Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

These Shoes Are Made for Walking: Sensitivity Performance Evaluation of Commercial Activity Monitors under the Expected Conditions and Circumstances Required to Achieve the International Daily Step Goal of 10,000 Steps

  • Sandra O’Connell,

    Affiliations Discipline of Physiology, School of Medicine, NUI Galway, Galway, Ireland, Electrical & Electronic Engineering, School of Engineering & Informatics, NUI Galway, Galway, Ireland

  • Gearóid ÓLaighin ,

    gearoid.olaighin@nuigalway.ie

    Affiliations Electrical & Electronic Engineering, School of Engineering & Informatics, NUI Galway, Galway, Ireland, National Centre for Biomedical Engineering Science, NUI Galway, Galway, Ireland, CÚRAM, Center for Medical Device Research, NUI Galway, Galway, Ireland

  • Lisa Kelly,

    Affiliation Discipline of Physiology, School of Medicine, NUI Galway, Galway, Ireland

  • Elaine Murphy,

    Affiliation Discipline of Physiology, School of Medicine, NUI Galway, Galway, Ireland

  • Sorcha Beirne,

    Affiliation Discipline of Physiology, School of Medicine, NUI Galway, Galway, Ireland

  • Niall Burke,

    Affiliation Discipline of Physiology, School of Medicine, NUI Galway, Galway, Ireland

  • Orlaith Kilgannon,

    Affiliation Discipline of Physiology, School of Medicine, NUI Galway, Galway, Ireland

  • Leo R. Quinlan

    Affiliations Discipline of Physiology, School of Medicine, NUI Galway, Galway, Ireland, CÚRAM, Center for Medical Device Research, NUI Galway, Galway, Ireland

These Shoes Are Made for Walking: Sensitivity Performance Evaluation of Commercial Activity Monitors under the Expected Conditions and Circumstances Required to Achieve the International Daily Step Goal of 10,000 Steps

  • Sandra O’Connell, 
  • Gearóid ÓLaighin, 
  • Lisa Kelly, 
  • Elaine Murphy, 
  • Sorcha Beirne, 
  • Niall Burke, 
  • Orlaith Kilgannon, 
  • Leo R. Quinlan
PLOS
x

Abstract

Introduction

Physical activity is a vitally important part of a healthy lifestyle, and is of major benefit to both physical and mental health. A daily step count of 10,000 steps is recommended globally to achieve an appropriate level of physical activity. Accurate quantification of physical activity during conditions reflecting those needed to achieve the recommended daily step count of 10,000 steps is essential. As such, we aimed to assess four commercial activity monitors for their sensitivity/accuracy in a prescribed walking route that reflects a range of surfaces that would typically be used to achieve the recommended daily step count, in two types of footwear expected to be used throughout the day when aiming to achieve the recommended daily step count, and in a timeframe required to do so.

Methods

Four commercial activity monitors were worn simultaneously by participants (n = 15) during a prescribed walking route reflective of surfaces typically encountered while achieving the daily recommended 10,000 steps. Activity monitors tested were the Garmin Vivofit , New Lifestyles’ NL-2000 pedometer, Withings Smart Activity Monitor Tracker (Pulse O2) , and Fitbit One .

Results

All activity monitors tested were accurate in their step detection over the variety of different surfaces tested (natural lawn grass, gravel, ceramic tile, tarmacadam/asphalt, linoleum), when wearing both running shoes and hard-soled dress shoes.

Conclusion

All activity monitors tested were accurate in their step detection sensitivity and are valid monitors for physical activity quantification over the variety of different surfaces tested, when wearing both running shoes and hard-soled dress shoes, and over a timeframe necessary for accumulating the recommended daily step count of 10,000 steps. However, it is important to consider the accuracy of activity monitors, particularly when physical activity in the form of stepping activities is prescribed as an intervention in the treatment or prevention of a disease state.

Introduction

Physical activity is an essential part of a healthy lifestyle, playing an important role in improving and maintaining both physical and mental health [14]. Indeed, lack of physical activity has been recognized as the fourth leading risk factor for global mortality, associated with 6% of deaths worldwide [5]. Since the benefits of physical activity have been recognized globally, governments internationally have highlighted to their citizens the need to be physically active on a regular basis. In 1996, the US Surgeon General recommended that as part of a healthy lifestyle, people of all ages should partake in at least 30 minutes of moderate-intensity physical activity, such as brisk walking, on a daily basis [6]. The 2008 ‘Physical Activity Guidelines for Americans’ from the Centers for Disease Control (CDC) recommended that adults need 2 hours and 30 minutes of moderate intensity aerobic physical activity, for example brisk walking, per week [7]. These two recommendations correspond closely with the 2010 WHO recommendation that adults should do at least 150 minutes of moderate intensity aerobic physical activity throughout the week [5].

An important consideration in assisting persons to change their physical activity behavior and to adhere to these public health recommendations is providing them with the ability to easily quantify the extent of physical activity completed in a given day. The development of commercial pedometers in the late 1960s provided a convenient, low cost method of quantifying one form of physical activity, namely walking. These devices provided straightforward feedback to the user of the quantity of physical activity completed, specifically their step count. A series of pedometer-based studies, from 2004 to 2011, evaluated the relationship between step count and adherence to physical activity guidelines and reported that less than 7,500 steps per day represented sedentary or “low-active” behavior [8, 9] and taking 10,000 steps per day was consistent with a physical activity level associated with person who is “active” [8, 10, 11]. Furthermore, a study carried out in an overweight population from the Lower Mississippi Delta of the U.S.A. reported a step count of 9,154 steps per day to equate to 30 minutes of moderate-to-vigorous physical activity. The authors concluded that, in this population, a step count of 8,300 to 9,100 steps a day should be accumulated in order to achieve recommended physical activity guidelines [12].

As part of their physical activity guidelines, the WHO states that physical activity in adults includes recreational or leisure-time physical activity, occupational (i.e. work), household chores, transportation (e.g. walking or cycling), play, games, planned exercise or sports in the context of daily, family, and community activities [5]. Thus the recommended “quantity” of physical activity of 10,000 steps being proposed internationally would only typically be achieved through a combination of these activities and under the varied circumstances associated with these activities. In other words, the recommended step count of 10,000 steps would typically be achieved, not only through a specific programme of exercise, but also through a combination of activities in the home, in the work place, and while getting to and from work. These activities would typically involve the person walking on a wide variety of surfaces from footpaths, indoor floors with varied surface types, and outdoor natural walking surfaces such as grass pathways. Indeed, one group have examined the effect of walking surface on step count sensitivity/accuracy of a pedometer and found surface type significantly affected step count [13]. Additionally, a number of physical activity device manufacturer websites state that surface type can affect step count sensitivity/accuracy of their devices [14, 15].

Furthermore, the range of activities conducted in achieving the recommended step count of 10,000 steps would also typically involve the person changing their footwear from more formal “dress” shoes in the workplace to more comfortable runner-type shoes in the home and while exercising. Therefore, a physical activity monitor being used to quantify adherence to the recommended daily step count must be able to deal with: (i) a variety of walking surfaces, (ii) different footwear, and (iii) extended periods of walking, where the accumulated number of steps is in the order of 10,000 steps.

These three variables have the potential to affect step count detection accuracy. Thus, it is the view of the authors that activity monitor testing must involve: (i) walking on different surfaces that reflect the range of surfaces that will typically be encountered to achieve the daily step count goal, (ii) walking with different footwear that reflects the different types of footwear expected to be used throughout the day, and (iii) walking for periods of time that properly reflect the time required to achieve the recommended daily step count. In fact, a review of the literature highlights that these considerations have not been widely adopted in a wide range of studies where the sensitivity of activity monitors has been assessed [1621]. Often, the types of surfaces and types of footwear tested are either not reported or not taken into consideration during testing. Furthermore, the timeframe of testing is often far less than would be required to achieve the recommended daily step count of 10,000 steps. For example, timeframes of between 11 and 25 minutes have previously been utilized and would be completely insufficient for accumulating 10,000 steps [1620].

In this paper we aimed to assess four commercial activity monitors: the Withings Smart Activity Monitor Tracker (Pulse O2) , the NL-2000 pedometer , the Garmin Vivofit , and the Fitbit One . We aimed to assess the step detection sensitivity of each activity monitor, i.e. the ability of the activity monitors to detect and count an actual step as a step, over a prescribed walking route that reflects a range of surfaces that would typically be used to achieve the recommended daily step count and in a timeframe required to do so. In addition, we also aimed to investigate two different types of footwear typically used throughout the day when aiming to achieve the recommended daily step count of 10,000 steps.

Methods

Participants

Fifteen healthy participants (8 female, 7 male) took part in this study, with a mean age of 21.1 ± 1.1 years. Males had a mean BMI of 23.60 ± 2.70kg/m2 while females had a mean BMI of 21.88 ± 1.81kg/m2. None of the participants had any history of cardiovascular disease or neurological disorders. Ethics committee approval was obtained from the Galway University Hospitals Research Ethics Committee, and all participants provided written, informed consent.

Study Protocol

All 15 participants completed a prescribed walking route. The activity monitors tested in this study were the Withings Smart Activity Monitor Tracker (Pulse O2) (Withings, Issy-les-Moulineaux, France), NL-2000 pedometer (New Lifestyles, Missouri, USA), Garmin Vivofit (Garmin, Kansas, USA), and Fitbit One (Fitbit, San Francisco, USA). (See Table 1). All activity monitors were put in place at the manufacturer’s recommended body location (Fig 1) by the investigators as per the manufacturers’ instructions. All four activity monitors were worn simultaneously on each participant for the duration of testing. The Garmin Vivofit was worn on the non-dominant wrist, with the NL-2000 and Withings Smart Activity Monitor Tracker (Pulse O2) worn on opposite sides of the waist. The Fitbit activity monitor was clipped onto clothing at the level of the chest. Participants were video recorded throughout the study with a hand-held camcorder. The true step count was extracted manually from the recorded video in real time and was then compared to the step count registered by each activity monitor. Additionally, the ActivPAL micro (PAL Technologies Ltd., Glasgow, UK) was worn by each participant as a reference device for measuring the overall total step count for the prescribed walking route. The ActivPAL was attached to the thigh using Tegaderm transparent dressing (3M Health Care, Minnesota, USA).

thumbnail
Fig 1. Position of the activity monitors.

The ActivPAL micro activity monitor on the thigh (right), the NL-2000 pedometer on the waist (right), the Garmin Vivofit on the wrist (left), the Withings Smart Activity Monitor Tracker (Pulse O2) on the waist (left), the Fitbit One activity monitor on the chest.

https://doi.org/10.1371/journal.pone.0154956.g001

Participants completed a prescribed flat walking route consisting of five different walking surfaces: linoleum (800m), natural lawn grass (900m), gravel (990m), ceramic tile (400m) and tarmacadam/asphalt (880m). In addition, the prescribed walking route included stair walking (49 steps up and 49 steps down, step height 16cm) and ramp walking (240m up and 240m down at an incline of 4.05%) (Fig 2).

thumbnail
Fig 2. Prescribed walking route.

Breakdown of surfaces and distances throughout the walking route.

https://doi.org/10.1371/journal.pone.0154956.g002

Each participant completed the walking route twice: the first time wearing standard running shoes, the second time wearing hard-soled dress shoes. At all times the participants were asked to walk at their normal walking pace, in other words, self-selected walking speed.

Inter-device Reliability

There were between three and five of each device type used during testing. Prior to testing, each unit of each device type was tested over a 400m flat walk of uniform ceramic tile surface while wearing hard-soled shoes. One investigator wore each unit of a device type simultaneously for the 400m walk. The step counts on each unit of that device type were then compared. There was found to be no difference in step counts between the Fitbit units. The largest difference between the units of the remaining activity monitors was 0.19% for the NL-2000 units, 1.55% for the ActivPAL units, and 13.67% for the Garmin units. For the Garmin units, any step counts outside of ±2 SD of the observed count were excluded from analysis.

Statistical Analysis

All statistical analyses were carried out using SPSS (SPSS for Mac, version 20, IBM Corporation). Sample size was chosen by selecting a type I error rate (α) of 5%, a power (1—β) of 0.80, a sampling ratio of 1 (κ = nA / nB), and a standard deviation of 250 steps (2.5% of the recommended daily step count of 10,000 steps). The Shapiro-Wilk test was used to analyze normality of data. Data were then fitted to a repeated measures model as this study followed a repeated measures design. Repeated-measures analysis of variance and post-hoc follow-up were used to detect differences in step count between the true step count extracted manually from the video recordings and the step count registered by each activity monitor and to detect step count detection errors associated with each surface type during the walking route. The Greenhouse-Geisser correction was used to correct any violations of the assumption of sphericity. The mean absolute percentage error (MAPE) was calculated for each activity monitor using Eq 1 with N the number of steps as extracted from video analysis and Ñ the number of steps as recorded from each activity monitor.

(1)

There is no standard for acceptable error in step detection in activity monitors, but we have chosen to select a 5% error zone as being acceptable. Bland-Altman plots were plotted to examine the agreement in step count between the observed step count and the step count recorded on the individual activity monitors. An unpaired t test was carried out to assess any differences in step count between participants classed as ‘fast’ and ‘medium to slow’ speed walkers.

Results

Step counts obtained from video analysis are referred to as the observed step count and were compared to output counts from each activity monitor. Surface type did not affect step count detection on the four activity monitors tested. The mean step counts for the four activity monitors tested over each section of the walking route are given in Table 2. There was no statistically significant difference in activity monitor detected step count vs. the observed step count for the Withings Smart Activity Monitor Tracker (Pulse O2) , NL-2000 pedometer , Garmin Vivofit , or Fitbit One activity monitors on the linoleum (P = 0.106), gravel (P = 0.131), natural lawn grass (P = 0.195), tarmacadam/asphalt (P = 0.286), or ceramic tile (P = 0.457) surfaces. There was also no statistically significant difference in activity monitor detected step count versus the observed count for the Withings Smart Activity Monitor Tracker (Pulse O2) , NL-2000 pedometer , Garmin Vivofit , and Fitbit One activity monitors on the ramp section of the walking route (P = 0.591). Stairs walking did not affect step count, with no statistically significant difference in activity monitor detected step count vs. the observed step count for any activity monitor (P>0.05 for all). Over the entire walking route, independent of surface type, there was no statistically significant difference in step count between the observed count and the four activity monitors tested (P > 0.05; Fig 3). There was also no statistically significant difference in step count between the step count on the ActivPAL reference device and the Withings Smart Activity Monitor Tracker (Pulse O2) , NL-2000 pedometer , Garmin Vivofit , or Fitbit One activity monitors (P > 0.05 for all). There was no statistically significant difference in step count when the devices were compared to each other on any surface type or overall, independent of surface type (P > 0.05 for all). Mean absolute percentage errors for each of the four activity monitors are given in Table 3. Over the prescribed walking route, eight mean absolute percentage errors were found to be outside of our selected 5% error zone. The Withings Smart Activity Monitor Tracker (Pulse O2) had a step count error for stair walking only, with errors on descending stairs of 6.64% and on ascending stairs of 5.93%. The Garmin had step count errors on the ceramic tile of 5.43%, on ascending stairs of 11.33% and on descending stairs of 6.48%. The NL-2000 pedometer step count error on ceramic tile was 7.66% and on ascending stairs of 8.65%. The Fitbit had step count error on descending stairs of 6.03%.

thumbnail
Fig 3. Overall walking route step count.

Independent of surface type, there was no statistically significant difference in step count versus the observed step count or versus the ActivPAL activity monitor. All data represent mean ± SEM.

https://doi.org/10.1371/journal.pone.0154956.g003

thumbnail
Table 2. Mean step counts observed and registered on the four activity monitors tested over the different surface types of the walking route.

https://doi.org/10.1371/journal.pone.0154956.t002

thumbnail
Table 3. Mean absolute percentage error (MAPE) and error range for the four activity monitors tested irrespective of footwear type.

https://doi.org/10.1371/journal.pone.0154956.t003

All other mean absolute percentage errors remained within the 5% error zone. The absolute percentage error range is also is in Table 3. Although there was no significant difference in the step counts observed versus the step counts detected by the activity monitors, these values show the variation in the percentage error for the activity monitors (Table 3). For example, when walking down stairs, the Fitbit One activity monitor had a mean absolute percentage error of 6.03% and a mean absolute percentage error range of -16.36% to 99.16%.

When all values lying ±2SD from the mean observed step count were removed, only two mean absolute percentage errors were found to be outside the 5% error zone (Table 4). These were both for the Garmin device, with error on ascending stairs of 6.59% and on descending stairs of 5.03%.

thumbnail
Table 4. Mean absolute percentage error (MAPE) and error range for the four activity monitors tested: all values ± 2SD from the mean observed step count excluded.

https://doi.org/10.1371/journal.pone.0154956.t004

Bland-Altman plots for the number of steps showed average ± limits of agreement underestimation of 84±238, 44±735, 75±235, 35±232, and 254±488 for the Withings , Garmin , Fitbit , ActivPAL , and NL-2000 activity monitors respectively (Fig 4).

thumbnail
Fig 4. Bland Altman plots for step counts of the A. Withings , B. Garmin , C. Fitbit , D. ActivPAL , and E. NL-2000 activity monitors.

The solid line represents the mean step count difference between the step count observed from the video recordings and each activity monitor for the overall step count. The dashed lines indicate mean ± limits of agreement (1.96*SD).

https://doi.org/10.1371/journal.pone.0154956.g004

Footwear type did not affect step detection of any activity monitor; there was no statistically significant difference in step count between the observed count and any activity monitor (P>0.05 for all).

Participants took a median time of 4 hours and 4 minutes to complete the entire walking route. Seven participants were classed as being walkers of a ‘fast’ speed as they completed the walking route in a time of between 2 hours and 6 minutes and 3 hours and 2 minutes, a gap of 1 hour and 2 minutes from the median walking time. Eight participants were classed as being walkers of a ‘medium to slow’ speed as they completed the walking route in a time between the median time (4 hours and 4 minutes) and 4 hours and 38 minutes. There was a statistically significant difference in step count between the ‘medium to slow’ speed walkers and ‘fast’ speed walkers, with the ‘medium to slow’ speed walkers taking a greater number of steps (P = 0.002). Additionally, the mean walking speed for the ‘fast’ walkers was 0.84m/s compared to 0.74m/s for the ‘medium to slow’ speed walkers, a statistically significant difference (P = 0.031).

Discussion

This study assessed four commercial activity monitors: the Withings Smart Activity Monitor Tracker (Pulse O2) , NL-2000 pedometer , Garmin Vivofit , and Fitbit One for their step count detection accuracy over a prescribed walking route that reflected a range of surfaces that would typically be encountered to achieve the recommended daily step count and in a timeframe required to do so. In addition, we investigated two different types of footwear typically used throughout the day when aiming to achieve the recommended daily step count of 10,000 steps.

As stated by the WHO, physical activity includes a variety of activities, such as recreational or leisure-time physical activity, occupational (i.e. work), household chores, and transportation (e.g. walking or cycling) [5]. As such, the recommended daily step count of 10,000 steps would typically be achieved over a variety of different surfaces. Taking this into consideration, we tested a prescribed walking route consisting of various surfaces that would be encountered on a daily basis, namely linoleum, natural lawn grass, gravel, tarmacadam/asphalt, and ceramic tile surfaces. All activity monitors were found to be accurate in step detection, irrespective of surface type. This is an important finding as step count is one of the main parameters utilized in physical activity quantification. It is vital that persons changing their behavior to adhere to public health physical activity recommendations can be confident in the sensitivity of the activity monitors they are employing to achieve their physical activity goals and reach the recommended daily step count of 10,000 steps. Additionally, our results correspond with those obtained by Brown et al. when they investigated the accuracy of the ActiPed activity monitor over two different surface types, namely grass and concrete. Participants walked 1,010m on grass and 1,070m on concrete, with the authors finding the ActiPed to be accurate in step count detection regardless of surface type [36].

When considering step counts, as opposed to surface type, participants in our study exceeded the recommended daily step count of 10,000 steps, taking an average of 10,950 steps to complete the prescribed walking route. Over the course of the entire walking route, independent of surface type, all activity monitors were found to be sensitive in their step detection accuracy, on both flat surfaces and stairs walking. Indeed, it is important to consider the effect of stairs walking, which is an activity of daily living and part of the normal daily step count goal for many adults. We observed similar results to those obtained by Storm et al., who investigated the sensitivity of both the ActivPAL and Movemonitor activity monitors during stairs walking at a self-selected walking speed, finding both activity monitors to be accurate in step detection [19]. Accuracy is an important factor to consider when the quantification of physical activity is paramount, such as when physical activity may be prescribed as an intervention in the treatment or prevention of a disease state. Sensitivity in this case is the probability that a step will be detected when a step is actually performed and a measure of all true positives (TP; i.e. a step was performed and it was detected) divided by the true positive (TP) plus the false negatives (FN; i.e. a step was performed but not detected). Due to the nature of the data output from the activity monitors, calculating FN is not possible thus we presented the MAPE as a valid measure of accuracy.

When we investigated the effect of two different types of footwear: running shoes and hard-soled dress shoes, on the sensitivity of the activity monitors, we found no effect on step detection sensitivity. To the best of our knowledge, this is the first study to assess the effects of footwear on the step detection sensitivity of activity monitors and is an important and positive finding, as activities carried out to achieve the recommended step count of 10,000 will most likely be done in various different types of footwear. However, we only investigated two different types of footwear. It would also be of interest to investigate other types of footwear that would often be worn on a daily basis when accumulating the recommended daily step count of 10,000 steps. Examples include slippers, flip-flops, high-heeled shoes, and hiking boots. These different types of footwear differ in terms of their sole type, degree of foot contact with the ground/floor, and the way in which they are worn, i.e. loose fitting or tight fitting on the foot, all of which may play a role in the step detection sensitivity of activity monitors.

The authors do recognize that limitations exist with our study, namely we have tested the accuracy of four commercial activity monitors but have not evaluated their specificity. For example, it is also of vital importance to validate activity monitors during a variety of activities of daily living that do not involve direct stepping, such as driving, cycling, and swimming. Furthermore, acceleration signals recorded by activity monitors can differ depending on the location of attachment [37]. Thus, the attachment location for each activity monitor is important to consider. In this study, we assessed each activity monitor in only one location. However, we specifically chose attachment locations for each activity monitor that was specified by the manufacturers.

Additionally, it is important to consider walking speed when evaluating the sensitivity of activity monitors. Slow walking speeds, for example, have been shown to reduce the sensitivity of some activity monitors [18, 27, 3840]. In this study, we evaluated different distributions of walking speeds within the normal walking self-selected walking speeds and found a greater speed and lesser number of steps in the ‘fast’ speed walkers. However, evaluation on a range of specific walking speeds, such as on a treadmill, would be of benefit to further elucidate the effect of walking speed on activity monitor step detection sensitivity.

Conclusions

Our study provides a comprehensive evaluation of four commercial activity monitors when (i) walking on different surfaces that reflect the range of surfaces that will typically be used to achieve the daily step count goal, (ii) walking with different footwear that reflects the different types of footwear expected to be used throughout the day, and (iii) walking for periods of time that properly reflect the time required to achieve the recommended daily step count. All activity monitors tested were accurate in their step detection sensitivity and are valid monitors for physical activity quantification over the variety of different flat surfaces tested and when stairs and ramp walking, when wearing both running shoes and hard-soled dress shoes, and over a timeframe necessary for accumulating the recommended daily step count of 10,000 steps. It is important to consider accuracy for activity monitors, particularly when physical activity in the form of stepping activities is prescribed as an intervention in the treatment or prevention of a disease state. All aspects of daily physical activity that go towards achieving the daily recommended step count of 10,000 steps are necessary to consider in accurate physical activity quantification.

Author Contributions

Conceived and designed the experiments: LRQ SOC GOL. Performed the experiments: LK EM NB SB OK SOC. Analyzed the data: SOC LRQ. Contributed reagents/materials/analysis tools: LRQ. Wrote the paper: SOC LRQ GOL.

References

  1. 1. Karmisholt K, Gotzsche PC. Physical activity for secondary prevention of disease. Systematic reviews of randomised clinical trials. Dan Med Bull. 2005;52(2):90–4. pmid:16009053
  2. 2. Nunan D, Mahtani KR, Roberts N, Heneghan C. Physical activity for the prevention and treatment of major chronic disease: an overview of systematic reviews. Syst Rev. 2013;2:56. pmid:23837523
  3. 3. Penedo FJ, Dahn JR. Exercise and well-being: a review of mental and physical health benefits associated with physical activity. Curr Opin Psychiatry. 2005;18(2):189–93. pmid:16639173
  4. 4. Young DR, Reynolds K, Sidell M, Brar S, Ghai NR, Sternfeld B, et al. Effects of physical activity and sedentary time on the risk of heart failure. Circ Heart Fail. 2014;7(1):21–7. pmid:24449810
  5. 5. World Health Organization. Global Recommendations on Physical Activity for Health. 2010. Available from: http://whqlibdoc.who.int/publications/2010/9789241599979_eng.pdf?ua=1
  6. 6. U.S. Department of Health and Human Services. Physical activity and health: a report of the Surgeon General. Atlanta, GA: U.S. Department of Health and Human Services, Centers for Disease Control and Prevention, National Center for Chronic Disease Prevention and Health Promotion, 1996. Available from: http://www.cdc.gov/nccdphp/sgr/chapcon.htm
  7. 7. Centers for Disease Control. "2008 Physical Activity Guidelines for Americans." U.S. Department of Health and Human Services. Available from: http://www.health.gov/paguidelines/guidelines/summary.aspx
  8. 8. Tudor-Locke C, Bassett DR Jr. How many steps/day are enough? Preliminary pedometer indices for public health. Sports Med. 2004;34(1):1–8. pmid:14715035
  9. 9. Tudor-Locke C, Leonardi C, Johnson WD, Katzmarzyk PT, Church TS. Accelerometer steps/day translation of moderate-to-vigorous activity. Prev Med. 2011;53(1–2):31–3. pmid:21295063
  10. 10. Kang M, Marshall SJ, Barreira TV, Lee JO. Effect of pedometer-based physical activity interventions: a meta-analysis. Res Q Exerc Sport. 2009;80(3):648–55. pmid:19791652
  11. 11. Tudor-Locke C, Hatano Y, Pangrazi RP, Kang M. Revisiting "how many steps are enough?". Med Sci Sports Exerc. 2008;40(7 Suppl):S537–43. pmid:18562971
  12. 12. Harrington DM, Tudor-Locke C, Champagne CM, Broyles ST, Harsha DW, Kennedy BM, et al. Step-based translation of physical activity guidelines in the Lower Mississippi Delta. Appl Physiol Nutr Metab. 2011;36(4):583–5. pmid:21854159
  13. 13. Leicht AS, Crowther RG. Pedometer accuracy during walking over different surfaces. Med Sci Sports Exerc. 2007;39(10):1847–50. pmid:17909414
  14. 14. Fitbit [Internet]. "How does my tracker count steps?" [Last updated 2016 April 15; cited 2016 April 26]; Available from: https://help.fitbit.com/articles/en_US/Help_article/How-does-my-tracker-count-steps
  15. 15. New Lifestyles [Internet]. "NL-2000 pedometer User's Guide. Available from: http://www.thepedometercompany.com/instructions.html
  16. 16. Dahlgren G, Carlsson D, Moorhead A, Hager-Ross C, McDonough SM. Test-retest reliability of step counts with the ActivPAL device in common daily activities. Gait Posture. 2010;32(3):386–90. pmid:20655228
  17. 17. Dondzila CJ, Swartz AM, Miller NE, Lenz EK, Strath SJ. Accuracy of uploadable pedometers in laboratory, overground, and free-living conditions in young and older adults. Int J Behav Nutr Phys Act. 2012;9:143. pmid:23232036
  18. 18. Giannakidou DM, Kambas A, Ageloussis N, Fatouros I, Christoforidis C, Venetsanou F, et al. The validity of two Omron pedometers during treadmill walking is speed dependent. Eur J Appl Physiol. 2012;112(1):49–57. pmid:21479653
  19. 19. Storm FA, Heller BW, Mazza C. Step detection and activity recognition accuracy of seven physical activity monitors. PLoS One. 2015;10(3):e0118723. pmid:25789630
  20. 20. Takacs J, Pollock CL, Guenther JR, Bahar M, Napier C, Hunt MA. Validation of the Fitbit One activity monitor device during treadmill walking. J Sci Med Sport. 2014;17(5):496–500. pmid:24268570
  21. 21. Taraldsen K, Askim T, Sletvold O, Einarsen EK, Bjastad KG, Indredavik B, et al. Evaluation of a body-worn sensor system to measure physical activity in older people with impaired function. Phys Ther. 2011;91(2):277–85. pmid:21212377
  22. 22. Ryan CG, Grant PM, Tigbe WW, Granat MH. The validity and reliability of a novel activity monitor as a measure of walking. Br J Sports Med. 2006;40(9):779–84. pmid:16825270
  23. 23. Godfrey A, Culhane KM, Lyons GM. Comparison of the performance of the activPAL Professional physical activity logger to a discrete accelerometer-based activity monitor. Med Eng Phys. 2007;29(8):930–4. pmid:17134934
  24. 24. Maddocks M, Petrou A, Skipper L, Wilcock A. Validity of three accelerometers during treadmill walking and motor vehicle travel. Br J Sports Med. 2010;44(8):606–8. pmid:18701531
  25. 25. Grant PM, Dall PM, Mitchell SL, Granat MH. Activity-monitor accuracy in measuring step number and cadence in community-dwelling older adults. J Aging Phys Act. 2008;16(2):201–14. pmid:18483442
  26. 26. Busse ME, van Deursen RW, Wiles CM. Real-life step and activity measurement: reliability and validity. J Med Eng Technol. 2009;33(1):33–41. pmid:19116852
  27. 27. Harrington DM, Welk GJ, Donnelly AE. Validation of MET estimates and step measurement using the ActivPAL physical activity logger. J Sports Sci. 2011;29(6):627–33. pmid:21360402
  28. 28. Ferguson T, Rowlands AV, Olds T, Maher C. The validity of consumer-level, activity monitors in healthy adults worn in free-living conditions: a cross-sectional study. Int J Behav Nutr Phys Act. 2015;12:42. pmid:25890168
  29. 29. Klassen TD, Simpson LA, Lim SB, Louie DR, Parappilly B, Sakakibara BM, et al. "Stepping Up" Activity Poststroke: Ankle-Positioned Accelerometer Can Accurately Record Steps During Slow Walking. Phys Ther. 2015.
  30. 30. Simpson LA, Eng JJ, Klassen TD, Lim SB, Louie DR, Parappilly B, et al. Capturing step counts at slow walking speeds in older adults: Comparison of ankle and waist placement of measuring device. J Rehabil Med. 2015.
  31. 31. Diaz KM, Krupka DJ, Chang MJ, Peacock J, Ma Y, Goldsmith J, et al. Fitbit(R): An accurate and reliable device for wireless physical activity tracking. Int J Cardiol. 2015;185:138–40. pmid:25795203
  32. 32. Lee JM, Kim Y, Welk GJ. Validity of consumer-based physical activity monitors. Med Sci Sports Exerc. 2014;46(9):1840–8. pmid:24777201
  33. 33. Crouter SE, Schneider PL, Karabulut M, Bassett DR Jr. Validity of 10 electronic pedometers for measuring steps, distance, and energy cost. Med Sci Sports Exerc. 2003;35(8):1455–60. pmid:12900704
  34. 34. Schneider PL, Crouter SE, Lukajic O, Bassett DR Jr. Accuracy and reliability of 10 pedometers for measuring steps over a 400-m walk. Med Sci Sports Exerc. 2003;35(10):1779–84. pmid:14523320
  35. 35. Schneider PL, Crouter S, Bassett DR. Pedometer measures of free-living physical activity: comparison of 13 models. Med Sci Sports Exerc. 2004;36(2):331–5. pmid:14767259
  36. 36. Brown DK, Grimwade D, Martinez-Bussion D, Taylor MJ, Gladwell VF. The validity of the ActiPed for physical activity monitoring. Int J Sports Med. 2013;34(5):431–7. pmid:23184480
  37. 37. Murphy SL. Review of physical activity measurement using accelerometers in older adults: considerations for research design and conduct. Prev Med. 2009;48(2):108–14. pmid:19111780
  38. 38. Cyarto EV, Myers A, Tudor-Locke C. Pedometer accuracy in nursing home and community-dwelling older adults. Med Sci Sports Exerc. 2004;36(2):205–9. pmid:14767241
  39. 39. Johnson M, Meltz K, Hart K, Schmudlach M, Clarkson L, Borman K. Validity of the Actical activity monitor for assessing steps and energy expenditure during walking. J Sports Sci. 2015;33(8):769–76. pmid:25356920
  40. 40. Stansfield B, Hajarnis M, Sudarshan R. Characteristics of very slow stepping in healthy adults and validity of the activPAL3 activity monitor in detecting these steps. Med Eng Phys. 2015;37(1):42–7. pmid:25455167