Five uncertainties decide whether a promising health technology creates value in practice. Model performance is rarely the one that settles it.
Need and adoption asks whether the people it is built for will actually use it. Workflow and accountability asks whether it fits real work, and who acts on the output. Action logic asks what should happen once the system knows something. Commercial viability asks who benefits, and whether they are the ones who pay. Regulation and governance asks whether it is a medical device, how data will be used, and what the product must be built towards.
They are ordered by how often they are the thing that stops a project, not by how interesting they are to build. Most teams can answer the technical question long before they can answer any of these, which is why so many products are accurate and unused.
The five uncertainties are the structure of the Before You Build Method, and an Evidence Phase is the engagement in which they get tested.
Where this is explored in depth