Multiagent Systems in Primary Health and Hospital Settings

Multiagent Systems in Primary Health and Hospital Settings” is a practice-oriented coursework for designing, planning, implementing, and governing multiagent AI systems across real-world healthcare environments, with a special focus on primary care and hospitals. Written for clinicians, administrators, health system leaders, and developers, it assumes no programming background in Part I and then progresses to hands-on code-based labs in Part II. The book begins by situating agentic and multiagent AI within the broader evolution of digital health and clinical decision support, explaining core concepts such as the agent loop, orchestration models, safety patterns, interoperability, and technical architecture using healthcare-centric examples.

Across 19 chapters, it systematically covers foundations, system architecture, orchestration frameworks, workflow design, clinical applications in primary care and hospital settings, remote monitoring, and specialised use cases including oncology, mental health, virtual wards, and health systems transformation. Dedicated chapters address ethics, patient safety, accountability, regulation, data privacy, security, and equity, linking global frameworks (e.g., WHO ethics and governance guidance) to concrete governance mechanisms such as HITL/HOTL oversight, audit trails, circuit breakers, and institutional AI committees. An extensive annexure provides a practitioner’s toolkit with canvases and scorecards for MAS design, readiness assessment, governance framework building, ethical risk assessment, oversight level decisions, vendor evaluation, incident investigation, and capstone planning, making the book directly usable for course delivery, institutional rollouts, and policy or guideline development.

Positioned as a textbook and field manual, this work aims to be a primary resource for teaching a full coursework on multiagent systems in healthcare and for guiding organisations, especially in low-resource settings, to adopt agentic AI in a safe, human-centred, and systems-oriented way.