Keywords
How to Cite
Abstract
Artificial intelligence (AI) and health analytics are moving from isolated diagnostic tools toward infrastructure for learning health systems. This systematic review synthesizes recent evidence on how AI supports digital transformation, decision intelligence, and operational resilience across modern healthcare. A structured search of PubMed/MEDLINE and supplementary publisher- and index-based searches was conducted, emphasizing peer-reviewed studies published from 2020 through 2026 while retaining a small number of foundational studies from 2018-2019. Search concepts combined artificial intelligence, machine learning, deep learning, natural language processing, large language models, electronic health records, medical imaging, predictive analytics, clinical decision support, digital transformation, patient flow, and operational resilience. Sixty-eight candidate records were assembled, six duplicates were removed, 62 records were screened, 45 reports underwent full-text or complete-abstract eligibility assessment, and 40 core studies were included in the evidence matrix. The synthesis identified six interacting domains: predictive clinical analytics; electronic health record and language intelligence; imaging and multimodal diagnostics; precision, population, and privacy-preserving analytics; operational intelligence; and governance-enabled decision intelligence. Across these domains, AI improved risk stratification, diagnostic interpretation, documentation, triage, admission and discharge forecasting, and resource planning, but gains were frequently constrained by retrospective designs, local data dependence, dataset shift, limited external validation, workflow friction, and uncertainty about accountability. Recent work increasingly reframes high-performing models as only one component of trustworthy deployment, emphasizing fairness, traceability, explainability, robustness, usability, privacy, continuous monitoring, and human oversight. The review proposes an AI Healthcare Transformation Framework in which data infrastructure and AI models generate decision intelligence that is embedded into clinical and operational workflows under governance and organizational-readiness controls. Operational resilience emerges when these capabilities enable anticipation, absorption, adaptation, recovery, and organizational learning. The review concludes that the strategic value of AI depends less on stand-alone algorithmic accuracy than on socio-technical integration, trustworthy governance, interoperable data, and measured improvements in patient and system outcomes.