Asked why Italy had a high rate of deaths due to the coronavirus, government officials there said it was because they were using a broader definition for such deaths than other countries, counting any victims who had tested positive even if other illnesses were at fault.

In Illinois, officials announced last weekend that an infant died after testing positive for the virus, but said they were still determining the cause of death. And in Florida, two deaths were listed as virus-related in mid-March, then later removed from the official count.

Incomplete data and inconsistencies come as no surprise to medical examiners, coroners, and physicians who fill out death certificates for a living. Even under normal circumstances, determining the cause of death is an inexact science. With the surge in deaths from a cause no one had heard of three months ago, ironclad certainty may, for now, be even more elusive.

The U.S. Centers for Disease Control and Prevention has provided preliminary guidance for how to record virus-related deaths in case of uncertainty, such as when test results are not available. The National Association of Medical Examiners is adapting as well.

“I think that across the country, you’re going to see different interpretations,” said Meredith J. Buck, the coroner in Bucks County. “People are trying to standardize something that has no precedent.”

Ultimately, the results will be crucial in helping researchers determine the virus’ deadliness. In the United States, 2% of identified patients had died as of March 31, according to figures compiled by Johns Hopkins University. But epidemiologists predict the true percentage will turn out to be lower, as wider testing identifies many more infected people who survived. In a March 30 study in The Lancet Infectious Diseases, researchers used a statistical model to estimate that the mortality rate in China was less than 1%.

Consensus is emerging on how to handle several coronavirus-related scenarios, said Gregory McDonald, dean of the school of health sciences at the Philadelphia College of Osteopathic Medicine. Early in the pandemic, before the virus was at the front of every physician’s mind, it is likely some patients died of an infection without samples being taken for testing, he said. Such cases would be never be counted.

More recently, some patients have had samples taken for testing but died before the results came back, meaning the cause of death should be filled out as “pending." But in Pennsylvania, treating physicians are not allowed to fill out a death certificate that way; only medical examiners and coroners can do so — such as when the result of a drug test comes back after death. (Medical examiners, who are appointed in a few counties, such as Philadelphia, are physicians. Coroners, most of whom are elected, do not require medical degrees.)

If a test result comes back positive after the person is buried or cremated, physicians are urged to contact their county medical examiner or coroner so the cause can be updated, said McDonald, who is a physician and also deputy coroner in Montgomery County.

“We’re pulling together as a country,” he said. “I suspect that as physicians become more and more attuned to these deaths, accuracy will get better and better.”

A subtler issue is what to do when the patient has other serious medical conditions. If the person suffered from chronic lung disease, then became infected with the virus and died of pneumonia, the immediate or primary cause would be pneumonia as a result of COVID-19. The lung disease would be listed as a contributing condition, said Sally S. Aiken, president of the National Association of Medical Examiners.

In other scenarios, the reverse could be true, McDonald said. If a person suffers serious head trauma then becomes infected after being put on a ventilator at a hospital, the trauma might be listed as the primary cause, with COVID-19 as contributing factor, he said.

That kind of approach is what Italian authorities have said might account for the high rate of deaths in that country — more than 11% of those infected, as of Wednesday. But in Germany, where the death rate is far lower, officials say they are using a similarly expansive definition. Epidemiologists say Italy’s high death rate is more likely the result of its older population, as well as a surge in cases that overwhelmed the health-care system.

In Florida, where two deaths were listed as virus-related then unlisted, officials blamed the shift on errors in data entry. In Illinois, where the infant died after testing positive for the coronavirus, health officials said the death remained under investigation.

In some deaths, the role of the virus may go unnoticed, such as for people with heart disease. Other respiratory viruses, such as the flu, are known to increase the risk of a heart attack, possibly because increased inflammation makes the plaque in their arteries more unstable. Cardiologists say the coronavirus likely has the same effect, meaning a person could die of a heart attack and not be identified as a COVID-19 patient.

Among the various professionals who determine cause of death, there can be differences of opinion even under normal circumstances. But as in so many other professions, these times are not normal, said Buck, the coroner in Bucks County:

“Scenarios are coming up that people didn’t necessarily think about."