BMJ  2006;333 (29 July), doi:10.1136/bmj.333.7561.0-d

Don't believe the relative risks and odds ratios in abstracts

Significant P values for relative risks and odds ratios in abstracts of published studies should generally be disbelieved, says Gøtzsche on page 231. He compared the distribution of P values in abstracts of 260 randomised trials with that in 260 observational studies (130 cohort studies and 130 case control studies) and checked P values between 0.04 and 0.06. A high proportion of abstracts of randomised controlled trials and observational studies had significant results, although many of these results were derived from subgroup analyses or biased selection of results. Also, errors in analysis and reporting were common.

CiteULike    Complore    Connotea    Del.icio.us    Digg    Reddit    Technorati    What's this?

Relevant Article

Believability of relative risks and odds ratios in abstracts: cross sectional study

Peter C Gøtzsche
BMJ 2006 333: 231-234. [Abstract] [Full Text] [PDF]