This is a two-part seminar series. Part 1 of the seminar series will be held on April 16th by Doug Manuel and Wenshan Li from the Ottawa Hospital Research Institute. A recording of Part 1 will be available in the HAD JC MS Teams channel. If you would like to view the recording and do not have access, please email emmabrown1@ohri.ca to receive the link.
Learning Objectives:
By the end of the sessions, participants will:
- Be able to apply open science principles in developing predictive algorithms to enhance reproducible science and improve patient care quality
- Understand the approaches used by other participants, thereby improving researcher and IS collaboration for algorithm development and deployment
Part 1 – Foundations for reproducible predictive algorithms in healthcare
Objectives: review and discuss the essentials of reproducible and transparent AI, understanding its imperatives, best practices, and challenges in the healthcare context.
- Review and discuss open science from the perspective of predictive algorithms in health care.
- Examine case studies comparing open-source and proprietary workflows in algorithm development, discussing their implications for healthcare IT and patient care.
Part 2 – Practical application of open and reproducible predictive algorithms
Objectives: Engage in hands-on development and deployment of predictive algorithms using open-source tools, with a focus on real-world healthcare applications.
- Review a hands-on example of algorithm development and deployment using an open-source workflow (R Tidymodels and Plumber).
- Walk-through of algorithm development and deployment at the Project Big Life platform.