Statistical analysis is an important step in order to validate the hypothesis of any research. Different statistical tests are available for comparisons or correlations between data samples. The interpretation of the results should be based on the correct utilization of the statistical tests and the relevance of these results for clinical analysis. The aim of this review was to analyze the use and interpretation of statistics in the most cited articles published in orthopedic journals.

A Pubmed search was conducted for the most cited articles in orthopedics according to google scholar or Scopus. Articles were included if they were published later than 2005 and if they performed a statistical analysis for the interpretation of their collected data. Two factors were investigated: (1) the correct use of the statistical tests: verification of normality and equality of variances, (2) the correct interpretation of the outcome of these tests: when comparisons between samples were statistically significant, the values of the differences between the means were observed to see if they were sufficiently large in order to be considered clinically relevant.

In total, 44 articles were reviewed. The journals from which they were extracted were: Am J of Sports Med, Arthroscopy, Bone and Joint J, Clin Orthop Rel Res, Gait and Posture, JBJS Am, JBJS Br, JPO Br, JPO, Spine, Eur Spine J, The Spine J, Osteoarthritis Cartilage. Only 4 articles had mentioned the verification of normality before applying the statistical tests. None of the articles had mentioned if the equality of variances was checked. Six of 44 articles presented comparisons between samples with statistically significant results where the values of the between-samples differences were not large enough to be considered clinically relevant.

Since researchers tend to often neglect to verify assumptions about normality and equality of variances (Hoekstra R., Frontiers in Psychology, 2012), it can be assumed that most of the authors who did not mention these verifications could have misused their comparison tests, and therefore the validity of their results could be in question. Attention should be given before trusting statistical tests’ results when small differences show statistical significance. Such results may be misleading because they are not necessarily clinically relevant.