|
Affiliation
|
School of Public Health, Physiotherapy and Sports Science, University College Dublin, Dublin, Ireland. |
|
Abstract
|
Can a healthcare intervention appear to have a substantial impact on an important outcome metric without really making a difference?
Understanding whether an intervention can really increase the likelihood of a positive outcome is essential when translating research findings into clinical practice. Selectively reporting outcome metrics in randomised controlled trials (RCTs) can mislead readers to believe that a healthcare intervention is more effective than it truly is. Interpreting whether healthcare interventions should be prescribed by clinicians requires not only critical appraisal by the reader, but also transparent reporting of outcome metrics by researchers.1 This seems like the proverbial ‘no brainer’, particularly as there are recommendations for reporting outcome measures.2
Best practice journal policy requires that RCTs must be registered and have their outcome domains (eg, injury), outcome measures (eg, injury risk) and associated outcome metrics (eg, relative risk) listed very specifically.2 Sadly, journal editors rarely police these reporting behaviours so investigators can, and do, choose to selectively present their most compelling results in the most impressive manner. Cherry-picking outcome metrics can mislead the reader (and/or policy maker) to think that an intervention ‘worked’ very effectively when it fact it only worked to some extent ...
Language: en |