Using Social Media to Maximize the Research Impact of Surgeons: Exploratory Linguistic Analysis.

Background: Surgeons work in a progressive field where communicating research is vital to advancing health care and enabling meaningful interactions among clinicians. It also contributes to societal impact, increases access to information, and reduces misinformation. Additionally, there can be barriers to accessing papers. Social media enhances research impact through sharing scholarly work and improving its translation into clinical practice, but little is known about how to design specific posts to maximize research impact through language. Objective: The purpose of this study was to determine the linguistic cues that optimize research impact among surgeons through Twitter (subsequently rebranded X). Additionally, this research combines the linguistic features of the posts and article access to determine their unique contributions. Methods: An exploratory linguistic analysis of 84 posts extracted from Twitter was conducted, which shared scholarly activity by 17 of the most-followed surgeons. The linguistic cues were measured on a continuous scale, computed from the percentage of each linguistic cue used in the text, and reported as mean (SD). Regression analysis and analysis of covariance were conducted to determine which cues influenced research impact and to estimate the potential association with study accessibility (open vs restricted access). Results: Analyzed tweets were highly analytic (mean 94.77, SD 9.00), moderate in clout (mean 42.69, SD 19.84), low in tone (mean 20.06, SD 33.91), suggesting negative tone use, and low in authenticity (mean 19.52, SD 24.50). Results suggest that a high use of formal language negatively impacts readership and citations. Analytical language was indirectly associated with readership (β=-0.296, 95% CI -423.57 to -59.95; P=.01) and citations (β=-0.524, 95% CI -0.442 to -0.187; P<.001). Linguistic clout had a positive association with readership (β=0.260, 95% CI 8.58-186.91; P=.03), and tone in tweets had a negative association with readership (β=-0.317, 95% CI -138.52 to -5.39; P=.04). Negative language tone was found to increase the impact of research. With respect to linguistic cues and study accessibility, the results also suggest that the number of citations was impacted by readership (F1,66=4.11, 95% CI 2.459E-06 to 0.003; P=.047) and analytic linguistic cues (F1,66=18.77, 95% CI -0.402 to -0.149; P<.001) used in the post, but the association of open (mean 3.04, SE 1.062) versus restricted access (mean 1.83, SE 0.716) was not statistically significant (F1,66=0.877, 95% CI 0.405-3.266; P=.352). Conclusions: This research is the first to explore article accessibility and linguistic cues used in creating posts that share research on social media to determine their influence on research impact, making this study both innovative and unique relative to existing studies in the surgery field. Through language, the medical field can expand its impact and encourage dialogue between scientists and the public, thereby increasing scientific and societal contributions while reducing the negative effects of limited article access.