An investigation into the validity and reliability of mHealth devices for counting steps in chronic stroke survivors

Abstract

Objective:

To investigate the validity and test–retest reliability of mHealth devices (Google Fit, Health, STEPZ, Pacer, and Fitbit Ultra) to estimate the number of steps in individuals after chronic stroke and to compare whether the measurement of the number of steps is affected by their location on the body (paretic and non-paretic side).

Design:

Observational study with repeated measures.

Setting:

University laboratory.

Subjects:

Fifty-five community-dwelling individuals with chronic stroke.

Intervention:

Not applicable.

Main measures:

The number of steps was measured using mHealth devices (Google Fit, Health, STEPZ, Pacer, and Fitbit Ultra), and compared against criterion-standard measure during the Two-Minute Walk Test using habitual speed.

Results:

Our sample was 54.5% men, mean age of 62.5 years (SD 14.9) with a chronicity after stroke of 66.8 months (SD 55.9). There was a statistically significant association between the actual number of steps and those estimated by the Google Fit, STEPZ Iphone and Android applications, Pacer iphone and Android, and Fitbit Ultra (0.30 ⩽ r ⩾ 0.80). The Pacer iphone application demonstrated the highest reliability coefficient (ICC(2,1) = 0.80; P < 0.001). There were no statistically significant differences in device measurements that depended on body location.

Conclusions:

mHealth devices (Pacer–iphone, Fitbit Ultra, Google Fit, and Pacer–Android) are valid and reliable for step counting in chronic stroke survivors. Body location (paretic or non-paretic side) does not affect validity or reliability of the step count metric.

Keywords

Stroke steps gait community walkers validity

Introduction

Physical activity and exercise guideline recommendations after stroke highlight the importance of keeping the stroke survivor physically active; in part to offset the sedentary lifestyle and associated risk of a recurrent stroke.¹ An easy way to quantify physical activity during clinical practice is the number of steps. Although it is not clear how many steps provide the ideal benefit to health,² 7000 steps per day have been recommended for non-neurologically impaired adults.³ However, individuals after stroke present an average of 4355 steps per day;⁴ an average of 4278 steps on weekdays and 3801 on weekends.⁵

Many technologies currently have a step counting system, among these are mobile applications for different operating systems and other devices, such as wearable sensors.⁶ In general, mobile health technology (mHealth) generates reports often available for analysis by the professionals, as well as patients and relatives. These resources could offer external feedback to motivate an increase in number of steps per day.⁷ However, these technologies may not be sensitive when applied to people with neurological diseases, including stroke, because of abnormal gait patterns. Another limiting factor is the anatomical position of the resource, which varies in the literature⁷and, it is not known yet, which place is the most appropriate to capture physical activity levels of individuals after stroke. Therefore, the use of a step-by-step activity monitoring program can be useful to increase activity outside of a clinical facility in individuals after chronic stroke.^8,9

The Statement from the American Heart Association⁶ recommends investigation of the measurement properties of commercially available smartphone applications and wearable sensors, since this technology is largely already in use. However, there are few studies that have investigated the validity and reliability of mHealth devices in individuals after stroke. Thus, the aims of this study are:

Determine the validity of mHealth devices (Google Fit, Health, STEPZ, Pacer, and Fitbit Ultra) compared with the criterion-standard measure to estimate the number of steps in individuals chronically after stroke.

Determine the test–retest reliability of mHealth devices, and criterion measurement to estimate the number of steps in individuals chronically after stroke.

To compare whether the number of steps in individuals chronically after stroke measured using mHealth devices is affected by where the device is positioned (paretic and non-paretic leg) on the individual.

Methods

This is an investigation of the validity and reliability of mHealth devices approved by Ethics Committee of the Faculty of Medical Science of Minas Gerais (CAAE: 80069917.9.0000.5134) and Ethics Com-mittee of Physical Education and Sport School of São Paulo (CAAE: 10937219.0.0000.5391).

The study enrolled a sample of convenience in the chronic stage after stroke recruited from two facilities: Belo Horizonte and São Paulo- Brazil during the period between March 2018 and May 2019. Each participant engaged in a one-day assessment in the laboratory. Individuals were included who met the following criteria: (a) at least six months after stroke onset,¹⁰ (b) older than 18 years, (c) ⩾18 on the Mini Mental State Examination,¹¹ (d) walking speed at least 0.1 m/s during 10-meter walk test, (e) able to walk independently without assistive devices, (f) able to follow verbal commands, and (g) read, agreed and signed the informed consent.

Sociodemographic and clinical data were obtained through a questionnaire developed by the research team. Walking speed, in m/s, was assessed with the 10-meter walk test for habitual and maximal speed.^12,13 Bilateral strength of knee extensors and plantar flexors, in mmHg, was assessed using the modified sphygmomanometer test.¹⁴ Fear of falling was assessed with the Falls Efficacy Scale—International (FES-I).¹⁵ Physical activity level was assessed with the Adjusted Activity Score of Human Activity Profile (HAP).¹⁶ Functional capacity was assessed with the Duke Activity Status Index (DASI).^17,18 The tests and questionnaires were administered by trained researchers.

Number of steps was measured by mHealth devices, as smartphone applications (Google Fit, Health, STEPZ, and Pacer), and wearable sensors (Fitbit Inc^®), and criterion-standard measure (steps counted live and videotaped—for comparison—by two trained researchers) during a Two-Minute Walk Test with habitual speed.^9,19 Both iOS and Android OS were chosen as operating systems.

Fitbit Inc^® and Google Fit are already known to be accurate for step measurement in individuals after stroke.^9,19,20 To the best of our knowledge, there are no previous studies which have investigated the accuracy of Pacer, Health, and STEPZ for step measurement after stroke. The Health device was previously validated only in non-disabled individuals.²¹

All data were collected by two well-trained researchers, in a random order, during a one-day assessment. Participants were instructed to wear comfortable clothes and habitual shoes. After signing the consent term, the primary outcome (Two-Minute Walk Test) was performed.

The Two-Minute Walk Test was performed over 10 meters marked by two signaling cones.¹⁹ The volunteers waited for the signal to start and to stop the test, using all devices. The mHealth devices used were: Google Fit, Health, STEPZ, Pacer, and the Fitbit Inc^®. The smartphones used here were S480 Positivo and iphone 6s^®—iOS.

The position of the monitors is illustrated in Supplemental Figure S1; the order of measurement was randomized. Prior to the test, the smartphones were positioned in the front pockets of the individuals’ paretic and/or non-paretic lower limb^9,22 and the gender, body mass (Kg), and height (cm) of the individuals were input to the software system. The first test was performed with the Android OS phone (S480 Positivo) and Fitbit Inc^® on the paretic side, and the iOS phone (iphone 6s^®—iOS) on the non-paretic side. The second measure was performed with the inverse positioning. Rest intervals between measurement 1 and measurement 2 were provided between the tests until the heart rate returned to the basal levels.

The individuals walked at their habitual speed with monitors allocated on their body according to the measurement order. Participants were supervised for safety by a member of the research team. The tests were videotaped by iphone 6^® (Apple—iOS) on a tripod positioned 86 cm high and 1 meter away at the beginning of the test. The tripod was positioned at 40° from the beginning of the test. The actual steps were determined by a researcher, who counted the steps taken by the individuals live and from the videotape recordings on two occasions, separated by one or two weeks. We found excellent test–retest reliability (ICC (2,1) P < 0.001). Thus, the actual number of steps counted live was used as the criterion-standard measure.

Descriptive statistics and tests for normality (Shapiro–Wilk) were carried out for all variables. Pearson’s correlation coefficients were calculated to examine the associations between the criterion-standard measure (live actual steps) and those estimated by the devices (Google Fit, Health, STEPZ, Pacer, and Fitbit Ultra). Cut-off values²³ were: 0–0.25: little or none; 0.26–0.50: fair; 0.51–0.75: moderate to good; and > 0.75: good to excellent relationship. Intra-class correlation coefficients (ICC (2,1)) were calculated to investigate the reliability between the actual steps and those estimated by the five devices (Google Fit, Health, STEPZ, Pacer, and Fitbit Ultra). Bland and Altman plots were generated for steps count to assess the agreement between the tests 1 and 2. An ICC value < 0.4 was considered poor, 0.4 to 0.75 as fair to good, and > 0.75 as excellent.²⁴ The comparison between the device position was investigated using a t-test (95% confidence interval [95% CI]). The significance level was set at 5%. All analyses were performed using International Business Machines Statistical Package for the Social Sciences (IBM SPSS, IBM Corp, Armont, NY) (Version 19) by an independent researcher.

Results

Fifty-five individuals with chronic stroke participated (Figure 1); they were 54.5% men, with a mean age of 62.5 years and a mean time from stroke onset of 66.8 months. Forty-one percent of the participants were community walkers. The participants’ characteristics are summarized in Table 1.

Figure 1.

Flowchart of recruitment process.

Table 1.

Participants’ characteristics.

Characteristic	n = 55
Age (years), mean (SD)	62.5 (14.9)
Sex (men), n (%)	30 (54.5)
Chronicity of stroke (months), mean (SD)	66.8 (55.9)
Affected cerebral hemisphere (left), n (%)	29 (52.7)
Type of stroke (ischemic), n (%)	44 (80)
Physical activity before stroke (no), n (%)	28 (50.9)
Physical activity after stroke (yes), n (%)	32 (58.2)
Mini Mental State Examination (score), mean (SD)	24.9 (3.4)
Walking performance level (community ambulators, ⩾ 0.8 m/s), n (%)	23 (41.8)
Habitual speed (m/s), mean (SD)	0.7 (0.3)
Fast speed (m/s), mean (SD)	1.0 (0.5)
Strength of knee extensors on paretic side (mmHg), mean (SD)	172.3 (67.5)
Strength of knee extensors on non-paretic side (mmHg), mean (SD)	220.1 (53.3)
Strength of plantar flexors on paretic side (mmHg), mean (SD)	116.6 (50.5)
Strength of plantar flexors on non-paretic side (mmHg), mean (SD)	169.8 (44.5)
Falls Efficacy Scale—International (score), mean (SD)	29.2 (10.5)
Human Activity Profile (Adjusted Activity Score), mean (SD)	48.8 (20.0)
Duke Activity Status Index (score), mean (SD)	25.9 (13.4)

Validity

As Shapiro–Wilk test revealed a normally distributed data, the Pearson correlation was used for all analyses. It was found positive and statistically significant associations between the criterion-standard measure and the Google Fit, STEPZ iphone and Android applications, Pacer iphone and Android, and Fitbit Ultra. There was no association between the criterion-standard measure and Health iphone and Android applications (Table 2 and Figure 2).

Table 2.

Validity of the devices compared with the criterion-standard measure.

Device
Google Fit	r	0.66
Google Fit	P	< 0.01
Health (iphone)	r	0.18
Health (iphone)	P	0.21
Health (Android)	r	0.19
Health (Android)	P	0.19
STEPZ (iphone)	r	0.65
STEPZ (iphone)	P	< 0.01
STEPZ (Android)	r	0.30
STEPZ (Android)	P	< 0.05
Pacer (iphone)	r	0.80
Pacer (iphone)	P	< 0.01
Pacer (Android)	r	0.68
Pacer (Android)	P	< 0.01
Fitbit Ultra	r	0.67
Fitbit Ultra	P	< 0.01

Bold values represent statiscally significant.

Figure 2.

Scatter plots of the validity of the (a) Health (iphone) (r = 0.18) and (b) Pacer (iphone) (r = 0.80) with the criterion-standard measure (live actual steps).

Reliability

The ICC (2,1) analysis demonstrated no agreement between the Health (iphone) and criterion-standard measure, which showed excellent test–retest reliability (ICC = 0.99) (Table 3 and Figure 3). The other four devices/apps showed agreements ranging from poor (Health Android, STEPZ iphone, and STEPZ Android) to excellent (Google Fit, Pacer iphone, and Fitbit Ultra).

Table 3.

Reliability of the devices.

Device	Measure	Mean of steps (SD)	ICC	95% CI	Mean difference (95% CI)
Criterion-standard measure	Measure 1	201 (33)	0.99	0.98 to 0.99	−0.3 (−14 to 13)
Criterion-standard measure	Measure 2	201 (37)	0.99	0.98 to 0.99	−0.3 (−14 to 13)
Google Fit	Measure 1	175 (65)	0.76	0.58 to 0.86	−10 (−32 to 12)
Google Fit	Measure 2	185 (51)	0.76	0.58 to 0.86	−10 (−32 to 12)
Health (iphone)	Measure 1	140 (151)	−0.70	−2.02 to 0.03	60 (7 to 112)
Health (iphone)	Measure 2	80 (125)	−0.70	−2.02 to 0.03	60 (7 to 112)
Health (Android)	Measure 1	132 (275)	0.10	−0.58 to 0.49	−38 (−112 to 38)
Health (Android)	Measure 2	170 (64)	0.10	−0.58 to 0.49	−38 (−112 to 38)
STEPZ (iphone)	Measure 1	197 (47)	0.28	−0.26 to 0.59	−8 (−15 to 30)
STEPZ (iphone)	Measure 2	189 (69)	0.28	−0.26 to 0.59	−8 (−15 to 30)
STEPZ (Android)	Measure 1	183 (299)	0.20	−0.41 to 0.55	−11 (−80 to 100)
STEPZ (Android)	Measure 2	172 (158)	0.20	−0.41 to 0.55	−11 (−80 to 100)
Pacer (iphone)	Measure 1	189 (49)	0.80	0.64 to 0.88	9 (−25 to 8)
Pacer (iphone)	Measure 2	198 (37)	0.80	0.64 to 0.88	9 (−25 to 8)
Pacer (Android)	Measure 1	174 (98)	0.68	0.44 to 0.82	23 (−64 to 18)
Pacer (Android)	Measure 2	197 (120)	0.68	0.44 to 0.82	23 (−64 to 18)
Fitbit Ultra	Measure 1	180 (65)	0.79	0.63 to 0.88	1 (−22 to 21)
Fitbit Ultra	Measure 2	181 (50)	0.79	0.63 to 0.88	1 (−22 to 21)

ICC: intra-class correlation coefficients; 95% CI: 95% confidence interval; Measure 1: Fitbit and Android on paretic side; Measure 2: Fitbit and Android on non-paretic side.

Figure 3.

Bland and Altman plots. Individual difference in steps count between Measurement 1 and 2, plotted against individual mean (Measurement 1 and 2). Dashed line shows the group mean steps between Measurement 1 and 2. The 95% upper and lower limits of agreement represent 2 standard deviations above and below the group mean difference in steps between Measurement 1 and 2, for (a) criterion-standard measure, (b) Google Fit, (c) Health (iphone), (d) Health (Android), (e) STEPZ (iphone), (f) STEPZ (Android), (g) Pacer (iphone), (h) Pacer (Android), and (i) Fitbit Ultra.

Location on the body

No statistically significant differences were found for any of the devices in relation to location on the paretic or non-paretic side. There was considerable variance on 95% CI (Table 4).

Table 4.

Comparison of measurement of the devices in relation to positioning on the paretic and non-paretic sides.

Device	P value	95% CI
Google Fit	0.17	−24.7 to 4.6
Health (iphone)	0.06	−3.6 to 122.9
Health (Android)	0.34	−116.3 to 41.7
STEPZ (iphone)	0.50	−14.7 to 29.5
STEPZ (Android)	0.82	−81.2 to 102.1
Pacer (iphone)	0.09	−19.0 to 1.5
Pacer (Android)	0.14	−53.9 to 7.8
Fitbit Ultra	0.93	−14.4 to 13.4

95% CI: 95% confidence interval.

Discussion

Our main findings provide evidence that mHealth devices (Pacer–iphone, Fitbit Ultra, Google Fit, and Pacer–Android) that measure the number of steps over a given period of time are all valid and reliable in comparison with a criterion-standard measure, with the exception of the Health with iphone and Android. In addition, location on the body does not appear to affect the measurement when the smartphone is positioned on the paretic or non-paretic side at hip level. Therefore, the investigated mHealth devices could be used as a valid and reliable tool to promote an increase in the number of steps in chronic stroke survivors.

The Health application did not demonstrate adequate values for validity in individuals after chronic stroke in either operating system (iOS and Android OS). In contrast, a previous study with non-disabled individuals demonstrated validity for walking and running using smartphone application Health for Android OS.²¹ One explanation for this discrepancy may be associated with the relatively reduced walking speed of individuals after stroke, which is approximately half the rate of non-disabled individuals (1.4 m/s vs. 0.7 m/s). Indeed, faster speeds have been shown to have better accuracy using the Health application.²¹

Consistent with this idea, a previous study, which investigated the validity of Health and Pacer in both smartphone systems (iOS and Android OS) in non-disabled individuals with a Smartphone (iphone and Samsung Galaxy S6 Edge) positioned in different locations (pant pocket on dominant and non-dominant side, bag transverse in shoulder and backpack) found that the smartphone application was not valid for slow speeds in any position (0.4 m/s).²⁵ However, Health for Android iOS was validated for non-disabled individuals for fast speeds and running.²⁶ In addition, Health for iOS was not found to be valid for any speed, fast or slow (0.7 m/s, 1.4 m/s, 2.8 m/s) with iphone on the left hand in non-disabled individuals.²⁷ Together, the Health application is not recommended for step counting in individuals after stroke, especially those who have a low walking speed. Although STEPZ presented lower values of correlation and ICC than other devices, this device presented fair to good magnitude of correlation with criterion-standard measure (live actual steps). In addition, the mean of the steps detected by STEPZ was similar to criterion-standard measure. These measurements are similar to the other devices, which presented mean differences ⩽ 11 steps. On the other hand, Health presented mean differences of 38 and 60 steps between Measure 1 and Measure 2, respectively. Thus, although STEPZ provide data statistically supported of validity and reliability, if a clinician or a researcher wants to choose a device, it is recommended to use Pacer (iphone), Fitbit Ultra, Google Fit, and Pacer (Android).

Here, we found that the location of the device (paretic or non-paretic side, at hip level) did not influence the step count. The literature reports that the smartphone application placed at the ankle or hip of the non-paretic side during walking did not influence the step count in individuals after chronic stroke during the 6-minute walk test, ramps, or stairs.²⁸ When the smartphone application was positioned on the non-dominant hand and right pocket, there were no significant differences in step count, regardless of placement, in a study with non-disabled individuals.²¹ Thus, the individual should position the device on the most convenient side.

The results of this study have important clinical implications. It is well known that limiting sedentary time is associated with a lower risk of cardiovascular events and mortality.²⁹Free and easy-to-use tools/strategies to improve physical activity levels are necessary, since these devices are used on a daily-basis without the need for specialized training or calibration. The literature reports that mHealth technology can improve outcomes such as adherence to medical therapy, awareness of diet and exercise, and ability to reach exercise goals.³⁰ Therefore, the use of valid and reliable smartphone applications should be implemented during the rehabilitation epoch, in order to set goals for these individuals, follow their results, and promote healthy competition between participants to enhance motivation and adherence to treatment.

Here, the number of smartphone applications was small when compared with what is commercially available. Individuals were evaluated in a laboratory setting, on a flat and standardized surface, and only individuals who were in the chronic stage after stroke, and able to walk without assistive devices were included. Also, we investigated the measurement properties during short distances, what does not reflect long distances on the community. As such, the results of this study should not be generalized beyond those individuals with the same characteristics, or other population. Also, it is important to notice that the sample size was not large enough to extrapolate the data to the entire stroke population.

Our results indicate that mHealth devices (Pacer–iphone, Fitbit Ultra, Google Fit, and Pacer–Android) provide a valid and reliable measure of step count in chronic stroke survivors. Placement of the device on the body (paretic or non-paretic hip) does not appear to affect the measure of step count.

Clinical messages

Pacer (iphone), Fitbit Ultra, Google Fit, and Pacer (Android) are valid for step counting.

Pacer (iphone), Fitbit Ultra, Google Fit, and Pacer (Android) are reliable for step counting.

The location of the smartphone (paretic or non-paretic front pocket) does not influence the accuracy of the step count.

Supplemental Material

Supplemental_Material – Supplemental material for An investigation into the validity and reliability of mHealth devices for counting steps in chronic stroke survivors

Supplemental material, Supplemental_Material for An investigation into the validity and reliability of mHealth devices for counting steps in chronic stroke survivors by Pollyana Helena Vieira Costa, Thainá Paula Dias de Jesus, Carolee Winstein, Camila Torriani-Pasin and Janaine Cunha Polese in Clinical Rehabilitation

Footnotes

Author contributions

J.C.P. conceived the original idea. P.H.V.C. and T.P.D.J. carried out the recruitment process and the tests. C.T.P. and J.C.P. were involved in planning, writing, and supervising the work. C.J.W. provided editorial assistance and scientific oversight for the manuscript. All authors discussed the results and contributed to the final manuscript.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Support for this study was given by FAPEMIG (Fundação de Amparo a Pesquisa de Minas Gerais) and FAPESP (Fundação de Amparo à Pesquisa do Estado de São Paulo).

ORCID iD

Janaine Cunha Polese

Supplemental material

Supplemental material for this article is available online.

References

Billinger

Arena

Bernhardt

, et al. Physical activity and exercise recommendations for stroke survivors: a statement for healthcare professionals from the American Heart Association/American Stroke Association. Stroke 2014; 45(8): 2532–2553.

Kraus

Janz

Powell

, et al. Daily step counts for measuring physical activity exposure and its relation to health. Med Sci Sports Exerc 2019; 51(6): 1206–1212.

Garber

Blissmer

Deschenes

, et al. American College of Sports Medicine position stand. Quantity and quality of exercise for developing and maintaining cardiorespiratory, musculoskeletal, and neuromotor fitness in apparently healthy adults: guidance for prescribing exercise. Med Sci Sports Exerc 2011; 43(7): 1334–1359.

Field

Gebruers

Shanmuga

, et al. Physical activity after stroke: a systematic review and meta-analysis. ISRN Stroke 2013; 2013: 1–13.

Fini

Burge

Bernhardt

, et al. Two days of measurement provides reliable estimates of physical activity poststroke: an observational study. Arch Phys Med Rehabil 2019; 100(5): 883–890.

Burke

Azar

, et al. Current science on consumer use of mobile health for cardiovascular disease prevention: a scientific statement from the American Heart Association. Circulation 2015; 132(12): 1157–1213.

Chow

Thom

Wewege

, et al. Accuracy of step count measured by physical activity monitors: the effect of gait speed and anatomical placement site. Gait Posture 2017; 57: 199–203.

Danks

Roos

McCoy

, et al. A step activity monitoring program improves real world walking activity post stroke. Disabil Rehabil 2014; 36(26): 2233–2236.

Polese

E Faria

Ribeiro-Samora

, et al. Google fit smartphone application or Gt3X Actigraph: which is better for detecting the stepping activity of individuals with stroke? A validity study. J Bodyw Mov Ther 2019; 23(3): 461–465.

10.

Bernhardt

Hayward

Kwakkel

, et al. Agreed definitions and a shared vision for new standards in stroke recovery research: the Stroke Recovery and Rehabilitation Roundtable taskforce. Int J Stroke 2017; 12(5): 444–450.

11.

Bertolucci

PHF

Brucki

Campacci

, et al. O mini-exame do estado mental em uma população geral impacto da escolaridade. Arq Neuropsiquiatr 1994; 52(1): 1–7.

12.

Nagano

Hori

Muramatsu

. A comparison of at-home walking and 10-meter walking test parameters of individuals with post-stroke hemiparesis. J Phys Ther Sci 2015; 27(2): 357–359.

13.

Salbach

Mayo

Higgins

, et al. Responsiveness and predictability of gait speed and other disability measures in acute stroke. Arch Phys Med Rehabil 2001; 82(9): 1204–1212.

14.

Silva

Lacerda

CSA

Roza

, et al. Confiabilidade interexaminadores do teste e re-teste do esfigmomanômetro modificado em indivíduos saudáveis. Conexão Ciência 2017; 12(1): 38–45.

15.

Camargos

FFO

Dias

JMD

, et al. Adaptação transcultural e avaliação das propriedades psicométricas da Falls Efficacy Scale—International em idosos Brasileiros (FES-I-BRASIL). Rev Bras Fisioter 2010; 14(3): 237–243.

16.

Souza

Magalhães

Teixeira-Salmela

. Adaptação transcultural e análise das propriedades psicométricas da versão brasileira do Perfi l de Atividade Humana. Cad Saúde Pública 2006; 12(22): 2623–2636.

17.

Coutinho-Myrrha

Dias

Fernandes

, et al. Duke Activity Status Index for cardiovascular diseases: validation of the Portuguese translation. Arq Bras Cardiol 2014; 102(4): 383–390.

18.

Polese

Servio

Chaves

GSS

, et al. Relationships between self-reported and performance-based measures of functional capacity in individuals with chronic stroke. J Phys Ther Sci 2016; 28(4): 1208–1212.

19.

Fulk

Combs

Danks

, et al. Accuracy of 2 activity monitors in detecting steps in people with stroke and traumatic brain injury. Phys Ther 2014; 94(2): 222–229.

20.

Klassen

Semrau

Dukelow

, et al. Consumer-based physical activity monitor as a practical way to measure walking intensity during inpatient stroke rehabilitation. Stroke 2017; 48(9): 2614–2617.

21.

Johnson

Turek

Dornfeld

, et al. Validity of the Samsung Phone S Health application for assessing steps and energy expenditure during walking and running: does phone placement matter? Digit Health 2016; 2: 2055207616652747.

22.

Capela

Lemaire

Baddour

. Feature selection for wearable smartphone-based human activity recognition with able bodied, elderly, and stroke patients. PLoS ONE 2015; 10(4): e0124414.

23.

Portney

Watkins

. Foundations of clinical research. 3rd ed. Upper Saddle River, NJ: Prentice Hall, 2009.

24.

Shrout

Fleiss

. Intraclass correlations: uses in assessing rater reliability. Psychol Bull 1979; 86(2): 420–428.

25.

Hochsmann

Knaier

Eymann

, et al. Validity of activity trackers, smartphones, and phone applications to measure steps in various walking conditions. Scand J Med Sci Sports 2018; 28(7): 1818–1827.

26.

Beltran-Carrillo

Jimenez-Loaisa

Alarcon-Lopez

, et al. Validity of the “samsung health” application to measure steps: a study with two different samsung smartphones. J Sports Sci 2019; 37(7): 788–794.

27.

Duncan

Wunderlich

Zhao

, et al. Walk this way: validity evidence of iphone health application step count in laboratory and free-living conditions. J Sports Sci 2018; 36(15): 1695–1704.

28.

Duclos

Aguiar

Aissaoui

, et al. Activity monitor placed at the nonparetic ankle is accurate in measuring step counts during community walking in post-stroke individuals: a validation study. PM R 2019; 11(9): 963–971.

29.

Benjamin

Muntner

Alonso

, et al. Heart disease and stroke statistics—2019 update: a report from the American Heart Association. Circulation 2019; 139(10): e56–e528.

30.

Gandhi

Chen

Hong

, et al. Effect of mobile health interventions on the secondary prevention of cardiovascular disease: systematic review and meta-analysis. Can J Cardiol 2017; 33(2): 219–231.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.28 MB