T-CAIREM’s Health Data Nexus was intended as a resource for researchers who needed real-world information to develop algorithms and encourage innovative data analyses. But it is also proving to be a boon to professors in classroom settings.
Dr. Azadeh Kushki is a senior scientist at the Holland-Bloorview Rehabilitation Hospital in Toronto. She’s also the first instructor to use the Health Data Nexus for her class, "Introduction to Data Science for Biomedical Engineers”, offered through the Institute of Biomedical Engineering. We caught up with her to learn more about her experience.
What inspired you to use datasets from the Health Data Nexus in your graduate class?
This course focuses on the application of data science tools in the context of health. There are many challenges that are unique to health data when it comes to data science and machine learning, so it was important that students have an opportunity to get hands-on experience with this type of data.
How did the students adapt to the use of real-world data?
Going from working on simple data to real-world datasets is always challenging. The Health Data Nexus team was very helpful in facilitating this transition by providing an overview of the dataset, as well as great documentation, and tutorials. Another advantage was that the data were already cleaned, preprocessed, and ready for exploration. It was great to see that the platform addressed issues related to personal health information protection and research ethics.
Was there anything that surprised you about the way the students interacted with the datasets to create their models?
Given that this is a biomedical engineering course, I was thrilled to see that students brought clinical insight into their work – from understanding the context of the data to formulating clinically-relevant research questions to interpreting their results in a clinical context.
What advice would you offer to other professors who might be interested in using datasets from the Health Data Nexus?
I highly recommend it. The Health Data Nexus gives the students the opportunity to gain experience not only with real-world data, but also with cloud computing.
What excites you the most about the possibilities of machine learning and AI in healthcare?
I was, and continue to be, intrigued by the complexity of the brain and the many ways in which it develops. The data we use to characterize this is multi-faceted and highly heterogeneous; from a technical perspective, there are many open questions about how data science approaches can be meaningfully applied to these datasets. I am also hopeful about the potential of this work to change outcomes for children.
