Flow Design for continuous remote patient monitoring using ML algorithms
Deliverables
Diving into the Black Box
Medicine had previously been unable to accurately depict recovery from post-operative patients once they left the hospital. Like a black box, the time spent at home resulted in re-admission rates that were sometimes difficult to manage and predict.
In this project, my colleague and I researched and designed the flow for monitoring patients remotely post-surgery. This project was followed with a subsequent research where we simulated hospital staffing as the new technology became implemented.
Key Modeling Challenges
Researching the flow was not as straight forward as anticipated. For patient safety, it was important to cover all cases, and ensure no situation was left unexamined, since the system would thereon operate without human supervision, alerting staff only when programmed to do so. For this reason, we developed many iterations and extensively consulted with the medical team, ensuring the accuracy of our model.
Escalation Points
The most important item for us to identify and define in the system were escalation points. At various moments, activities such as titration may occur. Some activities may be performed automatically, others require an alarm towards a nurse, or an escalation towards a physician or surgeon. This order of operations is critical to map correctly, in order to ensure patient safety.
Automating the system
The subsequent phase of this research covered the automation of this flow, into a queueing simulation. You can read about here.
GitHub Project (Public)
Repository (Private)
My Role
Researcher
Project Manager