The design section has nine chapters, from defining the clinical problem to weighing the tool’s effect on society. You can read them in order, but expect to loop back: what you learn about data or regulation often changes the problem statement.
Prerequisites¶
| Prerequisite | Importance | Level |
|---|---|---|
| Understanding of clinical workflows | Helpful | Beginner |
| Familiarity with research ethics processes | Helpful | Beginner |
| Basic understanding of machine learning | Helpful | Beginner |
Summary¶
By the end of this section, you should be able to:
state a clinical problem with success criteria you can measure
map the people affected by the tool and plan how to involve them
document intended users, clinical settings, explainability needs and use cases
find out which ethical review and data governance steps your project needs
plan data collection so that it represents the patients the tool will serve
identify the regulatory frameworks that apply and the legal basis for using patient data
open a risk management file
list likely sources of bias and the social effects of the tool