Statistical analysis and study design in plastic and reconstructive surgical research.

Appropriate study design and proper statistical analysis are necessary ingredients for improving the quality and reliability of the information in journal articles. General surgery and plastic surgery articles were compared for principal author's academic degree, a Ph.D.'s presence as a coauthor, the study type, the presence of statistical analysis, the analysis' appropriateness, and the types of errors in study design or statistical analysis. Ph. D. authorship was associated with increased percentage of articles using statistical analysis. When compared with general surgery articles, plastic surgery articles performed four times fewer statistical analyses. However, when statistical analyses were performed, there were few differences between these two specialties. Although there were no differences in the types of statistical analysis errors, there were differences in the types of study design errors. The causes of these discrepancies may lie in the nature of plastic surgery; they may be reduced by adherence to Feinstein's principles of study design and result interpretation.