Hi Anoop,

An interesting question in several ways:

1 A reflection on how we react when evidence conflicts with deeply ingrained beliefs: What, over weight people live longer than 'ideal weight' people? There must be something wrong with this study. 

2 How we non-epidemiologists (evidence-based clinicians as end-users of epidemiology) approach disagreement between experts in epidemiology. Can we take a stand? We can and must even if it is only to acknowledge uncertainty.

3 An appraisal of the evidence itself.

Picking up number 3, I had a look at the paper and did not find anything to justify Willett's assertions that there was no consideration of confounding factors. For example, under data extraction, the authors stated;

"In studies that only presented results stratified by smoking or health condition, we selected results for nonsmokers or never smokers or for those without the health condition."

…and in considering the quality of the analyses in the studies they stated:

"We considered the results adequately adjusted if they were adjusted for age, sex, and smoking and not adjusted for factors in the causal pathway between obesity and mortality, or if they had reported or demonstrated that adjustments or exclusions to avoid bias had shown little effect on their findings."

Now there may be some finer points that elude me so I look forward to hearing from those with more epidemiological knowledge. However, I remain unconvinced that there was no confounding because I am led to ask, What did they do about studies that did not report stratified data or give adequate descriptions of adjustment? And I don't like the fact that they included publications with some overlap in populations. So returning to points 1 & 2 above, my conclusion is that there is probably something wrong with this study but I cannot dismiss it as a pile of anything.

BTW, I can email a copy of this paper for those who want it so long as it is for personal study only.

Dr Kev (Kevork) Hopayian, MD FRCGP
General Practitioner, Leiston, Suffolk
Hon Sen Lecturer, Norwich Medical School, University of East Anglia
Primary Care Tutor, Suffolk

RCGP Clinical Skills Assessment examiner

http://www.angliangp.org.uk/

On 24 Mar 2013, at 04:27, Anoop Balachandran <[log in to unmask]> wrote:

I am curious to know anybody has any comments about Flegal's controversial paper on mortality and BMI:

Here is the link: http://jama.jamanetwork.com/article.aspx?articleid=1555137

Harvard epidemiologist Walter Willett says that the paper is a "pile of rubbish":http://www.hsph.harvard.edu/nutritionsource/questions/ask-the-expert-does-being-overweight-really-decrease-mortality-no/