Are network growth and the contributions to congresses associated with publication success? A pediatric oncology model

Background: The consistent focus of ‘Advances in Neuroblastoma Research’ congresses on the topic neuroblastoma sets it as a model for a circumscribed scientific community. Methods: The contributions of authors, institutions and countries to congress abstracts and their collaborations were compared to the Hirsch index (h-index) calculated from the Web of Science publication output on the topic ‘neuroblastoma’. Results: From 1975 to 2016, 18 congresses were held. 8459 authors affiliated to 553 institutions of 53 countries presented 3,993 abstracts. The number of coauthors increased over the years from 2 to 7. A considerable proportion of authors, institutions and countries presented only once (53.7%/25.7%/13.2%). Authors with a high number of abstracts and with a large local network were often among those with a higher publication rate and success (R2 = 0.508 for Pearson’s correlation between weight and h-index, R2 = 0.474 for degree centrality, R2 = 0.364 for lobby-index). Closeness and betweenness centralities were less correlated (R2 = 0.127/R2 = 0.33, resp.). The institutions showed a similar impact of local interactions on publication success (degree centrality R2 = 0.417, weight R2 = 0.308), while countries demonstrated a higher correlation of betweenness centrality and h-Index (R2 = 0.704) emphasizing their brokerage role. Of 553 institutions, 520 collaborated within 13 communities and belonged to the large scientific network. 33 satellite institutions had no connections to the central network. They attended 1–4 congresses over a period of 1–16 years. Conclusion: A large scientific network has been developed during the recent 42 years. Growth and interaction at congresses were correlated to publication success. Weight is suggested as a useful and simple estimate.