This presentation delves into the potential of using synthetic data to assess population health, posing the question of whether this innovative approach represents the future of AI in healthcare or is simply overhyped. Synthetic data, or artificially generated data, offers a realistic approximation of original health datasets without compromising the privacy of individuals. It mirrors the characteristics of genuine data sources, enabling the modeling of patient pathways through the healthcare system, the evaluation of intervention impacts on populations, and facilitating the secure and ethical sharing of data. Furthermore, synthetic data holds the promise of driving innovation by providing a sandbox for testing and development. Our exploration focuses on the application of synthetic data in modeling the health trajectories and treatment outcomes of two specific populations in Alberta, Canada: opioid users and individuals with diabetes. By generating artificial cohorts that closely resemble these real-world groups, we aim to assess the effectiveness of synthetic data in replicating complex health dynamics and outcomes. The presentation will critically examine the capacity of AI systems to produce synthetic data that is not only accurate and realistic but also actionable for patients, clinicians, and policymakers. Through a detailed analysis of the synthetic representations of opioid users and diabetic patients, we aim to uncover the potential of synthetic data as a tool for innovation in healthcare, while also scrutinizing the limitations and challenges that accompany its use. Is synthetic data truly a groundbreaking advancement in health research and policy-making, or does it fall short.
I am a renowned clinician scientist with a prominent international reputation, leveraging my expertise to advance public and population health. Renouncing my role as Canada Research Chair to further my university's equity, diversity, and inclusion efforts, I have amassed over 250 scholarly articles, with more than 100 published in the last five years in high-impact journals. My research is recognized for influencing policy within Health Canada and the FDA. My commitment extends to community work, particularly with Alberta's First Nations, enhancing chronic disease management. Additionally, my work integrates artificial intelligence and machine learning to innovate population health practices, using large health databases as a springboard for health AI advancements. In leadership, I serve as a principal investigator at the Alberta Diabetes Institute and have been integral to its research strategies for a decade. As the Director of the MSc/PhD programs in the School of Public Health, I am shaping the next generation of research training. Provincially, I was a funding member of Alberta's SPOR Support Unit and championed career development in research methods and health services. Nationally, I am a longstanding member of Diabetes Canada's National Research Council, contributing to Type 2 Diabetes Clinical Practice Guidelines and collaborating with government and regulatory bodies to enhance health outcomes.