In recent years, sleep disorders have gained importance due to their high prevalence and impact on daily life, affecting people's ability to perform daily tasks and reducing quality of life. These disorders include difficulties falling asleep, respiratory interruptions, and poor sleep quality, with sleep-related breathing disorders (SRBD), such as obstructive sleep apnea (OSA), being particularly significant. OSA, which involves repeated airway obstructions during sleep, is especially common in older adults, individuals with obesity, and men, but it remains frequently underdiagnosed.
SRBD not only disrupts sleep but also increases the risk of chronic conditions like diabetes, hypertension, and strokes while creating an economic burden due to higher demand for medical resources. Their effects on physical and mental health lead to fatigue, reduced productivity, workplace accidents, and even disability, highlighting the need for more efficient diagnostic and management tools.
While polysomnography (PSG) is the gold standard for diagnosing sleep disorders, its high cost and invasive nature limit its accessibility. Wearable devices, such as wristbands and watches, offer a more accessible and non-invasive alternative, providing real-time data on sleep, heart rate, and activity. Though promising, these devices still require further research to confirm their accuracy in detecting SRBD. This project aims to evaluate the effectiveness of wearables as complementary tools in diagnosing and managing these disorders. Specifically, it has the following specific objectives: (1) To assess the accuracy, specificity, and sensitivity of wearable devices, such as wristbands and watches, in measuring blood oxygen saturation, heart rate, and activity, compared to nocturnal polygraphy. (2) To analyze the effectiveness of these devices in identifying individuals with potential sleep-related breathing disorders (SRBD) using unsupervised learning techniques. (3) To evaluate the impact and performance of an Artificial Intelligence model for detecting and classifying potential SRBD.