It’s often said in public health that when we do our jobs right, no one knows about it. If we miss one thing, everyone knows. And worse, our credibility comes crashing down.
But because biology — especially disease — is a moving target, we can never have it 100% right, so we try to be prepared for everything. But sometimes, there’s a mistake.
The CDC is receiving a lot of criticism for not testing a now-confirmed California patient who did not meet the case definition of COVID-19. In an outbreak, public health experts first develop a case definition. This is essential in an epidemiological investigation and basic Epi 101. It can’t be too narrow, and it can’t too broad. And importantly, it must be flexible as new information emerges. And sometimes it’s not perfect.
Until a day ago, the case definition was fever and cough with a history of travel and/or contact with someone with a history of travel to known affected areas.
Remember this is flu season in the northern hemisphere.
There are limited tests available. This is an issue which must be addressed, but for now, we have limited resources. Even with more money and a complete CDC staff, the tests were only developed in the past 10 weeks. It’s a novel virus, there weren’t test kits sitting on the shelves. It takes time to ramp up.
Now, imagine testing every single person who came in to a clinic with a fever and cough during flu season, but no history of travel or known contact with someone who traveled. Not only would the cost and use of human resources be astronomical, they’d run out of tests real fast, leaving none available for those who did meet the case definition with travel history. Note that when Zika first appeared on the scene, travel history was also part of the criteria for defining a case.
It’s important to understand how CDC testing works. If it’s January in the US and you have a fever and cough, your doctor may do a rapid test for influenza. Or they may not. They may just prescribe Tamiflu if you’re in the window and tell you to stay home until you’re better. If they do test, in most clinics, that sample goes nowhere other than to confirm for the patient that they have flu. However, we have “sentinel clinics.” These are clinics who test regularly and send those samples to the CDC for further analysis. There is no way CDC could test every sample in every clinic, and no way they can even test what they receive. So they test a subset, and extrapolate from there. From these tests we know which flu strains are circulating.
In late March 2009, two sick teenagers went to the clinic at Randolph AFB for “influenza-like-illness.” Fortunately, Randolph’s clinic was a sentinel site. The samples taken were sent to the CDC and a novel virus, H1N1, was identified. Those first cases became the start of the H1N1 pandemic. Had they gone to most other private clinics in the city, it might not have been identified for months and they would have just presumed it was seasonal flu. We got “lucky” in 2009, though there were likely many cases undiagnosed.
But no one knew that.
In hindsight, the CDC decision not to test the California patient was unfortunately wrong, but not unreasonable and not without epidemiological merit.
We have a saying in public health. If you hear hoofbeats, think horses, not zebras (unless you’re in sub-Saharan Africa). But sometimes, a zebra appears.
The CDC is now revising its case definition and states will have more access to test kits once their capacity to test is verified. The experts are doing exactly what they should be doing in an outbreak.
So give the CDC and public health folks some love. And money. And staff.
But don’t hug them. Just wave. And cough into your elbow. And wash your hands.