| 1714 A review of the reported methodological quality of split-mouth studies | ||
|
E. LESAFFRE1, M.-J. GARCIA ZATTERA1, C. REDMOND2, H. HUBER2, and I. NEEDLEMAN3, 1Catholic University of Leuven, Belgium, 2University of Pittsburgh, USA, 3UCL Eastman Dental Institute, London, England, Uk Objectives: To review the current dental literature on split-mouth studies with respect to the reported quality of the statistical methodology used. Methods: PubMed was used as the selection pool of papers published in 2004 containing the keyword “split mouth”. All dental journals were considered. From the 88 papers that were initially selected 34 were considered for comprehensive review. Results: Nearly twenty years ago, the review by Hujoel and Moulton (1988) revealed that more than one-half of the reviewed papers used an incorrect statistical method. In subsequent reviews the same authors considered the validity and efficiency of split-mouth periodontal studies. Here, we have (re)evaluated the methodological quality of split-mouth papers both in terms of the methods employed and the quality of reporting utilizing an extension of the CONSORT guidelines developed for cluster randomized designs. We have found that many papers lack essential qualities of good reporting, for instance only 5 out of 34 papers gave the rationale of why a split-mouth design study was chosen. Only 4 papers contained a sample size calculation, and 1 paper took into account the correlation among the responses. Details on randomization, concealment of the randomization procedure and blinding were given in less than one-half of the papers. About one-half of the papers reported a statistical approach appropriate for paired designs. Only one paper reported on the correlation among the primary response variables within a subject. Five papers indicated that there might be a possible carry-across effect. Finally, as in many dental and medical papers no mention was made of the multiple testing problem. Conclusion: Despite the progress, there is still a substantial need for improvement in the methods employed. For example, none of the papers with a survival outcome used the correct methodology taking into account the paired nature of the data. | ||
| Seq #194 - Caries Diagnostic, Lasers, Fluorosis 2:00 PM-3:15 PM, Friday, March 23, 2007 Ernest N. Morial Convention Center Exhibit Hall I2-J | ||
|
Back to the Cariology Research Program
| ||