HRT is useful for some women, but it can also cause problems (like increase in cardiovascular incidents); however, when researchers looked at data from the Nurses’ Health Study, which just looked at lots and lots of nurses, observed what they did and tried to crunch the maths to figure out how they got on, those researchers thought they caught something: women who took HRT were generally much healthier than women who didn’t.
“Ah ha!”, they thought, “these women are being made really healthy by their HRT! We should tell all women to take it immediately!”
Sadly, they mistook cause from effect. What this really showed was that women who looked after themselves, who ate well, who exercised and had time to think about their health ALSO happened to be women who didn’t smoke, didn’t drink too much, exercised, heeded their doctors more carefully, looked at the latest treatments available - one of which was HRT! After all, plenty of these women who took HRT were doing so such that the HRT could help them to continue their active, healthy lifestyles. Women who were prepared to settle for the sofa and some cigarettes were less interested in such treatments.
So it wasn’t that HRT was making these superwomen healthy: it was that their general healthy attitudes were leading them to be early adopters of HRT. And, indeed, they were such a healthy bunch that their high health hid the fact that HRT, given to women at large and indiscriminately, actually could cause health problems, particular ones to do with heart incidents.
When proper intervention studies were done on HRT, the effects became clear - those Nurses had led to the opposite conclusion that should have been, and women should not have been told to take HRT indiscriminately as, for a while, they had been. It’s been calculated that this bad, broad-brush advice based on flawed studies like the Nurses’ Health, led to the death of thousands of women prematurely. So this stuff matters!
The moral of the story is: if you only have epidemiological, observational data, which mixes millions of variables, and you can’t tell cause from effect, it’s not just pseudoscience, it’s DANGEROUS PSEUDOSCIENCE to draw any firm conclusion. Remember that next time someone tells you red meat gives you cancer or any other such nonsense too!
If you want to read a more detailed summary, have a look at this link:
HRT and study bias