•  5 March 2024
     17:00 - 18:00

Join Leo Celi, Constanza Vasquez, and Mwavu Rogers in discussion with Katie MacLure (BCS Health & Care Scotland) for our live BCS BMJ HCI Journal Club hosted by Digital Health.

‌

Synopsis

This journal club is based on a paper entitled, How to organise a datathon for bridging between data science and healthcare? Insights from the Technion-Rambam machine learning in healthcare datathon event

In this education report, the authors provide a checklist and guidance to assist in organising a health related datathon event. From advance planning, timing, logistics, projects to focus on to clinical and data science mentors, they also share their reflections and learning points.

Datathons offer real-world problems in which the participants build models using an iterative process of understanding the problem, designing a study, exploring the dataset, re-designing the study based on the exploration, and re-running the models with a better understanding of the problem each time. This hands-on nature of data analysis makes these events a fertile ground for experiential learning, where the participants go through sequential phases of concrete experience, reflective observation, abstract conceptualization, and active experimentation.

The interprofessional nature of datathons also provides a platform for cross-disciplinary education, where professionals with different expertise learn to solve problems as a team. With the increasing application of big data and machine learning for healthcare and other sectors, datathons can provide educational opportunities to enhance data science skills and contribute to a cross-disciplinary ecosystem that leverages data to better understand and improve the world that we live in.

Responsible AI will not come from us perched in our ivory towers. It requires engagement of patients who are disproportionately burdened by diseases and disasters, as well as their caregivers. But they will need to have some basic understanding of AI in order to be able to drive the course of this technology, i.e., citizen data science. But how can we teach a technology that is advancing so rapidly? There will be no experts in this field. There will only be collective wisdom. We have to learn together. This is the ethos of the datathons that we have been organising around the world. In these events, participants with different expertise, perspectives, and across generations, take turns teaching and learning in the context of a research question using a local health dataset.

AI cannot be taught in the classrooms without bringing students from across departments. It is likewise crucial to partner with community colleges and minority-serving institutions as the students and faculty there have a deeper understanding of the burden of diseases and disasters.