Risk of bias in research in oral and maxillofacial surgery.

The risk of bias is important in the interpretation of the results of research. The aim of this review was to evaluate the risk of bias in randomised controlled trials (RCTs) in oral and maxillofacial surgery (OMFS) over a 10-year period. We searched databases of publications for RCTs published between January 2000 and January 2010. Papers were assessed with 2 up-to-date logical quality lists, the Delphi list and the Jadad scale. Those papers with a low risk of bias were given a Jadad score >/=4 (range 0-5) and a Delphi score >/=6 (range 0-9). A total of 230 papers met the inclusion criteria, and only 41 (18%) were assessed as being at low risk. Most of those included did not correctly describe such important items for risk of bias as method of randomisation (n=124, 54%), concealment of allocation (n=143, 62%), blinding (n=175, 76%), and intention-to-treat analyses (n=182, 79%). In the fields of implantology, traumatology, obstructive sleep apnoea syndrome, and extractions, no paper had a low risk of bias. This systematic review has shown a shortage of research in OMFS with a low risk of bias published over a 10-year period. Further research should concentrate on better describing items at important risk of bias.