Cradle
Building Modular Workflow Interfaces for Clinical Applications: A Component-Based Approach
Timeline
January 2025– April 2025
Role
Product Owner / Developer
Cradle is a digital health platform designed for use in low-resource settings by healthcare workers who often have limited medical training. Its early workflow support was hardcoded and difficult to scale. This made it challenging to adapt care pathways, incorporate research protocols, or support local clinical practices.
I led the design of Cradle's workflow system architecture and interface, creating a flexible, modular foundation that could guide healthcare workers through clinical tasks like assessments, form entry, and referrals in a structured and usable way.
Previously...
Workflow logic was embedded directly in code which made workflows:
- ☒ hard to update
- ☒ difficult to scale
- ☒ inaccessible to non-developers
From Hardcoded Logic to Modular, Data-Driven Design
How might we define a scalable, decoupled architecture?
Template-Instance Model
Workflow Templates
define the general logic and structure (steps, transitions, conditions)
Workflow Instances
are runtime instances tied to specific users, tracking execution and data state
This separation mirrors the distinction between clinical guidelines and their real-world applications, allowing for consistency in logic while supporting individualized care.
By storing workflow structure and logic as structured data (JSON-like objects) rather than hardcoded logic, the system becomes:
- ☑ configurable without code
- ☑ easier to test
- ☑ more maintainable