The misuse of 'no significant difference' in British orthopaedic literature.

INTRODUCTION

Many studies published in medical journals do not consider the statistical power required to detect a meaningful difference between study groups. As a result, these studies are often underpowered: the sample size may not be large enough to pick up a statistically significant difference (or other effect of interest) of a given size between the study groups. Therefore, the conclusion that there is no statistically significant difference between groups cannot be made unless a study has been shown to have sufficient power. The aim of this study was to establish the prevalence of negative studies with inadequate statistical power in British journals to which orthopaedic surgeons regularly submit.

MATERIALS AND METHODS

We assessed all papers in the last consecutive six issues prior to the start of the study (April 2005) in The Journal of Bone and Joint Surgery (British), Injury, and Annals of the Royal College of Surgeons of England. We sought published evidence that a power analysis had been performed in association with the main hypothesis of the paper.

RESULTS

There were a total of 170 papers in which a statistical comparison of two or more groups was undertaken. Of these 170 papers, 49 (28.8%) stated as their primary conclusion that there was no statistically significant difference between the groups studied. Of these 49 papers, only 3 (6.1%) had performed a power analysis demonstrating adequate sample size.

CONCLUSIONS

These results demonstrate that the majority of negative studies in the British orthopaedic literature that we have looked at have not performed the statistical analysis necessary to reach their stated conclusions. In order to remedy this, we recommend that the journals sampled include the following guidance in their instructions to authors: the statement 'no statistically significant difference was found between study groups' should be accompanied by the results of a power analysis.