Some homes reduce virus deaths after new state definition

In the last month, several Pennsylvania nursing homes have reduced the number of COVID-19-related deaths they had previously reported after the state told them residents who died after testing positive for the disease did not have to be counted if they had “recovered” from the disease before they died.

Among the four nursing homes that reduced their COVID-19 death count in their self-reported data in the state’s June 23 report was Brighton Rehabilitation and Wellness Center, the Beaver County nursing home that has had the

That change allowed Brighton — which has had the most COVID-19-related deaths of any nursing home in the state since the state began releasing nursing home figures — to self-report that it had fewer COVID-19-related deaths than one other nursing home, Conestoga View in Lancaster, which now has 75.

Brighton’s management said in an emailed answer to questions that when the pandemic began “any resident who contracted COVID-19, regardless of their health status at the time of death, was counted as a COVID-19 death. As tracking practices evolved, the Department of Health clarified their guidance to count only those who have an active COVID-19 status at the time of passing.”

The state Department of Health said this new interpretation of who should and should not be considered a COVID-19-related death using the “recovered” definition — what Brighton refers to as “active COVID-19 status” — was based on guidance from the federal Centers for Disease Control and Prevention.

But the CDC said Friday that it has never issued any guidance connecting its definition of “recovered” — which was created to help determine if a person should be removed from room isolation after testing positive — to any interpretation of whether that person later died of a COVID-19-related cause.

Among the four nursing homes that reduced their COVID-19 death count in their self-reported data in the state’s June 23 report was Brighton Rehabilitation and Wellness Center, the Beaver County nursing home that has had the worst outbreak of any nursing home in the state. In the June 23 report, Brighton reduced its number of deaths from 82 to 73.

She referenced a section of the CDC’s guidance from April on certifying deaths due to COVID-19 that includes a section that says: “When reporting cause of death on a death certificate, use any information available, such as medical history, medical records, laboratory tests, an autopsy report, or other sources of relevant information. Similar to many other diagnoses, a cause-of-death statement is an informed medical opinion that should be based on sound medical judgment drawn from clinical training and experience, as well as knowledge of current disease states and local trends.”

The CDC definition of “recovered” that the state is telling nursing homes to follow was created by the CDC early in the pandemic, and updated May 2, to help health care facilities to determine when someone was well enough to be removed from isolation status, where they had been placed in hopes they would not infect anyone else.

That definition says that if a person tested positive for COVID-19 and showed symptoms, they were considered to be “recovered” and could be removed from isolation if at least three days had passed without a fever or with an improvement in any respiratory symptoms.

If they tested positive and did not show symptoms, they could be removed from isolation if at least 10 days had passed since their last test.

And both symptomatic and asymptomatic people could be removed from isolation if they had two negative COVID-19 tests in a row.

Experts say using one aspect of a person’s health — in this case whether or not they met a definition of “recovered” that was not intended to be used in determining the cause of death — could potentially skew data on COVID-19 deaths if used broadly to exclude people who did die after contracting COVID-19.

“There are going to be cases where they recovered [from COVID-19] and later died of a stroke that was because of COVID-19’s impact on their health,” said Dr. Amesh Adalja, a Pittsburghbased infectious-disease physician and senior scholar at the Johns Hopkins Center for Health Security.

“You want to make sure changes [to a cause of death] are based on the complete medical record,” he said. “You have to look at these on a case-by-case basis.”

State Department of Health spokesman Nate Wardle said in an emailed statement that the state is following the CDC’s definitions: “PA DOH uses the CDC definitions of a COVID19 death. A death is considered a COVID-19 death if the medical certifier reported COVID-19 as the underlying cause or a significant contributor to the death . ... If a patient is COVID-19 positive but recovers from their illness, and then later dies, that death would not be classified as a COVID-19 death unless the medical certifier believes that the previous COVID-19 infection significantly contributed to the death of the individual.”

Mr. Wardle could not provide any state document or guideline, nor any CDC documents or guidelines, that connect the definition of “recovered” to the definition of what constitutes a COVID19-related death.

Instead, he wrote, that information was given to Brighton and other facilities by state Department of Health staff in phone conversations when they would call the state to ask questions about whose death could be considered COVID-19-related or not.

None of the other three Pennsylvania nursing homes that reduced their death counts returned phone messages left for them at their offices this past week. Three other nursing homes, whose data was reported by the state, not selfreported, also had their death counts reduced in what appear to be corrections to data by the state.

Even though Brighton in the state’s self-reported data now says it has 73 deaths, the state’s county-based nursing home data — which is based on reports of positive cases and deaths to the state’s infectious disease reporting system — still shows that Brighton has had 82 COVID19-related deaths, based on a Pittsburgh Post-Gazette analysis of that county data, which does not name the facilities.

Brighton has not yet tried to change any of the death certificates of the residents who died for whom COVID19 is listed as a factor in their death. To do that would take a court order and the cooperation of Beaver County Coroner David Gabauer, who said he could not comment about that. No such cases have yet been filed in Beaver County court.

Brighton management said in an email that it also has not notified any of the nine families that, at least in its self-reported data, it has changed the cause of death because it “only reports the passing of loved ones to family members.”