Asaf Kraus, founder and CEO of Dieta Health, suffers from irritable bowel syndrome (IBS), a dreadful, albeit non-fatal disease that is notoriously difficult to treat. In 2017, Kraus was employed as a data scientist for Uber when he participated in a clinical trial for experimental IBS treatment. As a participant in the trial he was required to peer at, remember, and subjectively classify his stool on the Bristol Scale, a tool gastroenterologists use to assess human stool, a.k.a poop, based on its shape and how formed or loose it is, to determine how effective an IBS treatment is for a particular patient.
Given his machine learning expertise, Kraus was convinced that artificial intelligence (AI) could do a better job. He founded Dieta Health and built a mobile app for iOS and Android. Patients using the app simply photograph each bowel movement, which is then instantly classified for an objective and accurate digestive health results. In addition to ranking stool on the Bristol Scale, the Dieta app also classifies multiple other stool characteristics including volume and fragmentation from a single image capture.
Go ahead, make every poop joke that comes to your mind, but this is a big step forward for people suffering with IBS. A clinical trial was run by the MAST Program (Medically Associated Science and Technology Program) team at Cedars-Sinai in Los Angeles to measure the effectiveness of the Dieta app. Patients recruited by Cedars-Sinai used Dieta’s app to capture stool images before, during, and after taking an experimental medication throughout 2021. Two expert gastroenterology professors then manually classified over 200 images of stool to create a “gold standard” on which to compare Dieta’s AI to the current standard of patient reported classifications.
The AI used by Dieta’s app was shown to be comparable to expert gastroenterologists and far better than patients in classifying stool on the Bristol Scale. Dieta’s AI was more sensitive and more specific when compared to patients (55% and 71% respectively). Additionally, Dieta’s AI stool classifications were more correlated (0.71) to the patients’ symptom severity scores as compared to even the patients’ own classifications (0.46).
“Dieta’s stool image recognition objectified constipation, diarrhea and normal stools so there was less subjective patient influence on trial outcomes,” said Dr. Mark Pimentel M.D., a world-renowned gastroenterologist and researcher who is credited with discovery of many innovative IBS treatments and diagnostics. “This could be a game-changer in terms of more clearly seeing true effects on patient stool outcomes and side effects of drugs in clinical trials for gastroenterology and beyond.”
The results of the groundbreaking human clinical trial validating Dieta’s stool image recognition technology were published today in the American Journal of Gastroenterology.
In addition to the proprietary stool imaging AI, Dieta’s platform for patients offers much more. Patients can also snap photos of meals, track medications and symptoms, and get guidance on choosing interventions (treatments) as they work towards improving their digestive health. Patients can also easily share their data with their clinicians to view on Dieta’s Clinician Portal.
Dieta’s applications have been used by over 7,000 patients. Next month, Dieta is launching a patient care service, where patients who use the platform can also connect with a unified team of gastroenterologists, dietitians, and coaches to get guidance and care directly in the Dieta app.
Dieta Health raised $1.9 million in its pre-seed funding round, and is backed by Techstars, and investors/partners including Cedars-Sinai, an international leader in gastroenterology, and UnitedHealthcare, the world’s largest healthcare payer.