Risk-Oriented Analysis and Classification of Network Attacks Affecting Hospital Information Systems: Evidence from a Patient-Monitoring Testbed

Vollständiger Abstract

Worum geht es in dieser Arbeit?

Hospital Information Systems increasingly rely on connected patient-monitoring devices that transmit biometric data through hospital networks. This study analyzed the WUSTL-EHMS-2020 testbed and assessed the potential risk of its recorded attacks to HIS. A quantitative descriptive secondary analysis retained all 16,318 verified records. Record-level summaries described Normal, Spoofing, and Data Alteration traffic; the primary comparison aggregated ordered records into 169 contiguous binary-label blocks (85 Normal and 84 attack). Nine network-flow and eight biometric variables were summarized, and Cliff's delta with 95% percentile block-bootstrap confidence intervals was estimated using 10,000 repetitions and seed 20260727. The dataset contained 14,272 Normal records and 2,046 attack records, comprising 1,124 Spoofing and 922 Data Alteration records. Network-flow variables showed strong block-level separation (absolute delta 0.7891–0.9655; all intervals excluded zero), while all biometric effects were negligible (absolute delta at most 0.0594; all intervals included zero). Researcher-defined assessment classified Spoofing (score 12) and Data Alteration (score 15) as High risk. The study provides a reproducible bridge from verified testbed evidence to a separate, transparent HIS risk assessment; its conclusions remain limited to the controlled dataset and its undocumented attack-category field.