Aller au contenu
BIBLIOTHÈQUE DE RECHERCHE

Capteurs et mesures

Validation et limites des appareils.

131articles

10années de publication

Sources vérifiées · Méthode de sélection

Publication years in this collection
Counts and coverage
  • 2017 · 10
  • 2018 · 10
  • 2019 · 11
  • 2020 · 12
  • 2021 · 11
  • 2022 · 11
  • 2023 · 10
  • 2024 · 13
  • 2025 · 20
  • 2026 · 23

Counts describe this library, not the strength of evidence.

Comprendre ce type de preuve

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.

Parcourir les articles

131 articles

Réinitialiser
·Étude de cohorte

Chronic low back pain: low pain in the moment, and more activity than sensors show

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.

JOR spine

Lecture structuréeLire l’étude →
·Étude transversale

Why chronic disease patients in China use wearables: usefulness, ease, social influence

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.

Frontiers in public health

Lecture structuréeLire l’étude →
·Revue des preuves

Wearable motion sensors for gait and fall risk: 30 studies, mostly wrist-worn

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.

JMIR mHealth and uHealth

Lecture structuréeLire l’étude →
·Revue des preuves

Detecting anxiety with wearables: combining signals beat single signals, 81.94% vs 76.85%

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.

Communications medicine

Lecture structuréeLire l’étude →
·Revue des preuves

Remote monitoring with wearables: mostly basic statistics, little AI, thin evidence

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.

Digital health

Lecture structuréeLire l’étude →
·Revue des preuves

Youth who move more by accelerometer have slightly fewer depression and anxiety symptoms

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.

BMC pediatrics

Lecture structuréeLire l’étude →
·Validation : étude/revue

Where a wrist or arm heart rate sensor sits changes its accuracy, especially in burpees

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.

Sensors (Basel, Switzerland)

Lecture structuréeLire l’étude →
·Essai randomisé

Older cardiac rehab patients who wore a Fitbit consistently walked farther in 6 minutes

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.

Journal of medical Internet research

Lecture structuréeLire l’étude →
·Étude de cohorte

A weak day-night activity rhythm on a wrist tracker linked to higher heart disease risk

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).

Journal of medical Internet research

Lecture structuréeLire l’étude →
·Article de recherche

Parents misjudge infant night wakings: a wearable shows they undercount short ones

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.

Pediatric research

Lecture structuréeLire l’étude →
·Validation : étude/revue

In 20 elite endurance athletes, readiness ratings matched wearable data but varied by person

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.

Sports medicine - open

Lecture structuréeLire l’étude →
·Article de recherche

AI wearables for heart care: good at arrhythmia, evidence small

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.

Cureus

Lecture structuréeLire l’étude →
·Revue des preuves

Wearables for older adults: 109 studies mapped, mostly design and observational work

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.

Journal of medical Internet research

Lecture structuréeLire l’étude →
·Essai randomisé

Remote cardiac rehab with a Fitbit raised fitness; online coaching added activity, not fitness

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.

JMIR mHealth and uHealth

Lecture structuréeLire l’étude →