Meta-analysis in dental research: a paradigm for performance and interpretation.

Applications of meta-analysis have begun to appear with some regularity in the dental research literature. Recently, a meta-analysis played a high-profile role in the ongoing academic debate concerning the relative anticaries efficacy of dentifrices containing fluoride as sodium fluoride and as sodium monofluorophosphate, engendering a controversy concerning the methodology employed. This has given rise to the need for a careful consideration of the principles involved in the meta-analytic process to provide a basis of understanding upon which such controversies may be resolved. The present report endeavors to meet this need by enumerating the steps involved in a meta-analysis, and contrasting them with the analogous steps in the more widely-understood "usual" inferential process, bringing to light both the similarities and differences between these two types of analyses. Within this context, the conflicting meta-analyses associated with the current controversy are discussed and compared, and the source of the difference between them is readily identified. This process reaffirms the need for close collaboration between statisticians and other scientists in the performance of a meta-analysis, and the value of such close collaboration in the evaluation and interpretation of a meta-analysis performed by others, as well.