Stakeholder requirements, explainable AI techniques, HCI design, and evaluation of AI-driven clinical decision support systems supporting shared decision-making: a scoping review

Counts differ because each database indexes a different set of publications. We treat OpenAlex as the canonical count; Google Scholar is not shown (no API, and crawling it violates its ToS).

Artificial intelligence (AI) and specifically Machine Learning (ML) have created a paradigm shift in healthcare by enhancing the capabilities of Clinical Decision Support Systems (CDSS). However, the adoption of CDSS in healthcare remains limited due to challenges related to transparency, trust, and meaningful human collaboration. Four areas of literature respond to this challenge: stakeholder taxonomy and requirements, explainable AI (XAI) techniques, human-computer interaction (HCI) design for explanation delivery, and evaluation of XAI and CDSS. Each of these areas is largely self-contained, with limited integration at the point where an AI-driven CDSS must support clinician-patient shared decision-making (SDM). This scoping review will map the extent, range, and nature of peer-reviewed and grey literature (2021-2026) at the intersection of these four areas. The goal of this review is to identify conceptual and methodological gaps and to generate the evidentiary foundation for a subsequent framework for developing AI-driven CDSS, which will be published separately. The review uses JBI scoping review methodology as a methodological framework only; it is not intended for publication in JBI Evidence Synthesis and is therefore not bound by JBI's editorial templates or its internal protocol-review process.