Balancing the Positives and Negatives: A Bibliometric Analysis of Positive and Negative Study Publication Patterns in High-Impact General Medical Journals

Despite calls to publish negative studies in prominent medical journals, greater submission and acceptance of positive results remains an issue. We aimed to quantify the degree to which high-impact general medical journals publish negative study results. We searched MEDLINE/PubMed for all randomized controlled trials published in five high-impact general medicine journals: Annals of Internal Medicine, the British Medical Journal (BMJ), the Journal of the American Medical Association (JAMA), the Lancet, and the New England Journal of Medicine (NEJM). Our search spanned a 10-year period from 2014 to 2023, which included data from before and after the emergence of the COVID-19 global pandemic. We performed single-author data extraction via abstract review to determine study positivity, defined as statistical significance for the primary outcome, flagging abstracts for secondary review if positivity was not clear. Two authors reviewed all flagged abstracts. We calculated the proportion of negative studies (i.e., not meeting statistical significance for the primary outcome) overall, by journal, and by publication year. We used logistic regression to model the odds of a study reporting a negative result by journal and year. Our search yielded 3722 individual citations, with screening resulting in 3600 randomized controlled trials for review, with 31% of studies reporting negative results. The proportion of negative studies varied, ranging from 22% in the Lancet to 51% in BMJ and JAMA. The proportion of negative studies remained consistent over time. High-impact general medical journals vary widely in the percentage of negative studies that they publish but did not change over time, even during and after a global pandemic. Further study is needed to determine factors influencing this phenomenon and what can be done to minimize publication bias.