Opinion | Mass testing has its problems. They're nothing compared to not testing. - The... - 1 views
www.washingtonpost.com/...a-970a-64c73a1c2392_story.html
testing mass statistics innumeracy numeracy probability probabilistic thinking
![](/images/link.gif)
-
Short of a vaccine, mass testing is among the most plausible paths back to some kind of normalcy.
-
Imagine a world where you could stop at a drive-through testing center, get a result in 20 minutes or so, and then motor onward to your dinner party with a “negative” certificate in hand. Imagine outdoor kiosks at airports, with a negative test result required to get inside. Imagine offices, even restaurants or bars, with a nurse stationed in the parking lot.
-
Would this be annoying and cumbersome? Yes. Would it be a vast improvement on what we are doing now? Also yes. It even seems plausible since Abbott Labs has won emergency approval for a test that costs $5, returns an answer in 15 minutes, requires no specialized equipment and can be produced in bulk. Mass deployment of this test, or others like it, could fundamentally change how we approach covid-19.
- ...10 more annotations...
-
No test is perfectly accurate; all generate false positives and false negatives. How many varies by the test — the new Abbott Labs test, for example, has a false negative rate of about 3 percent and a false positive rate of about 1.5 percent.
-
We’d also require public understanding that test results are a risk-assessment tool rather than a definitive answer. And shifting the public’s thinking might be harder than shifting Trump.
-
For mass testing to work, people will need to understand that even an event where all attendees just tested negative isn’t necessarily covid-free, because some false negatives might have slipped through
-
Giving hundreds of millions of citizens this crash course in elementary statistics would be very difficult. Teaching people how to think about a positive result would be even harder — especially if mass testing does its job of reducing caseloads. Because as case numbers fall, false positives will become a bigger and bigger problem.
-
Say we’re testing 200 people, 10 of whom have covid-19. Given its rate of false negatives, the Abbott test would probably catch all 10. But with false positives, it could also tell one or two people who don’t have covid-19 that they’re infected. That’s not ideal — as good citizens, they’d have to go home and quarantine. But a more sensitive follow-up test could substantially mitigate this problem, shortening such quarantines to a day or so. It would be worth it to get those 10 true positives out of circulation.
-
Doing this, we might eventually reduce the share of the population that’s infected to, say, 0.5 percent from 5 percent. Unfortunately, the false positive rate won’t budge, so now for every one true positive uncovered, roughly three false positives would still be generated.
-
That’s inherently costly: Every false positive means skipping your flight, rehearsal or birthday party. But it would be catastrophic if people didn’t understand that a positive covid-19 test is a guideline, not a guarantee
-
People who think that they’ve had covid-19, when they haven’t, are apt to go out and engage in risky behavior, maybe a lot of it. And if mass testing ever becomes common, there could be a lot of those people.
-
So mass testing isn’t just a matter of getting a test that is cheap enough and plentiful enough; the administration and the public must be educated to use this bounty wisely. That’s a hard messaging problem. Given the alternative, though, it’s the kind of problem we’d really like to have.