Search tips
Search criteria 


Logo of bmjThis ArticleThe BMJ
BMJ. 2007 September 8; 335(7618): 472.
PMCID: PMC1971183
Shortcuts from Other Journals

Evidence for guidelines often lacks external validity

Guidelines for clinical practice have proliferated recently and most say they are backed up by research findings. A study of 338 recommendations from nine guidelines for management of cardiovascular risk found, however, that fewer than one third were supported by high quality evidence, as assessed by a grading scheme that considered the external relevance of the research to the target group and outcome, as well as its internal validity.

The study examined Canadian, European, and American guidelines for diabetes mellitus, dyslipidaemia, and hypertension and excluded recommendations for diagnosis and prevention, as well as those for pregnant women and people admitted to hospital. The quality of evidence was graded with a four point scale using the Canadian hypertension education programme (CHEP) scheme.

Overall, 231 (68%) of the recommendations were backed by randomised controlled trials, but in only 105 (45%) of these was the evidence from trials judged to be of high quality. Concern about the applicability of the study to the target population was the most common reason that evidence from randomised controlled trials was given a suboptimal grade under the CHEP scheme (64 of 126 recommendations, 51%). The use of surrogate outcomes, rather than patient oriented outcomes, led to downgrading in 59 of the 126 recommendations (47%).


  • PLoS Med 2007;4:1325-32

Articles from The BMJ are provided here courtesy of BMJ Publishing Group