Counts describe this library, not the strength of evidence.
Cómo interpretar la evidencia
Validation is specific to a metric, model, reference and setting. A useful heart-rate measurement does not validate calorie estimates or sleep stages. Even processing the same accelerometer data differently can change estimated activity–disease relationships.
In 989 adults with chronic low back pain who logged pain three times a day for 7 days, the median pain intensity was only 1 on a 0-10 scale, and people rated their activity higher than wrist, waist and back sensors recorded. Participants were a large research-center cohort tracked for just one week, so these profiles may not reflect pain over months.
In a survey of 432 Chinese patients with chronic diseases, expected usefulness, ease of use, social influence and facilitating conditions each positively predicted intention to use wearables, which in turn predicted actual use. This is a self-reported cross-sectional survey, so it shows associations in a model, not what would change behavior.
In 30 studies of wearable sensors in clinical and real-world settings, 67% (20/30) used inertial measurement units, and 4 studies (13%) reported high accuracy for fall risk and mobility decline (AUC up to 0.97). The studies were mostly observational, with varied sensor placement and methods, so the best-case accuracy may not hold in everyday use.
Across 26 studies, wearable approaches combining several signals had a pooled accuracy of 81.94% for detecting anxiety, versus 76.85% for single signals, with electrocardiography the best single signal (80.34% across 12 studies). Few single-sensor studies and varied methods mean the comparison is rough, and accuracy in research settings may not carry over to daily life.
A systematic review of 55 studies from 2020 to 2025 found that wearable remote monitoring still leans on traditional statistics rather than advanced AI, and that research on mental and cognitive well-being lags behind physical monitoring. The abstract reports no pooled effect on patient outcomes, so it cannot say how well these systems improve care.
Across 30 studies of 40,334 young people aged 5-18, more moderate-to-vigorous activity was linked to fewer depression (partial r = -0.17) and anxiety (partial r = -0.21) symptoms, small associations. Half the studies were cross-sectional and accelerometer methods varied widely, so the pooled size is uncertain and direction of cause is unproven.
In 28 healthy young adults, optical heart rate monitors agreed closely with a chest strap at rest (CCC > 0.99), but agreement fell during burpees (CCC < 0.50), and three identical Whoop 4.0 units worn at different spots performed differently. The test used a small group of healthy people and a short exercise protocol, so it may not match everyday wear.
Among 271 older adults in a mobile cardiac rehabilitation program, the consistently high-wear cluster improved 6-minute walk distance by 38.5 meters (95% CI 2.2-74.7; P=.04) more than the low-adherence cluster in the 8-cluster analysis. The 2-cluster result was only a trend (24.6 meters, P=.06), so the evidence is weak and not causal.
Among 84,095 UK Biobank participants followed for a median of 7.90 years, those in the lowest quintile of rest-activity rhythm amplitude had a higher hazard of ischemic heart disease (HR 1.20, 95% CI 1.12-1.30). The wrist-device data and genetic analyses suggest a causal link, but Mendelian randomization depends on assumptions, and its result was borderline (P=.047).
This mixed-methods review of 48 studies found that wearables and interactive interfaces can help older adults engage with managing their health, but usability, low digital literacy, privacy and cost hold back adoption. It reports no effect sizes, and study quality varied considerably, with only some studies rigorous.
In 104 infants measured over 424 nights with a wearable, parents' sleep-quality ratings correlated moderately with both reported arousals (rho = -0.47) and measured arousals (rho = -0.40). Parents underestimated shorter arousals and overestimated longer ones, but the sample is small and the wearable was not checked against a gold-standard sleep test in this abstract.
Over one year in 20 national-team endurance athletes, each Readiness Advisor app question correlated with its counterpart measure (r = 0.39 to 0.81), but time-series analysis showed marked individual variability. With only 20 athletes, the results cannot be assumed to hold for other athletes or for recreational exercisers.
A scoping review of 2385 records found only 19 studies that used wearables with AI for real-time detection or prediction of cardiovascular events, and most relied on public datasets rather than real-world validation. It maps what exists and does not measure how well any of these systems work.
A narrative review found high diagnostic accuracy for arrhythmia detection and promising results in heart failure monitoring from AI-enabled wearables, particularly FDA- or CE-approved devices. Most studies were small, short-term and in controlled settings, and the review reports no pooled numbers.
In 72 studies with 7949 participants, wearables used at home to monitor movement were generally well accepted, with compliance of 70% or more and strong agreement with clinical assessments. Only half of the studies were controlled trials and just 5.6% were randomised, so evidence that they improve care is limited.
A scoping review of 109 studies mapped wearables used for health promotion in older adults, with mobility, mental health, falls, arrhythmia detection, activity recognition and sleep monitoring the most common targets. Most studies were application design or observational work, so the review shows what has been studied rather than what works.
In a 50-patient pilot trial of remote cardiac rehabilitation with a Fitbit, peak oxygen consumption rose from baseline in both the coached group (+1.6 mL/kg/min, P<.001) and the wearable-only group (+0.6 mL/kg/min, P=.008), with no difference between groups (P=.65). The trial was small, single-center and open-label, so it cannot show that coaching improves fitness.
A scoping review of 46 studies found that wearable AI for sleep disorders was used mainly to screen and diagnose sleep apnea, with none testing it for treatment; 30 of 46 (65%) used commercial devices. It maps what exists and does not pool accuracy, so it cannot say how well any of these systems work.