Analysis of the distribution of trial sites in South Korea using social network analysis.

Abstract

Location of trial sites can be a potential source of study bias. Considering that clinical trials have been mostly conducted in urban areas, the distribution of trial sites need to be evaluated. We analyzed clinical trial approval data using social network analysis to quantitatively assess the site-by-site connections. The approval list of clinical trials from the Ministry of Food and Drug Safety database between 2014 and 2021 was analyzed. The number of clinical trials per trial site was counted according to the approval year and study phase and evaluated for distribution using empirical cumulative distribution function plots. Trial sites and conducts of a clinical trial were mapped into nodes and edges in the social network analysis, and basic network parameters were obtained. The clinical trials were concentrated at several trial sites. Forty-nine to 60.6% of phase 1 and up to 30% of the other study phases of clinical trials were at the top 5 trial sites. The annual distribution of the number of clinical trials per site was comparable across the study period. Connections among the trial sites in the metropolitan area were prominent. Graph size and density were higher in phase 3 trials than in the other phases. We demonstrated that the conduct of clinical trials was concentrated in the Seoul Metropolitan Area in both number of trials and connections using social network analysis.