Reporting bias causes rare diseases to seem common

[Claesand]

Just thought I'd share a thought I've had about looking up stuff on the Internet and health anxiety... how it can mess with us assessing the likelihood of things.

There is a reporting bias that causes serious diseases to seem much more common than they are. Also, there is a psychological effect from reading case reports that is quite insidious - the same that causes us to buy lottery tickets against very long odds.

Medical science is of course concerned with serious diseases. So, if 100,000 people visit the doctor, and 10 turn out to have some disease, guess which we are going to read about? Those 10 of course... "The patient, XY a male of 32 years presented with a slight dizziness..." and so on. Could be you or me right? What it doesn't tell you is that the other 99,990 guys with a slight dizziness had just that: a slight dizziness. It's like seeing pictures every week of big lottery winners - feels like it's much more likely than it really is.

So if you search for "lump in the throat" or "dizziness" or "numbness" or something you'll turn up a list of diseases, and you are led to believe that you must have one of them. You don't. Medicine lists classified diseases. You should of course see your doctor on a regular basis... but be assured that they see all sorts of things, and the vast majority of all the things they see turn out to be harmless.

[bluebell68]

Hi that very true.. the daily newspapers are full of such stories and headlines like '"Hairdye linked to" .... (insert some scary disease).. or "drinking wine can double your risk of...." etc etc....

They refer to 'a study having shown that something is increasing our risk of this or that and never refer to the numerous other studies that show no such link..... this is scary, they quote percentage risk increases etc that are so misleading and scare mongering at its worst....like one yesterday saying if you use hair dye you have increased your risk of something or other by x%...can't remember the details cos its all rubbish anyway....

People on the this forum will all advise Never Google Symptoms!!....
.... well i would add to that by saying never believe any story in a newspaper that starts with the words "a recent study has shown....." cos next week there will be a new one telling us the opposite is true.
Best Wishes
Rachel

[Claesand]

Yes, very true... Media makes money off of peoples' health anxiety. I'm a scientist myself, so I use statistics quite a bit, and I also know both the good and bad drivers in science.

One is that you are more or less forced to publish a lot, that's how you build a CV as a scientist. The consequence is that scientists publish also when they have absolutely nothing worthwhile to say! Making statistical correlations in databases is a treasure trove from which to create an endless stream of articles... so scientists do this the whole time. It happens, for sure, that useful things are found this way, but most of it is junk... and as you say, the next day someone "proves" the opposite.

Also, statisticians typically get nightmares from looking at how medical and social scientists use statistics on data. There is no pressure to be thorough: it's a lot of back scratching there, you go easy on me and I'll go easy on you... we all need to publish.