Legal Infoveillance of Unlicensed Medical Practices in South Korea Through Criminal Court Decisions Using Machine...

Unlicensed medical practices (UMPs) pose a substantial threat to patient safety and public health, but their clandestine nature makes them difficult to monitor through conventional surveillance systems. Legal epidemiology offers a framework for using judicial data to study hidden health-related misconduct, and machine learning (ML) may help convert unstructured legal texts into analyzable public health information.

This study aimed to characterize prosecuted UMP cases in South Korea using a legal infoveillance framework and evaluate the utility of ML-assisted extraction from criminal court decisions for public health surveillance.

We conducted a retrospective observational study of 1532 criminal court decisions involving UMP-related convictions in South Korea between 2005 and 2023. Using an ML-assisted extraction pipeline with human-in-the-loop verification, we transformed unstructured judicial texts into structured legal and medical variables. Analyses were conducted at the case, charge, defendant category, and legal ruling levels. In addition to descriptive analyses, we performed exploratory inferential analyses to examine factors associated with legal rulings and professionals' involvement.

Of 1718 charge entries, 987 (57.5%) were related to Article 5 of the Act on Special Measures for the Control of Public Health Crimes, and 731 (42.5%) were related to Article 27(1) of the Medical Service Act. Profit motive was coded in 91.6% (1404/1532) of the cases. At the legal ruling level (n=2004 entries), suspended sentences, meaning sentences whose execution was conditionally suspended under Korean criminal law, were the most common outcome (1261/2004, 62.9%), followed by fines (421/2004, 21%) and imprisonment without suspension (209/2004, 10.4%). Of 1716 defendant category entries, ordinary persons accounted for 1294 (75.4%), health care professionals accounted for 264 (15.4%), and health care providers accounted for 158 (9.2%). Physicians were the largest subgroup among health care professionals. In exploratory multinomial models, licensed personnel-only and mixed ordinary person and licensed personnel cases were more likely than ordinary person-only cases to result in fines or imprisonment without suspension rather than suspended sentences. Secondary exploratory analyses also suggested distinctive patterns of professional involvement and possible scope-of-practice or delegation-related boundary violations.

ML-assisted analysis of criminal court decisions can serve as a useful supplementary surveillance method for hidden UMPs. In South Korea, prosecuted UMPs were predominantly profit driven and involved both ordinary persons and licensed personnel. The findings support closer monitoring of scope-of-practice and delegation-related violations and demonstrate the value of judicial records as a source of public health intelligence.