OpenEvidence Launches EvidenceGrade AI Feature, Expands NewYork-Presbyterian Partnership
Advertisement
The new feature is designed to help clinicians assess the strength of evidence behind AI-generated answers in real time.
OpenEvidence has launched EvidenceGrade, a new AI-powered feature that evaluates and visualizes the quality of medical evidence supporting responses to clinical questions, alongside announcing a deployment agreement with NewYork-Presbyterian and its affiliated medical schools, Columbia University Vagelos College of Physicians and Surgeons and Weill Cornell Medicine.
The new feature is designed to help clinicians assess the strength of evidence behind AI-generated answers in real time. Unlike conventional AI systems that summarize multiple sources without differentiating their quality, EvidenceGrade grades the underlying evidence to provide greater transparency during clinical decision-making.
EvidenceGrade is based on the GRADE (Grading of Recommendations Assessment, Development and Evaluation) framework, a globally recognized methodology used by organizations including the World Health Organization, Cochrane, and developers of major clinical guidelines to assess the certainty of medical evidence.
According to OpenEvidence, the system first determines whether a clinical question is suitable for grading before evaluating retrieved studies for quality, certainty, and relevance. It considers factors such as study design, consistency of findings, precision, and applicability to the clinical question.
The grading system assigns an "A" to evidence supported by high-quality randomized controlled trials, systematic reviews, or strong clinical guidelines. A "B" reflects moderate-quality evidence with certain limitations, while "C" indicates weaker evidence, including expert opinion or studies with significant limitations. A "D" is assigned to minimal evidence, such as case reports or preclinical data, while "U" is used when evidence cannot be graded.
OpenEvidence said the feature was developed by its medical AI and machine learning teams led by physician-scientists Sam Finlayson, MD, PhD, and Travis Zack, MD, PhD, along with AI researchers Evan Hernandez, PhD, and Eric Lehman, PhD.
Separately, the company announced that OpenEvidence will be deployed across all hospitals and care sites within NewYork-Presbyterian and will also be available to clinical staff at Columbia and Weill Cornell Medicine.
OpenEvidence said more than 915,000 licensed U.S. clinicians, including over 690,000 physicians, currently use its AI-powered medical search platform. The company has also established collaborations with health systems including Mount Sinai, Sutter Health, and Cedars-Sinai as it expands enterprise adoption of its clinical AI platform.
Stay tuned for more such updates on Digital Health News
