Accessing laboratory test results is the most common activity on patient portals, yet many patients—especially older adults with multiple chronic conditions—struggle to interpret their meaning. Despite ongoing efforts to improve how results are presented, few approaches offer personalized support grounded in an individual’s health context. The AHRQ-funded LabGenie project addresses this critical gap by developing a web-based, user-centered tool powered by large language models (LLMs) to enhance patient understanding and engagement.
In this webinar, we will present a series of studies conducted as part of LabGenie, including:
- a benchmark evaluation of four state-of-the-art LLMs in responding to lab-related questions sourced from a public health forum;
- the use of retrieval-augmented generation (RAG) to determine condition-specific reference ranges for lab tests;
- LLM-driven prediction of differential diagnoses from clinical vignettes; and
- the generation of personalized follow-up questions to support patient-provider communication.
By combining generative AI with user-centered design, LabGenie aims to make lab results more interpretable and actionable—especially for patients with limited health literacy—ultimately fostering informed decision-making and improved healthcare experiences.
Presenter