HealthLeap

HealthLeap builds AI-powered clinical screening tools that read hospital EHR data - labs, vitals, structured records and clinical notes processed with natural language processing - to detect malnutrition and other conditions in hospitalized patients. Its malnutrition screening tool, the only validated such AI tool on the market, screens 100% of patients daily and produces real-time risk scores at the point of care, surfacing conditions that would otherwise go unnoticed. The engineering problem spans machine learning, predictive risk scoring, EHR data processing and clinical decision support, applied to billions of EHR data points.

The results are measurable. The model achieves an AUROC of 95% during patient stays and identifies malnutrition four days earlier than manual screening, with 88% higher sensitivity than nurse-administered instruments. The clinical case is substantial: malnutrition affects 20–50% of hospitalized patients, yet historically fewer than 9% are diagnosed. Institutions including Cedars-Sinai, Penn Medicine, Houston Methodist, Emory and Intermountain Health have adopted the tool.

HealthLeap operates in the United States and works at the intersection of healthcare, hospital operations, clinical care and clinical nutrition. The company was founded by siblings and is led by registered dietitians, with an explicit safety-net approach: screening every hospitalized patient so that no patient falls through the cracks. Engineers joining the team work directly on models whose outputs are consumed in live clinical settings, where accuracy, timeliness and interpretability carry real weight.