AI news ·
From Personal Loss to Lifesaving AI: How One Researcher’s Journey is Revolutionizing Heart Failure Prediction and Equity in Healthcare
After losing his mother to heart failure, Blessing Ogbuokiri uses AI to predict risks for older patients. His model aims for fair, early intervention and better outcomes.
A Moment of Loss Sparks AI Research to Predict Heart Failure Risks
After losing his mother to a heart condition, Blessing Ogbuokiri, Assistant Professor of Computer Science at Brock University, turned personal tragedy into a mission: using artificial intelligence to help others at risk of heart failure. His work focuses on developing a machine learning model that predicts if older heart patients are likely to be hospitalized or face fatal outcomes due to heart failure.
Ogbuokiri leads Brock’s Responsible and Applied Machine Learning Laboratory (RAML Lab) and, along with his student team, secured funding from Brock University’s Black Scholar Research Grant to advance this project. The grant supports research with meaningful impact, especially for Niagara’s aging population.
How the Model Works
The team trains their machine learning model using data from the Canadian Longitudinal Study on Aging. The model analyzes multiple factors including:
- Medical history
- Smoking status
- Physical activity
- Socioeconomic status
- Presence of chronic conditions like diabetes
Once trained, the model identifies patterns and predicts the likelihood, for example, of a patient having a 50% chance of heart failure and potential hospital admission. The goal is to build a tool accessible to both patients and healthcare providers that can deliver risk assessments quickly—“at the click of a button.”
Practical Benefits for Patients and Healthcare Systems
This predictive tool can encourage patients to adopt healthier habits, such as increasing exercise or quitting smoking. For healthcare professionals, it offers a way to intervene early and manage patient care proactively, potentially reducing sudden surges in hospital admissions.
Addressing Bias and Equity in AI Healthcare
Ogbuokiri highlights an important challenge in healthcare AI: bias. Patients from Black and other equity-seeking communities often face higher risks of heart failure but may encounter systemic biases that affect diagnosis and treatment.
Bias in AI models can arise if training data reflects existing inequalities, leading to underprediction of risk for certain groups. To counter this, the research team applies bias mitigation techniques during data preprocessing and evaluates model fairness using specific metrics. Their aim is to ensure the model performs equitably across all demographic groups.
This approach not only improves prediction accuracy but also supports fairer health outcomes for underserved populations.
For those interested in advancing AI skills relevant to projects like this, exploring specialized AI courses can provide practical knowledge on machine learning and ethical AI development. Visit Complete AI Training for resources on AI courses tailored to various skill levels and applications.