Stop the Spin: Confronting Reporting Bias in Multiple Sclerosis Research

What is spin? The concept of spin was formally introduced by Boutron and colleagues in a seminal 2010 paper.1 They defined “spin” as “…the use of specific reporting strategies…to highlight that the experimental treatment is beneficial, despite a statistically nonsignificant difference for the primary outcome….” The investigators reviewed 72 trials published in December 2006 with negative primary outcomes and found that over 40% contained spin in 2 or more sections. Since then, other investigators from a variety of disciplines have identified the common use of spin in reports of clinical trials with negative primary outcomes.2-6 In this issue of Annals of Neurology, Mascareñas-García and colleagues find that multiple sclerosis (MS) research is no exception.7 Boutron et al divided spin into 4 categories, which often are called types A to D. Type A spin focuses on statistically significant secondary outcomes, subgroup analyses, or within-group analyses while downplaying the negative primary outcome. Type B spin inappropriately interprets nonsignificant differences in a superiority trial as demonstrating equivalency between 2 treatments. Importantly, non-inferiority requires a different trial design than a superiority trial. Type C spin claims the treatment had beneficial effects even though the trial had nonsignificant results for the primary outcomes. Type D is all other spin strategies. Spin results from conscious or unconscious bias of the authors who obfuscate the negative results for the primary outcomes and amplify other results from the trial. This can result in readers concluding incorrectly that the treatment is effective. Mascareñas-García et al sought to determine whether spin occurred in reports of MS clinical trials. They followed the Boutron et al classification of spin into 4 types. To help understand spin better, they provided a useful Table 1, reproduced below, illustrating the 3 main types of spin.7 The investigators identified 40 phase 3 or 4 trials of a treatment for MS that had nonsignificant primary outcomes and were published between 2013 and 2024. Using established methodology, 2 investigators independently examined the articles for spin in the title, abstract, results, discussion, and conclusion/last paragraph sections. Spin occurred in at least one section in 25 of 40 (62.5%) and in 3 or more sections in 19 of 40 (47.5%) of the articles. The most frequent places for spin to appear were the abstract, discussion, and conclusions and type A spin was the most commonly used strategy. Spin was more likely to occur in non-Q1 journals, articles with first authors not from Europe or the United States, and studies with small sample size. Yet spin also appeared in publications in high-tier journals from established research centers in the United States and Europe. Spin thus may be commonly used in MS clinical trial publications and as suggested by others, perhaps is a wide-spread problem in neurology and other medical disciplines.2-6 Why does spin matter? Since the release of the first MS disease modifying therapy in 1993, use of appropriate clinical trial designs has been critical to the development and regulatory approval of over 20 MS therapies. A key part of clinical trial design is the selection of the primary outcome.8 Selection of one or sometimes 2 primary outcomes is done as the clinical trial is being planned. The primary outcome serves several essential functions in trial design. First, using primary outcomes, rather than having multiple outcomes with equal importance, avoids investigators “cherry picking” favorable results. In addition, the primary outcome is critical to determining the sample size of the study: investigators use the expected changes in the primary outcome among an untreated population and the effect size of the intervention to determine the sample size. This needs to be done before the trial is commenced and helps ensure that the study will be large enough to avoid false negative results. Finally, at the conclusion of the study, the primary outcome result determines whether the study is positive or negative. Secondary outcomes and other analyses can provide important information and new insights but cannot establish the efficacy of a treatment. This is partly because clinical trials often analyze many secondary outcomes, requiring statistical adjustment for multiple comparisons to avoid false-positive results. These are basic “rules” of modern clinical trial design and are critical to obtaining reliable results. When investigators use spin strategies to imply positive benefits for a therapy despite the trial having nonsignificant primary outcome results, they violate the basic principles of modern clinical trials. Spin also can lead clinicians who read the publications to believe that the therapy is efficacious. This is particularly problematic with drugs that are approved for other indications. Such medications are commonly tested in MS as symptomatic therapies and for repurposing as disease modifying therapies. Clinicians may prescribe these therapies “off-label” in the United States and some other countries if spin has convinced them and their patients of efficacy, despite the lack of supporting evidence. Spin therefore carries the risk of encouraging physicians and their patients to use treatments of unproven benefit. Why does spin occur? Although we may not know the exact drivers of spin, the widespread prevalence must be viewed within the broader context of scientific publishing. From 1990 to 2007, papers reporting positive results increased by 22%, reaching 85% by 2007, reflecting, in part, a decline in reporting of negative trials.9 In MS research specifically, 36% of phase 3 and phase 4 trials remained unpublished 6.5 years after trial completion, and those with positive primary outcomes were significantly more likely to be published.10 Failure to publish trials with negative primary outcomes may result from journals being reluctant to publish negative clinical trials or researchers choosing to not write or submit manuscripts describing their negative investigation. Academic careers, institutional support, and grant funding depend heavily on publications. The adage “Publish or perish” remains true. In this environment, spin may be an adaptive, if unethical, tool for surviving in a system that rewards positive findings and penalizes failure. Moving forward, we must strive to prevent the use of spin in publications of clinical trials with nonsignificant primary outcomes. The editorial staff of journals should flag spin strategies for reviewers and educate staff and authors about the importance of recognizing and avoiding spin strategies. We also need to confront the root causes of spin. Recently, approaches to ease the publication of negative clinical trials have been undertaken. As an example, the Clinical Trials Platform in The Oncologist encourages submission of negative clinical trials.11 A new approach to publishing clinical trials, called Registered Reports, offers a pathway to assuring publication of a clinical trial regardless of the outcome. Over 300 journals, including Scientific Reports and PLoS One, have a Registered Reports system.12, 13 Journals, funding agencies, and universities also must value well-conducted clinical trials that have negative results, recognizing that science is advanced by both positive and negative outcomes. Finally, academic promotion should reward methodologic rigor over positive findings. Until we value both positive and negative clinical trials, we will continue a system that incentivizes using spin to “sell” clinical trial reports for the sake of publication. At a time when medical research faces increasing public skepticism, we must adhere to the highest reporting standards to maintain our credibility. The findings of Mascareñas-García and colleagues should serve as a call to action in all of neurology to expunge spin from clinical trial reports. A Department of Veterans Affairs Rehabilitation Research and Development Career Development Award (5IK2RX003407) provides support to Dr Silbermann. Claude AI by Anthropic was used to assist with assessment of flow and review grammar. Dennis Bourdette: Conceptualization; writing – original draft; writing – review and editing. Elizabeth Silbermann: Writing – review and editing; conceptualization. The authors declare no conflicts of interest.