- synopsis of parts of an article in forthcoming book, Evidence, Inference and Enquiry à In Praise of Randomization
- Sausage a Day can increase bowel cancer risk.
- 2007 report, Food Nutrition, Physical Activity and the Prevention of Cancer, a Global Perspective.
- Never been a randomized trial to test the carcinogenicity of bacon
o how strong is the evidence
o turns out, surprisingly flimsy.
In Praise of Randomization
- confusion of sequence and consequence
- confusion of correlation and cause
- Lady tasting Tea experiment (tea test)
- health studies
o case control
§ least satisfactory à retrospective
o cohort
§ better à prospective
§ big effect à can give very good indication
o Randomised controlled trials
§ best
· only RCT can demonstrate causality
- randomization
o can remove in statistical sense à bias that might result from all sources you weren't aware of.
o if you are aware of a bias, measure it.
o guarantees freedom from bias in long run statistical sense.
- case of hormone replacement therapy
o Harvard Nurses Study à prospective chort
o Women's Health Initiate Study à randomized double blind trial
Case of Processed Meat
- recommendations only make sense insofar as various dietary factor cause cancer.
- If association is not causal, changing diet won't help.
- 12 propective cohort studies showed increased risk for highest intake group compared to lowest
o statistically sig only in three studies
o tendency for relative risk to be above one, though not by much
- Meta analysis possible on 5 studies
o relative risk of 1.21
§ CI 1.04 to 1.42
- consistency suggests real association, but this cannot be taken as evidence for causality
o observational studies on HRT were just as consistent, but were wrong.
- outcome from vast number of observations only just reaches p = 0.05 level of statistical sig
- 2 more criteria might help
o good relationship between dose and response
o plausible mechanism
- no dose – response relationship apparent
- study on vegetarians and cancer
- pizza may simply represent a general and aspecific indicator of a favourable Mediterranean diet.
Is the Observed Association Ever Real
- real associations likely to be exaggerated in size à why
What do Randomized Studies Tell us
So Does Processed Meat Give You Cancer
- only sound guide to causality is properly randomized trial
- only exception when effects are really big
- own inclination is to ignore any relative risk based on observational data if it less than about 2
Dangers of being too pre cautionary
- ignore information that has sound basis
- excessive medicalisation of everyday life
- brings science into disrepute.
RESPONSES
observational epidemiology
Dr John Briffa
Michael Marmot
risks of second hand smoking
it doesn't matter if association is real if it isn't causal, because if it is not causal, it can't lead to useful action.
may be better to say we don't know
different types of sausages
whole idea of meta analysis is to look at all data
in observational studies, the statistical significance is not nearly as meaningful as the cofounders – those other factors that might explain the statistically significant effect observed.
how might high meat eaters differ from low meat eater à good point.
the reason to do randomized trials is to render irrelevant all these possible cofounders
it always struck me that the very fact of having to do a meta analysis is pretty convincing evidence that the effect you are trying to nail down isn't real.
no amount of maths will extract information that isn't in the data
huge numbers of subjects helps in getting significant effects but with the risk you may end up simply detecting smaller and smaller associations that are ever more susceptible to misinterpretation because of cofounders.
information extend data à Bayesian
jaynes book à logic
No comments:
Post a Comment