Introduction

The link between physical activity (PA) and health status has been evaluated in various epidemiological studies. Physical inactivity poses a behavioral risk factor for some types of non-communicable diseases (NCDs) including cardiovascular diseases, stroke, high blood pressure, non–insulin-dependent diabetes mellitus, osteoporosis, and certain types of malignancies.^1-5 Nevertheless, only a small proportion of individuals gets sufficiently adequate PA.⁶

PA measurement is hard to quantify due to its complex nature.^7,8 In data collection at the population level, selecting appropriate and precise estimation method to measure PA as a variable is a challenging task for investigators.^5-9 In comparison, a variety of methods have been used to assess PA, such as self-reports (interviews, diaries, and questionnaires) and doubly-labeled water at population-based studies. Self-reported PA questionnaires are usually chosen as the most feasible methods to assess PA. As such questionnaires are inexpensive, brief, and have general acceptance, they can be applied to measure the frequency, duration, and intensity of PA.^10,11

The AZAR Cohort Study, initiated in 2014, is a population-based follow-up study conducted in East Azerbaijan province, Iran. Investigating the level of PA and its effects on the health status of the population in this prospective cohort study has been considered as a crucial task. Hence, a valid and reliable questionnaire was required to determine the habitual PA of the people in this large-scale epidemiological study.⁵ In general, several questionnaires have been validated to assess PA among adults, but each instrument has some disadvantages. ⁸

In order to conduct this cohort study, there was a need for a scale developed to measure PA by questionnaire only, which particularly designed for a large epidemiological study.¹² This self-reported PA questionnaire was designed, developed, and validated by Aadahl and Jorgensen in 2003, and it is a PA scale with a simple usage to measure the level of PA among sedentary adults. This questionnaire was previously modified and validated among Danish adults.^12,13 In different settings, the accuracy of the self-reported PA questionnaires depends on the diversity of the sociocultural and ethnic context. Therefore, in order to assure an accurate measurement of PA level in a specific population there is a need to examine the validation of the tool being used for the study.⁴

Although the Persian translation of this questionnaire was previously used for the Childhood & Adolescence Surveillance and Prevention of Adult Non-communicable Diseases (CASPIAN) study,^14,15 among adult population, no validated Iranian version has been reported to date.

Previous studies reported the concurrent validity of this self-report PA questionnaire.^12,13 In those studies the findings of the questionnaire were compared with accelerometer, PA dairy questionnaire and Vo_2max. Additional studies are needed to validate it among other target populations. In addition, the factor structure of the Persian version of the questionnaire needs to be established. Searching the literature no published study was found on the factor structure of this PA scale among Iranian adult population. The present study aimed to assess the validity of this self-reported PA questionnaire by examining the construct, discriminate and concurrent validities. Temporal reliability and internal consistency were also examined. Therefore, the primary objectives of this study were to explore the dimensionality and evaluate psychometric properties (validity and reliability) of the self-report PA questionnaire in the pilot phase of the AZAR Cohort Study conducted on 35-70 years old population in Khameneh in 2014

Materials and Methods

Results

Factorial (construct) validity

EFA was conducted in the scale base to identify the factor model using all the observations. Based on the origin structure and the preliminary analyses of the items, it was found that the extracted factors may be divided into four sub-scales as detailed below:

1. Active leisure time, 4 items (PH2-PI-PF1-PE3)
2. Inactive leisure time, 5 items (PB1-PB2-PA1-PA2-PA3)
3. Household/gardening work, 3 items (PD1-PE1-PG2)
4. Occupation work, 11 items (PC1-PC2-PC3-PD2-PF4-PH1-PE2-PG1-PF2-PF3-PG3).

More specifically, the dimension reduction process was implemented for all the domains, separately, to identify underling potential factor (sub-scales). Table 2 shows the results of the scale based test of the item convergence validity. The final model found to be with four factors and 21 items (two items did not load on any factor [factor loadings < 0.2] and was removed): Factor 1 with 4 items and 2 sub-factors (variance = 52.73%), factor 2 with 6 items and 2 sub-factors (variance = 38.68%), factor 3 with 3 items (variance = 38.66%), and factor 4 with 10 items (variance = 12.67%). This model indicated that the extracted factors were suitable for the factor model in the observed dataset. The extracted sub-factors were named as sports (2 items), walking & bicycling (2 items), sitting (3 items), and sleep (3 items). The other recognized factors were given the same names as the basic factors of the underlying domains. In addition, the eigenvalues of all the domains were more than 1. The factor analysis results showed the value of the KMO measure of the sampling adequacy to be 0.69, and Bartlett’s test of sphericity showed the adequacy of the model (P < 0.001).

Table 2. Exploratory factor loading (principal axis factoring extraction with varimax rotation) of the questionnaire itemsa
Factor	Subfactor	Item	Loading
Factor 1. Recreational leisure time	Sport	PH2	0.729
		PI	0.710
	Commuting (Walking/bicycle)	PE3	0.781
		PF1	0.581
Factor 2. Sedentary leisure time	Sitting	PC3	0.769
		PB2	-0.565
		PB1	0.390
	Sleep	PA1	0.728
		PA2	0.617
		PA3	0.330
Factor 3. Household/gardening work		PG2	-0.687
		PD1	0.683
		PE1	0.471
Factor 4. Occupation work		PC1	-0.720
		PD2	0.504
		PG1	0.467
		PF4	0.338
		PH1	0.221
		PE2	0.231
		PF2	0.212
		PC2	-0.211
aAll loadings above 0.20 are presented; the negative values show indirect relation between an item and a scale.

In order to achieve a CFA model with a good external validity, it is highly recommended to perform the CFA in a random subsample (a random sample drown from the main sample is normally satisfied).²⁰ To do so, the CFA was conducted on the 21 items of the questionnaire with AMOS 23.0 software to test the fit of the final four-factor model. Sixty percent of the participants were considered as the sub-sample; through random sampling the data of 60% out of all the participants in the SPSS software were included in the CFA analysis. CFA supported the four-factor structure and displayed appropriate good fit to the data (χ2 [163] = 462.139, P < 0.001; CFI = 0.98; GFI = 0.936; AGFI = 0.90; RMSEA [90% CI] = 0.057 [(0.51-0.063]). Moreover, all the standardized coefficients in the factor showed moderate correlations between the latent factors (Figure 1).

Figure 1. Confirmatory factor analysis.

Discussion

Insufficient PA is considered as a behavioral risk factor for NCDs. Without a valid instrument, the associations between PA and health status may not be accurately evaluated and identified.^22,23 The present study sought to determine whether the selected self-reported PA questionnaire was a valid scale to assess PA patterns among the adults elected for the AZAR Cohort Study. From the results of the present study confirmed the internal consistency of the scale. Moreover, moderate temporal stability of the questionnaire during two separate occasions (correlation coefficient = 0.60) was found. These findings were in line with those obtained in the study that assessed the PA among Iranian young adults (correlation coefficient = 0.87).¹⁴

Although the ICC value showed moderate temporal stability, the 95% CI was wide for the scores. This wide 95% CI for the ICC value could be due to the fact that the PA has a multinomial nature and it is not a stable behavior.^7,8 However, it does not seem that the actual changes in the PA pattern of the subjects occurred during 2-weeks interval between the test-retest. The moderate coefficient of the correlation between the two occasions provided evidence for temporal validity of the self-reported PA questionnaire.

This is the first investigation to address the EFA and CFA approach of the self-reported PA questionnaire. The results confirmed the factorial structure of the questionnaire in a sample of 952 adults. The sample size was adequate for factor analysis in the present study as the proportion of the sample size was based on more than 5 participants per item.²⁴

The findings of the present study suggested that the questionnaire might have a four factors structure for the instrument, including recreational leisure time, sedentary leisure time, household/gardening work, and occupation work. The results showed that the factor loadings of the items PG3 and PF3 were less than 0.2. As, the considered cut-off value to retain an item in the scale was 0.2, these items were not included in the final model. Also, the CFA provided evidence to support the factor structure represented by these items. Therefore, the questionnaire assessed PA in four factors (domains) and four sub-factors with 21 items. A concern in the analysis of this instrument was its construct validity which had not been investigated previously. So, the identified model should be further assessed. However, the preliminary factor structure found in the present study was not different from those found in the original domains. Previously, some validation studies on the questionnaire have been conducted in Denmark and Iran, but the demographics and geographic contexts of the studies were different.^13,14

In the present study, the IPAQ, as a subjective measure, was considered as a criterion measure for concurrent validity. In consistent with the findings of the prior research in Denmark,¹³ a significant but not so strong correlation was found between the two self-reported questionnaires (r = 0.62, P < 0.001). However the significant moderate correlation confirmed the concurrent validity of the PA questionnaire. Concurrent validity of the questionnaire was assessed among Danish adults by Aadahl and Jorgensen in 2003. They found a high correlation between the scale and a PA dairy questionnaire (r = 0.74) and a poor correlation between the scale and an accelerometer (r = 0.20, P = 0.04); so the self-reported PA scale was approved as a valid instrument to assess the PA among the adult with sedentary to moderately active populations.¹³ In a study to validate the self-report PA questionnaire against maximal oxygen uptake (Vo_2max testing), correlation between the two scales was assessed and it was found that the activity scale had an acceptable validity.¹² In this study, the total amount of PA was not significantly associated with Vo_2max (r²= 0.69, P = 0.098), but the amount of daily vigorous intensity PA and Vo_2max had a strong and significant association (r²= 0.76, P < 0.001).¹² As a matter for validation, using subjective methods to assess criterion validity can be considered as a limitation for the present study. The PA questionnaires as an objective measure are prone to recall and desirability biases. Hence, the result of subjective methods should be paid attention in terms of misclassification while assessing the PA habits.^12,25,26

The Bland-Altman plot was used to verify the agreement between the two questionnaires, but a non-constant bias was observed over the whole range of the instruments. Therefore, as recommended by Bland-Altman, log transformation approach was applied in the present study.²⁷ Nevertheless this method could not improve the agreement between the two scales. As shown in the Bland-Altman plot (Figure 2), the discrepancy between the two scales was obvious in lower and higher values (a trend line for bias) and it is increased with increasing/decreasing total MET-time values. This discrepancy infers that activities with moderate intensity are being measured more accurately compared to the light/vigorous intensity activities. The results of our study are consistent with those found in the previous study.¹³

Figure 2. Bland-Altman plot: Difference in MET-time scores from the activity scale versus MET-time scores from the IPAQ (y-axis), plotted against the average MET-time scores of the two measures (x-axis). The central horizontal line represents the mean difference and the flanking lines represent the 95% limits of agreement. Mean difference: -0.02±2, SD = -0.2 to 0.24.

In another aspect of the results of the present study, one may note that the mean difference between the two methods was small, but the 95% limits of the agreement in the mean scores of MET-time values between the two self-reported questionnaires were wide and affected by four outliers. This would be an evidence for bias between the activity scales and, therefore, the agreement between the instruments may not be a gratifying result. Although the objective methods may provide more accurate information, it is not feasible to be used in population-base surveys.^10,11 Additional validation studies of the self-reported PA questionnaire against objective methods is recommended in pilot studies with smaller sample size. In the rest of the cohort profile, exploring the predictive validity of the self-reported PA questionnaire is suggested. This exploration may be conducted through examining the relationships among self-reported time spent to PA and several health outcomes related to chronic diseases, such as blood pressure (BP) and High-density lipoprotein (HDL) and low-density lipoprotein (LDL) cholesterol.

As another finding, when the known group validity analysis was conducted, the significant differences of total MET-time scores were observed in certain subgroups (e.g., marital, education and employment status) suggesting an acceptable achieved discriminative validity for the self-reported PA questionnaire. However, no significant difference was found in the activity score by age, gender and BMI, which was similar with those found in the previous studies.^28,29

The known group comparisons were not assessed in the level of subdomains which may be a reason for the non-discrepancy found in the results. As an evidence for this claim, the previous studies showed known groups discrepancy by relating subgroups to each domain of self-report PA.^15,30 Despite these issues, it may indicate that the validity of the self-reported PA questionnaire is not influenced by age, gender, and BMI. As the participants were already engaged in the AZAR Cohort study (with Azari culture), this may limit the generalizability of our results to the general Iranian adults. However, the significant association of the MET-time score was not affected by age, gender, and BMI groups in the sample this limitation to be slight. Further research is warranted in a variety of settings, as there was not found any validation studies evaluating the self-reported PA questionnaire among adults with a cultural adaptation in Iran.

Conclusion

The results of the present study suggested that the self-reported PA questionnaire has adequate psychometric properties for assessing PA in Khameneh adults. The modest reliability found for the instrument suggests that the self-reported PA questionnaire is internally consistent, stable, and valid. Although this instrument was applied in an Azari population in Iran, additional studies will be needed to better comprehend the psychometric properties of the scale among different populations.

Acknowledgments

This research was conducted under a thesis grant for the Master’s degree from the Department of Statistics & Epidemiology, Tabriz University of Medical Sciences. We gratefully acknowledge the researchers of the AZAR Cohort Study for their close collaboration during the research process.

Ethical approval

Ethics Committee in Tabriz University of Medical Sciences provided permission to conduct this survey. Signed informed consent was obtained from all participants of the AZAR Cohort Study prior to the data collection.

Competing interests

The authors declare that there is no conflict of interest.

Authors’ contributions

SD and NA-S contributed in original idea and protocol, conception of the work, conducting the study, revising the draft, approval of the final version of the manuscript, and agreed for all aspects of the work. MA-J contributed in the design of the work, doing the analysis, revising the draft and approval of the final version of the manuscript. FBA contributed in conception of the work and approval of the final version of the manuscript. MM contributed in conception of the work, conducting the study, wrote and editing of this manuscript.

References

1. Barwais FA, Cuddihy TF, Washington T, Tomson LM, Brymer E. Development and validation of a new self-report instrument for measuring sedentary behaviors and light-intensity physical activity in adults. J Phys Act Health 2014;1:1097-104. doi: 10.1123/jpah.2012-0118. [Crossref]
2. Warburton D, Charlesworth S, Ivey A, Nettlefold L, Bredin S. A systematic review of the evidence for Canada’s Physical Activity Guidelines for Adults. Int J Behav Nutr Phys Act 2010;7:39. doi: 10.1186/1479-5868-7-39. [Crossref]
3. Gomersall SR, Olds TS, Ridley K. Development and evaluation of an adult use-of-time instrument with an energy expenditure focus. J Sci Med Sport 2011;14:143-8. doi: 10.1016/j.jsams.2010.08.006. [Crossref]
4. Jurj AL, Wen W, Xiang YB, Matthews CE, Liu D, Zheng W, et al. Reproducibility and validity of the Shanghai Men’s Health Study physical activity questionnaire. Am J Epidemiol 2007;165:1124-33. doi: 10.1093/aje/kwk119. [Crossref]
5. Hu B, Lin L, Zhuang M, Yuan Z, Li S, Yang Y, et al. Reliability and relative validity of three physical activity questionnaires in Taizhou population of China: the Taizhou Longitudinal Study. Public Health 2015;129:1211-7. doi: 10.1016/j.puhe.2015.03.024. [Crossref]
6. Ishii K, Shibata A, Oka K. Environmental, psychological, and social influences on physical activity among Japanese adults: structural equation modelling. Int J Behav Nutr Phys Act 2010;7:61. doi: 10.1186/1479-5868-7-61. [Crossref]
7. Laporte RE, Montoye HJ, Caspersen CJ. Assessment of physical activity in epidemiologic research: problems and prospects. Public Health Rep 1985;100:131-46.
8. Hagströmer M, Oja P, Sjöström M. The International Physical Activity Questionnaire (IPAQ): a study of concurrent and construct validity. Public Health Nutr 2006;9:755-62. doi: 10.1079/PHN2005898. [Crossref]
9. Pettee GK, McClain JJ, Lee CD, Swan PD, Alvar BA, Mitros MR, et al. Evaluation of physical activity measures used in middle-aged women. Med Sci Sport Exer 2009;41:1403-12.doi: 10.1249/MSS.0b013e31819b2482. [Crossref]
10. Martínez-González MA, López-Fontana C, Varo JJ, Sánchez-Villegas A, Martinez JA. Validation of the Spanish version of the physical activity questionnaire used in the Nurses’ Health Study and the Health Professionals’ Follow-up Study. Public Health Nutr 2005;8:920-7. doi: 10.1079/PHN2005745. [Crossref]
11. Prince SA, Adamo KB, Hamel ME, Hardt J, Gorber SC, Tremblay M. A comparison of direct versus self-report measures for assessing physical activity in adults: a systematic review. Int J Behav Nutr Phys Act 2008;5:56. doi: 10.1186/1479-5868-5-56. [Crossref]
12. Aadahl M, Kjær M, Kristensen JH, Mollerup B, Jørgensen T. Self-reported physical activity compared with maximal oxygen uptake in adults. Eur J Cardiovasc Prev Rehabil 2007;14:422-8. doi: 10.1097/HJR.0b013e3280128d00. [Crossref]
13. Aadahl M, Jørgensen T. Validation of a new self-report instrument for measuring physical activity. Med Sci Sports Exer 2003;35:1196-202. doi: 10.1249/01.MSS.0000074446.02192.14. [Crossref]
14. Kelishadi R, Rabiee K, Khosravi A, Famori F, Sadeghi M, Roohafza H. Assessment of physical activity in adolescents of Isfahan. J Shahrekord Uni Med Sci 2004;3:55-65. [In Persian].
15. Kelishadi R, Ardalan G, Gheiratmand R, Gouya MM, Razaghi EM, Delavari A, et al. Association of physical activity and dietary behaviours in relation to the body mass index in a national sample of Iranian children and adolescents: CASPIAN Study. Bull World Health Organ 2007;85:19-26. doi: 10.1590/S0042-96862007000100008. [Crossref]
16. Moghaddam MH, Aghdam FB, Jafarabadi MA, Allahverdipour H, Nikookheslat SD, Safarpour S. The Iranian Version of International Physical Activity Questionnaire (IPAQ) in Iran: content and construct validity, factor structure, internal consistency and stability. World Appl Sci 2012;18:1073-80. doi: 10.5829/idosi.wasj.2012.18.08.754. [Crossref]
17. Bartko JJ. Intraclass correlation coef ficient as a measure of reliability. Psychol Rep 1966;19(1):3-11.
18. Gudbergsen H, Bartels EM, Krusager P, Wæhrens EE, Christensen R, Danneskiold-Samsøe B, et al. Test-retest of computerized health status questionnaires frequently used in the monitoring of knee osteoarthritis: a randomized crossover trial. BMC Musculoskelet Disord 2011;12:190.doi: 10.1186/1471-2474-12-190. [Crossref]
19. Rencher AC. Methods of Multivariate Analysis. New York: John Wiley & Sons, Inc; 2003.
20. Tinsley HE, Brown SD. Handbook of Applied Multivariate Statistics and Mathematical Modeling. New York: Academic Press; 2000.
21. Dishman RK, Saunders RP, McIver KL, Dowda M, Pate RR. Construct validity of selected measures of physical activity beliefs and motives in fifth and sixth grade boys and girls. J Pediatr Psychol 2013;38:563-76. doi: 10.1093/jpepsy/jst013. [Crossref]
22. Oyeyemi AL, Bello UM, Philemon ST, Aliyu HN, Majidadi RW, Oyeyemi AY. Examining the reliability and validity of a modified version of the International Physical Activity Questionnaire, long form (IPAQ-LF) in Nigeria: a cross-sectional study. BMJ Open 2014;4:e005820. doi: 10.1136/bmjopen-2014-005820. [Crossref]
23. Van Holle V, De Bourdeaudhuij I, Deforche B, Van Cauwenberg J, Van Dyck D. Assessment of physical activity in older Belgian adults: validity and reliability of an adapted interview version of the long International Physical Activity Questionnaire (IPAQ-L). BMC Public Health 2015;15:433. doi: 10.1186/s12889-015-1785-3. [Crossref]
24. DeVellis RF. Scale Development: Theory and Applications. 2nd ed. Thousand Oaks, CA: Sage; 2003.
25. Sallis JF, Saelens BE. Assessment of physical activity by self-report: status, limitations, and future directions. Res Q Exercise Sport 2000;71(2 Suppl):1-14. doi: 10.1080/02701367.2000.11082780. [Crossref]
26. Orsini N, Bellocco R, Bottai M, Hagströmer M, Sjöström M, Pagano M, et al. Validity of self-reported total physical activity questionnaire among older women. Eur J Epidemiol 2008;23:661-7. doi: 10.1007/s10654-008-9273-z. [Crossref]
27. Bland JM, Altman DG. Measuring agreement in method comparison studies. Stat Methods Med Res 1999;8(2):135-60.
28. Harada ND, Chiu V, King AC, Stewart AL. An evaluation of three self-report physical activity instruments for older adults. Med Sci Sport Exer 2001;33:962-970. doi: 10.1097/00005768-200105001-01412. [Crossref]
29. Matthiessen J, Biltoft-Jensen A, Rasmussen LB, Hels O, Fagt S, Groth MV. Comparison of the Danish physical activity questionnaire with a validated position and motion instrument. Eur J Epidemiol 2008;23:311-322. doi: 10.1007/s10654-008-9228-4. [Crossref]
30. Vasheghani-Farahani A, Tahmasbi M, Asheri H, Ashraf H, Nedjat S, Kordi R. The Persian, last 7-day, long form of the International Physical Activity Questionnaire: translation and validation study. Asian J Sports Med 2011;2:106-16.doi: 10.5812/asjsm.34781. [Crossref]