����ɫ��

Using data from tens of thousands of patients to pinpoint diagnosis, best treatment and likely outcomes for individuals might sound like something out a science fiction novel, but researchers are doing just that.

Dr. Joon Lee, PhD, a health data science and machine learning researcher at the ����ɫ��’s Cumming School of Medicine (CSM), is passionate about using artificial intelligence (AI) technology in health care to improve lives.  

As director of the , Lee harnesses the power of data science, machine learning and AI to address complex health problems at the population and individual level.

Lee’s lab is unique in that it focuses on using unconventional digital health data sources, such as information garnered from Twitter and other digital media, to understand – and even predict – health outcomes.

“A lot of our projects focus on creating models that can predict certain events or outcomes in the future,” says Lee. “For example, I detect what might happen to a patient, given what we know at the moment.”

Knowing how a patient might fare is a potentially powerful tool as it allows physicians to intervene and individuals to make lifestyle choices—such as being more physically active and adopting a healthy diet— proven to decrease the risk of future health problems.  

According to Lee, although using advanced computational techniques for prediction has become increasingly commonplace, it is much more difficult to make a clinical impact with this work.

“Once you have a good dataset, it is relatively straightforward to build a machine learning model,” says Lee. “But if you want to apply that and really make an impact to patients, that is unchartered territory.”

It’s arguably a critical task, especially in cardiovascular health. According to the , heart disease is the leading cause of death globally, accounting for 45 per cent of all non-communicable disease deaths in the world. Yet up to 80 per cent of these deaths can be prevented with healthy lifestyle choices.

Lee and his trainees and staff are working on three projects in this area. His lab is partnering with the Alberta Provincial Project for Outcome Assessment in Coronary Heart Disease (APPROACH) team on a that uses machine learning to enable precision care for patients with coronary artery disease, which can lead to heart attacks.

The project applies machine learning to clinical data from the APPROACH Registry and administrative health data from Alberta Health Services. APPROACH has followed more than 300,000 Albertans who have undergone invasive coronary angiography since 1995, with the goal of creating a rich data repository to help inform decisions about the best ways to treat patients with plaque buildup in their arteries.

“We are creating a tool that can predict if an individual receiving bypass surgery may require another procedure, or if they may go into heart failure, for example.” said Lee. 

In , in partnership with the Public Health Agency of Canada, the lab has developed a machine learning model that detects Tweets about physical activity, sedentary behaviour and sleep.

It wasn’t an easy task, as it required manual coding of more than 120,000 Tweets from Canada, the US, the UK and Australia. An example is a Tweet about going on a hike, or one about binge watching Netflix.

The information from this unique data source can be used to predict future health at the population level and to drill down to specifics like, “which Canadian province has the most physically active population?” The lab is for use by all researchers.

A involves a custom built AI system that monitors unhealthy food marketing to children on digital media. The project was created in partnership with Libin Cardiovascular Institute researcher, Dr. Dana Olstad, PhD, and Health Canada as a way of informing food marketing policy.

The overall goal of the project is to decrease the risk of cardiovascular conditions at a population level.

Dr. Joon Lee is an associate professor in the departments of Community Health Sciences and Cardiac Sciences at the Cumming School of Medicine (CSM) and a member of the Libin Cardiovascular Institute and the O’Brien Institute for Population Health.