Removing bias in surgical trialsBMJ 1997; 314 doi: https://doi.org/10.1136/bmj.314.7085.916 (Published 29 March 1997) Cite this as: BMJ 1997;314:916
New surgical procedures will continue to lack credibility unless assessed by properly randomised trials with objective outcome measures
- A G Johnson, Professor of surgerya,
- J Michael Dixon, Honorary senior lecturer in surgeryb
- a Department of Surgical and Anaesthetic Sciences, Royal Hallamshire Hospital, Sheffield S10 2JF
- b Breast Unit, Western General Hospital, Edinburgh EH4 2XU
Before a new drug is introduced into clinical practice it has to be shown, by randomised double blind clinical trials, to be at least as effective as currently available treatments. The same has not been true of new surgical procedures. The safety and efficacy register of new interventional procedures (SERNIP),1 recently set up by Britain's Academy of Royal Medical Colleges, provides a framework that should go some way to remedying the defect, but it is still voluntary rather than compulsory. Although comments such as “the personal attributes of surgeons differ from those needed for collaborative multicentre research”2 are unhelpful generalisations, there has been a tendency for surgeons to rely on series of operations without objective assessment of their value. In a careful study of 10 international journals from 1988 to 1994, Hall et al found that, of the few randomised controlled trials that were published, less than half included objective methods for assessing outcome.3 For surgical trials to have credibility, it is vital that assessments are seen to be unbiased and that those taking part are appropriately blinded to which procedure has been done.
Blinding of patients and doctors is necessary in the following three situations. Firstly, …