Abstract
A commonly misunderstood principle in medical literature is statistical significance. Often, statistically non-significant or negative results are thought to be evidence for equivalence; mistakenly validating treatment modalities and putting patients at risk. This study examines the prevalence of misinterpretation of negative results of superiority trials in orthopaedic literature and outlines the need for a non-inferiority or equivalence research design.
Four orthopaedic journals – Journal of Paediatric Orthopaedics A, Journal of Bone and Joint Surgery American Volume, Journal of Arthroplasty and Journal of Shoulder and Elbow Surgery – were hand searched to identify all randomised control trials (RCTs) published within the time periods 2002–2003, 2007–2008 and 2012–2013. The identified RCTs were read and classified by study methodology, results obtained, and interpretation of results.
A total of 237 RCTs were identified. When analysing the primary outcomes, 117 (49.4%) studies yielded negative results and 120 (50.8%) yielded positive results. Out of the 237 articles, 231 (97.5%) used superiority methodology and 6 (2.5%) used non-inferiority or equivalence methodology. Of the 231 studies that used superiority methodology, 115 (49.8%) obtained negative results; and 45 (39.1%) of those misinterpreted the negative results for equivalence. While no statistical differences were seen, there was an upward trend in utilising non-inferiority and equivalence methodologies over time.
Given the frequency of misinterpreted negative results, there is an evident need for a more appropriate research methodology that shows equivalence of treatment methods. A non-inferiority or equivalence study design can address orthopaedic clinical dilemmas more suitably when trying to show one treatment is no worse or is equal to another treatment. Regarding orthopaedic treatment modalities as equivalent when studies show negative statistical results can be detrimental to patients and their clinical outcomes. A non-inferiority methodology can be used to accurately depict no difference between treatment methods rather than attempting to show one treatment method as superior.
