Hacker News new | ask | show | jobs
by ManBlanket 1877 days ago
The first, "alternative explanation" the statistician in me wants to mention is perhaps the way influenza data is gathered has changed.

"Only people who get tested for influenzalike illnesses—typically about 5 percent of individuals who fall ill—are tallied."

If there were a change in the way these tests were administered, for example a blaring medical bias toward another disease, that would present a significant sampling problem. That same statistician also wanted me to mention the simplest and most boring answers are usually closest to the truth.

3 comments

A bias towards testing for Covid wouldn't explain fewer cases of the flu unless those cases of the flu were coming back as Covid falsely instead of so further diagnosis was stopped. If you have a bad enough flue to seek treatment, then once you get that Covid test and it says "negative", you would move on to the next step for treatment. (Consistent with past years - nobody was getting flu tested just cause they had a runny nose in 2019, it was just the people who needed treatment).

There could be other explanations, such as: a desire to avoid Covid causes people to avoid seeking treatment, so more flu cases self resolve. OR: a fear of Covid causes more people to get tested when they're sick, and some of them may then go for flu testing before they would otherwise after coming back negative for Covid...?

Seems like the simplest answer is just "actions that have reduced the spread of one disease have also reduced the spread of another, that's historically less widespread already."

How is this assessment affected by limited testing capacity and generally overwhelmed medical services?
Maybe in the first few weeks there were some cases assumed to be covid that could not be confirmed, but testing capacity quickly caught up.
> Seems like the simplest answer is just "actions that have reduced the spread of one disease have also reduced the spread of another, that's historically less widespread already."

This is not the simplest answer, and the evidence that the measures have really slowed spread is extremely low except in places like New Zealand and Australia which are small islands in Oceania.

---

> A bias towards testing for Covid wouldn't explain fewer cases of the flu unless those cases of the flu were coming back as Covid falsely instead of so further diagnosis was stopped.

Exactly this. I'm partial to the viral interference hypothesis, but what you don't seem to realize is that if you get infected for SARS-CoV-2 and recover in 7-14 days, you will still test PCR+ for months after. This goes into the widespread mistuning of the cycle threshold. Case in point: They tested George Floyd's corpse for COVID-19 and he was PCR+, despite having recovered from COVID-19 a couple months before. The test hit on the remnant viral debris from his long-gone infection.

> This is not the simplest answer, and the evidence that the measures have really slowed spread is extremely low except in places like New Zealand and Australia which are small islands in Oceania.

Have you ever heard of this small Asian country called CHINA?

With a lot of people getting tested for COVID "just in case" and hypochondriacs constantly checking in with their doctors, I'd guess that the percent of flu cases tallied might actually be a little bit higher than usual due to COVID
I wouldn't be surprised if there was more flu testing this year. I've never intentionally taking a flu test, but some of my COVID tests also screened for influenza a/b at the same time.