Factors associated with altmetric attention scores for randomized phase III cancer clinical trials.

Abstract

11014 Background: The rise of the internet and social media has created new opportunities for the rapid sharing of scientific findings among professional and public communities. Altmetric Attention Scores (Altmetrics) are real-time measures of scientific impact and attention through various public outlets, including news, blogs, and social media. Distributions and patterns of Almetrics for high-impact clinical cancer research remain under-reported. Our goal was to describe the relationship between Altmetrics and other characteristics of published trials. Methods: We identified all two-arm superiority phase III cancer randomized controlled trials (RCTs) with a publication date between 2015-2020 from HemOnc.org. For included trials, we tabulated the followed the data: Altmetric score as of February 11, 2022; positive study (yes/no); FDA registration trial (yes/no); primary endpoint (categorized as overall survival (OS), progression-free survival (PFS), event-free/disease-free survival (EFS/DFS), or other); journal tier (defined as tier 1: NEJM, Lancet, JAMA; tier 2: JCO, Blood, Lancet Oncology, JAMA Oncology, Annals of Oncology or Tier 3: other). Univariable testing was performed with Pearson correlation, Wilcoxon rank sum, and Kruskal-Wallis testing to identify associations between predictor variables and Altmetrics. For factors with significant univariable associations, we generated a multiple linear regression model to identify clinical trial factors predictive of Altmetric attention scores, with alpha = 0.05 defining statistical significance. Results: Of 673 publications identified per inclusion/exclusion criteria, Altmetric scores were found for 638 (95%). The median score was 38.9 (IQR 12.4-126.7) – well above the 95th percentile for all Altmetric scores. On univariable testing, positive studies (P < 0.001), registration trials (P < 0.001), those published in tier 1 journals (P < 0.001), and trials with primary endpoint of OS (P = 0.002) were associated with increased Altmetric scores. On multivariable testing, positive studies (P = 0.006), registration trials (P < 0.001), studies published tier 1 journals (P < 0.001), and those reporting on OS (P = 0.03) persisted as significantly associated with increased engagement as measured by Altmetric scores. Conclusions: Social media and other online platforms represent an increasingly important avenue for the sharing of scientific and medical research. In this cross-sectional observational study, we found that certain characteristics of cancer clinical trials and publications are predictive of impact as measured by Altmetric scores. Specifically, positive registration trials published in traditionally defined high-impact journals and those reporting on OS as a primary endpoint tend to have the greatest Altmetric attention scores. The significance and consequences of these relationships warrant further investigation.