HERMOSA: Understanding the Perimenopausal Health Experience of Latina Women Using Digital Biomarkers from Apple Watch.

About This Grant

Perimenopause remains one of the most under-researched, under-diagnosed, and under-treated conditions in medicine. Vasomotor symptoms (VMS)—including hot flashes and night sweats—affect approximately 70% of women and often persist for over seven years, significantly disrupting sleep, mood, and quality of life. Latina women, in particular, experience more frequent, severe, and longerlasting VMS than other women, yet they remain less represented in menopause-related clinical research and less likely to access evidence-based care such as hormone therapy. There is an urgent need for personalized and scalable tools to evaluate and manage menopause-related symptoms. This Phase I SBIR project will support the development of digital biomarkers to detect and monitor VMS in perimenopausal Latina women using physiological data collected via the Apple Watch. Amissa Health has developed the first—and only—clinical-grade perimenopausal health monitoring application with access to raw Apple Watch sensor data through Apple’s SensorKit. This platform enables continuous, passive data collection—including skin temperature, heart rate, movement, and sleep—alongside selfreported symptom tracking via a mobile app. Unlike traditional symptom diaries and in-clinic assessments, this approach provides high-frequency, real-world data that supports more accurate and inclusive symptom detection, even in remote or underserved settings. The project includes two specific aims. Aim 1 will recruit a cohort of 120 perimenopausal women—focusing on Latina participants—to wear Apple Watches and log hot flashes and night sweats using the Amissa app. Participants will be selected based on menstrual history, age, and VMS incidence. Data collected will include skin temperature, heart rate variability, oxygen saturation, sleep patterns, physical activity, and contextual data. Aim 2 will develop and validate machine learning models that fuse wearable sensor data with self-reports to detect and predict the occurrence, intensity, and timing of VMS. Models will be trained using supervised learning approaches, with performance evaluated using metrics such as sensitivity, specificity, and AUC. This project addresses major NIH priorities related to digital innovation and women’s health. By integrating wearable technology with AI-powered analytics, Amissa’s platform will support research and enable more personalized treatment strategies. Long-term, the data generated will enable the delivery of highly personalized, evidence-based, hormone replacement therapy interventions that improve quality of life for Latina women. Furthermore, this data will support research to measure cardiovascular disease and Alzheimer’s risks in women.