Dandelion Health launches data library for GLP-1 drug insights

The rising prominence of glucagon-like peptide 1 (GLP-1) medications is creating significant opportunities for the biopharmaceutical sector, life sciences companies, and researchers. This trend signifies a major shift in the treatment of obesity and diabetes within healthcare systems. The industry is now eager to explore the effects of GLP-1 agonists on various conditions and populations and examine their potential secondary uses for cardiovascular diseases and liver disorders.

Dandelion Health, a health tech startup, identified an opportunity to harness its extensive, de-identified data set to advance precision medicine for GLP-1s. The company has established a GLP-1 data library, which is a multimodal real-world clinical data set designed to uncover insights and opportunities related to the GLP-1 receptor agonist drug class. This data library is built using information collected from a consortium of three major health system partners: Sharp HealthCare, Sanford Health in Sioux Falls, South Dakota, and Texas Health Resources. These partners provided structured and unstructured data for over 10 million patients, representing a broad range of populations and patient journeys.

Dandelion’s GLP-1 library includes all available clinical data for millions of patients, including over 200,000 who are currently taking a GLP-1 medication. Elliott Green, co-founder and CEO of Dandelion Health, highlighted that nearly 80% of clinical data is unstructured and typically inaccessible without AI tools that can structure it for analysis. This data encompasses clinician notes in electronic health records, imaging in picture archiving systems, and echocardiogram waveforms in vendor systems.

Dandelion Health has aggregated both structured and unstructured data into a curated GLP-1 data library. This allows researchers, life sciences companies, and healthcare organizations to advance precision medicine for cardiometabolic diseases while exploring new potential uses for GLP-1s. Green noted that advancements in obesity care have been significant, but there remains a gap between current cardiometabolic care and the high-innovation, personalized care seen in fields like immunology and oncology. Data has been crucial in advancing these fields by revealing underlying disease mechanisms and individual responses to therapies.

Founded four years ago, Dandelion Health was established to advance clinical AI. The company focused on creating a de-identified data set for developers to test AI algorithms. While AI has the potential to enhance healthcare delivery and outcomes, many AI products are developed using narrow data sets that do not reflect the diversity of the U.S. population. Dandelion Health’s co-founders, including Chief Scientific Officer Ziad Obermeyer, aimed to address this gap. Obermeyer, an associate professor at the Berkeley School of Public Health, researches bias in healthcare. Last year, the company tested algorithms using electrocardiograms to predict heart conditions and assess AI performance across diverse racial, ethnic, and geographic groups.

Currently, Dandelion Health collaborates with three health system partners chosen for their diverse patient populations and comprehensive data coverage. Two additional health systems will join this year, expanding the dataset to include between 18 million and 20 million patients. Through the GLP-1 data library, researchers and life sciences companies can use AI applications on the Dandelion platform to answer questions about GLP-1s' impact on body composition and cardiovascular disease risk reduction.

Healthcare organizations can leverage the GLP-1 data library to evaluate weight loss quality through body scans or compare treatment efficacy across various real-world measures. The data library can also demonstrate GLP-1’s therapeutic effects beyond current uses, including secondary benefits and side effects. A precision medicine approach can help match patients to appropriate treatments.

Shivaani Prakash, Ph.D., head of data at Dandelion, emphasized that the goal was to curate a dataset reflecting the full range of benefits from emerging GLP-1 treatments, based on a broader and more representative patient population than typically seen in clinical trials. By providing rich, multi-dimensional data connected to real-world treatment patterns, the library enables users to answer key questions about GLP-1 treatments' role in clinical care.

Dandelion is collaborating with AI developers and researchers on proofs of concept using the GLP-1 data library. For example, two researchers from a major academic medical center developed an algorithm to segment abdominal CTs and quantify fat loss and muscle and bone preservation. Additionally, Dandelion is conducting its own research using the library’s data to evaluate GLP-1 efficacy across real-world patient cohorts and demographic groups and the impact on other comorbidities. These findings will be published in a scientific preprint in the second quarter.