408 Network Analysis of Neurosurgical Publication Coauthorship: Centrality Ranking of Authors and Institutions with Correlation to Clinical Activity and Industry Payments

Abstract

INTRODUCTION: Network analysis has been applied in a number of fields to understand relationships between members of different communities and organizations. METHODS: All publications from January 1, 2017, to April 30, 2020, were queried from 4 neurosurgical journals: Neurosurgery, Journal of Neurosurgery, Journal of Neurosurgery: Pediatrics, and Journal of Neurosurgery: Spine. Publication type excluded editorials, letters, and reviews. Authors with at least 7 publications in these journals were included. Measures of network centrality were calculated after conversion of authorship into a nondirected graph with links weighted by frequency of coauthorship. Latent communities in the network were detected with the Walktrap algorithm. RESULTS: Based on 811 included authors, network topology broadly demonstrated 4 poles related to subspecialties of vascular, spine, pediatrics, and radiosurgery/tumor. Of the top 100 coauthorship links, 42% of the authors were involved in 2 large study groups. We present rankings of neurosurgical authors based on network measures of centrality and evaluate how centrality metrics correlate to individual author h-index, Medicare utilization, and payments from industry sources. We further evaluate network centrality of institutions and examine institutional connectivity to discovered communities within the network. CONCLUSION: While traditional metrics like h-index demonstrate long-term research involvement, network analysis can be applied to understand current relationships between academic authors and evaluate centrality within a publishing network during a given timeframe.