Secondary Logo

Share this article on:

Accuracy of Wearable Devices for Determining Physiological Measures during Different Physical Activities: 2670 Board #190 June 2 930 AM - 1100 AM

Boudreaux, Benjamin D.; Cormier, Corinne L.; Williams, Brian M.; Hebert, Edward P.; Kraemer, Robert R. FACSM

Medicine & Science in Sports & Exercise: May 2017 - Volume 49 - Issue 5S - p 762
doi: 10.1249/01.mss.0000519027.48408.29
E-33 Free Communication/Poster - Monitoring Friday, June 2, 2017, 7:30 AM - 12:30 PM Room: Hall F

Southeastern Louisiana University, Hammond, LA. (Sponsor: Robert R. Kraemer, FACSM)


(No relationships reported)

PURPOSE: Wearable technology is ACSM’s number one 2016-2017 fitness trend. The accuracy of these devices has yet to be firmly established. The objective of the study was to determine the validity of wearable devices’ assessment of step count, heart rate (HR) and caloric expenditure (KCAL) during various physical activities.

METHODS: Thirty college students (19 female, 11 male) engaged in 3 activity sessions (sweeping with a broom and dustpan, climbing stairs, and walking 1/4 mile) while wearing an Apple Watch Sport, Fitbit Charge HR, Accusplit Hip Pedometer, and iPhone 6 Plus. Step counts from devices were compared to those objectively recorded using a tally counter. Subjects also completed a treadmill graded exercise test during which HR and KCAL were reported by an Apple Watch Sport, Fitbit Charge HR, and Polar T-31. Values were compared to those from a six-lead ECG and metabolic analyzer. HR was recorded at rest and during each stage. KCAL was determined at the end of the protocol.

RESULTS: Correlations between objective step counts and from the devices were: walking (.08 to .84), stair climbing (.12 to .90), and sweeping (.12 to .70). No device was accurate across all activities. The most accurate devices for activities were: walking (Apple Watch, r=.84); stair climbing (iPhone, r=.90); sweeping (Fitbit, r=.70). During the treadmill test, correlations between HR assessed via ECG and devices were: Apple Watch (.76 to .99), Polar T-31 (.72 to .94), and Fitbit (.19 to .98). Heart rate accuracy across the session was highest in the Apple Watch. KCAL from neither the Apple Watch (r=.63) nor Fitbit (r=.48) had a high correlational value to that from the metabolic analyzer.

CONCLUSIONS: Fitness-related values provided by wearable devices had varying levels of accuracy when compared to objective step counts, HR and KCAL assessed by calibrated scientific equipment. Accuracy of step counts varied by activity and was higher across activities in the iPhone. HR reported from wearable devices, similarly, had varying levels of accuracy, with the Apple Watch being most accurate across a graded exercise test. Fitness-related information from wearables that are not medical devices and should be considered as estimates and used for motivation. Further validation of these devices should include a variety of physical activity modalities.

© 2017 American College of Sports Medicine