!
Using Apple Watch to Estimate
Six-Minute Walk Distance
May 2021
Overview 3 ..........................................................................................................................................
Introduction 3 ......................................................................................................................................................
In-Clinic Six-Minute Walk Test 3 .....................................................................................................................
Remotely Administered Six-Minute Walk Test 3 .............................................................................................
Estimated Six-Minute Walk Distance 4 .................................................................................................
Metric Description 4 ............................................................................................................................................
Development 5 ....................................................................................................................................................
Results 6 .............................................................................................................................................................
Discussion 8 ........................................................................................................................................................
Conclusions 9 ......................................................................................................................................
References 9.......................................................................................................................................
Using Apple Watch to Estimate Six-Minute Walk Distance
May 2021
2
Contents
Overview
With watchOS 7, Apple Watch Series 3 and later provide an estimate of the distance
users are able to walk in six minutes, a measurement that may be used in health
applications as an overall marker of an individual’s cardiovascular fitness and mobility.
This paper provides a detailed understanding of how the metric is estimated on "
Apple Watch, including testing and validation.
Introduction
In-Clinic Six-Minute Walk Test
The six-minute walk test (6MWT) is a simple method of assessing an individual’s functional capacity and
endurance. First introduced in the early 1980s as a more tolerable and similarly valid and reliable substitute
for more rigorous and expensive forms of exercise testing, the 6MWT is most often performed in a clinic
1
setting, where an individual is observed walking up and down a hallway of known length for six minutes.
The test result is the total distance walked during the six minutes, referred to as the six-minute walk
distance (6MWD). Over the approximately 40 years since its introduction, the test has been wellcharacterized in many age groups, ethnicities, and states of health and disease. It has been used as a
clinical trial endpoint for studying the efficacy of different interventions on disease, as well as a metric for
risk stratification across broad populations.
2
Despite the advantages and utility of the 6MWT, there are drawbacks to the administration and interpretation
of the test in routine practice. First, the office or clinic layout must provide an environment — typically a
hallway — of adequate length (generally 10 to 30 meters) where an individual can walk uninterrupted without
disrupting normal clinic workflows. Second, the time and associated costs required for the test can be
significant, as best practice normally calls for at least one practice test to familiarize the user with the
protocol, followed by an hour rest, then followed by a second test that generally yields a more accurate
result.
, , ,
3 4 5 6
Remotely Administered Six-Minute Walk Test
In a busy clinical setting, administration of a 6MWT may not be practical. Home-administered 6MWTs "
are strongly correlated with supervised, in-clinic exams and may be a suitable substitute. Measurements
7
of 6MWD using smartphones correlate with in-clinic measurements; however, in a study setting, only "
60 percent of at-home 6MWTs administered and measured with a smartphone were completed. Home
8
testing, while promising, may not be feasible at scale or over long time periods given low participant
compliance.
Apple Watch is changing how 6MWD is measured, and it’s helping overcome the above practical
challenges of both home testing and in-clinic exams. Using data passively collected from Apple Watch
Series 3 and later with watchOS 7, users will be able to visualize and share an estimate of their 6MWD. The
estimated six-minute walk distance (e6MWD) is based on multimodal sensor signals passively observed
over long time periods of a user’s normal behaviors, rather than a direct, single point-in-time measurement
of a six-minute walking bout. In most cases, an e6MWD will be updated weekly and available for users to
view in the Health app on an iPhone paired to their Apple Watch.
Using Apple Watch to Estimate Six-Minute Walk Distance
May 2021
3
Estimated Six-Minute Walk Distance
This paper describes the development and validation of the e6MWD metric. The intended audience is
researchers, healthcare providers, and developers who might want to incorporate this metric into their
work, as well as customers who would like to know more about the e6MWD metric and how it’s collected
and validated. Developers can visit developer.apple.com/documentation/healthkit for information on how "
to access and share health and fitness data — like e6MWD — while maintaining user privacy and control.
Metric Description
The e6MWD metric estimates the distance a user would walk if asked to perform a 6MWT in a proctored
setting. This estimate is based on motion and activity data measured during the four weeks preceding the
estimate. An example of what a user might see in the Health app on iPhone is shown in figure 1. The value "
is reported in meters and, for most users, will be updated weekly. To generate an e6MWD, Apple Watch
requires at least three days with more than eight hours of wear time during normal daily behaviors (for
example, waking hours while engaging in light household activities) within the past week, as well as a total
of 10 days meeting the eight-hour threshold over the preceding four weeks. An outdoor-walk workout
tracked with the Workout app on Apple Watch in the prior week may reduce these requirements.
Apple’s HealthKit API allows users to share this information with apps installed on iPhone using Apple’s
HealthKit API. Each estimate has accompanying metadata that report the device calibration status, which
9
may impact estimate accuracy and is described further in the discussion. These metadata are included
when other apps access estimates using the HealthKit API.
Using Apple Watch to Estimate Six-Minute Walk Distance
May 2021
4
Figure 1: Estimated six-minute walk distance in the Health app in iOS 14
Development
To design and validate the e6MWD metric, Apple collected data from multiple studies, approved by an
institutional review board (IRB), involving adults ages 65 and older who live either in the community or "
in independent living housing and had consented to the collection and use of their data for this purpose.
Participants were selected to ensure a diversity of activity levels and functional status; however, all
participants resided in Santa Clara Valley, California.
Study participants completed supervised 6MWTs consistent with published guidelines2 on a linear and "
flat “out and back” course ranging from 15 to 30 meters in length while wearing Apple Watch Series 4 and
carrying an iPhone 8 or later. Protocols for reference 6MWTs for gathering design and validation data were
identical. Participants completed up to five 6MWTs over the course of their participation in the studies, with
no repeated tests occurring on the same day. Reference test results were verified using additional sensor
measurements to ensure that recorded lengths were accurate. Tests that failed these verification steps
weren’t used in algorithm development.
Participants were then asked to wear their Apple Watch and carry their iPhone during normal day-to-day
activities throughout the length of the study. Data from a variety of Apple Watch and iPhone sensors "
were collected during the study period and used to design the e6MWD algorithm. Inputs into the e6MWD
algorithm are sensor data (accelerometer, gyroscope, barometer, and GPS) and other metrics measured by
Apple Watch, such as flights climbed, steps, exercise minutes, walking distances, estimated step length,
and walking speed.
Participant data was split into design and validation data sets to balance age, gender, and functional level
across both sets; the design set was used to develop the e6MWD algorithm. Algorithm performance was
determined by comparing e6MWD to participant reference 6MWD results. The validation data set was then
used to confirm algorithm performance.
The validity of the e6MWD metric was computed as the mean and standard deviation of errors between "
the weekly e6MWD estimates and the participant’s temporally proximate reference test. The reliability was
evaluated by calculating the absolute agreement among measurements and was reported as the intraclass
correlation coefficient (ICC). The consistency of the e6MWD metric is expressed as the median and 90th
percentile standard deviation of per-subject weekly e6MWD estimates for participants who had at least
three estimates. The availability or yield of the e6MWD metric was evaluated as the fraction of total weeks
meeting the minimum Apple Watch wear time requirements that yielded estimates and the fraction of
participants who received estimates at least 75 percent of weeks.
Using Apple Watch to Estimate Six-Minute Walk Distance
May 2021
5
Results
Baseline characteristics of the participants whose data was used for design and validation are summarized
in table 1.
The median and 90th percentile standard deviation of reference 6MWDs per participant were 16 meters
and 41 meters, respectively, and the ICC was 0.926 [0.921–0.931 CI]. When restricting to reference tests
that passed verification for accuracy, the median and 90th percentile standard deviation was 15 meters and
37 meters, and the ICC was 0.939 [0.934–0.943 CI]. Studies have reported similar ICC values ranging from
0.82 to 0.99. !
10
Table 1. Participant characteristics
Design (N = 930)
Val id ati on ( N = 449)
Gender — number (%)
Female
578 (62)
263 (59)
Male
349 (38)
184 (41)
Other/Unknown
3 (0)
2 (0)
Age — years*
82 ± 7
78 ± 7
Reference 6MWD — meters (mean ± SD)
375 ± 98
399 ± 102
Length of observation — days (mean ± SD)
389 ± 48
359 ± 67
Comorbidities — number (%)
Arthritis (hip or knee)
244 (26)
80 (18)
Diabetes
62 (7)
30 (7)
Coronary artery disease
79 (9)
31 (7)
COPD
37 (4)
7 (2)
Assistive-device usage (during 6MWT) — number (%)
None
718 (77)
395 (88)
Cane
73 (8)
26 (6)
Walker
145 (16)
31 (7)
Other/Unknown
20 (2)
5 (1)
BMI category — number (%)
Underweight (BMI < 18.5)
13 (1)
2 (0)
Normal weight (18.5 ≤ BMI < 25.0)
379 (40)
158 (35)
Overweight (25.0 ≤ BMI < 30.0)
352 (38)
191 (43)
Obese (BMI ≥ 30.0)
185 (20)
98 (22)
*Participants over the age of 90 didn’t have their exact ages recorded to avoid potential reidentification. The mean age was calculated
with these participants’ ages set to 90 years.
Using Apple Watch to Estimate Six-Minute Walk Distance
May 2021
6
"
Table 2 shows algorithm performance for the design and validation data sets. Figure 2 shows the specificity
and sensitivity of using e6MWD to classify users with respect to a threshold of 360 meters, along with "
the corresponding ROC curve (AUC 0.946). Although the 6MWT doesn’t yet have widely accepted risk
stratification thresholds, several studies have demonstrated relationships between 6MWD (or the
corresponding mean walking speed) and health outcomes at similar thresholds.2
Ta b le 2 . e 6 M W D pe r f o r m an c e
Metric
Description
Design (N = 930
participants, !
35,890 weeks)
Validation (N = 449
participants, !
15,223 weeks)
Validity
Error (weekly e6MWD — nearest reference
test) — meters (mean ± SD)
1 ± 55
1 ± 51
Reliability
ICC [confidence interval]
0.925 [0.922–0.928]
0.913 [0.909–0.916]
Consistency
SD of e6MWD per user — meters
Median
211790th percentile
40
35
Availability
Percentage of weeks* yielding e6MWD
92
94
Percentage of subjects who received
e6MWD at least 75% of weeks*
89
92
*Weeks with sufficient wear time (that is, those meeting the threshold of at least three days with over eight hours of data in the most "
recent week).
Using Apple Watch to Estimate Six-Minute Walk Distance
May 2021
7
AUC 0.946
Figure 2: ROC curve, sensitivity, and specificity for e6MWD classification at 360-meter threshold !
(TP = true positive, FP = false positive)
Discussion
With the release of watchOS 7 and iOS 14, Apple Watch users will be able to view an estimate of their
6MWD in the Health app on their iPhone. Reliable weekly estimation of 6MWD may remove the need for
users to perform at-home 6MWTs and, given potentially low compliance with at-home testing, improve
long-term monitoring of functional capacity and endurance for at-risk individuals.
The e6MWD described here had comparable accuracy across design and validation data sets, with
consistency comparable to those of reference tests in the current study and reported by others. Under
11
normal, unprompted participant behavior (that is, undirected time spent wearing Apple Watch and normal
activity levels), the metric availability was over 90 percent (see table 2). For study participants for whom
continuous data were available (n = 703), the mean time between estimates was approximately nine days,
and 94 percent of participants had a mean time between estimates of less than two weeks. This suggests
that most users should have approximately weekly estimates, particularly in a setting such as a monitored
research study where participants are reminded to wear Apple Watch.
Users can optimize yield by wearing a calibrated Apple Watch daily across a representative range of
activities. In the absence of this device calibration, the iPhone-based walking speed metric or GPS from
both iPhone and Apple Watch can also be used to calibrate this metric. Most users (75 percent) achieved
this calibration within four weeks in the described studies. More information on calibrating Apple Watch "
can be found at support.apple.com/en-us/HT204516.
We developed the 6MWD algorithm to provide accurate estimates on low-capacity, potentially at-risk
individuals by designing and validating the algorithm in a population of older adults with reference 6MWD
within a clinically useful range and in the absence of any single comorbidity. The study populations used for
design and validation had prevalences of diabetes and osteoarthritis in line with those of the similarly aged
U.S. population; COPD and coronary artery disease may have been underweighted relative to expected
12
prevalence. Assistive-device usage during reference testing overall approximated that of the U.S.
13
population; although cane usage may have been slightly underrepresented.
14
Clinical applications of 6WMT (such as risk stratification and therapeutic response measurement) typically
utilize distances less than 500 meters across a variety of conditions, including heart failure, pulmonary
15
artery hypertension, chronic obstructive pulmonary disease, and cancer.8 Based on their level of fitness "
16 17
or age, many users may consistently have values of 500 meters, the maximum estimate supported by this
algorithm.
While equations predicting 6MWD are typically based on an individual’s height, weight, age, and potentially
other characteristics,
,
the e6MWD described here is an individualized estimation based on direct, sensor-
18 19
derived measurements; comparing e6MWD to reference equation-based predictions to inform a “percentage
of expected” may hold value beyond that of the e6MWD value in isolation. The absolute e6MWD may"
20
be useful for risk stratification in older adults. Yazdanyar et al. found that among community-dwelling older
adults, a 6MWD of less than 338 meters was associated with an increased risk of all-cause mortality.
2.
The approach and results presented aren’t without limitations. The data used was from proctored 6MWTs
using different course lengths to develop the e6MWD. While some reports suggest that course length has "
a statistically significant effect on 6MWD, most studies show that this difference isn’t clinically relevant
21
and that consistent course layout (for example, oval compared with “out and back”) has more impact on
6MWD.
,
Increased variability in low-capacity users was present, particularly those using walkers (data
22 23
not shown). Racial and ethnic diversity in the study population didn’t approximate the U.S. population;
however, prior studies in similarly aged populations haven’t found independent differences in 6MWD based
on race or ethnicity and haven’t recommended adjusting expected values based on these factors.
24
Using Apple Watch to Estimate Six-Minute Walk Distance
May 2021
8
The studies used for design and validation of the 6MWD algorithm weren’t designed to capture significant
changes in participants’ 6MWD over the course of observation; however, changes based on health events,
natural aging, or progression of comorbidities may have occurred. For clinic-based measurements of
6MWD, “clinically significant” change may range from approximately 15 to 50 meters depending on any
underlying pathology affecting the population being studied.4, Future work demonstrating the ability to
25
detect change on this scale may enable researchers and developers to use this metric for monitoring the
course of disease or recovery from a health event.
Conclusions
Apple Watch can provide a weekly estimate of 6MWD, giving users a new metric to measure and monitor
multiple dimensions of their health over time. This data may also be useful to researchers and healthcare
providers. Wearing a calibrated Apple Watch across a range of activities that are representative of a user’s
capacity is the best way to assure that an accurate estimate is recorded each week.
References
Butland RJ, Pang J, Gross ER, Woodcock AA, Geddes DM. Two-, six-, and 12-minute walking tests in respiratory disease. British Medical
1
Journal (Clinical Research Edition). 1982; 284(6329): 1607–1608. doi: 10.1136/bmj.284.6329.1607.
Yazdanyar A, Aziz MM, Enright PL, et al. Association Between 6-Minute Walk Test and All-Cause Mortality, Coronary Heart Disease–
2
Specific Mortality, and Incident Coronary Heart Disease. Journal of Aging and Health. 2014; 26(4): 583–599. "
doi: 10.1177/0898264314525665.
ATS Committee on Proficiency Standards for Clinical Pulmonary Function Laboratories. ATS statement: guidelines for the six-minute walk
3
test. American Journal of Respiratory and Critical Care Medicine. 2002; 166(1): 111–117. doi: 10.1164/ajrccm.166.1.at1102.
Solway S, Brooks D, Lacasse Y, Thomas S. A qualitative systematic overview of the measurement properties of functional walk tests used
4
in the cardiorespiratory domain. Chest. 2001; 119(1): 256–270. doi: 10.1378/chest.119.1.256.
Guyatt GH, Pugsley SO, Sullivan MJ, et al. Effect of encouragement on walking test performance. Thorax. 1984; 39(11): 818–822. "
5
doi: 10.1136/thx.39.11.818.
Guyatt GH, Sullivan MJ, Thompson PJ, et al. The 6-minute walk: a new measure of exercise capacity in patients with chronic heart failure.
6
Canadian Medical Association Journal. 1985; 132(8): 919–923.
Du H, Davidson PM, Everett B, et al. Correlation between a self-administered walk test and a standardised Six Minute Walk Test in adults.
7
Nursing & Health Sciences. 2011; 13(2): 114–117. doi: 10.1111/j.1442-2018.2011.00605.x.
Douma JAJ, Verheul HMW, Buffart LM. Feasibility, validity and reliability of objective smartphone measurements of physical activity and
8
fitness in patients with cancer. BMC Cancer. 2018; 18(1): 1052. doi: 10.1186/s12885-018-4983-4.
developer.apple.com/documentation/healthkit.
9
Holland AE, Spruit MA, Troosters T, et al. An official European Respiratory Society/American Thoracic Society technical standard: field
10
walking tests in chronic respiratory disease. European Respiratory Journal. 2014; 44(6): 1428–1446. doi: 10.1183/09031936.00150314.
Gibbons WJ, Fruchter N, Sloan S, Levy RD. Reference values for a multiple repetition 6-minute walk test in healthy adults older than "
11
20 years. Journal of Cardiopulmonary Rehabilitation. 2001; 21(2): 87–93. doi: 10.1097/00008483-200103000-00005.
Centers for Disease Control and Prevention (website). U.S. Department of Health & Human Services. Accessed September 2, 2020.
12
cdc.gov.
Older Americans & Cardiovascular Diseases: Statistical Fact Sheet. American Heart Association. heart.org/idc/groups/heart-public/
13
@wcm/@sop/@smd/documents/downloadable/ucm_483970.pdf.
Gell NM, Wallace RB, LaCroix AZ, Mroz TM, Patel KV. Mobility device use in older adults and incidence of falls and worry about falling:
14
findings from the 2011–2012 national health and aging trends study. Journal of the American Geriatrics Society. 2015; 63(5): 853–859.
doi: 10.1111/jgs.13393.
Using Apple Watch to Estimate Six-Minute Walk Distance
May 2021
9
Yap J, Lim FY, Gao F, Teo LL, Lam CSP, Yeo KK. Correlation of the New York Heart Association Classification and the 6-Minute Walk
15
Distance: A Systematic Review. Clinical Cardiology. 2015; 38(10): 621–628. doi: 10.1002/clc.22468.
Boucly A, Weatherald J, Savale L, et al. Risk assessment, prognosis and guideline implementation in pulmonary arterial hypertension. "
16
European Respiratory Journal. 2017; 50(2): 1700889. doi: 10.1183/13993003.00889-2017.
Cote CG, Casanova C, Marin JM, et al. Validation and comparison of reference equations for the 6-min walk distance test. European
17
Respiratory Journal. 2008; 31(3): 571–578. doi: 10.1183/09031936.00104507.
Zou H, Zhu X, Zhang J, et al. Reference equations for the six-minute walk distance in the healthy Chinese population aged 18–59 years.
18
PLOS ONE. 2017; 12(9): e0184669. doi: 10.1371/journal.pone.0184669.
Enright PL, Sherrill DL. Reference equations for the six-minute walk in healthy adults. American Journal of Respiratory and Critical Care
19
Medicine. 1998; 158(5 Pt 1): 1384–1387. doi: 10.1164/ajrccm.158.5.9710086.
Troosters T, Gosselink R, Decramer M. Six minute walking distance in healthy elderly subjects. European Respiratory Journal. 1999;
20
14(2): 270–274. doi: 10.1034/j.1399-3003.1999.14b06.x.
Almeida VP, Ferreira AS, Guimarães FS, Papathanasiou J, Lopes AJ. Predictive models for the six-minute walk test considering the
21
walking course and physical activity level. European Journal of Physical and Rehabilitation Medicine. 2019; 55(6): 824–833. "
doi: 10.23736/S1973-9087.19.05687-9.
Heinz PDR, Gulart AA, Klein SR, et al. A performance comparison of the 20 and 30 meter six-minute walk tests among middle aged "
22
and older adults. Physiotherapy Theory and Practice. 2019; 1: 1–9. doi: 10.1080/09593985.2019.1645251.
Sciurba F, Criner GJ, Lee SM, et al. Six-minute walk distance in chronic obstructive pulmonary disease: reproducibility and effect of
23
walking course layout and length. American Journal of Respiratory and Critical Care Medicine. 2003; 167(11): 1522–1527. "
doi: 10.1164/rccm.200203-166OC.
Enright PL, McBurnie MA, Bittner V, et al. The 6-min walk test: A quick measure of functional status in elderly adults. Chest. 2003;
24
123(2): 387–398. doi: 10.1378/chest.123.2.387.
Bohannon RW, Crouch R. Minimal clinically important difference for change in 6-minute walk test distance of adults with pathology: "
25
a systematic review. Journal of Evaluation in Clinical Practice. 2017; 23(2): 377–381. doi: 10.1111/jep.12629.
Using Apple Watch to Estimate Six-Minute Walk Distance
May 2021
10
© 2021 Apple Inc. All rights reserved. Apple, the Apple logo, Apple Watch, HealthKit, iPhone, and watchOS are trademarks of Apple Inc., registered in the U.S. and
other countries. IOS is a trademark or registered trademark of Cisco in the U.S. and other countries and is used under license. Other product and company names mentioned
herein may be trademarks of their respective companies.
Loading...