OpenEvidence is a Mayo Clinic Platform portfolio company. Information overload is a prevalent challenge in many high-value domains. It is an extreme challenge in medicine. In the year 2000, just shy of 2,300 clinical trials were conducted; by 2022, that number increased more than 10x, with 32,000 clinical trials started in 2022 alone. The amount of medical research published annually is doubling every 5 years. Such a rate of change makes it nearly impossible for the world's physicians, medical scientists, and healthcare professionals to monitor and understand all the latest research literature and clinical evidence relevant to their work. Biomedical Natural Language Processing (NLP) is highly domain-sensitive, and differs from more general domains in which NLP has been successfully applied. Large language models (LLMs) have demonstrated impressive capabilities in natural language understanding and generation, but the quality bar for medical and clinical applications is high and safety is critical; LLM alignment is arguably more important in medicine than in any other application of LLMs. OpenEvidence is pursuing foundational work in self-supervised learning and human-in-the-loop machine reading applied to biomedical text, to address substantial challenges in biomedicine that have been under-explored and unsolved in mainstream NLP. By analyzing medical text and extracting biomedical entities and relations from the entire history of published medical science, OpenEvidence can facilitate better real-world, evidence-based clinical decision support and medical knowledge synthesis. In so doing, OpenEvidence is fulfilling its mission of organizing the world’s medical knowledge and making it more useful.
