Commentary: Noninferiority trial: The devil is…

Central MessageInvestigators of noninferiority trials need thorough understanding statistical and technical peculiarities for their results to have valuable clinical influence.See Article page 2119. Investigators influence. See There are 3 main types based on the hypothesis tested. Superiority compare 2 treatments determine which is better, equivalence test whether a new treatment as effective standard treatment, aim demonstrate that experimental not unacceptably worse than treatment. These latter studies being more embraced1Suda K.J. Hurley A.M. McKibbin T. Motl Moroney S.E. Publication trials: changes over 20-year interval.Pharmacotherapy. 2011; 31: 833-839Crossref PubMed Scopus (29) Google Scholar because they allow investigators study can bring an advantage in terms costs, patient compliance medication, fewer adverse effects, or greater availability.2Piaggio G. Elbourne D.R. Pocock S.J. Evans Altman D.G. CONSORT GroupReporting randomized extension 2010 statement.JAMA. 2012; 308: 2594-2604Crossref (759) Falk Friede3Falk V. Friede Noninferiority what's clinically (ir) relevant?.J Thorac Cardiovasc Surg. 2021; 161: 2119-2123Abstract Full Text PDF (4) focus aspects with prodigious attention details. The authors describe critical step choosing margin, value effect measure must exceed when compared prove its noninferiority. margin should be oriented from both point view prospectively calculated clearly reported protocol study.4European Medicines AgencyGuideline choice non-inferiority margin.www.ema.europa.eu/pdfs/human/ewp/215899en.pdfDate accessed: April 6, 2020Google Once set, trial null will difference between strategy equal assumed whereas alternative state efficacy differ by no margin. It important acknowledge different hypotheses tested superiority inferiority interpret accordingly. In trial, acceptance hypothesis—the superior one—does us draw conclusion this would lead biased estimate. Conversely, rejected, we assume noninferior comparator only irrelevant effects. But what does mean patients? A practical example may help. Would you accept buy house located less-safe neighborhood just it less expensive another one safer area? If crime rates was 0.1%, reasonable save 10% cost house, but if savings were 0.1% crimes rate higher hard fairly choose less-expensive house. This applies patients our daily practice. As suggested European Agency,4European there areas where mortality other irreversible morbidities (eg, stroke) any inferiority. extremely difficult ethically excess deaths minor For example, QID drug has proved decades safe efficient preventing death, stroke, myocardial infarction. QD B positively accepted anticipated better long cause deaths, strokes, infarctions A. Here comes intriguing paradox: increase reasonably anticipate relevant translates into outcome goes beyond pure evidence having comfortable regimen patient. So could proved. represent rigorous method widen options proper settings, physicians carefully read understand deepest pieces them verify chosen assess extent events. What's (ir)relevant?The Journal Thoracic Cardiovascular SurgeryVol. 161Issue 6PreviewThe debates cardiovascular science increasingly dominated trials. Although doubt these fundamental generate make decisions, equally clinicians principles, benefits, limitations underlying design statistics. Full-Text