Machine learning models in clinical decision support systems: a scoping review

Abstract

Machine learning (ML) has increasingly been integrated into clinical decision support systems (CDSS) to improve diagnostic accuracy and clinical decision-making. However, variability in performance, adoption, and implementation across healthcare settings necessitates comprehensive evidence synthesis. This scoping review aimed to map and synthesise the existing evidence on ML models in CDSS, focusing on their performance, clinical utility, adoption, explainability, and implementation challenges. This review followed PRISMA-ScR guidelines. A systematic search was conducted across PubMed, Scopus, Web of Science, IEEE Xplore, and CINAHL from inception to 30th March 2026. Studies were included based on the Population–Concept–Context (PCC) framework, focusing on ML applications within clinical decision support systems. Data were extracted and synthesised thematically. Twenty studies met the inclusion criteria. Five key themes emerged: (1) high diagnostic and predictive performance of ML models, (2) enhanced clinical decision-making and workflow efficiency, (3) the importance of user trust and acceptance, (4) the role of explainability in improving interpretability, and (5) significant implementation barriers including infrastructure and regulatory challenges. While ML-CDSS show strong potential, their effectiveness is influenced by contextual and system-level factors. ML-driven CDSS offer substantial benefits for healthcare delivery, but their successful integration requires addressing challenges related to trust, explainability, and health system readiness. Future research should prioritise real-world validation and context-sensitive implementation strategies. Not applicable.