Seven Decades of Chemotherapy Clinical Trials: A Pan-Cancer Social Network Analysis

Abstract

AbstractBackgroundClinical trials establish the standard of care for cancer and other diseases. While social network analysis has been applied to basic sciences, the social component of clinical trial research is not well characterized. We examined the social network of cancer clinical trialists and its dynamic development over more than 70 years, including the roles of subspecialization and gender in relation to traditional and network-based metrics of productivity.MethodsWe conducted a social network analysis of authors publishing chemotherapy-based prospective trials from 1946-2018, based on the curated knowledge base HemOnc.org, examining: 1) network density; 2) modularity; 3) assortativity; 4) betweenness centrality; 5) PageRank; and 6) the proportion of co-authors sharing the same primary cancer subspecialty designation. Individual author impact and productive period were analyzed as a function of gender and subspecialty.FindingsFrom 1946-2018, the network grew to 29,197 authors and 697,084 co-authors. While 99.4% of authors were directly or indirectly connected as of 2018, the network had very few connections and was very siloed by cancer subspecialty. Small numbers of individuals were highly connected and had disproportionate impact (scale-free effects). Women were under-represented and likelier to have lower impact, shorter productive periods (P<0.001 for both comparisons), less centrality, and a greater proportion of co-authors in their same subspecialty. The past 30 years were characterized by a trend towards increased authorship by women, with new author parity anticipated in 2032. However, women remain a distinct minority of first/last authors, with parity not anticipated for 50+ years.InterpretationThe network of cancer clinical trialists is best characterized as a strategic or “mixed-motive” network, with cooperative and competitive elements influencing its appearance.Network effects e.g., low centrality, which may limit access to high-profile individuals, likely contribute to ongoing disparities.FundingVanderbilt Initiative for Interdisciplinary Research; National Institutes of Health; National Science FoundationResearch in contextEvidence before this studyWe reviewed the literature on social networks from the 1800’s to 2018. Additionally, MEDLINE was searched for (“Social Networking”[Mesh] OR “Social Network Analysis”) AND (“Clinical Trials as Topic”[Mesh] OR “Hematology”[Mesh] OR “Medical Oncology”[Mesh]) without date restriction. The MEDLINE search yielded 43 results, of which 8 were relevant; none considered gender nor temporality in their analyses. To our knowledge, there has not been any similar study of the dynamic social network of clinical trialists from the inception of the fields of medical oncology and hematology to the present.Added value of this studyThis is the first dynamic social network analysis of cancer clinical trialists. We found that the network was sparse and siloed with a small number of authors having disproportionate impact and influence as measured by network metrics such as PageRank; these metrics have become more disproportionate over time. Women were under-represented and likelier to have lower impact, shorter productive periods, less network centrality, and a greater proportion of co-authors in their same cancer subspecialty.Implications of all the available evidenceWhile gender disparities have been demonstrated in many fields including hematology/oncology, our analysis is the first to show that network factors themselves are significantly implicated in gender disparity. The increasing coalescence of the network by traditional cancer type and around a small number of high-impact individuals implies challenges when the field pivots from traditionally disease-oriented subspecialties to a precision oncology paradigm. New mechanisms are needed to ensure diversity of clinical trialists.