A sepsis alert on a ward involves a data scientist who trained the model, a nurse who receives the alert, a manager who bought or approved it and a patient whose care changes. The ELSA Way is written for all of them, because no single role can make AI in healthcare work responsibly on its own.
AI developers and data scientists¶
You build the technical system. The ELSA Way helps you look past accuracy figures to whether your system is fair, robust, explainable and fit for clinical use. It covers data management, privacy, bias mitigation, uncertainty estimates, documentation and preparing for clinical validation.
Start with: Problem definition, Development
Healthcare professionals¶
You know the clinical domain and you are often the end user. The ELSA Way shows which questions to ask about an AI tool, how to judge its clinical value and safety, and how to bring your clinical knowledge into its development. You do not need to be a data scientist to contribute.
Start with: Usability and user experience, Pathway for healthcare professionals
Researchers¶
You study, validate or publish AI systems in clinical settings. The ELSA Way helps you design evaluations, validate across sites, report bias and follow international reporting guidelines.
Start with: Evaluation, FUTURE-AI principles
Policy makers and managers¶
You decide on procurement, governance and adoption of AI in your organisation. The ELSA Way helps you question vendors, set up governance and make sure AI systems in your organisation are monitored and someone answers for them.
Start with: Governance and accountability, Pathway for policy makers and managers
Patients and communities¶
AI in healthcare affects you most directly, yet you are often the least involved in building it. The ELSA Way explains in plain language how AI tools are built and used, and how you can bring in your experiences and views.
Start with: Pathway for patients and communities, Stakeholder engagement