Advancing Healthcare AI Governance: A Comprehensive Maturity Model Based on Systematic Review

Abstract

Abstract Artificial Intelligence (AI) deployment in healthcare is accelerating, yet comprehensive governance frameworks remain fragmented and often assume extensive resources. Through a systematic review of 22 frameworks published between 2019-2024, we identified seven critical domains of healthcare AI governance: organizational structure, problem formulation, external product evaluation, algorithm development, model evaluation, deployment integration, and monitoring maintenance. While existing frameworks provide valuable guidance, they frequently target only large academic medical centers, creating barriers for smaller healthcare organizations. To address this gap, we propose the Healthcare AI governance Readiness Assessment (HAIRA), a five-level maturity model that provides actionable governance pathways based on organizational resources and capabilities. HAIRA spans from Level 1 (Initial / Ad Hoc) suitable for small practices to Level 5 (Leading) for major academic centers, with specific benchmarks across all seven governance domains. This tiered approach enables healthcare organizations to assess their current AI governance capabilities and establish appropriate advancement targets. Our framework addresses a critical need for adaptive governance strategies that can support AI-enabled healthcare value across diverse settings and ensures that AI implementation delivers tangible benefits to healthcare systems of varying sizes and resource levels.

Ähnliche Arbeiten