Data Analysis and Presentation: Writing a Paper for Publication

Abstract

The extent to which your clinical trial will contribute to the greater scientific good will depend, to a large degree, on the quality of the presentation and dissemination of the results. Your trial is likely to be one of many that address the research question you have posed. In some cases, the treatment effect will be overestimated and results, particularly from small trials, will be contradictory. The results from a number of trials will probably have to be combined to get a true picture of the effectiveness of a new molecular entity (NME). Ideally, the report of your trial will be of sufficient quality to be included in a metaanalysis and demonstrate the effectiveness of your intervention in the treatment of osteoporosis. There are, unfortunately, a number of limitations that are common when writing up trials that lead to bias and the exclusion of studies from subsequent metaanalysis, including the following: 1

• 1. Use of multiple endpoints (if 20 items are measured on a subject, one is bound to be significant—result: a publication)

• 2. Use of surrogate endpoints [e.g. bone mineral density (BMD) as a surrogate marker of fracture risk]

• 3. Too many subgroup analyses

• 4. Incorrect analysis of repeated measures

• 5. Too many treatment groups in one study

• 6. Small study numbers

• 7. Underreporting of nonsignificant results.