Artificial Intelligence (AI)-driven Primary Care company K Health announces the publication of its peer-reviewed research in Mayo Clinic Proceedings: Digital Health, which demonstrates how K Health’s AI can provide information that allows for more personalized hypertension treatment in patient-facing settings. According to the CDC, hypertension affects nearly half of American adults and is a major risk factor for heart attack and stroke. Optimal treatment has a proven role in reducing the risk of advanced disease and death, but historically, guidelines have been difficult to implement in day-to-day practice.
The article, “Causal Deep Neural Network-Based Model for First-line Hypertension Management,” describes how the model, which was created by K Health using de-identified data from Mayo Clinic, has successfully been used by physicians to stabilize blood pressure in hypertension patients. Co-authors of this paper include Francisco Lopez-Jimenez, M.D., M.S., MBA., Chair of the Division of Preventive Cardiology at Mayo Clinic, Lee Herzog, M.D., Director of Medical Sciences at K Health, and Ran Ilan Ber, Ph.D., Vice President of Data Science at K Health.
Key highlights from the publication:
- The model—which was built using anonymized patient data from 16,917 real-world primary care clinical cases—was designed and validated to help physicians personalize treatment and stabilize patients with hypertension at the point of care.
- The model is proven effective in predicting optimal medication regimen based on gender, age, ethnicity, chronic conditions, blood pressure, and relevant laboratory results.
- Treatment recommendations from the model aligned with established clinical JNC 8 hypertension medication guidelines 95.7 percent of the time, compared with actual physician practice of 77.9 percent.
“This is a major step in demonstrating how AI can completely change how we deliver care to millions of people living with chronic illness,” said Ran Shaul, Co-Founder and Chief Product Officer of K Health. “By referencing de-identified data from real Mayo Clinic patients, we’ve been able to build a way for patients in any primary care setting to directly benefit from the Mayo Clinic standard of care.”
“The key findings of this study demonstrate how machine learning can accurately predict a patient’s optimal antihypertensive treatment while minimizing their side effects and closely adhering to established treatment guidelines,” said Jon O. Ebbert, M.D., Professor of Medicine and Primary Care Physician at Mayo Clinic. “This research demonstrates the clinical utility of using AI at the point of care to control high blood pressure, one of the leading causes of preventable deaths in the world.”
“This is an example of how AI is directly contributing to a decrease in the rate of serious complications such as heart attack and stroke, while allowing doctors to practice at the top of their license,” said Stephanie Foley, M.D., Head of Cedars-Sinai Connect Medical Practice and Senior Medical Director at K Health. “The model makes recommendations in seconds that are aligned or even better than a practicing physician’s, and I cannot wait to see what else AI can do for patients and providers on an even wider scale.”
Disclosures: Dr. Herzog, Dr. Ilan Ber, Dr. Zehavi Horowitz-Kugler, Ms. Yardena Rabi, Mr. Ilan Brufman, and Dr. Yehuda Edo Paz were paid employees and had stock options from K Health, the algorithm’s developer, during the development and conduct of this study. Dr. Lopez-Jimenez owns intellectual property related to algorithms using ECG AI and the algorithms have been licensed to companies. None of these algorithms relate to hypertension or its treatment. Dr. Lopez-Jimenez is also a member of the scientific advisory boards for Anumana and Novo Nordisk.