Statistics on mortality little more than guess

Margin of error on death certificates is estimated to be 20 to 50 percent.

As it stands now, there’s no way to know for sure.

The numbers that state agencies, nonprofits, and leagues of researchers use to determine what kills people and how often have a huge margin of error, according to studies cited by the county coroner and an expert in pathology at Ohio State University.

An attempt to put numbers to causes of death here proved highly speculative. It turns out the best data available isn’t very reliable.

“It is a problem,” said Dr. Charles Hitchcock, an associate professor in the pathology department at OSU Medical Center. “And it has implications, too. ... Sometimes it has financial implications for a family.”

The implications may be even more widespread.

“You base health statistics on this,” said Dr. Richard Marsh, Clark County’s coroner. “That’s how we figure out why people are dying. If we’re going to direct (research) money, we better be certain we have the right data behind it.”

For example, in 2008, the Ohio Department of Health produced a county-by-county “cancer profile,” relying on data of cancer incidence and mortality rates. The mortality rates listed in the much-cited docuements may not be accurate.

Various scientific studies place the error rate for the “cause of death” box on death certificates between 20 and 50 percent.

Why? It’s incredibly hard to accurately describe why someone has died based solely on their medical history, Hitchcock said.

Those huge error rates are calculated by allowing a doctor to fill out a death certificate with how he or she thinks the person died based on medical history. Then autopsies are performed on the study group to determine the actual cause of death.

Autopsies themselves have a rate of error, but “they’re a lot better than the clinical (determination),” Marsh said.

“The autopsy is considered the gold standard,” he said.

The Ohio Department of Health has determined that 23.5 percent of Clark County residents currently smoke — very close to the rate in the rest of the state, at 23.4 percent.

But the rate of error in reporting causes of death skews any data about how often Clark Countians die from smoking-related diseases. Or any disease, for that matter.

Contact this reporter at (937) 328-0353 or at bsmith@coxohio.com.

X