Log in

No account? Create an account
Off in the distance
my journal
May 2016

The Bellinghman
Date: 2006-09-11 10:27
Subject: A badscience nomination ...
Security: Public
Mandarins 'cut liver cancer risk'.

So, there are two studies. One is conducted by 'the National Institute of Fruit Tree Science'. One wonders if they might have a certain expectation of the study's results.

The second is a study of 30 - yes, just 30 - to see how many had got liver cancer after a year, compared to a control group of 45. So, not exactly a very big study. Indeed:
Ed Yong, of the charity Cancer Research UK, welcomed the research but said the sample size needed to be larger.

He said: "With only 75 people, this study is far too small to tell us anything conclusive about the links between mandarin oranges and liver cancer.
So not exactly very useful. There's an awful lot of bad science to do with nutrition. For example:
Contrary to what the pill-peddlers would tell you, the evidence for omega 3 pills being beneficial in children is really rather thin: only a handful of small trials have been published in proper journals, and at last count 3 were positive, 2 were negative, and none were in mainstream children.
So Omega 3 pills aren't particularly useful, either. Or perhaps they are. But without actually having decent quality studies, we can't know whether it mayn't even be harmful.

Similar caveats apply to high cholesterol foods - a high level of (certain types of) cholesterol in your bloodstream is bad for you, but most of your cholesterol is synthesized in your body, not gained from the food you eat.
Post A Comment | 2 Comments | | Flag | Link

User: perdita_fysh
Date: 2006-09-12 08:25 (UTC)
Subject: (no subject)
Isn't that what various schools have tried to establish this year?
Reply | Thread | Link

The Bellinghman
User: bellinghman
Date: 2006-09-12 08:33 (UTC)
Subject: (no subject)
Ah, yes, the Durham "trial".

It'd be nice if that was a properly structured trial, with double blind trialling and all. But, despite the press reports, it's apparently not.

(As one commentator has said, if it might work, then please do your double blinding at the mocks stage, not at the point when it would tilt the playing field.)
Reply | Parent | Thread | Link