HOW CAN WE HELP YOU? Call 1-800-TRY-CHOP
Campbell Laboratory
We are hiring!
Are you passionate about using AI to improve child health? We're looking for data scientists and machine learning engineers to change the practice of medicine. Apply today at Careers at CHOP.
Clinicians spend an exorbitant amount of time sifting through countless electronic health record (EHR) screens while juggling a busy schedule and numerous tasks. The team of specialized researchers in the Campbell Lab is dedicated to developing technology that gives care providers more time to focus on their patients rather than paperwork.
The Campbell Lab created Hope Alpha, an artificial intelligence (AI) model fully trained on patient data from the EHR at Children's Hospital of Philadelphia. Researchers are using the model to better understand study participants. Future applications could include improving hospital operations and direct patient care.
CHOP researchers are encouraged to experiment with Hope Alpha and find how it can help you manage your research projects. Reach out to the Campbell Lab team to learn more.
Combining Human Expertise with Advanced AI
Commercially available large language models (LLMs) powered with AI show promise for everyday tasks. However, for healthcare-related needs, they often start from a blank slate — these models don't "know" the patient unless we feed them context every time.
Instead of relying on lengthy prompts each time, Hope Alpha already holds a secure, comprehensive, contextual understanding of more than 1.6 million patients. It can piece together information from notes, labs, and conversations to help clinicians and researchers see the bigger picture, faster.
By combining human expertise with advanced AI, we can make healthcare more personalized, efficient, and empathetic. Our ultimate goal is to see Hope Alpha empower CHOP pediatricians to spend more time interacting with children and families, and less time buried in files.
Advanced AI Model Progress
Privacy and Security
Hope Alpha handles patient data with the same rigor of privacy and security as the EHR. The platform runs on our secure, encrypted cloud environment and is fully HIPAA compliant. The CHOP Institutional Review Board supervises our work, and the CHOP AI Governance Committee ensures patient privacy at every step.
Massive Training Effort
We began with a refined open-source model and trained Hope Alpha for more than 65,000 hours with specialized AI-accelerating hardware called tensor processing units. That's more calculations than a typical laptop could do in a millennium.
Comprehensive Knowledge
Hope Alpha can integrate decades of notes, labs, and care plans into a single chronological view of a patient's medical journey.
Hope Alpha, like other LLMs, sometimes makes mistakes or "hallucinates." We're working on new methods to detect inaccuracies and refine responses. We are also exploring ways to continuously update Hope Alpha with new patient information and the latest medical breakthroughs.
Help us shape the future of pediatric healthcare. Donate to CHOP’s AI efforts today.
Leader
Ian M. Cambell, MD, PhD
Assistant Professor of Pediatrics
Dr. Campbell is interested in approaches to improve diagnosis of children with rare genetic diseases. He also focuses on better understanding of the full spectrum of symptoms associated with genetic diseases as well as discovering new diseases.