A Memorable Memorial Day Weekend!

This year, for obvious reasons, there were no friends gathering around grilles for a nice neighborhood BBQ, nor any ACBL bridge tournament to spend the long weekend at. But my friend Prof. Ying Ding at the School of Information UT Austin invited me to be a judge for the AI Health Data Challenge she organized for graduate students at the school. What an amazing way to spend the weekend going through exciting projects, presentations and code repositories included! I could not help myself write this post to talk about this experience and about the work undertaken in a short period of time by the awesome participants.

The projects were focused on apps or data analytics tools for data-driven, evidence-based patient care and was an opportunity for students to improve their skills in organizing and executing as teams. The top five teams presented their projects online to a panel of judges comprised of eminent doctors, academicians, technologists and business folks on Tuesday, May 21. 

A note on the goals of this challenge. The primary goal was to bridge disciplines and promote data-driven, evidence based care. There is a lot written about the challenges to decision driven healthcare, chief among them are interoperability of systems and data, lack of standards for data exchange/storage/access, plethora of apps for patient self help and seeking healthcare information on ailments but lack of well thought through and well implemented digital applications in the provider/payer community for care management, care co-ordination and care giver productivity that can effectively lead to evidence based care.

For most such challenges, it is a bigger challenge to access datasets for such research or application building. The MIMIC-III database, an openly available dataset developed by the MIT Lab for Computational Physiology, comprising de-identified health data associated with ~60,000 intensive care unit admissions and includes demographics, vital signs, laboratory tests, medications, and more. It is a great dataset for such challenges. However, we need more such datasets not just for students of medicine and data science to apply their creativity to come up with solutions, but to encourage good, quality research that is not encumbered by lack of data availability.

Another note on the participants and their projects. As I went through the submissions, I was impressed with the enthusiasm, depth of research and the ability to turn real life problems in healthcare into tangible data technology based solutions. The winning FHIR based app for access to analyzed SMART data for decision making or the runner up, clinically accurate NLP based automated report generating application were just some of the highly innovative work done by the students. There were many others, which I hope the students pursue further, based on the judges feedback, because I believe they all have the kernel of innovation that can lead to successful innovative startups in med-tech, using data science.

I wish all the participants great success in their future endeavors, and the judges from whom I learnt a lot about the practical aspects of bringing some of these solutions to life. Finally, I thank Prof. Ying Ding for leading this wonderful challenge and inviting me to be a part of it and make my Memorial Day weekend, memorable!