|
|
 |
STATISTICAL GUIDELINES
For randomized clinical trials and other comparative studies, authors should carefully follow the ICMJE and CONSORT statement guidelines. To produce consistent results, please refer to the JCO’s CONSORT template.
- Submit a CONSORT diagram for studies in which two or more groups are compared.
- For randomized clinical trials, report two-sided P values.
- If the trial design was Bayesian, used one-sided P values, or was not based on a standard two-sided test, the primary result and conclusion may be described in terms of the original design. However, also include the two-sided P value of a standard test.
- Report actual P values, not, for example, P >.05. Two digits are sufficient for P > .01.
- Report outcomes with both absolute and relative differences and confidence intervals where appropriate.
Authors may report differences in terms of odds ratios or hazard ratios, but should also include a comparison of an absolute difference of the primary measure (eg, difference in survival at 5 years or difference in proportions of successes).
When studies report prognostic markers (based on clinical, etiologic, or genetic factors), JCO gives highest priority to studies in which an algorithm obtained in a training set is applied exactly the same way in the validation set as in the training set. That is, the factors included and the cutoffs must be determined in the training set and applied to each patient in the validation set. Authors should clearly identify further statistical analyses as exploratory.
JCO gives lower priority to prognostic studies that report the results of an algorithm from the same data that were used to fit the algorithm. However, inclusion of cross-validation analyses and a clear statement of the limitations of the study will enhance the article’s priority.
JCO considers reports describing data from multiple studies to be a true meta-analysis, only if the author analyzed the data separately and attempted to include study data that were not reported. Studies that apply meta-analytic techniques to published data will have lower priority, and authors should include a discussion of the limitations.
Reports of studies with multiple or secondary end points should address the multiple comparison issues and describe the exploratory nature of the studies.
|
