Do Cochrane and non-Cochrane editors and authors prefer reporting statements based on statistically significant differences or do they prefer non-binary options?
Demián Glujovsky, Institute for Clinical Effectiveness and Health Policy (IECS)
2Reproductive Medicine, CEGYR (Centro de Estudios en Genética y Reproducción), Argentina
Background:
Background: For decades the p value-based interpretation and reporting of results dominated the publications. Nowadays, the scientific community agrees that this binary approach is not enough. A systemic reform is necessary for moving from this binary approach to some other non-binary options. Some of them, that were discussed in the last Cochrane Handbook version (2019), are the point estimate, the confidence interval, and the minimal important difference, suggesting some narrative statements.
However, it is not clear how Cochrane and non-Cochrane editors and authors agree with this approach and, even more important, if all retrieve the same conclusions when they look at some specific results.
Objectives: To evaluate how Cochrane and non-Cochrane editors and authors interpret results and which preferred reporting style is for them.
Methods: We are conducting an online survey among healthcare professionals and consumers. These stakeholders have to choose the binary or non-binary option that better expresses the results for the following scenario:
After exhaustive literature searches, a systematic review identified only two pivotal randomized controlled trials (RCTs) that evaluated the effectiveness of drug X versus placebo (P) in patients with a rare genetic disease. The risk of bias for all domains was low in both RCTs (assessed using the Cochrane RoB-2 tool), and there was no methodological, clinical, or statistical heterogeneity between studies. The meta-analysis yielded the following results:
Mortality risk: X 26%(10/39) and P 45%(18/40)
Risk difference: X 19% lower (95% CI 40% lower to 1% higher)
Risk Ratio: 0.57 (95% CI 0.30 a 1.08) P = 0.0721
*95% CI: 95% Confidence Interval
Results: will be shown at the colloquium.
Conclusions: It will be interesting to see which preferred reporting statements are for the Cochrane and non-Cochrane communities, and to compare the preferences between authors and editors. Besides, we will know the level of agreement or disagreement in each group, which is critical for the efficiency of the editorial process.
Patient, public and/or healthcare consumer involvement: Proper interpretation of study results is critical for the healthcare process