow substantially when the state health department publishes its count of probable deaths — the number of people suspected to have died from the virus without testing positive.
Public health experts say the government’s official death tally of 5,595 in Michigan is likely an undercount and that adding the presumptive deaths would be a more accurate representation of the toll of the virus.
This is in part because testing for the disease COVID-19 has been limited, and the official count in Michigan has so far only included deaths where the individual tested positive for the virus. At least a dozen other states are already disclosing probable coronavirus deaths and cases.
"Reporting these probable deaths is very important, and it will improve that data quality," said Nilanjan Chatterjee, a professor at Johns Hopkins University's Department of Biostatistics. "It's a step in the right direction."
The count of probable deaths "provides a more complete picture about how COVID-19 has impacted the state," said Lynn Sutfin, a spokeswoman for the Michigan Department of Health and Human Services.
A timeline for the release of data remains unclear, but Sutfin reaffirmed Wednesday the department's plan to release counts of both probable deaths and cases "soon."
Michigan is overdue to start reporting presumptive COVID deaths, as well as probable cases of the virus, said Dr. Teena Chopra, an infectious disease specialist in Detroit.
She noted the federal Centers for Disease Control and Prevention expanded its reporting to include the "probable" classifications almost a month ago on May 6.
"There is a huge number of patients who are clinically clear-cut COVID cases, but haven’t been reported," said Chopra, a professor at Wayne State University.
"There’s no reason why they should not be counted with the total deaths. Because when we are seeing these patients, clinically there is clearly no alternative explanation for them dying or even being admitted to the ICU," she added.
"Why would the state not want to report the probable cases? I don't understand that."
At least 12 states are already reporting probable COVID deaths and cases to the Centers for Disease Control and Prevention, including hardest-hit New York, which began reporting presumptive cases in April. New York City had nearly 17,000 confirmed deaths and another 4,760 probable deaths, through Thursday.
Classifying probable deaths
Probable deaths include individuals without a COVID-positive test who were presumed to be infected due to their symptoms and an epidemiological link, Sutfin said. The definition also covers those whose death certificate listed COVID-19 disease as a cause of death or a significant condition contributing to death, she said.
An example would be a symptomatic household member of a COVID-positive person where the household member ended up dying, said Brian Hartl, an epidemiologist in the Kent County Health Department.
"That person never got tested, but we'd call that a probable death," Hartl said.
Other examples would be sudden deaths at home attributed to COVID, deaths at the hospital before the individual could test positive, and patients who falsely tested negative but who were "clearly" a COVID case, Chopra said.
"This test, as you know, has a 30% false-negative rate," she said.
The state's definition of "probable" death due to COVID-19 is based on a national standard established by the Council of State and Territorial Epidemiologists on April 6.
Michigan’s local health departments have been reporting both confirmed and probable COVID deaths to the state for weeks. Each local health department received guidance, along with recommendations to evaluate cases based on the council's standard, Sutfin said.
Additional information is necessary to define a probable death, requiring a report and then a public health investigation, so establishing this class of deaths takes longer, she said. Most would be found during contact monitoring of people potentially exposed to infection, Sutfin added.
Kent County has not gotten many reports of probable COVID deaths, Hartl said.
"The majority of reports we've had have had a positive test," he said. "I don't know if it might be because of a lack of awareness in the community to report those (probable deaths)."
Counting the probable cases is critical to understanding the "true burden" of the new coronavirus, Chopra said, including the impact it could have on patients' long-term health.
"When we talk about reopening, we need to know how close we are to herd immunity," Chopra said. "If we are not counting the (probable) cases, we are under-reporting … so we are also under-counting our herd immunity."
Deaths up overall
It is possible that a small fraction of the probable deaths are not from COVID-19, but an influenza-like illness instead, Johns Hopkins' Chatterjee said.
Adding probable deaths to the state's official COVID toll is still likely to be an undercount, he said.
A more reliable estimate of the total deaths due to the coronavirus would be to examine the so-called "excess" deaths statewide — those above and beyond the deaths that would typically have been expected in Michigan during the same season in an ordinary year, the biostatistician said.
"That will give us an idea of how many deaths we still have missed, even after adding the probable deaths," Chatterjee said.
Provisional data on deaths from COVID-19 and other causes, updated weekly by the Michigan Vital Records Office, show that more people died from cancer, heart disease, COPD (chronic obstructive pulmonary disease), pneumonia and flu during March and April than during the same months last year.
The numbers aren't yet complete due to lags in the reporting of death certificates by funeral directors across the state, said Jeff Duncan, Michigan's registrar of vital records.
"There have been increases in other causes of death, particularly early in the crisis when testing wasn’t widespread," Duncan told The Detroit News.
"A lot of that may have been COVID-related, but it wasn’t recognized by the doctors who completed the death certificate."
March deaths attributed to pneumonia and flu were more than 50% higher year over year, with 197 reported last March, compared with 299 this spring. For April, pneumonia and flu deaths jumped about 36% from 203 in 2019 to 276 this spring.
April heart disease deaths were up 20% over last year, from 1,984 in 2019, to 2,379 this year. Stroke deaths were up about 19% for April, from 421 that month last year to 499 this spring.
And septicemia, a serious blood infection, was blamed for 124 deaths this March compared with 88 such deaths during March last year, an increase of 41%.
"It may be that those deaths were from COVID, but it may indicate a strain on the health care system," Duncan said.
"Maybe it’s people not seeking care, not going to the emergency room because they had symptoms of the heart attack (and) because they were afraid.
"I’ve also read speculation that response times are up for emergencies because they have to put on all the PPE (personal protective equipment) and they have to decontaminate their vehicles after every call," he said.
"We have to have a lot more data and a lot of real analysis to come to any conclusions."
CDC data shows a sharp jump in excess deaths since early this year. The CDC predicted an estimated 5,922 to 7,208 excess deaths in Michigan since Feb. 1, when compared with historical averages for the state.
The CDC says the excess deaths could represent misclassified COVID and other deaths that could be indirectly related to the virus — that is, deaths attributed to other causes but that would not have occurred without health care shortages or overburdened health care systems caused by the pandemic.
The excess deaths in Michigan this year that are not directly attributed to COVID-19 could also be because of the virus, but it's unclear, said Samiran Ghosh, an associate professor of biostatistics at Wayne State University.
"Why is this going up? Because there’s no other big epidemic going on. Is it related to COVID? That’s the big question," Ghosh said. "The answer is it could be the case, but we don’t know."