Comment by Alex3917 | Hacker News Reader

Comment by Alex3917 | Hacker News Reader

Parent Full thread

Yes. For something to be statistically significant means there is some difference between the two groups that is most likely not attributable to random chance. For something to be clinically significant means the difference between the two groups is large enough to have noticeable health effects or impact one's decision-making about medical issues.

In the book The Emperor's New Drugs, about the efficacy of anti-depressants (or lack thereof), Kirsch gives the example that if you had a group of 500,000 people and found that smiling frequently increases life expectancy by 10 seconds on average, this may well be statistically significant because the study population is so large. However, it's not clinically significant because no one really cares if they live an extra 10 seconds, certainly not enough to change some major aspect of their behavior at least.

In this case, if the chances of having kids with autism or schizophrenia increase by/to? at most 2%, as the article states, then in practice this shouldn't really deter anyone from having kids. Thus, while the findings are statistically significant, they are not clinically significant.

I think 2% is actually a very high risk for something with rather severe consequences. Still, the desire to have kids is probably higher in most men.

"I think 2% is actually a very high risk for something with rather severe consequences."

Keep in mind that the overall risk of autism is 1 in 88, or 1.1%, and the overall risk of schizophrenia is about 1%. So it's not like you get a free pass if you have kids earlier. Also, there are other benefits to having a child later in life in terms of things that benefit the child, even if there are more neurodevelopment risks and disadvantages at the biological level.

Sure, I just meant in general 2% seems very high. Friends of mine got the suggestion to do a test with a 3% chance of killing the embryo (or what they call it - it was rather early stage). The doc apparently thought nothing of it. I thought it is crazy (and they thought so too). Added information that the mom was already 40, so not that many new attempts left.

So is clinical significance determined in part by perceived significance? I feel that a potential 2% risk of conditions as serious as those mentioned in the article could easily influence some people towards choosing to have kids at a younger age.

"So is clinical significance determined in part by perceived significance?"

It's determined entirely by perceived significance. For each type of drug, there is usually some agreed on metric within the field that determines what constitutes clinical significance. E.g. a blood pressure drug might have to lower blood pressure by 5% or 10% or whatever. And often times the metric isn't even relevant, as in the case of a reduction in blood pressure vs. a reduction in total mortality / morbidity.

However, at the end of the day what the academic community considers clinically significant doesn't really matter, it's your own preferences and values that should be your guide. Especially since in practice everyone is basically out to trick you in order to make more money, from pharma companies to newspapers to academic journals to professional organizations (AMA, APA, etc.), etc. That's why it's important to really understand this stuff forwards and backwards and to keep your wits about you.

Reply on news.ycombinator.com