A network analysis of faculty training and hiring in US academic neurosurgery departments

Assembling faculty complements with diverse clinical backgrounds is critical to advancing patient care and trainee education in academic medicine. Whereas institutional relationships in other fields can be evaluated via collaborative grants and publications, metrics are lacking to characterize clinical departments and institutional networks. The aim of this study was to characterize connectivity among academic neurosurgery departments using social network analysis of individual-level faculty training histories and institutional characteristics, and to develop novel metrics within the specialty landscape.

The authors performed a quantitative network analysis of US academic neurosurgery departments by integrating geographical, research funding, leadership, ranking, and faculty training information. Three metrics were developed to characterize each institution: 1) influence on network connectivity, 2) representation of different training backgrounds among faculty as a measure of institutional diversity, and 3) alumni career mobility.

The network of 115 programs and 1910 faculty members comprised a small group of strongly linked institutions surrounded by a larger sparsely connected group (modularity = 0.2402, skewness coefficient = 1.371). Clustering and influence were associated with geographic proximity, private institutional status, external rankings and department size (p < 0.001 for each). Institutional diversity and influence were negatively correlated (τ = -0.287, p < 0.001), suggesting that influential programs might be the most siloed. Mobility was significantly correlated with program leadership but not with influence or institutional diversity, suggesting that structural factors might not limit alumni trajectories. Programs clustered into one of 3 archetypes with respect to their relative strengths in influence, institutional diversity, and/or mobility; no programs had high scores in all 3 areas.

The network of academic neurosurgery programs is nonrandom and highly structured. Connectivity, institutional cross-pollination, and alumni mobility are limited and unevenly distributed. Characterizing academic programs using graph theoretical approaches has the potential to advance clinical practice, trainee education, transparency, and equity.