Sequoia, First Round and Hummingbird put $38M behind Healthleap's overnight reading of hospital notes

Healthleap's product reaches its user at the start of a shift. "Each morning, we write a risk score into the care team's existing workflow with a dashboard accessible that holds additional information about the patients' trends," CEO Josiah Meyer told TechCrunch ^[5]. Before that, the software has read every adult inpatient's record overnight, including lab results, medications, diet orders and clinicians' notes, he said ^[4].

The pitch is the notes. "Labs, weights, and vital signs sit in structured fields, but the most telling signs sit in clinicians' written notes: poor appetite, recent weight loss, muscle loss, trouble swallowing," Meyer said ^[3]. Language models pull those mentions out, and the result goes into a risk model alongside the labs and vitals ^[6]. Meyer described the method as "extracting affirmative or negated mentions of these clinical concepts" ^[9]. A note saying a patient denies weight loss has to count differently from one that records it.

What a hospital actually gets is a nightly screen and one number per patient. Healthleap says the software does not diagnose and only highlights patients for further review ^[6]. So any benefit starts when a clinician opens the flag and does something about it. The product began as a clinical nutrition tool that his sister and co-founder Jemima Meyer built for dietitians ^[7]. In my view dietitians are still the obvious owners of a malnutrition flag. The person who signs off on the return is someone else.

At the Hospital of the University of Pennsylvania, Healthleap puts its malnutrition program's annualized impact at $23.8 million ^[10]. About 74% of that, $17.5 million, is credited to shorter stays, and the other $6.3 million to additional reimbursement ^[16]. Meyer said the customer's finance team does the crediting. "We use the hard ROI that the hospital finance team validates and attributes to us as the measurable ROI. Based on that, we contractually ensure that we deliver multiples of the contract price," he said ^[14]. "To date, every customer has seen a 5x hard ROI or more, in some cases over 20x annual total ROI," he added ^[19]. Those figures come from the company.

The customer base is young. Meyer said Healthleap went from three hospital partners to more than 50 in the past year ^[12]. That means at least 94% of its hospitals signed within the last twelve months ^[17]. The other open number is volume. TechCrunch cites research finding that 20% to 50% of hospital inpatients are malnourished ^[15], so a screen of every adult inpatient can produce a long morning list. The account does not include how many flags a typical unit receives or how often one turns out to be wrong.

Meyer said the company wants to cover more than 40 major conditions and expand into outpatient and home care ^[11]. The screens running today are for malnutrition and delirium. Aspiration pneumonia, pressure ulcers and heart-failure readmission risk are still in further clinical validation ^[8]. A three-year contract signed this quarter buys the first two.

I'd sort the decision on two axes. The first is whether a named person owns the morning list and has time in the day to review each flag. The second is whether finance has written down, before go-live, how it will credit a shorter stay. Where both hold, the hospital has someone producing the savings and an agreed way to count them. An owner without a method may help patients, but the contract's guaranteed multiple goes unmeasured. Finance with a method and no owner has a formula and few acted-on flags to count. With neither, the hospital has bought a dashboard, priced by licensed bed ^[13].