Azara Blog: Another study of studies looking at cancer

Blog home page | Blog archive

Google   Bookmark and Share

Date published: 2007/10/31

The BBC says:

Even those who are not overweight should slim down if they want to cut their risk of cancer, a major international study has claimed.

The World Cancer Research Fund carried out the largest ever inquiry into lifestyle and cancer, and issued several stark recommendations.

They include not gaining weight as an adult, avoiding sugary drinks and alcohol, and not eating bacon or ham.

Everyone must also aim to be as thin as possible without becoming underweight.

People with a Body Mass Index (BMI), a calculation which takes into account height and weight, of between 18.5 and 25, are deemed to be within a "healthy" weight range.

There is no new research involved in this document: the panel examined 7,000 existing studies over five years.

The result, they say, is the most comprehensive investigation ever into the risks of certain lifestyle choices.

They see body fat as a key factor in the development of cancer, estimating its significance to be much higher than previously thought.
However, two-thirds of cancer cases are not thought to be related to lifestyle, and there is little people can do to prevent the disease in these circumstances.
Cancers of the colon and breast are some of the most common forms of the disease, and the report says the evidence is "convincing" that body fat plays a key role in the development of these tumours.

The report also links the kind of food consumed to cancers, especially colorectal ones.

In particular, researchers say people should stop eating processed meats, such as ham, bacon and salami, and limit the consumption of red meat to 500g a week - although this still means you could eat, for instance, five hamburgers each week.

From a cancer perspective, all alcohol should be avoided, although researchers accepted drinking small amounts could have protective benefits for other diseases.

The recommendation is therefore no more than two drinks a day for a man, and no more than one for a woman, slightly less than current UK government guidelines.

There are several worrying things about this study. First of all, they seem to have made a classic confusion between correlation and causation. In particular, many of the things they mention are also correlated with wealth (or lack of it). And health is well known to correlate with wealth. So are they just observing yet another correlation or are they really observing a causation? Hardly any health studies are done with randomised sampling. So it is unlikely that any causation has been proven. Of course, there could well be the causation they imply, and since a lot of people, the authors no doubt included, believe this causation holds, they can easily get away with claiming it without most people noticing there is no sound basis for the claim.

Another problem with this kind of study is that it looks at one thing in isolation, here cancer. It's quite possible that things that are bad for cancer are good for other things, as even the panel recognised with alcohol.

Another worrying sign is that the BBC claims the panel examined 7000 studies. Nobody can look at even a small fraction of that many studies with any element of care. So unless the panel contained several hundred people (and then who guaranteed a similar analysis by all concerned?), the analysis must have been fairly superficial.

All material not included from other sources is copyright For further information or questions email: info [at] cambridge2000 [dot] com (replace "[at]" with "@" and "[dot]" with ".").