How reliable is medical research?

[MummyToSteven]

Just a follow on out of curiosity from a couple of comments on the bfing threads, and I know that some people on here have research and/or healthcare backgrounds so would know the answers to this?

How reliable is medical research on humans?

What is a good minimum sample size?

Hmm... I thought about this a while ago and discussed it a bit.

Here and here are my thoughts on research (in this case relating to breastfeeding research, but they sum up my general thoughts on medical papers and research).

Warning: they are long and boring posts... ;)

as a researcher i would have to say that no research is without its faults no study is 100% perfect you have to make compromises. one of the key issues with pregnancy and child rearing is that unless you have twins and do one thing for one and another thing for another you can never be 100% certain of the results. If you used a TENS machine in labour say then you don't know if the labour would hav ebeen longer and more painful without because you can't repeat the same experiment again, even the if it was thesame women in labour 2 years later. Same with breastfeeding you can't reverse time and therefore see whether your baby would have been bigger stronger less prone to bugs if you had/hadn't breastfed.

However i do think research is vaild but you have to study the whole arguement not just chose the data which fits your thinking.

I am a pharmacologist and medical writer and have worked in the pharmaceutical industry for over 10 years. A lot of my work is assessing research. It is certainly very variable. Usually one study, no matter how good, is not enough proof. Research needs to repeatable with the same results.

There is no minimum sample size because it depends what the question is. You would not need many people to show that paracetamol reduces headache. But you need thousands to show that treating blood pressure prevents stroke. Researchers do a calculation based on what they think will happen to determine how many patietns are needed.

Statistics are very complicated and I would need a degree to understand them fully - luckily my BF is medical statistician! You should also have a control group. So this group of children have had MMR, this group haven't - what is the rate of autism in each group?

All in all, you can never say something is absolutely proven. Only that the balance of evidence suggests a theory is true or not. HTH

strawberry - you want to speak to Gobbledigook - she does the same sort of work!

As I've said before on other threads - the annoying thing is - and makes me disbelieve half the things I read - is the controversy over HRT - Research 1 says yes it does this - Reseeaerch 2 says no it doesn't - Research 3 says yes it does - for goodness sake how are we supposed to make a valued judgment on that kind of research.

I'd recommend the NELH Hitting the Headlines section, personally. I find it useful when a health story has been in the media because it at least tells you the details of the research being reported, without having to read the research paper itself. I think there was a piece about HRT - the confusion with that was mostly about HRT having a different balance of effects on women of different ages, IIRC. maybe not, it's a while since I read it.

For me the question is not so much 'how accurate is medical research' but 'what alternative is there'? I don't believe there is anything more accurate than the scientific method out there, but scientists are only human and so there is no such thing as a perfect research project.

Strawberry - hellooooo! I'm a pharmacologist too (and physiologist in a way - degree was combined studies physiology and pharmacology) and I do medical writing and pharmaceutical market research (so helping pharma co's make critical business decisions). Also used to work at the Medicines Control Agency/Committee on Safety of Medicines in pharmacovigilance (drug safety).

Do you work freelance as a writer?

Sorry to hijack thread a bit - just really exciting to hear of someone with similar background!

The site acnebride linked is really good. this is the organisation that sets clinical guidelines for the NHS - not beyond criticism but their guidance does go through a lengthy process of review. Also Drug and Therapeutics Bulletin is an independent review published by the Consumers' Association that analyses drugs and medical/surgical interventions ? it's aimed at healthcare professionals but they do publish treatment notes, a version for patients, here DTB .
There are lots of issues about medical research, starting from the design of individual studies, to publication bias (which studies are offered for and chosen for publication by peer-reviewed journals and therefore cited by other researchers), to the pharmaceutical industry suppressing negative studies or funding jollies for doctors to encourage prescribing of their drugs. DTB has a nice stat that new drugs always turn up with loads of evidence showing that they are enormously effective, yet after ten years research shows they are little better than a sugar pill - you can plot the decline on a graph and it holds good as a general rule.

Hi Gobbledigook. I have sent you a message via CAT.

Hi again - not got it yet but will obviously reply when I do!