Medical studies are seriously biased by interested funders and by tolerance for sloppy methods. Here are four examples.
1. A recent PLoS Medicine looked at 111 studies of soft drinks, juice, and milk that cited funding sources.
22% had all industry funding, 47% had no industry funding, and 32% had mixed funding. … the proportion with unfavorable [to industry] conclusions was 0% for all industry funding versus 37% for no industry funding .
2. Last February the Canadian Medical Association Journal reported that in 487 studies, those whose method left more room for fudging "found" higher accuracy of diagnostic tests:
The quality of reporting was poor in most of the studies. We found significantly higher estimates of diagnostic accuracy in studies with nonconsecutive inclusion of patients … and retrospective data collection … Studies that selected patients based on whether they had been referred for the index test, rather than on clinical symptoms, produced significantly lower estimates
3. In 1995, the Journal of American Medical Association reported that of 250 studies of treatments, those with easier fudging similarly "found" stronger effects:
Compared with trials in which authors reported adequately concealed treatment allocation, … Odds ratios were exaggerated by 41% for inadequately concealed trials and by 30% for unclearly concealed trials … Trials that were not double-blind also yielded … odds ratios being exaggerated by 17%
4. In 2005, the Journal of American Medical Association found that of medical studies since 1990 cited 1000 times or more, 1/3 were contradicted by replications, and 1/4 had no replication attempts:
Of 49 highly cited original clinical research studies, 45 claimed that the intervention was effective. Of these, 7 (16%) were contradicted by subsequent studies, 7 others (16%) had found effects that were stronger than those of subsequent studies, 20 (44%) were replicated, and 11 (24%) remained largely unchallenged. Five of 6 highly-cited nonrandomized studies had been contradicted or had found stronger effects vs 9 of 39 randomized controlled trials (P = .008).
The obvious question is: how can we produce medical estimates that correct for such biases? And why don’t we?