Mapping EGS Care Networks With Modularity Optimization

TOPLINE:
Modularity optimization (MO) provides superior localization of emergency general surgery (EGS) care regions and hospital communities than the traditional Dartmouth Health Referral Regions (HRRs) by accurately reflecting current patient utilization patterns and facilitating the development of coordinated regional care systems for high-risk patients.
METHODOLOGY:
Researchers used data from the Healthcare Cost and Utilization Project (2019) and the American Hospital Association survey (2020) and analyzed 1,244,868 adults with a nonelective admission for 12 common EGS conditions (405,493 from California and 839,375 from New York).
The Louvain community detection method, particularly MO, identified regional EGS networks (RENs) based on EGS patient flow and spatial accuracy and compared the results with those of traditional HRRs.
The spatial accuracy of the MO RENs and HRRs was assessed using six network analysis measures: Localization index (local care), market share index (patient volume from outside areas), net patient flow (patient movement patterns), connectivity (hospital transfers), compactness (regularity of region shape), and modularity (whole network connections).
TAKEAWAY:
MO identified 9 and 14 RENs in New York and California, respectively, vs 10 and 24 HRRs, showing improved network accuracy with MO.
MO was superior to HRRs in all spatial accuracy metrics, except compactness, with higher modularity scores in both New York (0.69 vs 0.63) and California (0.74 vs 0.69) and stronger connectivity in California (430.1 vs 215.6).
A higher localization index indicated greater local care in both New York (0.86 vs 0.74; P <.001) and California (0.83 vs 0.74; P <.001), and lower market share index scores indicated smoother patient flow in both New York (0.16 vs 0.32; P =.07) and California (0.19 vs 0.39; P < .001).
Compared with HRRs, MO reclassified several hospitals, improving network alignment in both New York (37 [26.6%] hospitals) and California (48 [14.3%] hospitals).
IN PRACTICE:
“The findings of this cross-sectional study suggest that network science methods, such as MO, offer opportunities to identify empirical EGS care regions that outperform HRRs and can be applied in the development of coordinated regional systems of care,” the authors wrote.
SOURCE:
The study was led by Jiuying Han, PhD, Department of Geography, University of Utah, Salt Lake City. It was published online on October 15, 2024, in JAMA Network Open.
LIMITATIONS:
The analysis relied on administrative claims data without clinician identifiers, limiting the verification of surgeon involvement. Geographic precision was restricted to zip code areas, which may not be suitable for densely populated urban regions. Additionally, single-state datasets were used, which may not have captured cross-border patient movement.
DISCLOSURES:
This study was supported by the One Utah Data Science Hub pilot seed grant program through the University of Utah Data Exploration and Learning for Precision Health Intelligence Initiative. No conflicts of interest were reported.
This article was created using several editorial tools, including AI, as part of the process. Human editors reviewed this content before publication.
 
Send comments and news tips to [email protected].