Wearable AI forecasts prolonged sitting in women with chronic pelvic pain

Researchers at the Icahn School of Medicine in New York have created an artificial intelligence (AI) system that utilizes data from wearable devices to predict extended periods of sitting among women suffering from chronic pelvic pain. Published in the September 30 issue of npw Women's Health, the study's findings could lead to personalized digital health tools that gently remind individuals to move at optimal times while reducing unnecessary alerts.

Chronic pelvic pain affects around one in seven women, often exacerbated by conditions like endometriosis, adenomyosis, and uterine fibroids. These conditions can lead to prolonged periods of sitting due to pain, fatigue, and other symptoms impacting daily life. Despite the benefits of regular movement, generic advice to "sit less and move more" often fails to consider the unique challenges faced by those living with these conditions.

The research team examined data collected from wearable devices worn by 134 women with chronic pelvic pain, primarily endometriosis, alongside a group of 61 healthy individuals. By analyzing approximately ten days of each participant's data, they trained personalized forecasting models to predict activity levels one hour in advance. They then tested these forecasts to identify 15-minute periods of sedentary behavior during waking hours, a critical timeframe for delivering brief movement breaks, termed "exercise snacks."

The study revealed that relatively simple, interpretable models could predict prolonged sitting with the same accuracy as more complex, computationally intensive deep-learning approaches. This discovery highlights that lightweight models can reliably forecast sedentary behavior while remaining practical for direct use on individuals' devices, enhancing privacy and reducing computational demands.

Furthermore, the models demonstrated robustness even when data were incomplete, suggesting their potential to provide meaningful predictions in real-world scenarios, not just controlled laboratory settings. The findings suggest that wearable data could serve as an early-warning system for prolonged sitting in women with chronic pelvic pain, enabling predictive and timely movement prompts.

The next phase involves determining whether delivering personalized movement prompts based on these predictions can genuinely reduce sedentary time, alleviate symptoms, and improve quality of life for individuals with chronic pelvic pain. While the initial results are promising, prospective clinical trials are necessary to validate the effectiveness of AI-guided movement prompts.

Beyond chronic pelvic pain, the forecasting framework could potentially benefit other chronic conditions where prolonged sitting contributes to poorer health outcomes. The research team is now working on integrating this forecasting framework into a just-in-time adaptive intervention, which will further test the efficacy of personalized, AI-guided movement prompts in improving the lives of women with chronic pelvic pain disorders.