The accurate measurement of physical activity is important to researchers seeking to quantify dose-response relationships between physical activity and its outcomes; to determine the merits of light, moderate, and vigorous physical activity; and to record the physical activity performed in intervention studies (1). One objective measure of physical activity is the number of steps an individual accumulates per day, which can be measured by a pedometer (3,23,24).
In general, there are two different categories of pedometers: spring-levered and piezoelectric (7). Traditional spring-levered pedometers rely on a horizontal spring-suspended pendulum arm that moves up and down with each step, which opens and closes an electrical circuit, whereas the piezoelectric pedometer measures acceleration at frequent time intervals and the number of peaks or zero crossings is used to count steps. Both types of pedometers are typically worn on the waist at the midline of the thigh. However, some pedometers now have the ability to be worn in the pocket (chest pocket of a shirt or front pocket of pants); placed in a bag; hung around the neck, around the ankle, or wrist; or even attached to the shoe. Because of the variability of pedometer position, it is important to investigate the validity and reliability of pedometers, which could be worn at different locations on the body.
Although it is important to determine the accuracy of pedometers in controlled, laboratory settings, it is even more important to determine the accuracy of pedometers under free-living conditions, where people actually use the devices. Pedometer accuracy can be determined by counting steps in controlled laboratory settings (2,7-14,26), but it is not feasible to assess pedometer accuracy in this manner during a 24-h period (19). Therefore, the use of a criterion device is necessary. The StepWatch-3 (SW) (Cymatech, Inc., Seattle, WA) is a valid and accurate device for measuring steps in laboratory settings (4,9,13,21,22) and can be used as a criterion device to validate waist-mounted pedometers during a 24-h period (13).
Omron Healthcare, Inc. (Bannockburn, IL) recently introduced a new pedometer (model HJ-720ITC). According to the manufacturer, it can be worn in a pocket, on a belt, or around the neck and still validly record steps (11,12,27). It has been validated for treadmill and overground walking at various speeds (12,27); however, the Omron HJ-720ITC has not been tested under free-living conditions for 24 h.
Thus, the purpose of this study was twofold: 1) to examine the validity and reliability of the Omron HJ-720ITC piezoelectric pedometer in multiple wear locations (on the waist, in a pants pocket, or attached to a string around the neck) and 2) to compare it with two other pedometers (SW and Yamax SW-200) in a free-living environment during a 24-h period.
A total of 62 volunteers (31 males, 31 females), age 18-69 yr, were recruited from the University of Tennessee and the surrounding community. Over the phone, each participant was asked a series of questions to determine eligibility and to screen out individuals with abnormal gait patterns (Do you use an assistive walking device? Are you pregnant?). Individuals who had an internal defibrillator or a pacemaker were also excluded from the study. Participants were required to read and sign an informed consent form before participating. The University of Tennessee Institutional Review Board approved the study protocol.
After each participant underwent screening, an initial visit was scheduled in the Applied Physiology Laboratory. Participants removed their shoes and socks, and height was measured to the nearest 0.1 cm using a stadiometer (Seca Corporation, Columbia, MD). Weight was assessed in light clothing on a physician's scale (Health-O-Meter, Inc., Bridgeview, IL) to the nearest 0.25 lb. Waist and hip circumference measurements were taken using a Gulick tape measure with a tension handle. Waist circumference was measured at the narrowest portion of the torso between the iliac crest and inferior rib. Hip measurements were taken at the maximal circumference of the buttocks, above the gluteal fold (6). Waist-to-hip ratio was determined by dividing the waist circumference (cm) by the hip circumference (cm). Body fat percentage was assessed using the Tanita body fat analyzer (model BC-418; Tanita Corp., Tokyo, Japan). Body composition estimated by bioelectrical impedance analysis is less accurate in overweight and obese subjects compared with normal-weight subjects (17); therefore, several measures of adiposity were used as verification to confirm that we had three groups of individuals (normal body mass index (BMI) 19-24.9 kg·m−2), overweight (BMI 25-29.9 kg·m−2), and obese (BMI > 30 kg·m−2)) with differing levels of adiposity.
Each participant's step length was determined by having the participant walk exactly 20 steps down an indoor hallway. The total distance was measured in feet and divided by 20. Step length was determined from the average of two measurements.
The Omron HJ-720ITC is a new pedometer that can be worn in a pocket, on a belt, or around the neck (11,12,27). This particular pedometer is inexpensive (∼$35) and has the capability to store data in 1-h epochs over 42 d.
The Yamax Digiwalker SW-200 (∼$20) is the most widely used pedometer in research studies (19,25). This pedometer is worn on the belt or waistband, and it uses a horizontal lever arm that is suspended by a coil spring, which bobs up-and-down when walking. This movement causes an electronic circuit to open and close, and steps are accumulated using a simple circuit board (7). The SW-200 model has a reset button to set the steps to "0"; it has no memory and no other functions.
The SW is an expensive research grade device ($525 + $1470 for docking station and software) that is worn at the ankle. Unlike many other devices, the accuracy of the SW is not affected by walking speed or BMI (13,21). To detect steps, the SW uses an accelerometer, which measures directional (horizontal and vertical) acceleration. It can be set to record data in epochs ranging from 3 to 255 s. The SW can collect 60 d worth of data when a 3-s epoch is used. The device does not have a digital display, so it must be downloaded to a computer to obtain the recorded information (13).
Participants were shown the proper placement of the pedometers. An instruction sheet with a labeled picture was also given to the participants to assist with the correct placement. Omron HJ-720ITC pedometers were placed on the waist in the midline of the right thigh, the right pants pocket, and around the neck in the center of the chest. Time of day, step length, and body weight were entered into to the Omron pedometer. In laboratory-based studies, the Omron pedometer has been found to be a valid and reliable pedometer for recording step counts at walking speeds ranging from 2.0 to 4.0 mph across multiple positions and in lean, overweight, and obese individuals (11,12,27). The Yamax SW-200 was placed on the waist, in the midline of the left thigh. The SW was set to each participant's height. As in previous studies that used individuals who tend to walk at normal speeds, use a moderate range of walking speeds, and have a leg motion during walking that would be described as normal in appearance, the default settings (normal) for "walking speed," "range of speeds," and "leg motion" were used (13). This standard mode is what the manufacturer recommends for most individuals. Participants were asked to wear the pedometers for the rest of the day, until they went to bed, so that they would become accustomed to the devices.
24-h study protocol.
On the second day, the participants were instructed to wear the pedometers from the time they woke up until they went to bed at night. They were encouraged to participate in normal daily activities but to avoid nonambulatory activities during the testing period such as using an elliptical machine, lifting weights, or using a rowing machine. The Omron HJ-720ITC and SW automatically store the number of steps in the instrument's memory and are reset to zero at midnight. Because the Yamax SW-200 has no internal memory, participants were asked to record the amount of steps taken on a log sheet when they removed the instruments before they retired.
The SPSS version 17.0 for Windows (SPSS, Inc., Chicago, IL) was used for statistical analysis. The percentage of SW-determined steps recorded by each pedometer model/placement site was determined. Step data were converted to "percentage of SW-determined steps" by dividing the pedometer values by the SW-determined value and multiplying by 100% (i.e., [(Omron Belt steps / SW steps) × 100%]). Several studies report the absolute percent error between actual steps and pedometer-determined steps as evidence for pedometer accuracy because it is theoretically possible that underestimations and overestimations could cancel each other out and still yield 100% of actual steps. However, in our study, we report the mean "percent of SW-determined steps" because all pedometers always underestimated the number of steps in comparison to the SW. We wanted to focus on the percentage of steps that were accurately detected by each of the pedometers. To guard against the possibility of concluding that a pedometer was valid when it was not, we also reported a measure of the variability in percent of SW-determined steps (SD). Two-way repeated-measures ANOVA (3 × 4) tested the interaction effect of BMI category (normal, overweight, obese) and pedometer model/placement site (Omron Belt, Omron Neck, Omron Pocket, Yamax Belt) on the percentage of SW-determined steps recorded by each device. In the case of significant main effects, we conducted post hoc pairwise comparisons using Bonferroni adjustments to locate specific differences between the devices within each BMI category. One-way ANOVA was used to determine whether there were differences in the percentage of SW-determined steps among the BMI categories, within each device. Where appropriate, Tukey post hoc analyses were performed to determine which BMI categories were different within each device. One-sample t-tests were used to identify pedometer model/placement sites that overestimated or underestimated steps, relative to the SW-determined steps. Bland-Altman plots were constructed to show the level of agreement between the pedometers and the criterion measure (5). In this manner, the mean error score for the pedometer can be illustrated and the 95% prediction interval (i.e., 95% confidence interval for the individual observations) can also be shown. Individual error scores that have a tight prediction interval around zero signify a greater level of agreement. Data points below zero signify overestimation, whereas data points above zero signify underestimation. Intraclass correlation coefficients were used to measure reliability of the step counts across the three attached Omron pedometers. For all statistical analyses, an α level of 0.05 was used to show significant differences, and all values are shown as mean ± SD.
Descriptive characteristics are shown in Table 1. Mean values for BMI were 22.2, 27.0, and 36.5 kg·m−2 for normal, overweight, and obese participants. There were significant differences among groups for all indicators of adiposity. The pedometer-determined step counts and percentage of SW-determined steps counts for each pedometer model/placement site by BMI category are listed in Figure 1 and Table 2, respectively.
There was a significant interaction between BMI category and pedometer model/placement site; therefore, each model/placement site was looked at separately to see how they differed by BMI. The Yamax pedometer demonstrated significantly less error than the Omron pedometers in normal and overweight individuals (P < 0.05). In obese individuals, however, the pocket location (Omron Pocket) demonstrated significantly less error than the other two locations (Omron Belt, Omron Neck). In obese individuals, the Yamax, increased its inaccuracy to the degree that it not differ from the Omron Pocket, Omron Belt, or Omron Neck. BMI had no effect on the Omron Pocket, but it did affect the other pedometer model/placement sites (Omron Belt, Omron Neck, Yamax). The Omron Belt, Omron Neck, and Yamax pedometers were less accurate in the obese individuals. Independent of BMI, the Yamax and the Omron HJ-720ITC significantly underestimated steps in all three locations for all BMI categories in comparison to the SW (P < 0.001; Table 2).
Figure 2 shows the Bland-Altman plots for each pedometer. The y axes are standardized for easier comparison and to highlight the differences in accuracy between pedometers. For the Omron Pocket, the Bland-Altman plot indicated a mean difference from the criterion of 3171 ± 4276 steps per day and limits of agreement ranging from −1105 to 7447 steps per day (Fig. 2A). For the Omron Belt, the Bland-Altman plot indicated a mean difference from the criterion of 3969 ± 4385 steps per day, with limits of agreement between the two ranging from −416 to 8354 steps per day (Fig. 2B). For the Omron Neck, the Bland-Altman plot indicated a mean difference from the criterion of 4289 ± 5395 steps per day, with limits of agreement ranging from −1106 to 9685 steps per day (Fig. 2C). The Yamax had a mean difference from the criterion of 2671 ± 3616 steps per day, with limits of agreement ranging from −945 to 6287 steps per day (Fig. 2D). The intraclass correlation coefficient across the Omron pedometers was 0.94.
To our knowledge, this is the first study to examine the Omron HJ-720ITC under free-living conditions during a 24-h period. Previous research has shown that the Omron HJ-720ITC is a valid and reliable pedometer for speeds ranging from 2.0 to 4.0 mph, across multiple positions and in lean, overweight, and obese individuals (11,12,27). It has even been suggested that the Omron device is suitable for epidemiological studies in which ambulatory activity is of interest because of its multiple mounting positions (12). Holbrook et al. (12) found that the Omron had mean absolute percent error scores of <3.2% for all wear locations on the treadmill and track, but in our study, it did not demonstrate the same high level of accuracy under free-living conditions. When worn in the pocket, the Omron recorded 68.3% of SW-determined steps in normal individuals, 70.2% in the overweight category, and 64.9% in the obese category. When it was worn on the belt, it recorded 64.3% of SW-determined steps per day in normal weight individuals, 63.1% in overweight individuals, and 57.8% in the obese category. The Omron was least accurate when worn around the neck, recording only 63.1% of SW-determined steps in normal, 63.2% in overweight, and 42.5% in obese individuals.
In the present study, the Omron was worn in three different locations and was compared with the StepWatch-3 and Yamax SW-200, which has not been done previously. This is also the first study to compare the Omron HJ-720ITC accuracy across three different BMI categories and in a free-living environment for 24 h. Holbrook et al. (12) examined the effects of walking speed (2.0, 3.0, and 4.0 mph and self-paced) and wear location (right and left hip, midback on the waist band, and right and left pockets and in a backpack) on the accuracy of this device, but they did not examine the effect of BMI. They found the Omron to be valid across the range of walking speeds and in each location except for those in the backpack. Our findings in the free-living environment contrast with findings from laboratory studies using a similar pedometer (11). In the free-living environment, the position of the device clearly affects device step counting. In obese individuals, the belt and neck positions yielded lower accuracy compared with the pocket position. When worn in the pants pocket, the Omron was unaffected by BMI category. This may be because the pocket keeps the face of the pedometer from tilting forward or backward and exceeding the tilt angle of 30°. The pocket location may prove useful for researchers examining obese individuals.
With regard to the Omron Belt and Neck positions, underestimation in step counting in obese individuals can be attributed to several factors. The manufacturer warns that a slow gait or tilt angle of the face of the instrument exceeding 30° causes error in step counting with the Omron device; however, these traits are also associated with spring-levered pedometers such as the Yamax SW-200 (7,15). Omron Healthcare, Inc., recommends that the face of their pedometer should not exceed a 30° angle from vertical. In obese individuals, it is possible that the tilt angle exceeded 30° when the Omron was worn on the belt. Our findings regarding the decreased accuracy of the Omron device in obese individuals contradict those of Hasson et al. (11), who reported that pedometer (Omron HJ-112) accuracy in obese individuals (BMI ≥ 30 kg·m−2) was maintained even when the vertical axis of the sensor was not aligned properly.
The Omron HJ-112 uses the same internal mechanism (two piezoelectric accelerometers oriented at 90° and multiple-position sensing technology) as the downloadable Omron HJ-720ITC. In a laboratory-based study, Hasson et al. (11) examined the validity of the Omron HJ-112 in different BMI groups during constant- and variable-speed walking. Participants wore an Omron device on the hip, in the pants pocket, in the chest shirt pocket, and around the neck and the Yamax Digiwalker SW-701 on the opposite hip. For the continuous-speed condition, participants walked on the treadmill for 12 min at 1.12, 1.34, and 1.56 m·s−1. The variable-speed condition involved walking bouts of various speeds (1.12, 1.34, 1.56, and 2.5 m·s−1) ranging from 2 to 4 min, interspersed by rest periods ranging from 10 to 60 s. The purpose of the variable-speed walking condition was to simulate actual walking behaviors, where speed, bout duration, and breaks between trips might change depending on the situation. Hasson et al. (11) concluded that the Omron HJ-112 pedometer validly recorded steps in different BMI groups when walking on the treadmill continuously at speeds between 1.12 and 2.5 m·s−1. This was not the case for the Yamax Digiwalker, which undercounted steps in obese individuals.
The most recent study conducted by Zhu and Lee (27) examined seven different wear locations (waist (front and back), pant's front pocket, left side shirt pocket, inside a shoulder bag carried at the side, around the neck, and in a backpack) of 10 Omron HJ-720-IT pedometers. This study examined the accuracy of the steps taken during self-selected pace, free-living-type activities in a controlled laboratory setting (walking 100 steps on a flat sidewalk), up and down stairs (80 steps), and a mixed condition (a set 1253-m course consisting of walking on a sidewalk, grass, up and down a low hill, and up a ramp). Zhu and Lee (27) concluded that, other than a small reduction in accuracy in the pant pockets, the Omron HJ-720-IT pedometers accurately recorded steps in different BMI groups when walking on a flat sidewalk and for walking up the stairs. When walking down stairs and during the mixed condition, the Omron pedometers recorded accurately for all locations across all BMI categories. Overall, the results of this study highlighted the location invariance and accuracy of Omron pedometers when used to measure continuous controlled bouts of walking.
In the study by Hasson et al. (11), all walking bouts were performed continuously for a minimum of 2 min. In the study by Zhu and Lee (27), all walking bouts were continuous from the start to the finish of each condition. The Omron pedometer has a 4-s filter that is incorporated into the device to avoid counting movements that are not technically considered walking (such as shuffling, standing up, sitting down, vibrations from a moving vehicle). The filter was designed to determine the movement pattern in each individual before recording steps. If a participant were to walk for <4 s and stop, the pedometer would not record any steps (16). Although the speed of the walking bouts was altered in the study of Hasson et al., the walking was performed continuously for a minimum of 2 min. Thus, the 4-s filter would not have influenced their results.
The filter in the Omron HJ-720ITC results in a very accurate device for step counting during continuous walking bouts; however, the device may miss a significant percentage of steps taken during intermittent lifestyle activities. Results from the present study are the first to show that the Omron underestimates steps in all BMI categories under free-living conditions, partly because of the 4-s filter. The most common walking bout in nondisabled employed free-living adults is four steps in a row (accounting for 17% of all bouts), and the second most common bout is six steps in a row (accounting for 10% of all bouts) (18). Because of the 4-s filter, the Omron HJ-720ITC would be unlikely to detect any of these steps, thus contributing to an underestimation of the total number of steps accumulated.
Researchers have shown that piezoelectric pedometers are generally more accurate than spring-lever arm electronic pedometers in laboratory settings (2,8,15,20). In particular, the Yamax SW-200 pedometer underestimates step counts at slow walking speeds (15). In contrast, in a free-living environment, the spring-levered, Yamax SW-200 demonstrated greater accuracy than the Omron in normal-weight and overweight individuals. However, the Yamax SW-200 was far less accurate in obese individuals than in normal-weight and overweight individuals, which is consistent with previous research done in laboratory settings (7,15,21).
As shown by the Bland-Altman plots, the Omron Pocket, Omron Belt, Omron Neck, and the Yamax all underestimated steps per day when compared with the StepWatch-3. The inaccuracy of the Omron devices (pocket, belt, and neck) was magnified in individuals who took more daily steps. When fewer steps were taken, the data points were more clustered tightly around the mean. As the daily number of steps increased, there was more variance in the inaccuracy of the Omron devices, causing the overall appearance of the plot to take on a funnel/horn shape. The Yamax pedometer had the lowest mean bias, and the data points were more tightly clustered around the mean, indicating that it was the most accurate device in the combined group of participants. However, the Yamax pedometer still underestimated the criterion by an average of 2671 steps per day.
In Figure 2, there are a few unusual values. Although they seemed to be outliers, they were not influential outliers. After checking the normality of the data, and determining that assumptions of normality were not grossly violated, parametric statistics were run. The analysis was run with and without the outliers, and there was no effect of pulling the outliers out or including them in the analysis.
The performance of the Omron and the Yamax pedometers was not compared with the criterion measure of investigator hand tally counter but to another device (StepWatch-3). The StepWatch-3 is considered to be a good criterion because of its high level of accuracy in step counting, regardless of BMI and walking speed (13,18). The StepWatch-3 mechanism has greater sensitivity for detecting steps than most pedometers, and it counts accurately even in slow walkers and obese individuals. However, there is still the potential for a small amount of error due to recording of erroneous steps. Karabulut et al. (13) demonstrated that heel tapping and leg swinging will cause some steps to be recorded with the StepWatch-3. However, those authors concluded that erroneous step counts associated with these types of activities would be unlikely to have a significant effect on the total daily number of steps recorded per day.
In summary, previous research has shown that the Omron HJ-720ITC pedometer is an accurate device for counting steps during continuous bouts of walking at 53.6-107.2 m·min−1 (12). However, this device has a 4-s filter that contributes to an underestimation of steps during intermittent activities. Despite this limitation, the Omron pedometer could be a useful device for clinical interventions that prescribe continuous bouts of moderate-intensity walking, but researchers should recognize that it does not measure all steps taken throughout the day.
No funding was received for this study. Dr. David Bassett received a grant from Omron to examine the validity of a different device, the HJ-303 triaxial pedometer.
The results of the present study do not constitute endorsement by the American College of Sports Medicine.
1. Bassett DR Jr. Validity
and reliability issues in objective monitoring of physical activity
. Res Q Exerc Sport
. 2000;71(2 suppl):S30-6.
2. Bassett DR Jr, Ainsworth BE, Leggett SR, et al. Accuracy of five electronic pedometers for measuring distance walked. Med Sci Sports Exerc
3. Bassett DR, Strath SJ. Use of pedometers to assess physical activity
. In: Welk GJ, editor. Physical Activity Assessments for Health-Related Research
. Champaign (IL): Human Kinetics; 2002. p. 163-8.
4. Bergman RJ, Bassett DR Jr, Muthukrishnan S, Klein DA. Validity
of 2 devices for measuring steps
taken by older adults in assisted-living facilities. J Phys Act Health
. 2008;5(1 suppl):S166-75.
5. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet
6. Callaway CW, Chumlea WC, Bouchard C, Himes JH, Lohman TG, Martin AD. Circumferences. In: Lohman TG, Roche AF, Martorell R, editors. Anthropometric Standardization Reference Manual
. Champaign (IL): Human Kinetics Books; 1988. p. 39-47.
7. Crouter SE, Schneider PL, Bassett DR Jr. Spring-levered versus piezo-electric pedometer accuracy in overweight and obese adults. Med Sci Sports Exerc
8. 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
9. Foster RC, Lanningham-Foster LM, Manohar C, et al. Precision and accuracy of an ankle-worn accelerometer-based pedometer in step counting and energy expenditure. Prev Med
10. Gayle R, Montoye HJ, Philpot J. Accuracy of pedometers for measuring distance walked. Res Q
11. Hasson RE, Haller J, Pober DM, Staudenmayer J, Freedson PS. Validity
of the Omron HJ-112 pedometer during treadmill walking
. Med Sci Sports Exerc
12. Holbrook EA, Barreira TV, Kang M. Validity
and reliability of Omron pedometers for prescribed and self-paced walking
. Med Sci Sports Exerc
13. Karabulut M, Crouter SE, Bassett DR Jr. Comparison of two waist-mounted and two ankle-mounted electronic pedometers. Eur J Appl Physiol
14. Kemper HC, Verschuur R. Validity
and reliability of pedometers in habitual activity research. Eur J Appl Physiol Occup Physiol
15. Melanson EL, Knoll JR, Bell ML, et al. Commercially available pedometers: considerations for accurate step counting. Prev Med
16. Nakae S, Oshima Y, Ishii K. Accuracy of spring-levered and piezo-electric pedometers in primary school Japanese children. J Physiol Anthropol
17. Neovius M, Hemmingsson E, Freyschuss B, Udden J. Bioelectrical impedance underestimates total and truncal fatness in abdominally obese women. Obesity
18. Orendurff MS, Schoen JA, Bernatz GC, Segal AD, Klute GK. How humans walk: bout duration, steps
per bout, and rest duration. J Rehabil Res Dev
19. Schneider PL, Crouter SE, Bassett DR. Pedometer measures of free-living physical activity
: comparison of 13 models. Med Sci Sports Exerc
20. 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
21. Shepherd EF, Toloza E, McClung CD, Schmalzried TP. Step activity monitor: increased accuracy in quantifying ambulatory activity. J Orthop Res
22. Storti KL, Pettee KK, Brach JS, Talkowski JB, Richardson CR, Kriska AM. Gait speed and step-count monitor accuracy in community-dwelling older adults. Med Sci Sports Exerc
23. Tudor-Locke C. Taking steps
towards increased physical activity
: using pedometers to measure and motivate. President's Council on Physical Fitness and Sports Research Digest
24. Tudor-Locke C, Myers AM. Methodological considerations for researchers and practitioners using pedometers to measure physical (ambulatory) activity. Res Q Exerc Sport
25. Tudor-Locke C, Sisson SB, Lee SM, Craig CL, Plotnikoff RC, Bauman A. Evaluation of quality of commercial pedometers. Can J Public Health
. 2006;97(1 suppl):S11-5.
26. Washburn R, Chin MK, Montoye HJ. Accuracy of pedometer in walking
and running. Res Q Exerc Sport
27. Zhu W, Lee M. Invariance of wearing location of Omron-BI pedometers: a validation study. J Phys Act Health