News.Health.001 News.Health.001
Context
Introduction
AI is playing an increasingly central role in disease prediction, helping healthcare providers shift from reactive to preventative care—but concerns remain about accuracy, fairness, and clinical integration. The National Human Genome Research Institute (NHGRI) is exploring how AI models trained on structured health data can assist clinicians in identifying individuals at risk for chronic conditions like heart disease.Role
The learner is a data science intern working with NHGRI’s Predictive Health Analytics team. They are tasked with building an interpretable model that predicts heart disease risk based on clinical variables and preparing a report on model performance, bias, and real-world feasibility. This role requires balancing technical outcomes with ethical considerations and practical limitations faced in clinical decision-making.Business Objectives
The goal is to build a heart disease risk prediction tool that can assist doctors in identifying high-risk patients for early intervention. The learner’s understanding of modeling and evaluation allows them to assess both predictive performance and fairness across demographic groups.Products
The final product is a short technical and ethical evaluation report accompanied by a prototype predictive model for heart disease risk. It demonstrates whether the AI tool meets clinical standards for utility and fairness—and whether it’s appropriate for real-world deployment.Codebook
ages, ex, cp, trestbps, chol, fbs, restecg, thalach, exang, oldpeak, slope, ca, thal, numDataset
License
Not Provided
Tags
- TEXT