Data Science & Machine Learning

The Ted Rogers Centre funds a trio of advanced data-driven initiatives that are transforming heart failure research and personalized care now and into the future.

UHN Toronto General Hospital
digital heart

They are led by a collaborative team of software developers, programmer analysts, biostatisticians, data scientists, data engineers, data managers, regulatory specialists, cardiologists, and trainees.

The Digital Cardiovascular Health Platform (DCHP) is our hallmark infrastructure, which enables researchers and clinicians to discover, integrate, store and share data on all heart patients at UHN. It also accelerates digital health innovations and data-driven research.

Our Ted Rogers Computational Program fuses data science, machine learning and traditional statistical methods to the benefit of patients, clinicians and investigators.

Experts in our Program in Data Science develop novel studies, programs and quality improvement projects, establish guidelines and best practices for statistical analyses, program hands-on training to all members, and collaborate with leading academic health sciences centres.

Areas of Impact

Unlocking, sharing and using deeply valuable data

The Digital Cardiovascular Health Platform unlocks siloed, un-linked, and even previously discarded data by capturing and integrating different streams of new patient data. This data lake pulls from many sources including EMRs, test results, clinical notes, research data, machine data, and readings from ECG machines, echocardiography images and wearable technologies.

This extraordinary innovation allows our clinical teams to perform advanced analytics into risk analysis (causes), prognosis and ideal treatment for each individual with heart failure.

Enabling groundbreaking research

This immense data lake – combined with AI and machine learning techniques to analyze it – allows us to streamline clinical and basic research with new knowledge of heart failure fed by millions of patient encounters and real-world evidence.

Our investigators can now study this disease from new angles of discovery and perform pragmatic clinical trials. At any given time, there are at least 200 active analytics/research data management projects underway.

An early highlight of this program was in successfully extracting breath-by-breath data from a cardiopulmonary exercise test machine, then integrating it with EMR data and analyzing the results. This model is able to calculate  a patient’s risk of progressing to end-stage heart failure in one year. It was over 84% accurate predicting outcomes – a rate higher than any validated model.

Advanced techniques to bolster person-centered care

Through our industry-leading analytics and advanced machine learning techniques, teams can perform complex – previously impossible – analyses to bring personalized medicine to the forefront.

On one hand, it generates new studies that bring us closer to predictive care, such as how weather conditions influence outcomes or how exercise therapy can improve the heart health of children and youth cancer survivors.

And on another, our AI-driven predictive analytics boost person-centred efficient care while driving higher and higher levels of patient safety and quality improvement.

Driving system-wide change

A testament to the power and potential of the DCHP is that it has changed research and care at Canada’s largest hospital network.

Our DCHP laid the groundwork for UHN’s mission to “unleash the power of technology and innovation” and was the foundation for a hospital-wide Digital Health Platform.

Over time, the DCHP’s innovative data environment will produce more effective predictive models, algorithms and novel findings that prevent heart failure before it arises.