Hospital AI Spots Hidden Illnesses, Saving Lives and Millions

Healthleap, a startup founded in 2022, just raised $38 million to expand its AI that helps hospitals catch conditions like malnutrition and delirium early. The AI reads patient records—including doctors' written notes—to flag those who might need a closer look. It's already used in over 50 hospitals, including Penn Medicine and Cedars-Sinai. This matters because malnutrition affects up to 50% of hospital patients and can lead to longer stays and complications. By catching it early, hospitals can treat patients faster, improving outcomes and saving money. For example, at Penn Medicine, the AI program led to $23.8 million in annual savings from shorter stays and extra reimbursements.

The AI works by plugging into a hospital's electronic health records. Every night, it analyzes all adult inpatients' data—labs, vitals, notes—and writes a risk score into the care team's workflow each morning. It doesn't diagnose; it just highlights patients for review. This is crucial because doctors are busy and might miss subtle signs like poor appetite or weight loss mentioned in notes. The AI acts like a safety net, ensuring no patient slips through the cracks. The company charges hospitals based on bed count and outcomes, guaranteeing at least 5x return on investment.

The catch is that the AI isn't perfect. It relies on the accuracy of notes and data, and it only flags potential issues—doctors still need to confirm. Also, it's currently limited to certain conditions like malnutrition and delirium, though the company plans to cover 40+ conditions and expand to outpatient care. Privacy is another concern: the AI accesses sensitive patient data, so hospitals must ensure strict security. Despite these limitations, the technology could significantly improve patient care and reduce healthcare costs, making it a win for everyone.