Quote:
|
Redlan, regarding weight loss I can point you to more rubust evidence, at least in my opinion, that LC beats Low fat by a mile.
|
Yes please
I've not read Taubes (and have no intention of buying his books) - however I understand that he first off challenges the diet-heart hypothesis, which is rather easy to do, as the evidence to support the notion that saturated fat intake causes heart disease is non-existent. He then goes on to hypothesis that carbohydrate is the cause of the current obesity/health crisis, which is where myself and Taubes stop agreeing with each other.
now onto to this "robust" evidence and why it isn't very robust at all, and why it does not get included in the kind of reviews that I cite.
The problem with the Cochrane report that I cited and the reason that only a few test papers with a handful of papers were included were because the studies were of too poor quality to be included.
In the original 2004 paper only 36 papers out of a total 287 papers were considered to be of high enough quality to be considered. These 36 papers covered only 18 studies. In the 2007 update an additional 68 papers were reviewed and NONE of them were considered to be of high enough quality to be included.
Why wouldn't a chart study (i.e. analysis of treatment by a clinic) be considered to be high enough quality to be included? The reason is simple - clinic studies automatically suffer from selection bias. There is selection bias in the people that present at the clinic, and there is selection bias in the treatment that the clinician decides to give that patient. Secondly there are no controls - there is nothing to measure the treatment you are offering, and this is perhaps the most important reason why clinic studies are considered to be too biased to be used.
Analogy....
There have in the past been many treatments suggested for the common cold, such as taking vitamin C. Once symptoms develop, a cold typically clears up in around 3 days. The average person normally seeks treatment when the symptoms are the worst, and buys flu magic powder no 9 (or whatever the latest remedy is called), and low and behold they get better in around 3 days. How they discovered that colds typically take around 3 days to clear up was by running studies in which they gave one half magic flu powder no 9, and the other half a placebo, and lo and behold they found it made no difference. The notion that vitamin C improved recovery times was slain this way.
So here you are, you've been struggling with your weight, and diabetes, and so you get recommended to go to a clinic, who show outstanding results (selection bias), you go there and they recommend a low carb diet (another selection bias) - the clinic does not recommend any other kind of diet - and low and behold you start losing weight, and your diabetes improves. You are hailed as a success by the clinic, and your results get published as clinic study, and the low-carb diet is hailed as the cause.
But was it the low-carb diet?
It could be something very much akin to the placebo effect. On your own, without the support of the clinic, you struggled. Maybe your calorie counting was off, maybe you had days when you slipped. You go to the clinic, you have faith in the doctors and specialists, your really believe the treatment will work. You don't want to let down your doctor, you're also aware of the financial investment in you going, so you try really hard, you stick to the diet, and don't deviate - you lose weight.
now what if the clinic you went to specialised in a low-fat diet instead?
The only way to be sure it wasn't some psychological effect from the doctor-patient relationship, is to run a control - one half (randomly assigned) gets the low-carb diet, the other half get the low-fat diet. If they turn out the same, then...
Clinic studies never have a control, and so will never meet the criteria for studies such as the cochrane review. They are simply too open to bias to give a reliable answer.
and finally...
Quote:
|
The trials you cite usually (not always) involve a hand full of test subjects overseen by a biased researcher for a few weeks.
|
I think that's rather inflammatory, and without justification. The Cochrane Collaboration have a good international reputation for providing systematic reviews of randomized controlled trials.
