Regulatory Compliance and AI in Healthcare: A Hospital Case Study

Thesis Abstract

The rapid integration of artificial intelligence (AI) tools within hospital settings presents transformative benefits for clinical decision-making, operational efficiency, and patient safety, yet raises significant regulatory and ethical concerns related to data governance, accountability, and compliance with health information laws. This study investigates how a mid-size tertiary hospital navigates regulatory compliance amid AI-enabled workflows, identifying the governance structures, risk controls, and stakeholder practices that influence lawful and ethical deployment. The aim is to develop a robust understanding of how regulatory mechanisms interact with AI adoption to affect clinical outcomes, patient safety, and privacy preservation. Specific objectives are to (1) map the regulatory landscape governing AI in healthcare at the hospital level, (2) assess the effectiveness of governance frameworks in ensuring data protection, accountability, and algorithmic transparency, (3) analyze clinician and IT governance practices affecting AI uptake, (4) evaluate the impact of compliance processes on clinical decision quality and patient safety indicators, and (5) propose a policy and practice blueprint for harmonising AI innovation with regulatory requirements. The study adopts a mixed-methods design, combining a cross-sectional survey of clinical, IT, and compliance staff (n = 180 respondents) with in-depth semi-structured interviews (n = 30) and a document analysis of internal policies, risk assessments, incident reports, and audit records over a 24-month period. The hospital chosen for the case study is a 600-bed institution affiliated with a national university, implementing AI-powered decision support, imaging analytics, and patient flow optimization. Data collection instruments include a structured questionnaire validated for reliability (Cronbach’s alpha > 0.80) and interview guides designed to elicit nuanced insights into regulatory interpretations, accountability mechanisms, and incident reporting cultures. Validity and reliability procedures involve pilot testing, triangulation across sources, and intercoder reliability checks (Cohen’s kappa ? 0.75) for qualitative coding. Quantitative data will be analyzed using multiple regression to test relationships between governance maturity, compliance processes, and outcomes such as incident rates and decision turnaround times, complemented by ANOVA to explore group differences by professional role. Qualitative data will be analyzed thematically via NVivo, guided by the institutional theory of regulation and the sociotechnical systems framework, with a priori codes aligned to data governance, risk management, and algorithmic transparency. A conceptual model will be developed to illustrate the interdependencies among regulatory input, AI system characteristics, and clinical performance outcomes. Anticipated findings indicate that mature governance structures—encompassing risk-based approvals, rigorous data quality controls, transparent model documentation, and clear accountability lines—are positively associated with reduced misdiagnosis risk, improved data integrity, and lower regulatory incident rates. The study expects to reveal gaps in clinician- and administrator-facing accountability mechanisms, uneven incident reporting, and variances in interpreted regulatory requirements across departments. It is anticipated that transparent, auditable AI systems with standardized risk assessments will correlate with higher clinician trust, better adoption rates, and more consistent patient safety outcomes. The contribution to knowledge lies in empirically linking regulatory compliance processes to AI-enabled clinical performance, offering a validated model of hospital governance that integrates legal compliance, ethical considerations, and technical safeguards with practical implications for policy and practice. The findings will extend current regulatory theory by applying institutional and socio-technical perspectives to AI in a real-world hospital environment, highlighting how organizational capabilities mediate regulatory impact on patient care. The study will conclude with evidence-based recommendations for hospital policymakers, regulators, and AI vendors, including (1) a standardized regulatory governance framework tailored to hospital AI deployments, (2) a set of audit-ready documentation practices and risk assessment protocols, (3) guidelines for model transparency, clinical accountability, and incident learning, (4) strategies to align clinician workflows with compliance requirements without compromising clinical efficacy, and (5) a roadmap for continuous regulatory readiness as AI technologies evolve. These recommendations aim to enhance patient safety, protect privacy, and accelerate responsible AI adoption in healthcare.