What is Outcome Reporting Bias?
The Cochrane Handbook defines it as ”The selective reporting of some outcomes but not others, depending on the nature and direction of the results.”
It identifies several types of Outcome Reporting Bias including:
- Selective omission of registered outcomes
- Selective choice of data for an outcome
- Selective reporting of different analyses of the outcomes
- Selective reporting of subsets of the data
- Selective underreporting of data with inadequate details such as simply non-significant
Result = Missing Data = Missing Knowledge
What is the Impact of Outcome Reporting Bias?
Outcome reporting bias not only impacts the validity of the single trial in which results are withheld or substituted. It has cumulative impact on the body of evidence.
How common is Outcome Reporting Bias?
Only 13% of trials published in the top 5 med journals (based on impact factor) from October 2015 to January 2016 matched all primary and secondary outcomes specified in their protocols or registries. Ioannidis BMJ 2017;356:j408 – The COMPare Trials Project
Learn more at the COMPare project website http://compare-trials.org/
How Do We Prevent Outcome Reporting Bias?
For researchers - Register all trials and ensure the protocols include outcome description and planned analysis following CONSORT standards. www.consort-statement.org
For Journals - Follow ICMJE standards that require trial registration at or before the time of first patient enrollment as a condition for publication. www.icmje.org
Preventing outcome reporting bias improves our ability to determine whether or not an intervention is effective and safe and also reduces research waste.