Accuracy and Calibration of Evidence-Grounded Biomedical Signal Extraction From Free-Text Symptom Descriptions: A Prospective Observational Registry Using the OpenGenome Automated Research Instrument

What this Trial Is About

This research evaluates the accuracy and consistency of an AI platform called OpenGenome that analyzes free-text symptom descriptions submitted voluntarily by adults. The study collects anonymous symptom data and matches it with real biomedical literature from PubMed and ClinicalTrials.gov to generate structured biological signal reports. It aims to assess how well the extracted signals align with sources, how confidence scores correlate with dataset size and symptom specificity, and the distribution of biological signals across a large population. Adults aged 18 or older can submit free-text symptom descriptions via the OpenGenome platform, which then uses an AI-assisted method to extract primary and secondary biological signals grounded in biomedical evidence. For each submission, the system retrieves up to 16 relevant sources and provides a confidence score and signal strength. No treatment or intervention is given, and no participant contact occurs during the observational registry. Participants are involved by voluntarily submitting symptom descriptions online through the platform, with all data anonymized at collection. The study continuously monitors the internal signal-source concordance rate over 12 months. There are no in-person visits or direct assessments, ensuring privacy and minimal burden. This ongoing data collection helps researchers understand the platform's accuracy and signal calibration in a broad anonymous population.

Biomedical Signal Extraction From Symptom Descriptions: An Observational Registry Using the OpenGenome Platform