See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/326226457

Accuracy of swimming wearable watches for estimating energy expenditure

Article in IJASS(International Journal of Applied Sports Sciences) · June 2018

DOI: 10.24985/ijass.2018.30.1.80

CITATIONS READS

12 1,485

3 authors, including:

Mihyun Lee Saejong Park

Korea institute of sports science Korea institute of Sport Science
27 PUBLICATIONS 163 CITATIONS 80 PUBLICATIONS 1,081 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Mihyun Lee on 11 July 2018.

The user has requested enhancement of the downloaded file.

International Journal of Applied Sports Sciences ISSN 2233-7946 (Online)
2018, Vol. 30, No. 1, 80-90. ISSN 1598-2939 (Print)
https://doi.org/10.24985/ijass.2018.30.1.80 ⓒ Korea Institute of Sport Science

Accuracy of swimming wearable watches

for estimating energy expenditure
Mihyun Lee1, Hyojin Lee2, & Saejong Park3*
1
Sungkyul University, Gyeonggi-do, Korea
2
Yongin University, Gyeonggi-do, Korea
3
Korea Institute of Sport Science, Seoul, Korea

Abstract
With the recent installation of waterproof function on wearable watches, various sports activities
including walking, running and even swimming are monitored. Commercially available swimming
wearable watches automatically identified stroke type, swim distance, stroke counts and energy
expenditure (EE). Although the accuracy of estimating EE of walking, bilking and activities of daily life
on activity monitors have been evaluated, it has not been examined for swimming. Thus, the purpose of
the study was to evaluate the accuracy of estimating EE for swimming wearable watches (Apple Watch
S2, Apple and Garmin Finex 3HR, Garmin). A total of 78 swimmers aged 20-59 years (female: 48%)
participated in the study. All the participants wore Apple and Garmin and completed a set of swimming
protocol comprising various speeds (0.4, 0.6, 0.8, 1.0, 1.2 m/s). At each swimming speed they were
asked to swim for four minutes. Lap counts, stroke counts and energy expenditure (EE) from the Apple
and Garmin were evaluated with the criterion measures. Lap counts and stroke counts were directly
counted by the research assistant. The portable respiratory gas analyzer (K4b2, Cosmed, Italy) and a
swimming snorkel (Aqua Trainer Snorkel, Cosmed, Italy) was used as the criterion measure of EE. The
mean absolute percentage error (MAPE) of lap counting and stroke counts at various swimming speed
were within 10% for Apple (lap counts: 0.5-6.1%, stroke counts: 6.2-9.3%) and about 20% for Garmin
(lap counts: 0-20.6%, stroke counts: 6.8-17.6%). However, the MAPE of EE was higher for Apple
(17.1%-151.7%) than for Garmin (17.9%-32.7%). The accuracy of estimating EE tended to improve with
increasing swimming speed for both Apple and Garmin. The EEs from Apple were outside the
equivalence zone except for at 1.2 m/s and were overestimated compared to the criteria. On the other
hand, EEs from Garmin were within the equivalence zone at all speeds except for 1.2 m/s. In
conclusion, Apple and Garmin wearable watches accurately measure lap counts and stroke counts.
However, the accuracy of estimating EE are poor at slow to medium swimming speed. Further
improvement is needed to estimate energy expenditure of swimming at various speed.

Key words: Wearable device, Swimming, Lap count, Stroke count, Energy expenditure

Introduction1 activity monitoring functions such as step counts, heart rate

and energy expenditure (EE) etc. Wearable watches are
Most wearable watches on the market have physical used as tools to maintain and/or promote health by
motivating physical activities (Chowdhury, Western,
Submitted : 28 March 2018 Nightingale, Peacock, & Thompson, 2017). Wearable-based
Revised : 24 May 2018
Accepted : 11 June 2018 behaviors changes may be effective to promote physical
Correspondence : [email protected]
 Accuracy of Swimming Wearable Watches 81

activity with instantaneous feedback and monitoring yet. Previously, the accuracy of step counts, heart rate and
exercise intensity. As the wearable devices are utilized as EE were examined for walking, running, cycling, and
intervention tools in clinical and research setting, the rowing (Brazeau et al., 2011; Case et al., 2015;
accuracy of the wearables are important. Accordingly, Chowdhury et al. 2017; Dooley et al. 2017; Erdogan et al.,
many studies on the accuracy of wearable devices are now 2010; Wallen et al., 2016; Stahl et al., 2016). Therefore,
emerging. Previous studies were limited to the validity of the purpose of the present study was to evaluate the
wearable devices that included measures of step counting, accuracy of the information on lap count, stroke count, and
heart rate and EE while walking and running. These EE provided by wearable devices (Apple and Garmin)
studies have reported that counting is relatively accurate during swimming.
(Case et al., 2015), however, the accuracy of heart rate and
EE vary depending on the device and the intensity of Research Method
exercise (Chowdhury et al. 2017; Dooley et al. 2017;
Wallen et al., 2016; Stahl et al., 2016). Participants
Swimming is one of the most popular cardiorespiratory
The participants of this study were healthy adults, aged
exercise involving rhythmic movements of large muscles
from 20 to 59 years, free from illnesses or injuries,
along with walking and running. Swimming can be
unrestricted physical function and able to swim at various
enjoyed through the lifespan regardless of age and health
speeds. Exclusion criteria were health risk factors such as
status. Swimming is widely recommended not only for
hypertension, or a dental implant because of problems
healthy people but also for overweight/obese people,
wearing a snorkel. The five participants (4 male and 1
patients with arthritis, and elderly persons with joint
female) were excluded due to missing data or dropping out
problems for health (American College of Sports Medicine,
in the experiment. In the study, a total of 78 participants
2013). With the recent installation of waterproof function
(40 male and 38 female) were collected for statistical
on wearable watches, swimming performance is easily
analysis. The demographic characteristics are shown in
monitored. The commercial wearable watches include
Table 1. All participants were given an overview of the
features such as stroke counts and swim speed, and types
procedures and signed the Institutional Review Board
of stroke performed atomically. Swimming wearable
approved informed consent document.
watches are primarily aimed for recreational swimmers as
opposed to elite swimmers. Thus, the accuracy of basic
Study Procedures
features such as lap counts, stroke counts and EE are
prominent to be used as an intervention tool for The experiment was conducted in a 25 m indoor pool
recreational swimmers to promote physical activity. with the water temperature was 28.8°C. Participants
Mooney et al. (2017) investigated the accuracy of the completed the Physical Activity Readiness Questionnaire
swimming movement recognition (stroke type, swim and resting heart rate, blood pressure, body height, weight
distance, stroke rate, stroke length, average speed etc.) of and body fat percentage were measured using standard
TM
wearable devices. The accuracy of Garmin Swim and methods.
Ⓡ
Finis Swimsense was tested on national swimmers, and All participants wore Apple and Garmin and completed
the results were reported to be similar. The authors, a set of swimming protocol (0.4, 0.6, 0.8, 1.0, 1.2 m/s)
however, proposed a study on recreational swimmers since comprising various speed. An experimental protocol was
their participants were elite swimmers whose swimming chosen based on the intensity used in a previous study
form is outstandingly consistent. Also one of the key related to swimming EE (Holmer et al., 1972; Pendergast
features of wearables, EEs of swimming is not evaluated et al.; 1977).
82 Mihyun Lee et al.

Table 1. Demographic characteristics of the participants (mean±SD)

Characteristics Male (n=40) Female (n=38) Total (n=78)

Age (year) 38.7±11.05 39.1±10.67 37.1±10.79
Height (cm) 173.7±5.99 161.0±5.72 167.5±8.66
Weight (kg) 73.9±7.20 57.8±6.20 66.0±10.49
2
BMI (kg/m ) 24.5±1.85 22.3±1.84 23.4±2.13
Body fat (%) 19.1±5.08 26.1±6.28 22.5±6.67
Swimming experience (year) 8.8±5.97 10.5±5.96 9.6±5.98
50m best record (s) 34.4±5.69 40.35±6.33 37.3±6.67
BMI: body mass index, weight (kg) divided by height (m)2, %BF: %body fat

Table 2. Objective and subjective exercise intensity by swimming speed (mean ± SD)

N Oxygen uptake, VO2 Metabolic equivalents Heart rate (bpm) Rating of perceived
(mL/kg/min) exertion
Resting 78 3.59±1.05 1.03±0.30 88.61±12.44 8.26±1.82
0.4m/s 46 13.28±2.89 3.80±0.83 115.26±11.06 9.67±2.09
0.6m/s 75 18.57±4.62 5.31±1.33 123.21±12.39 10.41±2.25
0.8m/s 68 25.03±5.29 7.15±1.51 142.00±15.65 12.33±1.99
1.0m/s 53 31.11±6.07 8.89±1.73 159.35±14.87 15.85±3.20
1.2m/s 10 40.77±6.13 11.65±1.75 165.20±9.11 16.64±1.80
bpm: beats per minute

To help the participant maintain appropriate time for Criterion lap counts and stroke counts were directly
each protocol, a research assistant provided feedback the counted by the research assistant. The lap counts were
time at every 25 m turn from the swimming pool floor. swimming back and forth once in a 50m pool (unit: lap
To standardize the amount of swimming, underwater counts/ bout) and stroke counts were one stroke was
swimming was controlled, and the side turn was the considered to be completed when one arm was rotated (unit:
only turn method permitted to use snorkel. stroke counts/ bout). Bout means individual speed of
In the present study, the total lap counts and stroke swimming consisting of four minutes. The portable
counts for four minutes of each protocol were used in the respiratory gas analyzer (K4b2, Cosmed, Italy) and a
analysis. To use the value of steady-state of each protocol, swimming snorkel (Aqua Trainer Snorkel, Cosmed, Italy)
the data of the last two minutes (3–4 minutes) for all was used as the criterion measure of EE. Cosmed K4b2 is
activities were used in data analysis. For the EE of each a light weight (925g) measuring instrument worn on the back,
wearable watches, EE per minute was calculated by which allows real-time monitoring of oxygen intake, carbon
dividing actual swimming time by minutes. dioxide emissions, and ventilation rate using the method. The
The descriptive statistics of 78 participants by protocol Aqua Trainer Snorkel is connected to the K4b2 analyzer and
including resting are shown in Table 2. Oxygen uptake, used for real-time gas analysis during swimming. There
metabolic equivalents, heart rate, and participative RPE liability of both criterion measurement instruments has been
were gradually increased as the speed increased. verified by many researchers (Eisenmann, Brisko, Shadrick,
& Welsh, 2003; McLaughlin, King, Howley, Bassett, &
Measuring instruments Ainsworth, 2001; Baldarietal., 2013; Keskinenetal., 2003).
The portable gas analyzer was calibrated according to its
Criterion measurement instruments manual before data collection.
 Accuracy of Swimming Wearable Watches 83

Statistical analysis
Wearable watches
Among the top five best-selling (international data The analyses were conducted using SPSS ver. 22.0 and
corporation; IDC, 2017) wearable watches (Xiaomi, Apple, Medcale ver. 14.0, and the specific data processing
Fitbit, Samsung, Garmin), two swim activity monitors, the methods were as follows.
Apple Watch S2 and Garmin Finex 3HR. Apple has had a Means and standard deviations were calculated of
swimming recognition function since 2016. Apple Watch S2 physical characteristics and all relevant data. Two-way
(Apple Inc, Cupertino, CA, USA), which was used in the analysis of variance (ANOVA) was used to test the effects
present study, provides information such as activity of swimming by speed and the interaction effect between
metabolism, EE, heart rate, stroke count, lap count, travel test instruments, and the p hoc test was performed using
distance, total time, and pace during swimming (Apple, 2017). the Bonferroni test. The levels of error observed on the
Garmin was established in 1989 and specialized in wearable watches were calculated into mean absolute
global positioning system (GPS). It began to produce percent error (MAPE %) based on the criterion measure:
exercise-related wearables in 2003, and a swimming-related [|measured valued observed in wearable devices – criterion
function was first installed on the Forerunner 910XT in measures | / criterion measures × 100%].
2012. The Garmin Finex 3HR (Garmin, Ltd, Schaffhausen, Mean differencewearable watches using Bland &
Switzerland), which was used in the present study, Altman plot. In addition, to assess measurement agreement
provides information such as EE, stroke count, travel between criterion measures and the measurements made by
distance, exercise duration, average pace, maximum pace, the Apple and Garmin devices, a 95% equivalence test was
average speed, top speed, and SWOLF (swim golf) etc. performed. For wearable watches to be considered
(Garmin, 2017). equivalent to the criterion with 95% precision, equivalence
In the present study, lap counts, stroke counts, and EE zone was determined as ± 10% of the criterion mean (Lee,
from the Apple and Garmin were evaluated with the Kim and Welk, 2014) this. All statistical significance
criterion measures. But Garmin does not present lap count levels were set at α=.05.
information, lap counts were used to calculated by dividing
the total distance traveled in four minutes by 50m. Results
Test of difference between criteria and
Others measures wearable watches
Height was measured with an accuracy of 0.1 cm using
an extensimeter (Jenix, DS-102, Korea). Weight and body Descriptive statistics of criterion measures of swimming
fat percentage were measured to 0.1 kg and 0.1% in light by speed and lap counts, stroke counts, and EE provided
clothing using a body composition analyzer (Biospace, by Apple and Garmin devices are shown in Table 3. The
InBody 720, Korea). The measured height and weight results of a two-way ANOVA on the effects of swimming
information was entered into the user information field of by speed and the interaction effect between test
each wearable watches and the portable respiratory gas instruments showed no statistically significant interaction
analyzer before the experiment. Heart rate was measured effect for lap counts and stroke counts (lap counts: F =
using a waterproof heart rate chest strap (Polar Electro Oy, 1.359, p = .211, stroke counts: F = .145, p =. 997), but
Polar V800, Finland), and the subjective exercise intensity statistically significant interaction effect (F = 9.395, p <
was measured on the Borg scale immediately after each .001) was found for EE. The post hoc test showed no
protocol. significant difference between Garmin and the criteria at
all speeds, but Apple was overestimating compared to the
84 Mihyun Lee et al.

Table 3. Descriptive statistics of criteria and wearable watches

Raw Score MAPE (%)

N Device
Mean SD Mean SD
Lap counts/bout
Criterion 4.00 .00
0.4m/s 41 Apple Watch S2 3.85 .57 6.10 13.44
Garmin Finex 3HR 3.38 1.05 20.63 22.57
Criterion 5.97 .23
0.6m/s 71 Apple Watch S2 6.00 .17 0.47 2.78
Garmin Finex 3HR 5.60 .81 8.89 12.07
Criterion 7.94 .58
0.8m/s 67 Apple Watch S2 8.04 .66 0.93 4.52
Garmin Finex 3HR 7.69 .97 5.89 10.19
Criterion 9.94 .74
1.0m/s 52 Apple Watch S2 9.88 .90 1.35 4.86
Garmin Finex 3HR 9.78 .95 3.34 7.00
Criterion 12.00 .47
1.2m/s 10 Apple Watch S2 11.78 .67 1.67 3.51
Garmin Finex 3HR 12.00 .53 0 0
Stroke count/bout
Criterion 57.71 14.82
0.4m/s 44 Apple Watch S2 56.30 17.00 9.33 11.95
Garmin Finex 3HR 63.53 18.25 17.59 20.76
Criterion 70.75 13.40
0.6m/s 71 Apple Watch S2 70.10 12.80 6.52 8.18
Garmin Finex 3HR* 76.69 12.11 9.01 12.77
Criterion 86.11 16.02
0.8m/s 63 Apple Watch S2 87.66 17.04 6.67 6.08
Garmin Finex 3HR 90.93 15.04 7.65 6.43
Criterion 109.00 21.43
1.0m/s 51 Apple Watch S2 108.51 23.39 6.23 7.06
Garmin Finex 3HR 114.06 17.85 7.95 8.51
Criterion 117.32 17.20
1.2m/s 10 Apple Watch S2 117.40 10.44 6.81 2.46
Garmin Finex 3HR 120.60 10.61 6.81 3.37
Energy expenditure (kcal/min)
Criterion 3.78 1.00
0.4m/s 33 Apple Watch S2 9.48 2.11 151.66 55.16
Garmin Finex 3HR 3.49 1.30 32.74 24.95
Criterion 5.79 1.89
0.6m/s 60 Apple Watch S2 10.88 2.24 100.09 55.25
Garmin Finex 3HR 6.28 1.74 30.50 30.23
Criterion 8.06 2.22
0.8m/s 59 Apple Watch S2 12.13 2.52 61.26 34.79
Garmin Finex 3HR 8.43 1.96 24.82 20.48
Criterion 10.59 2.71
1.0m/s 41 Apple Watch S2 13.03 2.56 32.70 22.51
Garmin Finex 3HR 10.97 1.91 18.94 18.70
Criterion 14.53 2.46
1.2m/s 9 Apple Watch S2 14.41 2.72 17.71 9.33
Garmin Finex 3HR 14.86 3.48 17.93 15.67
MAPE (%): mean absolute percentage error, Bout means individual speed of swimming consisting of four minutes.
 Accuracy of Swimming Wearable Watches 85

criteria and Garmin (p < .001). 1.2m/s. Garmin showed MAPE of 17.9%-32.7%, and both
Figure 1 shows criterion measures by swimming speeds wearable watches showed a tendency of gradually
compared to the MAPE of lap counts, stroke counts, and decreasing the MAPE as the intensity of exercise
EE provided by Apple and Garmin. For lap counts and increased.
stroke counts, the MAPE of Apple was within 10% (lap
counts: 0.5-6.1%, stroke counts: 6.2-9.3%) while that of Bland & Altman plot of criteria and
Gamin was about 21% (lap counts: 0-20.6%, stroke counts: wearable watches
6.8-17.6%), and both devices showed higher error rates
The 95% confidence interval of mean difference limits
when speed was slower. On the other hand, Apple
was determined with a Bland-Altman plot created by using
overestimated EE at all speeds in that the MAPE at the
criterion measures and averages and average differences of
speed of 0.4 m/s-1.0 m/s were vigorous (32.70%-151.66%)
lap counts, stroke counts, and EE provided by the Apple
but the MAPE (17.93%) became lower at the speed of
and Garmin devices (Figure 2). For lap counts, Apple
(Mean difference = 0, Low LoA-Upper LoA: -0.8-0.8)
showed a confidence interval with a smaller deviation than
Garmin (Mean difference = -0.3, Low LoA-Upper LoA:
-2.0 - 1.4). For the stroke counts, Apple (Mean difference
= -0.4, Low LoA-Upper LoA: -16.5 - 15.7) also showed a
confidence interval with a smaller deviation than Garmin
(Mean difference = 4.9, Low LoA-Upper LoA: 21.6-
-11.8). For EE, however, Garmin (Mean difference = 0.3,
Low LoA-Upper LoA: -4.1 - 4.7) showed a confidence
interval with a smaller deviation than Apple (Mean
difference = 4.2, Low LoA-Upper LoA: -0.7 – 9.0).

Test of equivalence between criteria and

wearable watches (95% CI)
To test measurement agreement between criterion
measures and the measurements of the Apple and Garmin
a 95% equivalence test was performed Figure 3. Lap
counts were within the equivalence limits (criterion
measure ± 10%) except for 0.4 m/s for Garmin and the
stroke counts was within the equivalence limits except for
Garmin at 0.4 m/s and 0.6m/s.
In the case of Apple, EE was within the equivalence
zone (90% CI = 13.08-15.93) at 1.2 m/s, but it was not
within the equivalence zone at other speeds and
Figure 1. Mean absolute percentage error (MAPE; %) for overestimated the EE compared to the criteria. In the case
wearable watches based on intensity: (A) Lab counts, (B) of Garmin, EE was within the equivalence zone at all
Stroke counts, (C) Energy expenditure. speeds except for 1.2 m/s at which it was outside the
Apple: Apple Watch S2, Garmin: Garmin Finex HR equivalence zone.
86 Mihyun Lee et al.

A. Lap counts

B. Stroke counts

C. Energy expenditure

Figure 2. Bland & Altman plot for wearable watches: (A) Lab counts, (B) Stroke counts, (C) Energy expenditure.
Apple : Apple Watch S2, Garmin : Garmin Finex HR
 Accuracy of Swimming Wearable Watches 87

technological development, wearable watch have been

recently released with which feedback on the amount of
exercise can be checked in real time during swimming.
The goal of ordinary swimmers is health promotion and
weight control, and objective evaluation of the amount of
physical activities can be useful for achieving such a goal.
Until now however, research has been lacking on the
accuracy of the stroke counts, lap counts, and EE at
different speeds, which are the main indicators of the
function of wearable swimming recognition devices.
The accuracy of swimming related data (lap count,
stroke count, EE) by both Apple and Garmin wearable
watches used in the present study was improved as the
intensity of exercise was increased. In addition, the
accuracy of the information was high for lap count and the
stroke count while the accuracy of EE varied depending on
the speed and the type of wearable watches. Recently,
wearable devices are being used as tools to maintain health
by measuring personal activities and increasing motivation
to participate in physical activities (Chowdhury, Western,
Nightingale, Peacock, & Thompson, 2017). Accordingly,
the necessity of research on the accuracy of wearable
devices that provide information on various physical
activities is increasing. The present study evaluated the
accuracy of swimming information provided by wearable
watches while swimming. The significance of the findings
from academic and practical perspectives based on the
results of the present study and other related studies can be
comparatively analyzed as follows.
Dooley et al., (2017) evaluated the accuracy of EE on
Figure 3. Equivalence test (95% CI) for wearable watches : treadmill activities measured by Fitbit, Apple, and Garmin
(A) Lab counts, (B) Stroke counts, (C) Energy expenditure. devices by exercise intensity. The results were consistent
Apple : Apple Watch S2, Garmin : Garmin Finex HR, Bout with the findings of the present study in that the accuracy
means individual speed of swimming consisting of four
minutes. improved as the exercise intensity increased and all three
wearable devices overestimated compared to the criterion
Discussion test instrument. The findings of the study by Wallen et al.
(2016) who evaluated EE measured by wearables such as
Swimming is a representative aerobic exercise and many
Apple Watch, Fitbit Charge HR, Samsung Gear S, and
international organizations related to health recommend
Mio Alpha, however, is in contrast to the findings of the
swimming as an exercise to improve and prevent
present study in that all test instruments in their study
hypertension and cardiovascular diseases. Owing to
tended to underestimate. Because Wallen et al. (2016)
88 Mihyun Lee et al.

analyzed the results of 58-minute circuit training (sitting, respectively, and reported that accelerometers can be
standing, treadmill exercise stress test, cycle exercise) reliably used as a means to measure the amount of
without differentiating exercise intensity, it is impossible to exercise during swimming, which partially agrees with the
interpret which part of the protocol caused the difference. findings of the present study. Both studies (Nordsborg et
In a study (Chowdhury et al. 2017) evaluating the EE al., 2014; Ganzevles et al., 2017) conducted accuracy test
accuracy of wearable devices (Microsoft Band, Apple at higher speeds than the protocol of the present study
Watch, Fitbit Charge HR, Jawbone UP 24, Bodymedia because the research protocol of Ganzevles et al. (2017)
Armband, Actiheart) in a laboratory and in daily life, all also used elite swimmers who could swim freestyle at a
devices were reported to underestimate compared to the speed of 1.32 m/s. A review on protocols used in previous
criterion in a 24-hour daily life setting. However, in a studies related to EE in swimming showed that Montpetiti
laboratory environment, the measurement by Apple was et al., (1988) used 1.0 - 1.25 m/s speed protocol and
similar to the criterion EE and the error rate decreased in presented swimming EE using speed. Barbosa et al.,
vigorous intensity activities such as jogging. In summary, (2006) tested differences in EE by the swimming style and
the accuracy of EE so far measured by commercialized the study protocol was also performed at 1.0 m/s - 1.6 m/s.
wearable devices has varied depending on the research Other studies on swimming EE mostly employed elite
design, and the accuracy increased when the intensity of swimmers and relatively high swimming speed (Craig &
the activity increased. Pendergast, 1979; Zamparo et al., 2005). The reason for
Apple Watch disclosed on its home page that lap count, low accuracy at light intensity activities appears to be
travel distance, and EE are estimated during indoor attributable to differences in swimming skills of the
swimming based on an accelerometer (Apple, 2017). participants, the characteristics of the instrument (such as
Among studies that used accelerometers to predict the accelerometer) installed on wearable devices, and the
amount of swimming, Mooney et al., (2015) investigated limitations in the internal algorithms.
the trend of using sensor technology during the training of Mooney et al., (2017) who evaluated the accuracy of
elite swimmers, and found that the most frequent location wearable devices in the recognition of swimming motion
of sensors during swimming was the wrist and the waist. stated that even though the accuracy of distance estimation
In addition, Nordsborg et al., (2014) investigated the was accurate, the accuracy of stroke speed, stroke length,
relationship between reference oxygen uptake and the size and average speed was affected by lap time and the stroke
of the triaxial accelerometer worn on the wrist, waist, and count, which makes it appropriate for general swimmers to
ankle. They found the strength of correlation in the order use it as a reference value but accuracy needs to be
of wrist (r = 0.77), ankle (r = 0.73), and waist (r = 0.46) improved to use them to improve performance.
and were able to confirm the possibility of using wearable In conclusion, the accuracy of the Apple was higher
devices for the prediction of swimming EE. than that of the Garmin in lap count and the stroke count,
The research protocol of Nordsborg et al., (2014) used but the performance of both wearable watches was similar
the speed of 1.6 m/s for elite swimmer and 1.3 m/s for if lower speed was excluded. The accuracy of the
general participants, which was faster than the speed evaluation of EE of Apple was lower than that of Garmin
employed in the protocol of the present study. Ganzevles and the accuracy was worse at lower speed, which
et al., (2017) compared the reliability and validity of lap indicates the necessity of continuous improvement in the
time and the stroke count by attaching an accelerometer to algorithm of swimming EE of wearable devices.
the back of swimming athletes and comparing the Accordingly, the information on EE of wearable watches
measurements to video recordings. They found the errors while swimming should be selectively used according to
for lap time and the stroke count to be ±2% and ±1%, the situation.
 Accuracy of Swimming Wearable Watches 89

This study has some limitations. First, the accuracy of speed. Further improvement is needed to estimate energy
swimming EEs was limited to freestyle. Because the expenditure of swimming at various speed.
protocol was selected with consideration given to the
maximum freestyle speed of the participants for 50 m, Reference
accuracy for speeds faster than 1.2 m/s were not evaluated
because there was no participant who could perform at a American College of Sports Medicine. (2013). ACSM’s
speed higher than this. In previous studies, participants guidelines for exercise testing and prescription.
were elite swimmers which might influence high accuracy Lippincott Williams & Wilkins.
of estimating energy expenditure. Future studies are needed Apple. (2017). About workout types. Retrieved from
using recreational speeds. In addition, wearable watches https://support.apple.com/en-us/HT207934
were limited to two types to minimize discomfort to Baldari, C., Fernandes, R. J., Meucci, M., Ribeiro, J.,
participants. Considering the fact that wearable watches, Vilas-Boas, J. P., & Guidetti, L. (2013). Is the
which provide swimming information, are continuously new AquaTrainer® snorkel valid for VO2
developing owing to recent technological development assessment in swimming?. International Journal of
device types, speed, and swimming skills (elite vs. Sports Medicine, 34(04), 336-344.
recreational swimmers) is needed in future studies. Barbosa, T. M., Fernandes, R. J., Keskinen, K. L.,
Colaço, P., Cardoso, C., Silva, J., & Vilas-Boas,
Conclusions J. P. (2006). Evaluation of the energy expenditure
in competitive swimming strokes. International
This is the first study to investigate the accuracy of Journal of Sports Medicine, 27(11), 894-899.
estimating energy expenditures from two commercially Brazeau, A. S., Karelis, A. D., Mignault, D., Lacroix,
available wearable watches (Garmin and Apple) of M. J., Prud'homme, D., & Rabasa-Lhoret, R.
freestyle swimming on swimmers. The error rate of lap (2011). Accuracy of the SenseWear ArmbandTM
counting and stroke counts at various swimming speed during ergocycling. International Journal of Sports
were within 10% for Apple and about 20% for Garmin. Medicine, 32(10), 761-764.
The criterion measurements and a 95% equivalence test Case, M. A., Burwick, H. A., Volpp, K. G., & Patel,
showed that the lap counts and the strokes counts recorded M. S. (2015). Accuracy of smartphone applications
by Apple were within the equivalence zone for all of the and wearable devices for tracking physical activity
exercise intensities measured. Bland-Altman plot showed data. JAMA, 313(6), 625-626.
confidence intervals with relatively small deviations in lap Chowdhury, E. A., Western, M. J., Nightingale, T. E.,
counts and the stroke counts for Apple, and EE for Peacock, O. J., & Thompson, D. (2017). Assessment
Garmin. But the error rate of estimating energy of laboratory and daily energy expenditure estimates
expenditure was higher for Apple than for Garmin. The from consumer multi-sensor physical activity
EEs of most swimming speeds from Apple were outside monitors. PloS One, 12(2), e0171720.
the equivalence zone except for at 1.2 m/s and were Craig, A. B., & Pendergast, D. R. (1979). Relationships
overestimated compared to the criteria. On the other hand, of stroke rate, distance per stroke, and velocity in
EEs of various swimming speeds from Garmin were within competitive swimming. Med Sci Sports, 11(3),
the equivalence zone except for 1.2 m/s. In conclusion, 278-283
Apple and Garmin wearable watches accurately measure Dooley, E. E., Golaszewski, N. M., & Bartholomew, J.
lap counts and stroke counts. However, the accuracy of B. (2017). Estimating accuracy at exercise
estimating EE are poor at slow to medium swimming intensities: A comparative study of self-monitoring
 90 Mihyun Lee et al.

heart rate and physical activity wearable devices. K4b2 portable metabolic system. International
JMIR mHealth and uHealth, 5(3), e34. Journal of Sports and Medicine, 22, 280–284.
Eisenmann, J. C., Brisko, N., Shadrick, D., & Welsh, Montpetit, R., Cazorla, G., & Lavoie, J. M. (1988).
S. (2003). Comparative analysis of the Cosmed Energy expenditure during front crawl swimming:
Quark b2 and K4b2 gas analysis systems during A comparison between males and females.
submaximal exercise. Journal of Sports Medicine Swimming Science V. Human Kinetics, Champaign,
and Physical Fitness, 43(2), 150. 101, 229–236.
Erdogan, A., Cetin, C., Karatosun, H., & Baydar, M. Mooney R, Corley G, Godfrey A, Osborough CD,
L. (2010). Accuracy of the Polar S810iTM heart Quinlan L,O Â Laighin G (2015) Application of
TM
rate monitor and the Sensewear Pro Armband video-based methods for competitive swimming
to estimate energy expenditure of indoor rowing analysis: a systematic review. Sports and Exercise
exercise in overweight and obese individuals. Medicine, 1, 133–150.
Journal of Sports Science & Medicine, 9(3), 508. Mooney, R., Quinlan, L. R., Corley, G., Godfrey, A.,
Evenson, K. R., Goto, M. M., & Furberg, R. D. Osborough, C., & ÓLaighin, G. (2017). Evaluation
(2015). Systematic review of the validity and of the Finis Swimsense® and the Garmin Swim™
reliability of consumer-wearable activity trackers. activity monitors for swimming performance and
International Journal of Behavioral Nutrition and stroke kinematics analysis. PloS One, 12(2),
Physical Activity, 12(1), 159. e0170902.
Ganzevles, S., Vullings, R., Beek, P. J., Daanen, H., & Nordsborg, N. B., Espinosa, H. G., & Thiel, D. V.
Truijens, M. (2017). Using tri-axial accelerometry in (2014). Estimating energy expenditure during front
daily elite swim training practice. Sensors, 17(5), crawl swimming using accelerometers. Procedia
990. Engineering, 72, 132–137.
Garmin (2017). Fenix 3HR. Retrieved from Pendergast, D. R., Di Prampero, P. E., Craig Jr, A. B.,
https://buy.garmin.com/en-US/US/p/545480. Wilson, D. R., & Rennie, D. W. (1977).
Holmer, I. (1972). Oxygen uptake during swimming in Quantitative analysis of the front crawl in men
man. Journal of Applied Physiology, 33(4), 502-509. and women. Journal of Applied Physiology, 43(3),
IDC. (2017). Xiaomi and Apple Tie for the Top Position 475–479.
as the Wearables Market Swells 17.9% During the Stahl, S. E., An, H. S., Dinkel, D. M., Noble, J. M., &
First Quarter, According to IDC. Retrieved from Lee, J. M. (2016). How accurate are the wrist-based
https://www.idc.com/getdoc.jsp?containerId=prUS427 heart rate monitors during walking and running
07517. activities? Are they accurate enough? BMJ Open
Lee, J. M., Kim, Y, Welk, G. J (2014). Validity of Sport & Exercise Medicine, 2(1), e000106.
consumer-based physical activity monitors. Med Wallen, M. P., Gomersall, S. R., Keating, S. E.,
Sci Sports Exerc, 46, 1840-1848. Wisløff, U., & Coombes, J. S. (2016). Accuracy
Keskinen, K. L., Rodríguez, F. A., & Keskinen, O. P. of heart rate watches: implications for weight
(2003). Respiratory snorkel and valve system for management. PLoS One, 11(5), e0154420.
breath‐by‐breath gas analysis in swimming. Zamparo, P., Bonifazi, M., Faina, M., Milan, A.,
Scandinavian Journal of Medicine & Science in Sardella, F., Schena, F., & Capelli, C. (2005).
Sports, 13(5), 322–329. Energy cost of swimming of elite long-distance
McLaughlin, J., King, G., Howley, E., Bassett, D., & swimmers. European Journal of Applied Physiology,
Ainsworth, B. (2001). Validation of the Cosmed 94(5–6), 697–704.

View publication stats