Control of bias in randomized controlled trials published in prosthodontic journals.

STATEMENT OF PROBLEM

Randomized controlled trials (RCTs) have become the gold standard for evaluating the effectiveness of treatment interventions. If not properly controlled, bias in the design of trial methodology can affect the validity of the study results.

PURPOSE

The purpose of this investigation was to assess the methodological quality of RCTs published in 3 prosthodontic journals over a 10-year period.

MATERIAL AND METHODS

Issues of The International Journal of Prosthodontics, The Journal of Prosthetic Dentistry, and The Journal of Prosthodontics published between 1988 and 1997 were searched manually to identify RCTs. Specific inclusion and exclusion criteria were established to identify articles about studies that qualified as RCTs. Two independent reviewers evaluated all qualified RCTs on the basis of how potential sources of bias in the trial methodology were controlled. Three areas-control of bias at entry, control of bias in assessment of outcome, and control of bias after entry-were evaluated with a scheme developed through the Cochrane Collaboration. A score of 1 or 0 was assigned for each of the 3 potential sources of bias, with the maximum quality score for an RCT being 3 (good bias control) and the minimum 0 (poor control). Frequencies were calculated for each dimension of trial methodology and overall quality scores of the RCTs.

RESULTS

Sixty-two RCTs were identified from 3631 articles screened. The method of randomization was explicit in only 47% of the RCTs. Forty percent of RCTs incorporated blinding in the assessment of outcome, and 76% accounted for all subjects at the end of the study. Overall quality scores revealed that only 16% of RCTs attempted to control bias in all 3 areas examined. Forty percent were deficient in 1 area, 34% were deficient in 2 areas, and 10% were deficient in all areas examined.

CONCLUSION

The quality of RCTs published in prosthodontic journals may be improved by minimizing potential sources of bias and adequately reporting trial methodology.