Background

Over the past decades, the proportion of individuals being overweight or obese has more than doubled. In 2014, 39% of adults aged 18 years and over were overweight and 13% were obese [1]. The Middle-East region and especially the United Arab Emirates (UAE) are facing a similar public health problem. Indeed, obesity is reaching particularly high levels in the UAE, with more than 40% of adult women being obese in 2010 [2]. It is well-established that overweight and obese individuals are more likely to develop serious health complications like type 2 diabetes, cardiovascular disease and cancer. Hence, prevention and treatment of obesity is a priority [3].

Physical Activity (PA) is defined as any bodily movement, made by the skeleton muscle, increasing the energy expenditure above the basal energy metabolism. The regular practice of PA contributes to the prevention and treatment of many chronic diseases, especially obesity and its related health complications [4,5], through a positive impact on body composition, blood lipid profile, insulin sensitivity and blood pressure [6].

The 2008 PA Guidelines stipulate that any adult should practice at least 30 minutes of moderate PA daily in combination with muscle strengthening exercise and that any child should practice at least 1 hour of PA daily, preferably vigorous, combined with muscle and bone strengthening exercise [7]. Nonetheless, most of people remain insufficiently active. Globally, in 2010, 23% of adults and 81% of adolescents were insufficiently active in the world [8]. It was recently highlighted that only 27% of university students satisfy the recommendations for aerobic activity [9]. Similarly, in the UAE, 25% of students, 12-24 years old, report a total sedentary lifestyle, with no PA and only 25% are involved in vigorous PA [10,11].

Hence, in the context of an alarming rate of obesity worldwide, the promotion of an active lifestyle is a necessity. Moreover, valid and reliable tools are needed to self-monitor PA. Pedometers are motion sensors that are typically worn on the waist to count the number of steps during locomotion and enable PA level measurement. They are designed to capture vertical motion of the hip joint with movement and are mainly used for self-monitoring, as motivational tool and to provide a feedback on daily PA levels [12,13]. Their validity has been demonstrated in adults to assess their PA level [14].

With the advances in technology, various novel methods have been developed to assess PA. Mobile technologies like smartphone applications, are now widely available and very popular, especially among young populations. These innovative tools offer a significant potential in increasing PA levels and thus, are highly recommended to include them into clinical practice, prevention strategies, interventions and research [15,16]. A great number of pedometers applications can be accessed on all smartphones, most of them being free. PACER is one of the bestrated pedometer application by users in Google Play and presents the advantage to be available on both iPhone and androids. The use of these applications has been associated to a higher level of PA and weight loss but with a moderate level of evidence in previous studies [17-19].

Reliability is the degree to which an assessment tool produces stable and consistent results while validity refers to how well a test measures what it is purported to measure [20]. Even though, both are necessary conditions for effective interventions and high level of scientific evidence in research, the great majority of the publication refer to validity only Besides, mainly iPhone-based applications were considered and in very different conditions of testing. Overall, available results on validity and reliability of PA smartphone applications are contradictory [21]. Finally, although overweight and obese people represent a significant part of the populations, only one study was conducted among obese individuals and in controlled conditions. A low accuracy of the application was reported [22].

The aim of this work was to investigate, the validity and reliability of the PACER smartphone pedometer application, in assessing number of steps, in both controlled and free-living conditions among overweight and obese young adults.

Materials and Methods

Participants

A sample of 40 overweight or obese students (20 females and 20 male) was recruited from a major university in the UAE. Weight and height were assessed at screening by using an impedance scale (TANITA DC-430M) and a stadiometer (SECA 217), respectively. An age between 18 and 25 years as well as a Body Mass Index (BMI) above 25kg/m² were required to be included. In case of any physical disability preventing from moving, or in case of pregnancy or chronic medical conditions, such arthritis, heart diseases, respiratory disorders, participants were excluded. The sample size allows for accurate, valid and representative results and was based on previous reliability and validity studies that have been carried out on smartphone applications [22-25]. In terms of reliability, a sample above 30 participants, with two repeated measurements, can be considered as sufficient to produce results within an acceptable subject error range and with stable validity [26,27].

The project was approved by the Al Ain Medical District Human Research Ethical Committee (Protocol 15/91 ERH_2015_3140). Each participant was informed about the protocol and signed a written informed consent form prior any data collection.

Experimental conditions

Participants used their own smartphones, regardless of the brand. In each trial, the smartphones were placed on the right pocket. Omron HJ-320E pedometer placed in the right hip was used as the reference in testing the validity of PACER to assess s number of steps, in both controlled and free-living conditions. OMRON pedometer has been previously shown to be a valid and reliable tool to count steps [28-30].

Test-retest reliability and validity of PACER application in controlled conditions

For this first trial, participants were asked to come to the testing site (the clinical nutrition laboratory of the university). They were requested to wear comfortable shoes and clothes to walk for 30 minutes on a treadmill, at a speed of 3 km/h. A video recording was conducted for each 30 minutes session. This trial was repeated after 1 week.

Test-retest reliability and validity of PACER application in free-living conditions

For this second trial, participants were used both PACER and Omron pedometer for one typical day. In addition, they were asked to fill a dairy indicating the times at which pedometer was on and off. This second trial was repeated after two weeks and was held on the same day of the week.

Statistical analysis

Data were analysed by using SPSS v.23 for Windows (SPSS Inc., Chicago, IL). Statistical significance was set at p<0.05. Means±s.e. were calculated and ANOVA was used to test gender differences. In reliability study, Intraclass Correlation Coefficients (ICC) and 95% confidence interval (95%CI) were calculated to determine the level of agreement between steps count measurement at the two separate occasions by using PACER, in controlled and free-living conditions. An ICC with a value of 0.40 was rated as poor agreement, with 0.40-0.75 as fair-to-good agreement and with 0.75 as excellent agreement [31].

In validity study, the means of the step counts obtained at the two separated occasions were considered, in both free-living and controlled conditions. Then, Spearman correlation coefficients were calculated. As correlational analysis alone may not reveal the potential bias, Bland-Altman method was used when any correlation coefficient was above 0.40 [32] and a Bland-Altman plot was generated. The differences between the results from the two methods of measurement (PACER vs OMRON, PACER vs videotape) were shown against the means of the results from the two methods of measurements. The mean +/- SD of the difference and the limits of agreement defined as mean +/-1.96 SD were calculated and added to the graphs.

Results

Table 1 represents the main characteristics of the sample. Among the 40 participants which were recruited, 5 males and 5 females dropped from the study. The mean age of the 30 participants who completed the study was 21.37±0.20 years old, with a BMI of 34.06±1.27 kg/m². Female participants had significantly (p<0.05) higher percentage of fat mass than their male counterparts. The resting heart rate was 93.47±2.72 bits/min at baseline.

Test-retest reliability of PACER

Table 2 shows the step count as obtained with PACER on the two separate occasions, in both controlled and free-living conditions. ICC (95%CI) are presented too. Based on the step-defined PA hierarchy which was proposed by Tudor-Locke C. in 2010 [20], 26.67% of male and 20% of female only were doing more than 10,000 steps/day according to PACER. Hence, near a quarter only of the sample could be considered as active. In controlled conditions, in the whole sample and in female, the level of agreement was fair to good. It became greater, but remained fair to good, in free-living conditions. The male group was associated with a poor agreement in controlled conditions and excellent agreement with free-living conditions.

Validity of PACER in counting steps

The correlations of step counts as recorded by PACER with steps counts as recorded from OMRON pedometer, in free-living conditions, are presented in Table 3.

Significant rho values ranging from 0.64 to 0.94 were obtained in the whole sample, male and female groups. Bland-Altman plots indicated two outliers in the whole sample and female group and only one in the male group (Figure 1). The limits of agreement were -3125.4 to 2812.6, -1280.4 to 1699.2 and -3125.4 to 2812.6 steps for the whole sample, male and female groups, respectively. The means of the differences were -156.4±1484.5, 209.4±744.9 and -552.2±1929.7 steps in the whole sample, male and female groups, respectively. This means that PACER tended to overestimate the number of steps compared to OMRON in the whole sample and female group whereas it tended to underestimate the number of steps compared to OMRON in the male group. However, in the three groups, the majority of the points (22, 27 and 28 out of 28, in the whole sample male and female groups, respectively) were within the range of the mean (3533.9-13487.5, 2205.4-15207.8 and 5321.8-11308 steps in the whole sample male and female groups, respectively). This is an agreement with the fair to good Spearman coefficients of correlation which were obtained here.

The correlations of steps counts as recorded by PACER with steps counts as recorded from OMRON pedometer and from video, in controlled conditions, are presented in Table 4. There was no significant correlation between PACER steps counts and OMRON. By contrast, there were significant correlations between PACER steps counts and manual counting (video) in the whole sample and female group (rho=0.48, p=0.01 and rho=0.72, p≤0.001, respectively). Bland-Altman plot was built for steps counts from PACER compared with manual steps counting by using the video in the whole sample and female group (Figure 2). In the whole sample, it shows that only two points were outliers: one was above the upper limit of agreement (1423.2 steps) and one was below the lower limit of agreement (-1588.4 steps). The mean of the difference ± SD was on average -82.6±752.9 steps, indicating that on average, the manual counting by using the video was lower than the steps counts from PACER. The majority of the points (23 out of 28) are within the range of the mean (2490.8-3333.0 steps) for step counts between PACER and manual counting. This agreement is considered as fair, according to Spearman correlations.

In the female group, only one point was outlier above the upper limit of agreement (1768.1 steps). The mean of the difference±SD was on average -154.9±806.9 steps, indicating, that the manual counting was higher than the steps counts from PACER. However, the majority of the points (28 out of 29) were within the range of the mean (2353.5-3307.5 steps) between manual counting and step counts from PACER. This indicates another fair Spearman correlation coefficient.

Discussion

Pedometer smartphone applications are very popular and are frequently used to assess PA in many research works and interventions. Hence, they represent a particularly relevant tool to promote PA, especially in populations in overweight or obese individuals. Reliability and validity of any assessment tool are pre-requisites to an effective use in research and interventions. However, there are few studies only that are related to the reliability and validity of pedometer smartphone applications in free-living conditions and results remain contradictory. Besides, while overweight and obesity keep increasing worldwide, it received relatively little attention and almost no research to-date. Our work demonstrated that PACER, one of the widely used smartphone pedometer applications, which is freely available for both iPhone and androids is reliable and valid in obese young adults, especially in free-living conditions.

Reliability is an important factor to be considered when selecting a tool to count steps. It is a condition for a long-term use, like in interventions, and a pre-requisite for validity study. In spite of that, data on pedometer applications’ reliability remain rare [22,23,25,33-35]. In our experiment, in free-living conditions, a fair to excellent reliability was observed. This means that PACER could be recommended to overweight and obese individuals to track their PA. The only one study that was conducted in similar natural conditions, considered three other pedometer smartphone applications and non-overweight and non-obese participants, which does not allow a fair comparison with our results [25]. Controlled conditions were associated with a lower reliability compared to free-living conditions in our experiment. This could be related to the movement of the treadmill belt, which is challenging participants’ equilibrium and walk regularity. This could also explained why our ICC values were lower compared to the values reported by Konharn, et al. [22], in overweight and obese adults who performed shorter, 3 minutes only, walking sessions on treadmill.

Importantly, PACER was shown as an accurate tool to count steps in overweight and obese young adults, against pedometer, in free-living conditions, with rho value ranging from 0.64 in female to 0.94 in male. To-date, there are no other studies conducted in free-living conditions among overweight or obese adult participants. The great majority of the studies, which were conducted in free-living conditions and in healthy individuals, reported a very poor validity of the smartphone pedometer applications [23,24,36]. It is known that weight status is likely to influence the accuracy of pedometers [37]. Indeed, due to the presence of a thicker subcutaneous fat layer, as in overweight and obese individuals, the movement may be more difficult to detect and may request a pedometer technology with a higher sensitivity to get accurate results. Our results indicate that PACER may satisfy this requirement in free-living conditions.

In controlled conditions, PACER was not valid against the OMRON pedometer, Here, a relatively slow speed of 3km/h was used on treadmill. This speed makes the walking session feasible for overweight and obese adults but has been associated with a lower accuracy with other pedometer applications and also electronic pedometers [22,25,36,38-40]. This is most probably due to the higher difficulty in detecting slow movements. Furthermore, in overweight and obese individuals, Konharn, et al., reported a lack of accuracy on treadmill, within the speed range of 3.2 -8.0 km/h [22]. This is of a great concern since indoor treadmill exercise is one of the most popular type of exercise and the way that is usually recommended by health professionals, mainly due to the climate conditions. In addition, it may not be realistic and achievable for overweight and obese individuals to exercise at higher speed than 3 km/h running speed.

However, in our study, PACER was valid against manual counting on video. Video recording allowed counting the steps several times and by different observers. It can reasonably be considered that the step count obtained by using this method is more accurate than the value given by the OMRON pedometer. By consequence, even in controlled conditions, PACER could be recommended as a valid tool to overweight and obese individuals to count their steps.

In this study, differences were observed between gender. The less motivation which was reported with male participants may explain the lower reliability which has been observed in controlled conditions in males compared to females. Researchers had to provide more support to male participants to complete the 30 minutes walking session on treadmill and limit the brief walking breaks. A good validity of PACER was reported in controlled conditions against manual counting, in the whole sample, in female but not in male. The way of carrying the smartphone has been shown to affect the quality of the movements detection [23,25,40,41].

In the present study, during the treadmill sessions female subjects wore tighter clothes than the male counterparts who were more likely to wear baggy shorts and carry their smartphone loosely in a pocket. This may have resulted in an overestimation of movement by PACER compared to both the OMRON pedometer and manual counting. There are some limitations of the present study. First, although the sample size was acceptable and comparable to other studies, a greater validity and ability to generalize our results would have been achieved with a larger sample size. Secondly, although clear instructions on the use of pedometers were given by the researchers to the participants, this did not prevent fully the non-compliance with the instructions.

Conclusion

The results of present study indicate that smartphone application, PACER, is a fairly reliable and valid tool to count steps under both free-living and controlled conditions, for overweight and obese young adults. Therefore, PACER can be recommended to overweight and obese young adults to self-monitor their PA levels. It may also play a role in future interventions designed to increase PA levels.

Acknowledgements

Not applicable

Authors contributions

Carine Platat: design of the study, coordination of the project, data interpretation, manuscript writing.

Amjad Jarrar: data collection, data analysis, manuscript review. Fatima Al Qshadi, Ghofran Kayed, Nour Abou Hussein: research assistants, screening of participants, data collection, review of the manuscript.

Habiba Ali: conception, obtaining funding, significant revision of the manuscript.

All authors read and approved the final manuscript.

No conflict of interest to be declared for any of the authors.

Funding

This study is part of the research project “Development, implementation and evaluation of technology mediated lifestyle intervention for promoting weight loss and improving nutrition behaviors among young adults” which received the Zayed Center for Health Sciences-based internal grant, from the United Arab Emirates University, in 2014.

References

1. World Health Organization (2004) Obesity: preventing and managing the global epidemic. Report of a WHO Consultation (WHO Technical Report Series 894). ISBN: 92 4 120894 5.
2. Mehio Sibai A, Nasreddine L, Mokdad AH, Adra N, Tabet M, et al. (2010) Nutrition transition and cardiovascular disease risk factors in Middle East and North Africa countries: reviewing the evidence. Ann Nutr Metab 57: 193-203.
3. Kopelman P (2007) Health risks associated with overweight and obesity. Obes Rev 8: 113-117.
4. Naci H, Ioannidis JP (2015) Comparative effectiveness of exercise and drug interventions on mortality outcomes: metaepidemiological study. Br J Sports Med 49: 1414-1422.
5. Pareja-Galeano H, Garatachea N, Lucia A (2015) Exercise as a Polypill for Chronic Diseases. Prog Mol Biol Transl Sci 135: 497-526.
6. Hills AP, Street SJ, Byrne NM (2015) PA and Health: "What is Old is New Again". Adv Food Nutr Res 5: 77-95.
7. World Health Organization (2010) Global recommendations on PA for health. ISBN 978 9241599979. WHO library.
8. World Health Organization (2014) Global status report on non-communicable diseases 2014. ISBN 978 9241564854. WHO library.
9. Katzmarzyk PT, Lee IM, Martin CK, Blair SN (2017) Epidemiology of PA and Exercise Training in the United States. Prog Cardiovasc Dis 60: 3-10.
10. Yammine K (2016) The prevalence of PA among the young population of UAE: a meta-analysis. Perspect Public Health. 137: 275-280.
11. Zaabi MA, Shah SM, Sheek-Hussein M, Abdulle A, Junaibi AA, et al. (2016) Results from the United Arab Emirates' 2016 Report Card on PA for Children and Youth. J Phys Act Health. 11: 299-306.
12. Al-Eisa E, Alghadir AH, Iqbal ZA (2016) Measurement of PA in obese persons: how and why? A review. J Phys Ther Sci 28: 2670-2674.
13. Bravata DM, Smith-Spangler C, Sundaram V, Gienger AL, Lin N, et al. (2007) Using pedometers to increase PA and improve health: a systematic review. JAMA 298: 2296-2304.
14. Tudor-Locke C, Bassett DR, Shipe MF, McClain JJ. (2011) Pedometry methods for assessing free-living adults. J Phys Act Health 8: 445-453.
15. Lewis BA, Napolitano MA, Buman MP, Williams DM, Nigg CR (2017) Future directions in PA intervention research: expanding our focus to sedentary behaviors, technology, and dissemination. J Behav Med 40: 112-126.
16. Sallis R, Franklin B, Joy L, Ross R, Sabgir D, et al. (2015) Strategies for promoting PA in clinical practice. Prog Cardiovasc Dis 57: 375-386.
17. Bacigalupo R, Cudd P, Littlewood C, Bissell P, Hawley MS, et al. (2013) Interventions employing mobile technology for overweight and obesity: an early systematic review of randomized controlled trials. Obes Rev 14: 279-291.
18. Bort-Roig J, Gilson ND, Puig-Ribera A, Contreras RS, Trost SG (2014) Measuring and influencing PA with smartphone technology: a systematic review. Sports Med 44: 671-686.
19. Schoeppe S, Alley S, Van Lippevelde W, Bray NA, Williams SL, et al. (2016) Efficacy of interventions that use apps to improve diet, PA and sedentary behaviour: a systematic review. Int J Behav Nutr Phys Act 13: 127.
20. Sullivan GM (2011) A primer on the validity of assessment instruments [published correction appears in J Grad Med Educ 3: 119-120.
21. Silva AG, Simões P, Queirós A, Rodrigues M, Rocha NP (2020) Mobile Apps to Quantify Aspects of PA: a Systematic Review on its Reliability and Validity. J Med Syst 44: 51.
22. Konharn K, Eungpinichpong W, Promdee K, Sangpara P, Nongharnpitak S, et al. (2016) Validity and Reliability of Smartphone Applications for the Assessment of Walking and Running in Normal-weight and Overweight/Obese Young Adults. J Phys Act Health 13: 1333-1340.
23. Akerberg A, Söderlund A, Lindén M (2014) Accuracy in pedometers: dependent on the technology for measurement?. Stud Health Technol Inform 200: 173-175.
24. Bergman RJ, Spellman JW, Hall ME, Bergman SM (2012) Is there a valid app for that? Validity of a free pedometer iPhone application. J Phys Act Health 9: 670-676.
25. Orr K, Howe HS, Omran J, Smith KA, Palmateer TM, et al. (2015) Validity of smartphone pedometer applications. BMC Research Notes 8: 733.
26. Hobart JC, Cano SJ, Warner TT, Thompson AJ (2012) What sample sizes for reliability and validity studies in neurology?. J Neurol 259: 2681-2694.
27. Shoukri M, Asyali M, Donner A (2004) Sample size requirements for the design of reliability study: review and new results. Stat Meth Med Res 13: 251-271.
28. De Cocker K, De Meyer J, De Bourdeaudhuij I, Cardon GM (2012) Non-traditional wearing positions of pedometers: Validity and reliability of the Omron HJ-203-ED pedometer under controlled and free-living conditions. J Sci Med Sport 15: 418-424.
29. Hasson RE, Haller J, Pober DM, Staudenmayer J, Freedson PS (2009) Validity of the Omron HJ-112 Pedometer during Treadmill Walking. Med Sci Sports Exerc 41: 805-809.
30. Holbrook E, Barreira TV, Kang M (2009) Validity and Reliability of Omron Pedometers for Prescribed and Self-Paced Walking. Med Sci Sports Exerc 41: 670-674.
31. Altman DG (1991) Inter-rater agreement. In Practical statistics for medical research. London: Chapman & Hall 403-409.
32. Schmidt ME, Steindorf K (2006) Statistical methods for the validation of questionnaires-discrepancy between theory and practice. Methods Inf Med 45: 409-413.
33. Benson A, Bruce L, Gordon B (2015) Reliability and validity of a GPS-enabled iPhone TM “app” to measure PA. Sport Sci 33: 1421-1428.
34. Fokkema T, Kooiman TJM, Krijnen WP, VAN DER Schans CP, Groot MDE (2017) Reliability and validity of ten consumer activity trackers depend on walking speed. Med Sci Sport Exerc 49: 793-800.
35. Kooiman TJM, Dontje ML, Sprenger SR, Krijnen WP, van der Schans CP, et al. (2015) Reliability and validity of ten consumer activity trackers. BMC Sports SciMed Rehabil 7: 24.
36. Leong JY, Wong JE (2017) Accuracy of three Android-based pedometer applications in laboratory and free-living conditions. J Sports Sci 35: 14-21.
37. Tuyo BM, Bassett DR, Coe DP, Feito Y, Thompson DL (2013) Effect of BMI on pedometers in early adolescents under free-living conditions. Med Sci Sports Exer 45: 569-573.
38. Crouter SE, Schneider PL, Karabulut M, Bassett Jr DR (2003) Validity of 10 electronic pedometers for measuring steps, distance, and energy cost. Med Sci Sports Exerc 35: 1455-1460.
39. Dondzila CJ, Swartz AM, Miller NE, Lenz EK, Strath SJ (2012) Accuracy of uploadable pedometers in laboratory, overground, and free-living conditions in young and older adults. Int J Behav Nutr Phys Act 9: 143.
40. Presset B, Laurenczy B, Malatesta D, Barral J (2018) Accuracy of a smartphone pedometer application according to different speeds and mobile phone locations in a laboratory context. J Exerc Sci Fit 16: 43-48.
41. Tudor-Locke C (2010) Steps to Better Cardiovascular Health: How Many Steps Does It Take to Achieve Good Health and How Confident Are We in This Number? Curr Cardiovasc Risk Rep 4: 271-276.