Clinically Weighted Zero Trust: Securing Healthcare AI Where Lives Depend on It

Mostafa Rahmany

Abstract: Health systems continue to be a significant focus of ransomware; when large-scale attacks (e.g., the Change Healthcare ransomware breach in 2024 (Kannarkat, 2024; Alder, 2025; TechCrunch, 2025)) seemed to have a domino effect of disrupting claims, eligibility, and pharmacy services across the country, with an unprecedented impact on patients. Current defences are over-reliant on flat networks, VPN trust, and endpoint agents that are incompatible with regulated medical devices. The NHS WannaCry post-incident review (National Audit Office, 2017) and the HSE Ireland post-incident review (HSE,2022) confirm the lack of patching, segmentation, and response orchestration. Design and test ClinDefend-ZTA, a clinical-safety-first Zero-Trust and deception hospital architecture that: (i) minimises both lateral movement and care harm; (ii) isolates any vulnerable devices flagged by SBOM scanners; and (iii) can contain breaches faster than legacy tools, observably reducing time-to-contain (TTC) and service outage. The policy model is selected that assigns device criticality to authentication, authorization, and micro-segmentation strengths; introduces decoys (credentials, DICOM shares, HL7 test endpoints) at convergence points, and incorporates runtime SBOM gates to steer device traffic. To perform the said evaluation, it is based on a lab testbed, a retrospective what-if replay of three real breaches (NHS, 2017; HSE, 2021; U.S. 2024 clearinghouse event). Prototypes are expected to achieve TTC reduction by 50+ per cent, impacted subnet reduction by 40+ per cent, and zero clinical safety incidents during simulated lock-downs, and this qualifies as the performance norm compared to perimeter+EDR baselines. Hospitals can implement ClinDefend-ZTA with existing identity, MFA, NAC, and micro-segmentation products, in line with FDA device cybersecurity recommendations and HHS/HICP best practices, without having to rely on ML/AI products and therefore being impeded by their validation.

Keywords: zero trust, hospital AI security, ransomware resilience, software bill of materials, patient safety

How to Cite?: Mostafa Rahmany, "Clinically Weighted Zero Trust: Securing Healthcare AI Where Lives Depend on It", Volume 14 Issue 9, September 2025, International Journal of Science and Research (IJSR), Pages: 582-594, https://www.ijsr.net/getabstract.php?paperid=SR25913190836, DOI: https://dx.doi.org/10.21275/SR25913190836