Social Media Attention and Citation Metrics as Predictors of U.S. Neurosurgery Residency Reputation

Social Media Attention and Citation Metrics as Predictors of U.S. Neurosurgery Residency Reputation | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Social Media Attention and Citation Metrics as Predictors of U.S. Neurosurgery Residency Reputation Patrick Pema, Akil Anthony, Youssef Atef AbdelAlim, Jack Kilgallon, and 7 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9204678/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 7 You are reading this latest preprint version Abstract Introduction Social media presence exerts a growing influence on visibility both at the individual and program level. In this study, we aim to assess the degree of influence that citation impact and social attention metrics assert by determining the degree to which these measures predict program reputation. Methods A paper level dataset for U.S. Neurosurgery residency program faculty was compiled from the Top 25 programs by reputation as listed by doximity.com, including each physician’s three most cited publications, and aggregated to the program level. The Altmetric tool was used to gather social media occurrences. We modeled program rank as an ordered outcome using ordinal logistic regression with three specifications: (A) aggregate attention + citations, (B) disaggregated attention (tweets, blogs, news) + citations, and (C) social-only. Model fit and generalization were compared via AIC, likelihood-ratio tests, and leave-one-out cross-validation. Predictive performance was summarized with Spearman correlation and mean absolute error (MAE). Results Models emphasizing social attention consistently outperformed citation-based approaches across all analytic metrics. The disaggregated social-attention model demonstrated the best overall performance, achieving an AIC of 198.5 versus 211.9 for aggregate attention + citations, and a significantly improved likelihood-ratio test (p < 0.001). Predictive accuracy was highest for social models (Spearman ρ ≈ 0.60–0.62; MAE ≈ 4.4). Conclusions Neurosurgery program reputation correlates more strongly with multi-platform social engagement than with reference volume. Integrating research dissemination with strategic digital communication may enhance both visibility and perceived academic prestige. Social media Altmetric neurosurgery residency reputation bibliometrics Figures Figure 1 Figure 2 Figure 3 Figure 4 Introduction Program rank is an important consideration for residency applicants each year and influences decision making throughout medical school. Selection of away rotations, involvement with research organizations, and submission of the residency application itself are actions that are all in part affected by the level of prestige of each target program in which a student may be interested. While such rankings change year to year and are multifactorial in nature, gaining insight into how medical culture shapes such program reputation can inform medical students, residents, and faculty alike. Recent studies demonstrate that research productivity, as measured by bibliometric indices (such as h-index, publication count, and citation metrics), is strongly correlated with neurosurgery program rankings and reputation[ 3 , 6 , 10 , 12 ]. Institutional bibliometric profiles, including 3 and 5-year publication and citation counts, are significantly associated with Doximity rankings and social media presence, reflecting the importance of research output in program evaluation[ 10 , 13 , 18 ]. The ih(5)-index and related metrics have been validated as robust tools for comparing academic productivity across programs and are less susceptible to distortion by outlier faculty, further supporting their use in ranking systems. Resident research productivity, as measured by h-index and publication count, also correlates strongly with faculty productivity and program ranking, highlighting the importance of a research-oriented environment in top programs[ 4 , 12 ]. Analyses also showed that programs with higher research output are more likely to be highly ranked, with greater emphasis placed onto clinical trials, dedicated research time for residents and faculty, and incoming applicant research acumen. These programs also tend to attract more competitive applicants; as such, research productivity is a key differentiator in the residency selection process. Thus, program reputation is both underscored and influenced heavily by academic output as judged by metrics such as unique citations, h-indices of faculty, and prestige of journals of publication. While objective measures such as h-index and citation count have existed for decades, social media platforms have only been active since the turn of the century. Use of sites like Facebook, Twitter, Instagram, Wikipedia, and more, can increase exposure for physicians, hospital systems, and residency programs alike[ 6 ]. Data shows that neurosurgery programs with strong social media presence have better patient reviews and are more likely to be ranked higher on sites such as Doximity[ 3 ]. As Doximity does not include a social media component, capturing this dimension of neurosurgery residency program exposure is critical to understand how reputation is shaped in the digital age. Altmetric is a research analytics tool that measures the online attention surrounding scholarly work in real time and is increasingly used in medical research, including neurosurgery, to complement traditional citation-based metrics. It systematically tracks mentions of academic works across multiple online platforms such as news outlets, policy documents, blogs, Wikipedia, and social media platforms like Twitter and Facebook, as well as academic bookmarking tools like Mendeley. Each source type contributes a weighted value to the Altmetric Attention Score (AS), with higher weights assigned to sources considered more influential, such as major news coverage or policy citations. The resulting score provides an indicator of the reach, dissemination, and societal engagement of a publication. In neurosurgery, Altmetric has been used to evaluate how clinical trials, surgical innovation papers, and systematic reviews gain visibility beyond traditional academic circles, often highlighting which topics attract broader public or policy interest. This makes Altmetric particularly valuable for identifying high-impact areas of translational research and assessing how new findings influence clinical discourse, patient advocacy, and public health awareness. Using this tool, we set out to assess the respective influence of research prowess and social media-related attention on program rank. Our purpose is to determine the degree of correlation that social media attention has on neurosurgery residency program reputation and determine if online presence is an independent predictor of program rank in the 21st century. Methods Data Acquisition This study used publicly available bibliometric and altmetric data and did not involve human participants, patient data, or animals. We gathered data from the top 25 neurosurgery residency programs as listed on Doximity.com. Doximity’s ranking is generated each year from three metrics: 1. Resident satisfaction data based on anonymized surveys of current residents and recent alumni, 2. Reputation data based on a national survey of board-certified neurosurgeons where each can nominate up to five programs they believe offer the best clinical training, and 3. Objective program data including a research-output score that combines alumni h-index, publication and clinical-trial participation rates, and research grants, as well as other metrics such as program size and subspecialty rate[18]. Members of the faculty of each program were identified via the program’s own website. Included faculty members had MD or DO degrees; physicians with PhD, MS, MPH, and other degrees were included only if an MD or DO degree was concurrently present. For each physician, bibliometric data was obtained from Scopus.com and included the number of unique publications and citation count for each paper. Given this direct impact of citational research output score on Doximity ranking, Altmetric data for the top 3 publications by citation count were gathered. Using the Altmetric tool[19], data for each publication was collected and included tweets, blogs, news occurences, Facebook posts, Wikipedia references, and patents. The data gathering stage yielded 2047 publications across 695 faculty from 25 programs. The relationship between online impact, Scopus references, and program reputation was quantified using a hierarchical aggregation pipeline, characterized as such: 1. Paper level → 2. Faculty level → 3. Program level. Regression modeling was conducted utilizing ordinal logistic regression, treating program rank as the ordered outcome. Paper level To represent each physician’s most visible work, the top 3 most cited papers from Scopus.com were retained. We mitigated author name ambiguity in Scopus by verifying each faculty member’s profile using institutional affiliation and specialty, and resolved discrepancies by cross-checking publication titles. For each distinct publication, year of publication and Scopus references were collected along with metrics measured by Altmetric: attention score, news, tweets, blogs, Facebook, Wikipedia, patents. A derived amplification ratio was computed for each paper as attention_score/(Scopus_citations +1) to describe attention per citation. If no Altmetric record was returned, it was treated as missing and omitted within aggregations. Faculty level For each faculty member, paper level metrics were averaged. Mean citations, mean attention score, mean Tweets, blogs, news, Facebook, Wikipedia, patents, and mean attention per citation were calculated. Faculty data regarding number of unique publications and h-index were also collected. Program level Faculty metrics were averaged across all faculty members in a program to ascertain program-level predictors. Mean citations, mean attention, and mean tweets, news, blogs, Facebook, Wikipedia, patents were calculated along with mean h-index. The outcome was the program’s Doximity reputation rank, with rank 1 = best. Modeling Strategy A proportional-odds ordinal logistic regression modeling system was used with program rank as the ordered response. Three model families were outlined to compare different representations of online attention: Aggregate model (A): rank ~ mean_attention + mean_citations Components model (B): rank ~ mean_tweets + mean_blogs + mean_news + mean_citations Social-only model (C): rank ~ mean_tweets + mean_blogs + mean_news For Models B and C, predictors were program-level means (averaged across faculty, then across each faculty member’s top three cited publications); thus, a 1-unit increase represents one additional mention per paper in that program-level mean for the specified source (such as blogs, news, or tweets). Model coefficients are reported on the log-odds scale, with Wald z-statistics from the model summary used for inference Model comparison and fit The Akaike Information Criterion (AIC) was used for primary model selection to guide whether social metrics demonstrate explanatory value beyond aggregate attention and citations. Effect Size Comparison Interpretability was enhanced by converting log-odds coefficients to odds ratio (OR) via exponentiation and calculated 95% Wald confidence intervals as exp(β ± 1.96 × SE). Cross-validation and Robustness Robustness was assessed using leave-one-out cross validation (LOO-CV). For each left out program, a model was re-fit on the remaining programs and AIC recorded. The mean and SD of LOO AIC across folds were used to compare the out-of-sample parsimony of model. Predictive Performance Checks To complement AIC-based comparisons, each fitted model produced category probabilities for ranks; these were collapsed to an expected rank (probability-weighted average of rank categories). We summarized in-sample predictive performance with Spearman correlation between expected and observed rank and mean absolute error (MAE). These metrics provide practical, unit-interpretable checks of alignment between model predictions and observed program ranks . Visualization and reporting Descriptive distributions (e.g., histograms of attention), bivariate scatterplots (attention vs citations) at paper/faculty/program levels, forest plots of ORs with 95% CIs (log scale), AIC bar charts (raw vs LOO means), and predicted-vs-observed rank scatterplots were used for presentation and model diagnostics in figures. Tabular results (OR tables) were rendered with gt[9]. Software All analyses were performed in R, using Modern Applied Statistics with S (MASS)[14] for ordinal models, dplyr[16]/tidyr[17] for data wrangling, ggplot2[15] for figures, and gt for tables. Results Study Sample Characteristics The cleaned dataset yielded 2047 publications across 695 physicians of the top 25 programs. Paper-level analysis Faculty level analysis Most faculty have relatively low mean attention scores, with an average attention score of 66.08 across 695 individuals. This dynamic resulted in a heavy right skew. A small minority of faculty had very high mean attention scores. These findings, depicted in Figure 2, demonstrate that attention is not evenly distributed among faculty members and that some physicians achieve a high level of online notoriety. Analysis yielded a mildly positive correlation between mean citations and attention score for each physician, suggesting that attention and citations are related but distinct measures of influence, and emphasizing that those metrics may be independent predictors of program rank. Program level analysis At the program level, mean citations and mean attention per faculty showed a moderate positive association (Figure 3). Programs with higher citation averages also tended to have higher attention averages. The spread indicates that attention is not a perfect proxy for citations. Some programs achieve higher attention relative to their citation output, suggesting they could be more effective at disseminating their work broadly, while others are more citation-heavy but draw less public/social attention. Ordinal Logistic Regression and Odds Ratios Ordinal logistic regression was used to test the relationship between academic and social media metrics and Doximity program reputation. Three models were compared. The aggregate model (A), which combined mean attention and citation counts, did not identify either variable as a significant predictor of program rank (AIC = 211.9). The component model (B) separated attention into tweets, blogs, and news mentions while controlling for citations. This model fit the data more accurately (AIC = 198.5). Blog mentions were a strong positive predictor of higher rank (OR 29.6, 95% CI 3.9–226.4, p < 0.01). News mentions showed a weak negative trend (OR 0.57, 95% CI 0.33-1.00). Tweets and citations were not significant. The social-only model (C), which excluded citations, showed the best overall fit (AIC = 196.6). Blog mentions again remained significant (OR 26.9, 95% CI 3.8-192.4). News mentions remained weakly negative, and tweets were not associated with rank. Cross-validation confirmed consistent results (mean LOO-CV AIC 187-189). Model performance was strongest for the component and social-only models (Spearman ρ = 0.60–0.62). Aggregate attention and citation-based measures alone did not explain variation in program reputation. Cross-validation and Robustness Leave-one-out cross validation (LOO-CV) confirmed the accuracy of Model C (social-only) when compared to Model A and B. Model results were: Aggregate (A): mean LOO-CV AIC = 201.45 (SD 0.35), Components (B): mean LOO-CV AIC = 188.66 (SD 2.60), Social-only (C): mean LOO-CV AIC = 186.87 (SD 2.57). These findings indicate that the social-only and components models fit the data better and generalize more effectively to programs with lesser attention. Predictive Performance In predictive accuracy analyses, the aggregate model (A) showed limited ability to reproduce observed Doximity reputation rankings. The correlation between predicted and observed ranks was low (Spearman ρ = 0.20), and the mean absolute error (MAE) was 5.85 ranks. Model B achieved a Spearman correlation of .62 and MAE of 4.42. Model C demonstrated Spearman correlation of .60 and MAE of 4.43. Thus, incorporating social media variables increased rank-order agreement roughly threefold and reduced prediction error by ~1.4 rank positions compared with the aggregate citation-based model. Discussion Neurosurgery as a specialty has placed high importance historically on the research output of its faculty, fellows, residents, and even prospective applicants[ 3 , 10 ]. Academic acumen, as gauged by factors such as h-index, citations, and unique publication counts, is a key feature of any neurosurgery residency application. Likewise, program ranking itself has been highly influenced by and correlated to objective measures of research output, with more prestigious programs often placing higher importance on the aforementioned factors[ 4 , 10 , 11 ]. In recent years, online visibility has played an increasingly significant role[ 3 ]. Neurosurgeons across the country have been able to promote both their individual practice and program through online visibility. Our goal through this analysis is to provide insight into the degree of impact this online attention has on program reputation. At the paper level, exposure to any given publication may occur through means such as PubMed, Google, Scopus, or other search as well as promotion through social media. Impactful literature is often posted, tweeted, and blogged through many different sites, apps, and profiles. Our paper-level analysis, showing skew towards a few publications, supported the idea that some papers and their content gain massive amounts of traction online. Papers that employ eye-catching titles and easy-to-understand terminology can appeal to readers who may not have an advanced level of medical knowledge, increasing like, share, and repost counts. Such exposure increases visibility not only for the authors but also for their associated institution. For example, Human hippocampal neurogenesis drops sharply in children to undetectable levels in adults by Sorrells et al. had one of the highest attention scores in our analysis (2291), with an associated 1183 tweets and 167 news articles. Its understandable and appealing title likely contributed to its extremely high tweet count, allowing their findings to spread fast and generating more visibility for the authors and the neurosurgery department at University of California, San Francisco. The faculty-level analysis demonstrated a similar dynamic. Certain authors possessed extremely high levels of online attention, while the mean attention store across all was low, leading to another right skew. This finding once again supports the notion that some physicians are able to gather a large amount of social attention while the average neurosurgeon's online attention remained lower. Additionally, the correlation between mean attention and citations was moderate, indicating that the impact of those two variables is related yet distinct. Regression analysis also showed that mean citations per faculty and mean attention per faculty were moderately correlated, indicating that programs with more citations also garnered more attention. Moderate correlation indicates that attention is related to, but not fully correlated to academic output. A possible explanation for this trend is that programs that output more research also strive to increase their attention online. Departments that promote their own literature through their respective social media channels can increase read and citation count via that channel, which could explain the moderate correlation between those factors. Relying on journal publication alone as means to report findings may result in both lower social exposure and total online attention. Regression modeling revealed interesting trends among the top 25 programs. In model A, aggregating social attention and citations failed to demonstrate significance, reinforcing the idea that those two variables influence rank separately. Model B, including all components of social attention, had a substantially lower AIC, revealing that evaluating total attention separately by social component and citations improves predictability. In the same manner, the social-only model had the best AIC; in particular, blog coverage was an extremely strong predictor. On the whole, social media engagement was a far stronger and more robust predictor of reputation rank, a conclusion which is reinforced by the results of the LOO cross-validation. Splitting attention into its two distinct components increased predictive value, reinforcing the idea that social attention is an entirely separate influencer of rank. Publications are at the heart of this analysis due to their historical impact on reputation, and the ways in which they exert that influence seem to be changing. While journal content via website and print edition lend traditional exposure to paper content, increasing usage of platforms like Facebook, Twitter, and more allows authors to promote their content through such sites as well. Advantages offered by presenting publication material in that manner include shorter reading time, lack of journal paywall, easy sharing/reposting options, and greater engagement with readers. Physician and program-specific pages automatically present posted content to followers, generating clicks and increasing engagement with said pages and the entities they represent. Programs that lean into this dynamic can generate greater amounts of exposure and increase their reputation by demonstrating their important literature in a concise and easy to read manner. With this dynamic in mind, dissemination of study results through social media before peer review has taken place can lead to misconceptions and premature conclusions. An example of such rapid spread was observed when the NASCIS II clinical trial, which reported that high-dose methylprednisolone administration within 8 hours of spinal cord injury improved neurologic recovery[ 2 ]. These results were widely promoted and implemented by institutions around the country. However, later review revealed that this conclusion was based on post-hoc analysis, and the actual benefit was controversial[ 5 , 8 ]. Today, modern guidelines do not recommend high dose steroids for acute spinal cord injury[ 1 ]. This instance reflects the direct effect that social attention has on clinical practice and should inspire caution when disseminating early preliminary study results. In similar fashion, blog coverage may also reflect deeper engagement within the neurosurgical community. Many posts highlight ongoing projects, clinical achievements, or new publications The University of Miami Department of Neurosurgery engages in this sort of activity with its blog Neurosurgery Blog: More Than Just Brain Surgery [ 20 ]. Their site encourages posts from their own faculty as well as physicians from other programs across the country, publishing several in-depth articles monthly. This type of communication can shape perceptions of productivity and leadership in ways that citation counts do not, allowing direct communication between physicians from different programs.The results suggest that program visibility, especially through consistent online engagement, may now play a measurable role in how reputation is formed. Limitations and Future Directions Several limitations should be acknowledged in this study. The analysis was restricted to the top 25 neurosurgery residency programs, which may limit generalizability to mid- or lower-ranked institutions. Only the top three publications per faculty member were included, potentially underrepresenting overall scholarly output and online impact. Social media metrics such as tweets, blogs, and news mentions are also inherently ambiguous and may reflect institutional promotion or general public interest rather than true program-driven engagement. Although robust modeling approaches were used, the relatively small sample size reduces the ability to detect more subtle associations. Expanding publication capture beyond each faculty member’s top three works would allow a more comprehensive assessment of scholarly activity and visibility. Additionally, evaluating emerging platforms such as podcasts, TikTok, LinkedIn, and YouTube may provide a fuller understanding of how modern digital presence influences residency program reputation. This study is an exploratory analysis of elite programs; such institutions have garnered a level of fame and interest through long-term academic reputation that may not generalize well towards newer and less academically focused programs. The observational design prevents causal inference, as higher program reputation may itself drive greater online visibility just as increased visibility may influence perceived reputation. Future projects may expand on these conclusions by quantifying a larger or even all-encompassing dataset of neurosurgery residency academic metrics. Lastly, Doximity rankings, while multifactorial in nature, include certain subjective characteristics which may introduce bias. Some authors argue that Doximity rankings may include a circular methodology to characterize programs, where reputation influences rank which in turn influences reputation later on. As such, Doximity rankings should be viewed as a metric of reputation perception rather than an objective assessment of program training quality. Despite such limitations, the Doximity system is one of the most prominent and widely viewed[ 7 ], hence its utilization in this study. Conclusion In the current era of digital communication, social media has become a central part of academic medicine, shaping how research is shared, reputations are formed, and professional networks expand. In neurosurgery, it has emerged as an important means for academic visibility and engagement. Platforms like Twitter, Instagram, LinkedIn, and YouTube allow surgeons to share new techniques, highlight clinical cases, discuss evidence-based updates, and disseminate research more rapidly than traditional journals alone. The results of this study suggest that social media activity contributes directly to program reputation. Among leading neurosurgery residency programs, online engagement, especially through blog mentions, was more strongly associated with higher Doximity rankings than traditional citation metrics. This reflects a shift in how academic influence is recognized: visibility and accessibility of research discussions online now play a meaningful role alongside established bibliometric indicators. This trend underscores the growing importance of digital presence in shaping academic identity. Programs that use social media strategically to communicate research, showcase achievements, and engage with the broader medical community may enhance both their visibility and perceived leadership. As social platforms continue to evolve, their integration into academic culture will likely remain essential for effective communication, collaboration, and public engagement in neurosurgery. Declarations Conflicts of Interest The authors have no conflicts of interests or disclosures to report. Competing Interests The authors declare no competing interests. Funding The authors received no financial support for the research, authorship, and/or publication of this article. Author Contribution Patrick Pema conceived and led the study. Patrick Pema directed study design, coordinated project execution, interpreted the findings, and wrote the first draft of the manuscript. Patrick Pema designed the figures and outlined tables. Akil Anthony and Michael Chaga performed the statistical analyses and contributed to interpretation of the data. Youssef Atef AbdelAlim, Jack Kilgallon, Kush Desai, Paxton Sweeney, and Daniel Monahan contributed to data collection and manuscript editing. Ira Goldstein, Shabbar Danish, and Nitesh Patel contributed to manuscript appraisal, critical revision, and editorial guidance. All authors reviewed previous versions of the manuscript, read and approved the final manuscript. Acknowledgements The authors would like to acknowledge the neurosurgery department at Jersey Shore University Medical Center and the library staff at Hackensack Meridian School of Medicine for their assistance with this project. Data Availability The data used in this study were derived from publicly available sources, including neurosurgery program rankings from Doximity.com, Scopus.com for author h indices and citation metrics, and Altmetric.com for social media occurrences using the Altmetric bookmarklet, a free tool available to the public. Current faculty were gathered from each program's own website. The processed dataset generated during the current study is available from the corresponding author on request. References American College of Surgeons Trauma Quality Programs Best Practices Project Team (2022) Best Practices Guidelines: Spine Injury. American College of Surgeons, Chicago, IL Bracken MB, Shepard MJ, Collins WF, Holford TR, Young W, Baskin DS, Eisenberg HM, Flamm E, Leo-Summers L, Maroon J, Marshall LF, Perot PL, Piepmeier J, Sonntag VKH, Wagner FC, Wilberger JE, Winn HR (1990) A Randomized, Controlled Trial of Methylprednisolone or Naloxone in the Treatment of Acute Spinal-Cord Injury. N Engl J Med 322:1405–1411. doi: 10.1056/NEJM199005173222001 Chang AN, Boyett D, Chou D, Chan AK (2025) Updated 5-Year Institutional Bibliometric Profiles for United States Neurosurgery Residency Programs and the Relationship Between Social Media Presence and Objective Departmental Metrics. Neurosurgery 96:1397–1409. doi: 10.1227/neu.0000000000003256 Cole KL, Carter A, Rawson C, Tenhoeve S, Orton C, Zeinali M, Karsy M (2025) The impact of NIH funding and program reputation score on research output and residency matches in neurosurgery: A bibliometrics analysis. J Natl Med Assoc 117:115–122. doi: 10.1016/j.jnma.2025.03.002 Coleman WP, Benzel D, Cahill DW, Ducker T, Geisler F, Green B, Gropper MR, Goffin J, Madsen PW, Maiman DJ, Ondra SL, Rosner M, Sasso RC, Trost GR, Zeidman S (2000) A critical appraisal of the reporting of the National Acute Spinal Cord Injury Studies (II and III) of methylprednisolone in acute spinal cord injury. J Spinal Disord 13:185–199. doi: 10.1097/00002517-200006000-00001 Colombo E, Höbner LM, Blom V, Berglar I, Alakmeh A, de Wilde D, El-Hajj VG, Regli L, Serra C, Staartjes VE, Burström G (2025) The Implementation of Social Media in Neurosurgery: A Systematic Review of the Literature. Acta Neurochir (Wien) 167:277. doi: 10.1007/s00701-025-06695-1 Feinstein MM, Niforatos JD, Mosteller L, Chelnick D, Raza S, Otteson T (2019) Association of Doximity Ranking and Residency Program Characteristics Across 16 Specialty Training Programs. J Grad Med Educ 11:580–584. doi: 10.4300/JGME-D-19-00336.1 Hurlbert RJ (2000) Methylprednisolone for acute spinal cord injury: an inappropriate standard of care. J Neurosurg 93:1–7. doi: 10.3171/spi.2000.93.1.0001 Iannone R, Cheng J, Schloerke B, Hughes E, Lauer A, Seo J, Brevoort K, Roy Ö (2025) gt: Easily Create Presentation-Ready Display Tables Lee RP, Venable GT, Roberts ML, Parikh KA, Taylor DR, Khan NR, Michael LM, Klimo P (2016) Five-Year Institutional Bibliometric Profiles for 119 North American Neurosurgical Residency Programs: An Update. World Neurosurg 95:565–575. doi: 10.1016/j.wneu.2016.07.006 Ponce FA, Lozano AM (2010) Academic impact and rankings of American and Canadian neurosurgical departments as assessed using the h index. J Neurosurg 113:447–457. doi: 10.3171/2010.3.JNS1032 Sarkiss CA, Riley KJ, Hernandez CM, Oermann EK, Ladner TR, Bederson JB, Shrivastava RK (2017) Academic Productivity of US Neurosurgery Residents as Measured by H-Index: Program Ranking with Correlation to Faculty Productivity. Neurosurgery 80:975–984. doi: 10.1093/neuros/nyx071 Taylor DR, Venable GT, Jones GM, Lepard JR, Roberts ML, Saleh N, Sidiqi SK, Moore A, Khan N, Selden NR, Michael LM, Klimo P (2015) Five-year institutional bibliometric profiles for 103 US neurosurgical residency programs. J Neurosurg 123:547–560. doi: 10.3171/2014.10.JNS141025 Venables WN, Ripley BD (2002) Modern Applied Statistics with S. Springer, New York Wickham H (2016) ggplot2: Elegant Graphics for Data Analysis. Springer-Verlag, New York Wickham H, François R, Henry L, Müller K, Vaughan D (2023) dplyr: A Grammar of Data Manipulation Wickham H, Vaughan D, Girlich M (2024) tidyr: Tidy Messy Data Doximity Residency Navigator. In: Doximity. https://www.doximity.com/residency/programs?specialtyKey=eb85a000-2e4f-4977-ac28-f5de6e72ebc9-neurological-surgery. Accessed 10 Feb 2026 Bookmarklet. In: Altmetric. https://www.altmetric.com/solutions/free-tools/bookmarklet/. Accessed 7 Dec 2025 Neurosurgery Blog. In: Neurosurg. Blog. https://www.neurosurgeryblog.org/. Accessed 26 Dec 2025 Additional Declarations No competing interests reported. Supplementary Files SupplementalTable1.docx Cite Share Download PDF Status: Under Review Version 1 posted Reviews received at journal 07 May, 2026 Reviewers agreed at journal 03 May, 2026 Reviewers agreed at journal 30 Apr, 2026 Reviewers invited by journal 29 Apr, 2026 Editor assigned by journal 29 Apr, 2026 Submission checks completed at journal 26 Mar, 2026 First submitted to journal 23 Mar, 2026 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-9204678","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":635894990,"identity":"5182c849-20ec-4ba8-8d55-0919cb360484","order_by":0,"name":"Patrick Pema","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA4ElEQVRIiWNgGAWjYLCCBCAEg48NMBZewIzQwjiTaC0MUIXMvMRoMWc/f+zDg5o0IKM78bPtjm155gzMxz5+waPFsieZeUbCsRwg4+xm6dwzt4stG9iSZ8vg0WJwIJmZIbGhgsHgRu4G6dy224kbDvAYM0vg03L+MVTL/bebf1sSpeUG2JYcIIN3mzQjVAvjB3x+mfHYmCHhWBqPZU/uNstekJbDbMnMeHQwmPMnPmb8UZMsZ85+dvONnyAtx5sPM/7A5zAozWMAFwJawcxDhBYEAwTw2jIKRsEoGAUjDgAAvbJQnDcPePsAAAAASUVORK5CYII=","orcid":"","institution":"Hackensack Meridian School of Medicine","correspondingAuthor":true,"prefix":"","firstName":"Patrick","middleName":"","lastName":"Pema","suffix":""},{"id":635894991,"identity":"f51fd49d-8c5e-4e3b-8617-806232b557f4","order_by":1,"name":"Akil Anthony","email":"","orcid":"","institution":"Jersey Shore University Medical Center","correspondingAuthor":false,"prefix":"","firstName":"Akil","middleName":"","lastName":"Anthony","suffix":""},{"id":635894992,"identity":"ae7cbf15-e9c0-4a37-a784-a7bc988f188b","order_by":2,"name":"Youssef Atef AbdelAlim","email":"","orcid":"","institution":"Hackensack Meridian School of Medicine","correspondingAuthor":false,"prefix":"","firstName":"Youssef","middleName":"Atef","lastName":"AbdelAlim","suffix":""},{"id":635894993,"identity":"536a31b5-10d5-4795-b603-f38b34d4a135","order_by":3,"name":"Jack Kilgallon","email":"","orcid":"","institution":"Hackensack Meridian School of Medicine","correspondingAuthor":false,"prefix":"","firstName":"Jack","middleName":"","lastName":"Kilgallon","suffix":""},{"id":635894995,"identity":"9d655067-183a-4b0f-a164-f503eeb3fc19","order_by":4,"name":"Kush Desai","email":"","orcid":"","institution":"Hackensack Meridian School of Medicine","correspondingAuthor":false,"prefix":"","firstName":"Kush","middleName":"","lastName":"Desai","suffix":""},{"id":635894996,"identity":"98831640-f613-4d97-bf99-d23830e4d1a2","order_by":5,"name":"Paxton Sweeney","email":"","orcid":"","institution":"Hackensack Meridian School of Medicine","correspondingAuthor":false,"prefix":"","firstName":"Paxton","middleName":"","lastName":"Sweeney","suffix":""},{"id":635894997,"identity":"83845f59-1de2-407d-98ca-082deb5e6cb9","order_by":6,"name":"Daniel Monahan","email":"","orcid":"","institution":"Hackensack Meridian School of Medicine","correspondingAuthor":false,"prefix":"","firstName":"Daniel","middleName":"","lastName":"Monahan","suffix":""},{"id":635894999,"identity":"8c80ac23-ca7d-44b7-95c7-24584e702cbe","order_by":7,"name":"Michael Chaga","email":"","orcid":"","institution":"Jersey Shore University Medical Center","correspondingAuthor":false,"prefix":"","firstName":"Michael","middleName":"","lastName":"Chaga","suffix":""},{"id":635895000,"identity":"aa89229e-33dc-40a9-908d-95470e93c0e3","order_by":8,"name":"Ira Goldstein","email":"","orcid":"","institution":"Jersey Shore University Medical Center","correspondingAuthor":false,"prefix":"","firstName":"Ira","middleName":"","lastName":"Goldstein","suffix":""},{"id":635895001,"identity":"f863de08-4eb2-4441-9c28-860fc1bc85c4","order_by":9,"name":"Shabbar Danish","email":"","orcid":"","institution":"Jersey Shore University Medical Center","correspondingAuthor":false,"prefix":"","firstName":"Shabbar","middleName":"","lastName":"Danish","suffix":""},{"id":635895002,"identity":"fba1e274-5135-4b49-82a9-77715524718e","order_by":10,"name":"Nitesh Patel","email":"","orcid":"","institution":"Jersey Shore University Medical Center","correspondingAuthor":false,"prefix":"","firstName":"Nitesh","middleName":"","lastName":"Patel","suffix":""}],"badges":[],"createdAt":"2026-03-23 21:53:15","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-9204678/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9204678/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":108850444,"identity":"b7cf576f-85c6-4d24-9f74-7d7bd60ad748","added_by":"auto","created_at":"2026-05-09 05:33:39","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":187733,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eA: Distribution of attention scores of individual papers. \u003c/strong\u003eThe vast majority of papers had an Altmetric attention score of less than 10 (1071 out of 2047), indicating low exposure on social media. Certain papers have an extremely large amount of social media attention, leading to a right skewed distribution. \u003cstrong\u003eB: Attention Score vs. Scopus Citations. \u003c/strong\u003eThe scatterplot indicates that, while a moderate correlation is observed between Scopus citations and attention score, these variables are not redundant. Many papers have low citations but high attention scores, and many others have high citations but minimal attention online. This finding supports the central hypothesis that attention score captures a different dimension of reader activity and that online impact matters independently of traditional metrics like citations.\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-9204678/v1/63191d58c67c74df7e74c056.png"},{"id":108977128,"identity":"ebb60fe7-2c48-4d53-84c6-cb77a5088e7c","added_by":"auto","created_at":"2026-05-11 11:30:30","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":220049,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eA: Distribution of faculty mean attention scores. B: Scatterplot of mean citations vs attention score for each physician.\u003c/strong\u003e The plot depicts a mildly positive correlation between citations and Scopus references (R\u003csup\u003e2 \u003c/sup\u003e= .2109).\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-9204678/v1/f1ae150abf57aa02986f33fc.png"},{"id":108850446,"identity":"99f79910-7ef8-4b01-846b-afb5dabf1f0c","added_by":"auto","created_at":"2026-05-09 05:33:39","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":81001,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eMean citations vs. Mean attention score per faculty of each program. \u003c/strong\u003e\u0026nbsp;The plot shows a moderate positive relationship (R2= .4209).\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-9204678/v1/46e3e50c20c52a96a3855241.png"},{"id":108850447,"identity":"b149813e-9777-424f-a597-6d34591fde86","added_by":"auto","created_at":"2026-05-09 05:33:39","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":261940,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eOdds Ratios with 95% CI for Models A, B, C. \u003c/strong\u003eBlog attention was the strongest predictor of program rank.\u003c/p\u003e","description":"","filename":"4.png","url":"https://assets-eu.researchsquare.com/files/rs-9204678/v1/ea7b91b90ca667eb3b09c434.png"},{"id":108979502,"identity":"b95757a0-32a9-42be-93d3-bdb07e53bccc","added_by":"auto","created_at":"2026-05-11 11:59:27","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":860327,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9204678/v1/d30e1516-2563-46e2-b461-216177a96560.pdf"},{"id":108850443,"identity":"31d57e5b-eb8a-424d-9873-b4deae017645","added_by":"auto","created_at":"2026-05-09 05:33:39","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":11498,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementalTable1.docx","url":"https://assets-eu.researchsquare.com/files/rs-9204678/v1/7d935acc3e31fd06450d4df4.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Social Media Attention and Citation Metrics as Predictors of U.S. Neurosurgery Residency Reputation","fulltext":[{"header":"Introduction","content":"\u003cp\u003eProgram rank is an important consideration for residency applicants each year and influences decision making throughout medical school. Selection of away rotations, involvement with research organizations, and submission of the residency application itself are actions that are all in part affected by the level of prestige of each target program in which a student may be interested. While such rankings change year to year and are multifactorial in nature, gaining insight into how medical culture shapes such program reputation can inform medical students, residents, and faculty alike.\u003c/p\u003e \u003cp\u003eRecent studies demonstrate that research productivity, as measured by bibliometric indices (such as h-index, publication count, and citation metrics), is strongly correlated with neurosurgery program rankings and reputation[\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e, \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e, \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e, \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e]. Institutional bibliometric profiles, including 3 and 5-year publication and citation counts, are significantly associated with Doximity rankings and social media presence, reflecting the importance of research output in program evaluation[\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e, \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e, \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e]. The ih(5)-index and related metrics have been validated as robust tools for comparing academic productivity across programs and are less susceptible to distortion by outlier faculty, further supporting their use in ranking systems. Resident research productivity, as measured by h-index and publication count, also correlates strongly with faculty productivity and program ranking, highlighting the importance of a research-oriented environment in top programs[\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e, \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eAnalyses also showed that programs with higher research output are more likely to be highly ranked, with greater emphasis placed onto clinical trials, dedicated research time for residents and faculty, and incoming applicant research acumen. These programs also tend to attract more competitive applicants; as such, research productivity is a key differentiator in the residency selection process. Thus, program reputation is both underscored and influenced heavily by academic output as judged by metrics such as unique citations, h-indices of faculty, and prestige of journals of publication.\u003c/p\u003e \u003cp\u003eWhile objective measures such as h-index and citation count have existed for decades, social media platforms have only been active since the turn of the century. Use of sites like Facebook, Twitter, Instagram, Wikipedia, and more, can increase exposure for physicians, hospital systems, and residency programs alike[\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e]. Data shows that neurosurgery programs with strong social media presence have better patient reviews and are more likely to be ranked higher on sites such as Doximity[\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e]. As Doximity does not include a social media component, capturing this dimension of neurosurgery residency program exposure is critical to understand how reputation is shaped in the digital age.\u003c/p\u003e \u003cp\u003eAltmetric is a research analytics tool that measures the online attention surrounding scholarly work in real time and is increasingly used in medical research, including neurosurgery, to complement traditional citation-based metrics. It systematically tracks mentions of academic works across multiple online platforms such as news outlets, policy documents, blogs, Wikipedia, and social media platforms like Twitter and Facebook, as well as academic bookmarking tools like Mendeley. Each source type contributes a weighted value to the Altmetric Attention Score (AS), with higher weights assigned to sources considered more influential, such as major news coverage or policy citations. The resulting score provides an indicator of the reach, dissemination, and societal engagement of a publication. In neurosurgery, Altmetric has been used to evaluate how clinical trials, surgical innovation papers, and systematic reviews gain visibility beyond traditional academic circles, often highlighting which topics attract broader public or policy interest. This makes Altmetric particularly valuable for identifying high-impact areas of translational research and assessing how new findings influence clinical discourse, patient advocacy, and public health awareness.\u003c/p\u003e \u003cp\u003eUsing this tool, we set out to assess the respective influence of research prowess and social media-related attention on program rank. Our purpose is to determine the degree of correlation that social media attention has on neurosurgery residency program reputation and determine if online presence is an independent predictor of program rank in the 21st century.\u003c/p\u003e"},{"header":"Methods","content":"\u003cp\u003e\u003cstrong\u003eData Acquisition\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study used publicly available bibliometric and altmetric data and did not involve human participants, patient data, or animals. We gathered data from the top 25 neurosurgery residency programs as listed on Doximity.com. Doximity\u0026rsquo;s ranking is generated each year from three metrics: 1. Resident satisfaction data based on anonymized surveys of current residents and recent alumni, 2. Reputation data based on a national survey of board-certified neurosurgeons where each can nominate up to five programs they believe offer the best clinical training, and 3. Objective program data including a research-output score that combines alumni h-index, publication and clinical-trial participation rates, and research grants, as well as other metrics such as program size and subspecialty rate[18]. \u003c/p\u003e\n\u003cp\u003eMembers of the faculty of each program were identified via the program\u0026rsquo;s own website. Included faculty members had MD or DO degrees; physicians with PhD, MS, MPH, and other degrees were included only if an MD or DO degree was concurrently present. \u003c/p\u003e\n\u003cp\u003eFor each physician, bibliometric data was obtained from Scopus.com and included the number of unique publications and citation count for each paper. Given this direct impact of citational research output score on Doximity ranking, Altmetric data for the top 3 publications by citation count were gathered. Using the Altmetric tool[19], data for each publication was collected and included tweets, blogs, news occurences, Facebook posts, Wikipedia references, and patents. The data gathering stage yielded 2047 publications across 695 faculty from 25 programs. \u003c/p\u003e\n\u003cp\u003eThe relationship between online impact, Scopus references, and program reputation was quantified using a hierarchical aggregation pipeline, characterized as such: 1. Paper level \u0026rarr; 2. Faculty level \u0026rarr; 3. Program level. Regression modeling was conducted utilizing ordinal logistic regression, treating program rank as the ordered outcome. \u003c/p\u003e\n\n\u003cp\u003e\u003cstrong\u003ePaper level\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo represent each physician\u0026rsquo;s most visible work, the top 3 most cited papers from Scopus.com were retained. We mitigated author name ambiguity in Scopus by verifying each faculty member\u0026rsquo;s profile using institutional affiliation and specialty, and resolved discrepancies by cross-checking publication titles. For each distinct publication, year of publication and Scopus references were collected along with metrics measured by Altmetric: attention score, news, tweets, blogs, Facebook, Wikipedia, patents. A derived amplification ratio was computed for each paper as attention_score/(Scopus_citations +1) to describe attention per citation. If no Altmetric record was returned, it was treated as missing and omitted within aggregations. \u003c/p\u003e\n\n\u003cp\u003e\u003cstrong\u003eFaculty level\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eFor each faculty member, paper level metrics were averaged. Mean citations, mean attention score, mean Tweets, blogs, news, Facebook, Wikipedia, patents, and mean attention per citation were calculated. Faculty data regarding number of unique publications and h-index were also collected. \u003c/p\u003e\n\n\u003cp\u003e\u003cstrong\u003eProgram level\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eFaculty metrics were averaged across all faculty members in a program to ascertain program-level predictors. Mean citations, mean attention, and mean tweets, news, blogs, Facebook, Wikipedia, patents were calculated along with mean h-index. The outcome was the program\u0026rsquo;s Doximity reputation rank, with rank 1 = best. \u003c/p\u003e\n\n\u003cp\u003e\u003cstrong\u003eModeling Strategy\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eA proportional-odds ordinal logistic regression modeling system was used with program rank as the ordered response. Three model families were outlined to compare different representations of online attention: \u003c/p\u003e\n\n\u003col start=\"1\" type=\"1\"\u003e\n\u003cli\u003e\u003cstrong\u003eAggregate model (A):\u003c/strong\u003e rank ~ mean_attention + mean_citations\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003eComponents model (B):\u003c/strong\u003e rank ~ mean_tweets + mean_blogs + mean_news + mean_citations\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003eSocial-only model (C): \u003c/strong\u003erank ~ mean_tweets + mean_blogs + mean_news\u003c/li\u003e\n\u003c/ol\u003e\n\n\u003cp\u003eFor Models B and C, predictors were program-level means (averaged across faculty, then across each faculty member\u0026rsquo;s top three cited publications); thus, a 1-unit increase represents one additional mention per paper in that program-level mean for the specified source (such as blogs, news, or tweets).\u003c/p\u003e\n\u003cp\u003eModel coefficients are reported on the log-odds scale, with Wald z-statistics from the model summary used for inference\u003c/p\u003e\n\n\u003cp\u003e\u003cstrong\u003eModel comparison and fit\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe Akaike Information Criterion (AIC) was used for primary model selection to guide whether social metrics demonstrate explanatory value beyond aggregate attention and citations.\u003c/p\u003e\n\n\u003cp\u003e\u003cstrong\u003eEffect Size Comparison\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eInterpretability was enhanced by converting log-odds coefficients to odds ratio (OR) via exponentiation and calculated 95% Wald confidence intervals as exp(\u0026beta; \u0026plusmn; 1.96 \u0026times; SE).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCross-validation and Robustness\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eRobustness was assessed using leave-one-out cross validation (LOO-CV). For each left out program, a model was re-fit on the remaining programs and AIC recorded. The mean and SD of LOO AIC across folds were used to compare the out-of-sample parsimony of model. \u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ePredictive Performance Checks\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo complement AIC-based comparisons, each fitted model produced category probabilities for ranks; these were collapsed to an expected rank (probability-weighted average of rank categories). We summarized in-sample predictive performance with Spearman correlation between expected and observed rank and mean absolute error (MAE). These metrics provide practical, unit-interpretable checks of alignment between model predictions and observed program ranks\u003cem\u003e.\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eVisualization and reporting\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eDescriptive distributions (e.g., histograms of attention), bivariate scatterplots (attention vs citations) at paper/faculty/program levels, forest plots of ORs with 95% CIs (log scale), AIC bar charts (raw vs LOO means), and predicted-vs-observed rank scatterplots were used for presentation and model diagnostics in figures. Tabular results (OR tables) were rendered with gt[9].\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSoftware\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll analyses were performed in R, using Modern Applied Statistics with S (MASS)[14] for ordinal models, dplyr[16]/tidyr[17] for data wrangling, ggplot2[15] for figures, and gt\u003csup\u003e \u003c/sup\u003efor tables.\u003c/p\u003e"},{"header":"Results","content":"\u003cp\u003e\u003cstrong\u003eStudy Sample Characteristics\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe cleaned dataset yielded 2047 publications across 695 physicians of the top 25 programs.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ePaper-level analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFaculty level analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eMost faculty have relatively low mean attention scores, with an average attention score of 66.08 across 695 individuals. This dynamic resulted in a heavy right skew. A small minority of faculty had very high mean attention scores. These findings, depicted in Figure 2, demonstrate that attention is not evenly distributed among faculty members and that some physicians achieve a high level of online notoriety. Analysis yielded a mildly positive correlation between mean citations and attention score for each physician, suggesting that attention and citations are related but distinct measures of influence, and emphasizing that those metrics may be independent predictors of program rank.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eProgram level analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAt the program level, mean citations and mean attention per faculty showed a moderate positive association (Figure 3). Programs with higher citation averages also tended to have higher attention averages. The spread indicates that attention is not a perfect proxy for citations. Some programs achieve higher attention relative to their citation output, suggesting they could be more effective at disseminating their work broadly, while others are more citation-heavy but draw less public/social attention.\u0026nbsp;\u003cbr\u003e\u003cstrong\u003eOrdinal Logistic Regression and Odds Ratios\u003cbr\u003e\u0026nbsp;\u003c/strong\u003eOrdinal logistic regression was used to test the relationship between academic and social media metrics and Doximity program reputation. Three models were compared. The aggregate model (A), which combined mean attention and citation counts, did not identify either variable as a significant predictor of program rank (AIC = 211.9). The component model (B) separated attention into tweets, blogs, and news mentions while controlling for citations. This model fit the data more accurately (AIC = 198.5). Blog mentions were a strong positive predictor of higher rank (OR 29.6, 95% CI 3.9\u0026ndash;226.4, \u003cem\u003ep\u003c/em\u003e \u0026lt; 0.01). News mentions showed a weak negative trend (OR 0.57, 95% CI 0.33-1.00). Tweets and citations were not significant.\u003c/p\u003e\n\u003cp\u003eThe social-only model (C), which excluded citations, showed the best overall fit (AIC = 196.6). Blog mentions again remained significant (OR 26.9, 95% CI 3.8-192.4). News mentions remained weakly negative, and tweets were not associated with rank. Cross-validation confirmed consistent results (mean LOO-CV AIC 187-189). Model performance was strongest for the component and social-only models (Spearman \u0026rho; = 0.60\u0026ndash;0.62). Aggregate attention and citation-based measures alone did not explain variation in program reputation.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCross-validation and Robustness\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eLeave-one-out cross validation (LOO-CV) confirmed the accuracy of Model C (social-only) when compared to Model A and B. Model results were: Aggregate (A): mean LOO-CV AIC = 201.45 (SD 0.35), Components (B): mean LOO-CV AIC = 188.66 (SD 2.60), Social-only (C): mean LOO-CV AIC = 186.87 (SD 2.57). These findings indicate that the social-only and components models fit the data better and generalize more effectively to programs with lesser attention. \u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ePredictive Performance\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eIn predictive accuracy analyses, the aggregate model (A) showed limited ability to reproduce observed Doximity reputation rankings. The correlation between predicted and observed ranks was low (Spearman \u0026rho; = 0.20), and the mean absolute error (MAE) was 5.85 ranks.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eModel B achieved a Spearman correlation of .62 and MAE of 4.42. Model C demonstrated Spearman correlation of .60 and MAE of 4.43. Thus, incorporating social media variables increased rank-order agreement roughly threefold and reduced prediction error by ~1.4 rank positions compared with the aggregate citation-based model.\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eNeurosurgery as a specialty has placed high importance historically on the research output of its faculty, fellows, residents, and even prospective applicants[\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e, \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e]. Academic acumen, as gauged by factors such as h-index, citations, and unique publication counts, is a key feature of any neurosurgery residency application. Likewise, program ranking itself has been highly influenced by and correlated to objective measures of research output, with more prestigious programs often placing higher importance on the aforementioned factors[\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e, \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e, \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e]. In recent years, online visibility has played an increasingly significant role[\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e]. Neurosurgeons across the country have been able to promote both their individual practice and program through online visibility. Our goal through this analysis is to provide insight into the degree of impact this online attention has on program reputation.\u003c/p\u003e \u003cp\u003eAt the paper level, exposure to any given publication may occur through means such as PubMed, Google, Scopus, or other search as well as promotion through social media. Impactful literature is often posted, tweeted, and blogged through many different sites, apps, and profiles. Our paper-level analysis, showing skew towards a few publications, supported the idea that some papers and their content gain massive amounts of traction online. Papers that employ eye-catching titles and easy-to-understand terminology can appeal to readers who may not have an advanced level of medical knowledge, increasing like, share, and repost counts. Such exposure increases visibility not only for the authors but also for their associated institution. For example, \u003cem\u003eHuman hippocampal neurogenesis drops sharply in children to undetectable levels in adults\u003c/em\u003e by Sorrells et al. had one of the highest attention scores in our analysis (2291), with an associated 1183 tweets and 167 news articles. Its understandable and appealing title likely contributed to its extremely high tweet count, allowing their findings to spread fast and generating more visibility for the authors and the neurosurgery department at University of California, San Francisco.\u003c/p\u003e \u003cp\u003eThe faculty-level analysis demonstrated a similar dynamic. Certain authors possessed extremely high levels of online attention, while the mean attention store across all was low, leading to another right skew. This finding once again supports the notion that some physicians are able to gather a large amount of social attention while the average neurosurgeon's online attention remained lower. Additionally, the correlation between mean attention and citations was moderate, indicating that the impact of those two variables is related yet distinct. Regression analysis also showed that mean citations per faculty and mean attention per faculty were moderately correlated, indicating that programs with more citations also garnered more attention. Moderate correlation indicates that attention is related to, but not fully correlated to academic output. A possible explanation for this trend is that programs that output more research also strive to increase their attention online. Departments that promote their own literature through their respective social media channels can increase read and citation count via that channel, which could explain the moderate correlation between those factors. Relying on journal publication alone as means to report findings may result in both lower social exposure and total online attention.\u003c/p\u003e \u003cp\u003eRegression modeling revealed interesting trends among the top 25 programs. In model A, aggregating social attention and citations failed to demonstrate significance, reinforcing the idea that those two variables influence rank separately. Model B, including all components of social attention, had a substantially lower AIC, revealing that evaluating total attention separately by social component and citations improves predictability. In the same manner, the social-only model had the best AIC; in particular, blog coverage was an extremely strong predictor. On the whole, social media engagement was a far stronger and more robust predictor of reputation rank, a conclusion which is reinforced by the results of the LOO cross-validation. Splitting attention into its two distinct components increased predictive value, reinforcing the idea that social attention is an entirely separate influencer of rank.\u003c/p\u003e \u003cp\u003ePublications are at the heart of this analysis due to their historical impact on reputation, and the ways in which they exert that influence seem to be changing. While journal content via website and print edition lend traditional exposure to paper content, increasing usage of platforms like Facebook, Twitter, and more allows authors to promote their content through such sites as well. Advantages offered by presenting publication material in that manner include shorter reading time, lack of journal paywall, easy sharing/reposting options, and greater engagement with readers. Physician and program-specific pages automatically present posted content to followers, generating clicks and increasing engagement with said pages and the entities they represent. Programs that lean into this dynamic can generate greater amounts of exposure and increase their reputation by demonstrating their important literature in a concise and easy to read manner.\u003c/p\u003e \u003cp\u003eWith this dynamic in mind, dissemination of study results through social media before peer review has taken place can lead to misconceptions and premature conclusions. An example of such rapid spread was observed when the NASCIS II clinical trial, which reported that high-dose methylprednisolone administration within 8 hours of spinal cord injury improved neurologic recovery[\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]. These results were widely promoted and implemented by institutions around the country. However, later review revealed that this conclusion was based on post-hoc analysis, and the actual benefit was controversial[\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e, \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e]. Today, modern guidelines do not recommend high dose steroids for acute spinal cord injury[\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e]. This instance reflects the direct effect that social attention has on clinical practice and should inspire caution when disseminating early preliminary study results.\u003c/p\u003e \u003cp\u003eIn similar fashion, blog coverage may also reflect deeper engagement within the neurosurgical community. Many posts highlight ongoing projects, clinical achievements, or new publications The University of Miami Department of Neurosurgery engages in this sort of activity with its blog \u003cem\u003eNeurosurgery Blog: More Than Just Brain Surgery\u003c/em\u003e[\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e]. Their site encourages posts from their own faculty as well as physicians from other programs across the country, publishing several in-depth articles monthly. This type of communication can shape perceptions of productivity and leadership in ways that citation counts do not, allowing direct communication between physicians from different programs.The results suggest that program visibility, especially through consistent online engagement, may now play a measurable role in how reputation is formed.\u003c/p\u003e \u003cdiv id=\"Sec23\" class=\"Section2\"\u003e \u003ch2\u003eLimitations and Future Directions\u003c/h2\u003e \u003cp\u003eSeveral limitations should be acknowledged in this study. The analysis was restricted to the top 25 neurosurgery residency programs, which may limit generalizability to mid- or lower-ranked institutions. Only the top three publications per faculty member were included, potentially underrepresenting overall scholarly output and online impact. Social media metrics such as tweets, blogs, and news mentions are also inherently ambiguous and may reflect institutional promotion or general public interest rather than true program-driven engagement. Although robust modeling approaches were used, the relatively small sample size reduces the ability to detect more subtle associations. Expanding publication capture beyond each faculty member\u0026rsquo;s top three works would allow a more comprehensive assessment of scholarly activity and visibility. Additionally, evaluating emerging platforms such as podcasts, TikTok, LinkedIn, and YouTube may provide a fuller understanding of how modern digital presence influences residency program reputation.\u003c/p\u003e \u003cp\u003eThis study is an exploratory analysis of elite programs; such institutions have garnered a level of fame and interest through long-term academic reputation that may not generalize well towards newer and less academically focused programs. The observational design prevents causal inference, as higher program reputation may itself drive greater online visibility just as increased visibility may influence perceived reputation. Future projects may expand on these conclusions by quantifying a larger or even all-encompassing dataset of neurosurgery residency academic metrics.\u003c/p\u003e \u003cp\u003eLastly, Doximity rankings, while multifactorial in nature, include certain subjective characteristics which may introduce bias. Some authors argue that Doximity rankings may include a circular methodology to characterize programs, where reputation influences rank which in turn influences reputation later on. As such, Doximity rankings should be viewed as a metric of reputation perception rather than an objective assessment of program training quality. Despite such limitations, the Doximity system is one of the most prominent and widely viewed[\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e], hence its utilization in this study.\u003c/p\u003e \u003c/div\u003e"},{"header":"Conclusion","content":"\u003cp\u003eIn the current era of digital communication, social media has become a central part of academic medicine, shaping how research is shared, reputations are formed, and professional networks expand. In neurosurgery, it has emerged as an important means for academic visibility and engagement. Platforms like Twitter, Instagram, LinkedIn, and YouTube allow surgeons to share new techniques, highlight clinical cases, discuss evidence-based updates, and disseminate research more rapidly than traditional journals alone.\u003c/p\u003e \u003cp\u003eThe results of this study suggest that social media activity contributes directly to program reputation. Among leading neurosurgery residency programs, online engagement, especially through blog mentions, was more strongly associated with higher Doximity rankings than traditional citation metrics. This reflects a shift in how academic influence is recognized: visibility and accessibility of research discussions online now play a meaningful role alongside established bibliometric indicators.\u003c/p\u003e \u003cp\u003eThis trend underscores the growing importance of digital presence in shaping academic identity. Programs that use social media strategically to communicate research, showcase achievements, and engage with the broader medical community may enhance both their visibility and perceived leadership. As social platforms continue to evolve, their integration into academic culture will likely remain essential for effective communication, collaboration, and public engagement in neurosurgery.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e \u003ch2\u003eConflicts of Interest\u003c/h2\u003e \u003cp\u003eThe authors have no conflicts of interests or disclosures to report.\u003c/p\u003e \u003c/p\u003e\u003cp\u003e \u003ch2\u003eCompeting Interests\u003c/h2\u003e \u003cp\u003eThe authors declare no competing interests.\u003c/p\u003e \u003c/p\u003e\u003ch2\u003eFunding\u003c/h2\u003e \u003cp\u003eThe authors received no financial support for the research, authorship, and/or publication of this article.\u003c/p\u003e\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\u003cp\u003ePatrick Pema conceived and led the study. Patrick Pema directed study design, coordinated project execution, interpreted the findings, and wrote the first draft of the manuscript. Patrick Pema designed the figures and outlined tables. Akil Anthony and Michael Chaga performed the statistical analyses and contributed to interpretation of the data. Youssef Atef AbdelAlim, Jack Kilgallon, Kush Desai, Paxton Sweeney, and Daniel Monahan contributed to data collection and manuscript editing. Ira Goldstein, Shabbar Danish, and Nitesh Patel contributed to manuscript appraisal, critical revision, and editorial guidance. All authors reviewed previous versions of the manuscript, read and approved the final manuscript.\u003c/p\u003e\u003ch2\u003eAcknowledgements\u003c/h2\u003e \u003cp\u003eThe authors would like to acknowledge the neurosurgery department at Jersey Shore University Medical Center and the library staff at Hackensack Meridian School of Medicine for their assistance with this project.\u003c/p\u003e\u003ch2\u003eData Availability\u003c/h2\u003e\u003cp\u003eThe data used in this study were derived from publicly available sources, including neurosurgery program rankings from Doximity.com, Scopus.com for author h indices and citation metrics, and Altmetric.com for social media occurrences using the Altmetric bookmarklet, a free tool available to the public. Current faculty were gathered from each program's own website. The processed dataset generated during the current study is available from the corresponding author on request.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eAmerican College of Surgeons Trauma Quality Programs Best Practices Project Team (2022) Best Practices Guidelines: Spine Injury. American College of Surgeons, Chicago, IL \u003c/li\u003e\n\u003cli\u003eBracken MB, Shepard MJ, Collins WF, Holford TR, Young W, Baskin DS, Eisenberg HM, Flamm E, Leo-Summers L, Maroon J, Marshall LF, Perot PL, Piepmeier J, Sonntag VKH, Wagner FC, Wilberger JE, Winn HR (1990) A Randomized, Controlled Trial of Methylprednisolone or Naloxone in the Treatment of Acute Spinal-Cord Injury. N Engl J Med 322:1405\u0026ndash;1411. doi: 10.1056/NEJM199005173222001 \u003c/li\u003e\n\u003cli\u003eChang AN, Boyett D, Chou D, Chan AK (2025) Updated 5-Year Institutional Bibliometric Profiles for United States Neurosurgery Residency Programs and the Relationship Between Social Media Presence and Objective Departmental Metrics. Neurosurgery 96:1397\u0026ndash;1409. doi: 10.1227/neu.0000000000003256 \u003c/li\u003e\n\u003cli\u003eCole KL, Carter A, Rawson C, Tenhoeve S, Orton C, Zeinali M, Karsy M (2025) The impact of NIH funding and program reputation score on research output and residency matches in neurosurgery: A bibliometrics analysis. J Natl Med Assoc 117:115\u0026ndash;122. doi: 10.1016/j.jnma.2025.03.002 \u003c/li\u003e\n\u003cli\u003eColeman WP, Benzel D, Cahill DW, Ducker T, Geisler F, Green B, Gropper MR, Goffin J, Madsen PW, Maiman DJ, Ondra SL, Rosner M, Sasso RC, Trost GR, Zeidman S (2000) A critical appraisal of the reporting of the National Acute Spinal Cord Injury Studies (II and III) of methylprednisolone in acute spinal cord injury. J Spinal Disord 13:185\u0026ndash;199. doi: 10.1097/00002517-200006000-00001 \u003c/li\u003e\n\u003cli\u003eColombo E, H\u0026ouml;bner LM, Blom V, Berglar I, Alakmeh A, de Wilde D, El-Hajj VG, Regli L, Serra C, Staartjes VE, Burstr\u0026ouml;m G (2025) The Implementation of Social Media in Neurosurgery: A Systematic Review of the Literature. Acta Neurochir (Wien) 167:277. doi: 10.1007/s00701-025-06695-1 \u003c/li\u003e\n\u003cli\u003eFeinstein MM, Niforatos JD, Mosteller L, Chelnick D, Raza S, Otteson T (2019) Association of Doximity Ranking and Residency Program Characteristics Across 16 Specialty Training Programs. J Grad Med Educ 11:580\u0026ndash;584. doi: 10.4300/JGME-D-19-00336.1 \u003c/li\u003e\n\u003cli\u003eHurlbert RJ (2000) Methylprednisolone for acute spinal cord injury: an inappropriate standard of care. J Neurosurg 93:1\u0026ndash;7. doi: 10.3171/spi.2000.93.1.0001 \u003c/li\u003e\n\u003cli\u003eIannone R, Cheng J, Schloerke B, Hughes E, Lauer A, Seo J, Brevoort K, Roy \u0026Ouml; (2025) gt: Easily Create Presentation-Ready Display Tables \u003c/li\u003e\n\u003cli\u003eLee RP, Venable GT, Roberts ML, Parikh KA, Taylor DR, Khan NR, Michael LM, Klimo P (2016) Five-Year Institutional Bibliometric Profiles for 119 North American Neurosurgical Residency Programs: An Update. World Neurosurg 95:565\u0026ndash;575. doi: 10.1016/j.wneu.2016.07.006 \u003c/li\u003e\n\u003cli\u003ePonce FA, Lozano AM (2010) Academic impact and rankings of American and Canadian neurosurgical departments as assessed using the h index. J Neurosurg 113:447\u0026ndash;457. doi: 10.3171/2010.3.JNS1032 \u003c/li\u003e\n\u003cli\u003eSarkiss CA, Riley KJ, Hernandez CM, Oermann EK, Ladner TR, Bederson JB, Shrivastava RK (2017) Academic Productivity of US Neurosurgery Residents as Measured by H-Index: Program Ranking with Correlation to Faculty Productivity. Neurosurgery 80:975\u0026ndash;984. doi: 10.1093/neuros/nyx071 \u003c/li\u003e\n\u003cli\u003eTaylor DR, Venable GT, Jones GM, Lepard JR, Roberts ML, Saleh N, Sidiqi SK, Moore A, Khan N, Selden NR, Michael LM, Klimo P (2015) Five-year institutional bibliometric profiles for 103 US neurosurgical residency programs. J Neurosurg 123:547\u0026ndash;560. doi: 10.3171/2014.10.JNS141025 \u003c/li\u003e\n\u003cli\u003eVenables WN, Ripley BD (2002) Modern Applied Statistics with S. Springer, New York \u003c/li\u003e\n\u003cli\u003eWickham H (2016) ggplot2: Elegant Graphics for Data Analysis. Springer-Verlag, New York \u003c/li\u003e\n\u003cli\u003eWickham H, Fran\u0026ccedil;ois R, Henry L, M\u0026uuml;ller K, Vaughan D (2023) dplyr: A Grammar of Data Manipulation \u003c/li\u003e\n\u003cli\u003eWickham H, Vaughan D, Girlich M (2024) tidyr: Tidy Messy Data \u003c/li\u003e\n\u003cli\u003eDoximity Residency Navigator. In: Doximity. https://www.doximity.com/residency/programs?specialtyKey=eb85a000-2e4f-4977-ac28-f5de6e72ebc9-neurological-surgery. Accessed 10 Feb 2026 \u003c/li\u003e\n\u003cli\u003eBookmarklet. In: Altmetric. https://www.altmetric.com/solutions/free-tools/bookmarklet/. Accessed 7 Dec 2025 \u003c/li\u003e\n\u003cli\u003eNeurosurgery Blog. In: Neurosurg. Blog. https://www.neurosurgeryblog.org/. Accessed 26 Dec 2025 \u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"neurosurgical-review","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"nrev","sideBox":"Learn more about [Neurosurgical Review](https://www.springer.com/journal/10143)","snPcode":"10143","submissionUrl":"https://submission.nature.com/new-submission/10143/3","title":"Neurosurgical Review","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"Social media, Altmetric, neurosurgery, residency reputation, bibliometrics","lastPublishedDoi":"10.21203/rs.3.rs-9204678/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9204678/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eIntroduction\u003c/h2\u003e \u003cp\u003eSocial media presence exerts a growing influence on visibility both at the individual and program level. In this study, we aim to assess the degree of influence that citation impact and social attention metrics assert by determining the degree to which these measures predict program reputation.\u003c/p\u003e\u003ch2\u003eMethods\u003c/h2\u003e \u003cp\u003eA paper level dataset for U.S. Neurosurgery residency program faculty was compiled from the Top 25 programs by reputation as listed by doximity.com, including each physician\u0026rsquo;s three most cited publications, and aggregated to the program level. The Altmetric tool was used to gather social media occurrences. We modeled program rank as an ordered outcome using ordinal logistic regression with three specifications: (A) aggregate attention\u0026thinsp;+\u0026thinsp;citations, (B) disaggregated attention (tweets, blogs, news) + citations, and (C) social-only. Model fit and generalization were compared via AIC, likelihood-ratio tests, and leave-one-out cross-validation. Predictive performance was summarized with Spearman correlation and mean absolute error (MAE).\u003c/p\u003e\u003ch2\u003eResults\u003c/h2\u003e \u003cp\u003eModels emphasizing social attention consistently outperformed citation-based approaches across all analytic metrics. The disaggregated social-attention model demonstrated the best overall performance, achieving an AIC of 198.5 versus 211.9 for aggregate attention\u0026thinsp;+\u0026thinsp;citations, and a significantly improved likelihood-ratio test (p\u0026thinsp;\u0026lt;\u0026thinsp;0.001). Predictive accuracy was highest for social models (Spearman ρ\u0026thinsp;\u0026asymp;\u0026thinsp;0.60\u0026ndash;0.62; MAE\u0026thinsp;\u0026asymp;\u0026thinsp;4.4).\u003c/p\u003e\u003ch2\u003eConclusions\u003c/h2\u003e \u003cp\u003eNeurosurgery program reputation correlates more strongly with multi-platform social engagement than with reference volume. Integrating research dissemination with strategic digital communication may enhance both visibility and perceived academic prestige.\u003c/p\u003e","manuscriptTitle":"Social Media Attention and Citation Metrics as Predictors of U.S. Neurosurgery Residency Reputation","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-05-09 05:33:35","doi":"10.21203/rs.3.rs-9204678/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"editorInvitedReview","content":"","date":"2026-05-07T04:54:20+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"295428192997512487429755901880183443431","date":"2026-05-04T01:17:17+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"103377787118577634129268168349502736378","date":"2026-04-30T09:12:13+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2026-04-30T00:01:11+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-04-29T15:58:53+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-03-26T05:44:15+00:00","index":"","fulltext":""},{"type":"submitted","content":"Neurosurgical Review","date":"2026-03-23T21:37:33+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"neurosurgical-review","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"nrev","sideBox":"Learn more about [Neurosurgical Review](https://www.springer.com/journal/10143)","snPcode":"10143","submissionUrl":"https://submission.nature.com/new-submission/10143/3","title":"Neurosurgical Review","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"18d204c0-9541-4ff6-91d3-307e9a3cd351","owner":[],"postedDate":"May 9th, 2026","published":true,"recentEditorialEvents":[{"type":"editorInvitedReview","content":"","date":"2026-05-07T04:54:20+00:00","index":38,"fulltext":""},{"type":"reviewerAgreed","content":"295428192997512487429755901880183443431","date":"2026-05-04T01:17:17+00:00","index":35,"fulltext":""},{"type":"reviewerAgreed","content":"103377787118577634129268168349502736378","date":"2026-04-30T09:12:13+00:00","index":30,"fulltext":""},{"type":"reviewersInvited","content":"24","date":"2026-04-30T00:01:11+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-04-29T15:58:53+00:00","index":"","fulltext":""}],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2026-05-09T05:33:35+00:00","versionOfRecord":[],"versionCreatedAt":"2026-05-09 05:33:35","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9204678","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9204678","identity":"rs-9204678","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}