Interoperability woes still plague healthcare in 2026. Here's what to do about it.
Center for AI and Biomedical Informatics (CAB)
Conducting and promoting impactful artificial intelligence and biomedical informatics research with practical and generalizable real-world implementations that focus on improving population health and efficiency.
About the Center
We are strategically positioned to drive academic research, enhance competitiveness for extramural funding, and foster multidisciplinary collaborations in the dynamic fields of biomedical informatics and artificial intelligence.
Our in-depth knowledge of and strong partnerships with health plans, care delivery systems, public health agencies, health information exchanges, and other health data aggregators allows access to large datasets and to testing of health informatics and AI innovations to improve population health outcomes.
We will achieve our mission through collaboration, demonstration, innovation, and evidence-generation using state-of-the-art methods and technologies.
Research
Our research is split into three main topic areas: data quality and interoperability, patient-centered decision support, and ArtificiaI Intelligence (AI) and Large Language Models (LLMs).
What We’re Working on to Advance Population Health
Research Project: Connxus Partnership: Real World Data from Health Information Exchanges
Connxus is the Health Information Exchange (HIE) for the Greater Austin/Central Texas area. HIEs organize and store electronic health information of patients across multiple providers and hospitals in one place, allowing doctors, pharmacists, EMTs and other providers to access a more complete longitudinal record of patients’ medical information in a quick, secure, consistent, and cost-effective manner.
Our team partners with Connxus to explore and test new ways for HIEs to safely and securely contribute to the generation of new knowledge by participating in population health research, public health surveillance systems, and Distributed Data Networks (DDN) such as the FDA's Sentinel Initiative.
Research Project: Supporting Public Health Response with LLMs: Processing Electronic Case Reports (eCRs) with Generative AI
This study is a collaborative pilot between HPHCI, Dallas County Health and Human Services, and Amazon to evaluate whether large language models (LLMs) can help public health teams more efficiently retrieve information from electronic case reports (eCRs) and interpret the data. ECRs often arrive to public health departments with inconsistent structure and incomplete clinical or epidemiologic details. Focusing on specific high‑impact data elements (such as pregnancy status and travel history), the project will evaluate and compare a human‑annotated reference set of eCRs with LLM‑assisted extraction of data. By assessing when key information is present in the free text, how reliably AI tools can retrieve it, and where eCRs are too sparse to support automation, the study aims to generate a practical benchmark for LLM performance in real-world public health workflows and inform national conversations on eCR quality, data completeness, and the value of augmenting case investigations with AI.
Research Project: Setting Up a National HIE Data Network for Population Health
Despite billions being spent in digitizing the health system, there is still not a single national-level data system that captures health data for the entire population. Funded by a Robert H. Ebert Career Development Award, the National HIE Data Network study has established a workgroup of HIE leaders from across the country to design a national HIE network pilot that leverages new technical innovations and updated health data regulations. This study is exploring the technical and operational ability of participating HIEs from different regions of the country to contribute data to common public health queries using a common data model (CDM) and shared data QA practices.
Research Projects
Our research and growing industry and academic partnerships emphasize a patient-centered approach to the use of AI that ensures privacy and patient safety.
Our Team
We represent a multidisciplinary team committed to developing a state-of-the-art program that fosters collaboration across the Harvard campus, external researchers, and industry experts.
Join our team!
We embody a wide collection of talents, expertise, and backgrounds, cultivating a dynamic and collaborative environment.
CAB Seminars
CAB invites experts to speak at events quarterly. While some events are internal meetings, many will be open to the public.
CAB Updates
CAB has been proudly selected for the Gemini Academic Program.
Anjum Khurshid has joined the Harvard Data Science Initiative at Kennedy School of Government as a Faculty Affiliate.
Contact Us
Email