By Ike Swetlitz

Neil Wenger, a professor of medicine at UCLA, was researching different ways of encouraging patients to make end-of-life care plans when he discovered something troubling — hundreds of patients who were seriously ill, according to the health system’s records, were actually dead.

This is a well-known but little-studied phenomenon, according to Wenger — until now. Wenger and his colleagues wrote up their findings in a short paper. They identified 676 patients from UCLA’s health system that were recorded as alive, but were actually dead, according to state data.

At face value, it sounds absurd. But there are many reasons why it might happen, Wenger says. While patients who die in the hospital are automatically recorded as deceased in that system’s database, patients who die at home are not. Same goes for patients who die at another health system with a different electronic records database.

“The health system continues to act as if they’re alive,” Wenger says. “If we don’t know they’re dead, we can’t do the right thing.”

The medical center might continue to send letters, make phone calls, and dispatch emails to patients who have died. It could hold appointments and process medication refills. And if a health system doesn’t know that a patient died, clinicians can’t offer care to the family.

“It could even harm the family if they felt that the health system was being insensitive,” Wenger says.

The clinicians also can’t learn from their mistakes. “Maybe it’s something we did,” Wenger says. “Maybe a drug isn’t working right. Who knows?”

Wenger was able to do this research because there’s another data source in California that has information about who has died — the state public health department’s death file. The team compared the death file to the hospital records, read obituaries on the internet, and looked for information from other health systems.

But the public death file doesn’t have enough information for hospitals to confidently match up the records automatically, Wenger says. More exact data exist, but haven’t been available to hospitals. That should change, he says.

“We think this is a really important finding that needs to be corrected,” Wenger says.