Data-driven insights in menopause: A scoping review of sensors and AI.

Menopause is a complex biological transition associated with physiological and psychosocial changes that can affect health and daily life. Yet, most current assessment methods are subjective and rely on occasional self-reporting. This scoping review maps emerging trends in sensing technologies and data-driven or artificial intelligence approaches for objective and continuous monitoring of menopause-related symptoms. Web of Science Core Collection, Scopus, and PubMed MEDLINE were searched for English-language journal articles published up to December 2025, with no lower date restriction. Additional studies were identified through forward and backward citation tracking. Study characteristics were charted using a standardised extraction table and synthesised descriptively. The database searches identified 81 records, with 12 additional records identified through citation tracking. Following screening and eligibility assessment, 58 studies met the inclusion criteria. Fourteen studies were conducted exclusively in laboratory settings, whereas forty-four incorporated both laboratory-based and free-living monitoring. The sensing technologies included skin-based physiological and thermal sensors, digital diaries, cardiovascular monitoring such as electrocardiography and heart rate variability, wearable activity sensors and actigraphy, electromyography-based systems, electroencephalography and polysomnography, and environmental sensing or exposure assessment approaches. Data-driven approaches show potential for improving symptom detection and early prediction, but their use remains limited by the lack of standardised ground truth. Studies rely on inconsistent symptom reports, physiological thresholds, and expert-labelled events, making validation and comparison difficult. Heart rate variability shows weak and inconsistent associations with symptom severity, while environmental factors mainly act as contextual modifiers. Overall, progress is further constrained by small datasets, inconsistent validation, and a strong focus on vasomotor symptoms. These limitations highlight the need for multimodal, real-world, and data-driven approaches to support more reliable and personalised menopause monitoring.