Implementing Artificial Intelligence in Clinical Decision Support, Diagnostics And Predictive Analytics In Healthcare: A domain-contingent AI implementation model

Objective: The study aimed to analyse Artificial Intelligence (AI) techniques for clinical decision support, diagnostic analysis and predictive analytics for better implementation in healthcare. Design: The study adopted a systematic literature review guided by Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines to identify 17 studies related to three domains: clinical decision support, diagnostic analysis and predictive analytics. Thematic analysis was conducted to propose a conceptual model, the Domain-Contingent AI Implementation Model in Healthcare, to suggest that AI effectiveness relies on the domain-specific drivers, constraints, and system-level enablers. Main outcome measures: The study evaluates AI implementation against the drivers and constraints of three domains, derived through thematic analysis in healthcare settings. Results: Findings emphasised the nature of AI as a domain-contingent capability across the three healthcare domains. AI effectiveness in healthcare is driven by interpretability, data intensity and forecasting power, while it is constrained by trust, black-box models and a validation gap in clinical decision support, diagnostic analysis and predictive analytics, respectively. Conclusion: The study concluded that AI’s effectiveness in healthcare implementation is driven by adaptation factors, including usability, infrastructure, and reliability.