Tracking recovery
Why Bay Area numbers on how many people have recovered from COVID-19 are so hard to come by — and not necessarily accurate.
Amid last week’s coronavirus statistics came one piece of seemingly good news: Globally, the number of people who have recovered has risen to more than 1 million. So what does the data on COVID19 recoveries look like for the Bay Area?
Many readers have asked The Chronicle this question, seeking a hopeful counterpoint to the case counts, hospitalizations and deaths updating every day in The Chronicle’s Coronavirus Tracker. But many of California’s largest counties, including most in the Bay Area, are not providing data on how many people have recovered from the virus.
“It’s probably the most ignored of the public metrics that are out there,” said Jeffrey Martin, professor of epidemiology and biostatistics at UCSF. Martin says the reason counties are deprioritizing this data is simple: “In short, it isn’t telling us that much.”
One key issue with data on recovered patients is that counties are defining what it means to be “recovered” in different ways. Martin lays this at the feet of the federal Centers for Disease Control and Prevention, which has issued “vague” criteria for the term. The CDC says a “recovered” person is someone whose fever goes away without the use of feverreducing medications, whose respiratory symptoms have improved and whose symptoms first appeared more than seven days ago.
Neetu Balram, public information manager for the Alameda County Public Health Department, said that public health agencies including the CDC and the World Health Organization do not offer guidance for reporting this data.
“Recovery is not systematically recorded by health care providers or reported to health departments the way that new cases, hospitalizations and deaths are,” Balram said.
She said that data is more important to health departments because those numbers show the direct impact of COVID19 on the health system.
While the majority of California counties are reporting data on recovered cases to the public, only three are doing so in the Bay Area: Marin (185 recovered, as of Saturday), Napa (32 recovered, as of Sunday) and Sonoma (128 recovered, as of Saturday night). Sacramento County displays recoveries as “likely recovered,” and was reporting 803 on Sunday.
Molly Rattigan, deputy county CEO for Napa, said her county is defining recovered cases as individuals diagnosed with COVID19 who meet the following criteria: At least seven days have passed since the onset of symptoms, or at least 14 days have passed since the diagnosis, and they no longer have symptoms. Individuals who were hospitalized must be discharged and no longer in isolation, she said.
To determine recovered cases requires quite a bit of legwork.
“Our case and contact team follows up daily with those that are at home,” Rattigan said.
Santa Cruz County also lists recoveries on its health department website. Jason Hoppin, communication manager for the county, said officials follow CDC guidelines to determine the number based on patient contacts.
“Our health officer reports all positives and negatives to us, and we have patient information through that process,” he said. “We follow up with the positives and check in on symptoms.”
Once the patient’s condition matches the guidelines, they are able to list that person as recovered.
San Francisco, Contra Costa, Alameda, Santa Clara, San Mateo and Solano do not share recovered cases on their websites. In Southern California, Los Angeles, Orange and San Diego counties do not report the data, either.
Experts say especially for larger counties, the process is too timeconsuming and would require resources that could be better spent elsewhere.
“To actually ascertain it, you need to track down patients and ask them how they’re doing, have they been tested?” Martin said. “Many people are choosing to not actually bother to test people once they are feeling better, and even if not, there’s no point clinically to do it.”
John Swartzberg, an infectious disease expert at UC Berkeley, said in theory, tracking recoveries seems simple. A person is confirmed to have gotten the virus, and then they get better. But he said it’s more complicated than that.
“So they’ve clinically recovered, but are they still shedding the virus?” Swartzberg said. “Most of the data we have is not robust enough . ... They may have recovered clinically but can’t answer that they are 100% recovered from the contagion.”
Another way to determine recoveries is via antibody tests, but Swartzberg said only a handful of companies offer accurate testing, and the rest have gone through little verification.
“We’re seeing a lot of false positives where the test was positive, but the person never was positive,” he said. “It’s a big problem that has to be ironed out. We have to have tests that are very specific.”
Martin said the recovery data being provided is “too crude” based on the CDC’s definition and is underestimated. He said counties shouldn’t spend time trying to gather it.
“It will never be done for everyone,” he said. “It’s incompletely being recorded, and it’s not reproducibly being recorded.”
Martin said the more important question down the line is seeing how patients look three, six and nine months out, which is one research effort at UCSF.
“Are you completely normal, or do you have some minor or major residual problems?” he said. “That’s going to come out in carefully coordinated epidemiological research in a carefully managed sample.”