Hacker News Reader

Comment by amluto | Hacker News Reader

This is interesting, but I don’t believe the study should be trusted. The authors determined toxoplasmosis status by asking the participants. This seems like it will introduce huge bias based on the bizarre subset of people who know their status.

A similar study that tested for toxoplasmosis antibodies would be much more useful.

View on HN

I agree it should not be taken as definitive. However, I do think it makes sense to have done this first, since it provides a relatively quick and low-cost way of finding out if there is anything that justifies a more expensive and time-consuming investigation (that yes, should involve a test of toxoplasmosis antibodies).

It appears that the answer is yes, it justifies such a study.

The authors addressed this issue at least somewhat:

"Toxoplasmosis status was self-reported by respondents, which may be viewed as a drawback. On the other hand, it has previously been demonstrated that information on toxoplasmosis status provided by 3,827 participants of another internet study nearly perfectly (99·5%) corresponded to information in our files on subjects tested for toxoplasmosis in our lab."

That doesn't make any sense: the CDC claims "in the United States it is estimated that 11% of the population 6 years and older have been infected with Toxoplasma" and Wikipedia claims "serological studies have estimated that 30–50% of the global population has been exposed to and may be chronically infected with latent toxoplasmosis"... and yet, other than pregnant women, does anyone ever get tested for it? The CDC even goes so far as to directly claim "most people who become infected with Toxoplasma gondii are not aware of it because they have no symptoms at all"!

Presumably, they tested a representative sample at some point and extrapolated

Like the various health authorities in the US tested a representative sample for Covid RNA, antigens, and antibodies to determine unbiased population level data?

/me runs.

You run, but it’s pretty obvious one reason that never happened is because you can use the phrase “best available evidence” to do whatever you want when you control the funding for evidence gathering.

That’s unnecessary. The CDC and the WHO are perfectly willing to make statements that contradict the best available evidence even without deliberately suppressing the production of better evidence. It’s easier to just ignore the evidence.

(I don’t think this is malice. A combination of being unwilling to admit mistakes and simply not being very good at one’s job seems more than adequate to explain this.)

Extrapolated what? I am taking what the CDC is saying here as truth, at which point it seems almost impossible that 99.5% of people would know correctly if they have toxoplasma... it would seem like your accuracy on guessing that should be extremely low. (FWIW, I am guessing that you somehow failed to read the context of the conversation and are saying the CDC tested a sample of the population and extrapolated the number of people who have it... "duh".)

So people who were tested for toxo knew they had it? Scratches head.

That’s not exactly the same issue though. The issue was that the subset who know their toxoplasmosis status is probably not representative for the population at large. This remains a potential problem even if they are perfectly right about their status.

Perhaps people that do know they have Toxoplasmosis, know for sure; but many people have it do not know.

So maybe the risk factor here isn't having Toxoplasmosis, but knowing you have Toxoplasmosis :-) (or more seriously: some other factor correlated with it, like being a veterinary or animal handler which may be required to test regularly)

How can that be possible? 99.5% accuracy in knowing status? That's absurdly high, considering, just for a start, more than 0.5% of people have cognitive disorders that would interfere with accurate surveying.

And how many of those 3827 Internet study participants got tested?

Reply on news.ycombinator.com