Background and objectives: Risk of bias is an important factor in interpreting research results. Therefore, risk of bias of randomized controlled trials (RCTs) in oral and maxillofacial surgery (OMS) over a ten year period was evaluated.

Methods: A literature search for RCTs published between January 2000 and January 2010 was done. Articles were assessed on important risk of bias items with two different quality lists, the Delphi List and the Jadad Scale. Regarded as low risk of bias were articles with a Jadad greater than or equal to 4 (range 0–5) and a Delphi score greater than or equal to 6 (range 0–9).

Results: A total of 230 articles met the inclusion criteria and were analyzed. 41 articles (17.8%) could be assessed as low risk of bias. The majority of the included RCTs did not correctly describe items as randomization (53.5%), allocation concealment (61.7%), blinding (76.1%) and intention to treat analyses (79.1%). Median number of included patients was 50 and mean scores for Jadad and Delphi were not improved over time ( P = 0.35 and P = 0.45 respectively).

Conclusion: There is a shortage of low risk of bias research in OMS. Important risk of bias items should be improved in further research.

Key words: risk of bias; evidence based medicine; methodological quality