Share this article on:

Validity of the International Physical Activity Questionnaire in the Arctic


Medicine & Science in Sports & Exercise: April 2013 - Volume 45 - Issue 4 - p 728–736
doi: 10.1249/MSS.0b013e31827a6b40

Purpose: Information about physical activity (PA) in Greenland is limited, partly because of a lack of validated instruments in countries with non-Western living conditions. We modified the long form of the International Physical Activity Questionnaire (IPAQ-L) to arctic living conditions. The aim of the study was to compare IPAQ-L estimates with combined accelerometry and heart rate monitoring (ACC + HR) in a population-based study of adult Inuit in Greenland.

Methods: Cross-sectional data were collected by face-to-face interview and ACC + HR monitoring among Inuit (18 yr and above) in Greenland during 2005–2010 (n = 1508). PA energy expenditure (PAEE) and time spent sedentary and on PA at moderate and vigorous intensity were derived from IPAQ-L and ACC + HR. Estimates were compared using Bland–Altman agreement analysis and Spearman correlations stratified by sex, place of residence (capital, towns, and villages), and age groups.

Results: Questionnaire-based PAEE was moderately correlated with objectively measured PAEE (r = 0.20–0.35, P < 0.01). Self-reported time spent at moderate- and vigorous-intensity PA and time spent sedentary were weakly correlated with the objective measure (r = 0.11–0.31). Agreement analyses showed relatively small median differences for all measures of PA; however, time spent at moderate-intensity PA was substantially overreported by IPAQ-L when including walking (>1.5 h·d−1, P < 0.001) but not when excluding walking.

Conclusions: The IPAQ-L adapted to arctic living conditions in Greenland had a moderate level of agreement with combined accelerometry and heart rate monitoring for total PAEE at population level, but it was less valid to measure different intensities of PA and sedentary activity. Validity did not differ markedly between rural and urban communities.

1National Institute of Public Health, University of Southern Denmark, Copenhagen, DENMARK; 2Steno Diabetes Center, Gentofte, DENMARK; and 3MRC Epidemiology Unit, Institute of Metabolic Science, Cambridge, UNITED KINGDOM

Address for correspondence: Inger Katrine Dahl-Petersen, MSPH National Institute of Public Health, University of Southern Denmark, Øster Farimagsgade 5A,2, DK 1353 Copenhagen K, Denmark; E-mail:

Submitted for publication August 2012.

Accepted for publication October 2012.

Inuit in Greenland have experienced a substantial increase in chronic lifestyle diseases such as type 2 diabetes along with the rapid cultural and social transition over the last 50 yr (5,22). Differences are found in physical activity (PA) patterns, suggesting a less physically active lifestyle in relation to the social change (13). Knowledge about the level of PA still remains limited mainly because of lack of validated instruments to assess everyday life PA in countries with non-Western living conditions as in Greenland. A need for further research in different cultural settings has been suggested (11). Measurement of PA by questionnaire is still the most commonly used method at population level because it is inexpensive, is feasible to use in large populations, and can provide information on PA patterns. The International Physical Activity Questionnaire (IPAQ) is a questionnaire developed for measuring PA in different cultural settings and is the most frequently used (34). It exists in a short (IPAQ-S) and a long (IPAQ-L) form. The short form is recommended for national monitoring (seven items), whereas the long version is more comprehensive (27 items) and assesses time spent at different intensities of PA within four domains of daily life: transportation, work, leisure time, and domestic activities (16). Both forms have been tested for reliability and validity in adult populations in various countries against accelerometers (7,11,17–19,24) and have shown fair to moderate validity, although lower in rural areas (23). To our knowledge, only the short form has been used in the Arctic and found to be significantly correlated to body fat and waist circumference among Liyiyiuch, a Cree community in Canada (14). We planned to modify IPAQ-L to arctic living conditions in Greenland and to compare its main PA variables with objective estimates from combined accelerometry and heart rate monitoring (ACC + HR). Estimates of PA energy expenditure (PAEE) from ACC + HR compare favorably to doubly labeled water (DLW) estimates of PAEE in free-living adults in urban and rural population samples (2) and also provide valid estimates of PA intensity (12,32,33). The aim of the study was to assess the validity of IPAQ-L against combined accelerometry and heart rate monitoring in both rural and urban communities in a country undergoing rapid social transition.

Back to Top | Article Outline


Back to Top | Article Outline

Study population.

Data for this population-based, cross-sectional study were collected in Greenland during 2005–2010. The total population of Greenland is 57,000, of which 90% are Inuit. Twenty-two communities, the capital (Nuuk, pop = 16,181), eight smaller towns (pop = 469–5571), and 13 villages (pop = 7–513) (25% of all communities) were selected as study areas being representative of each region in Greenland. Nuuk represents the most Westernized living conditions compared with smaller towns and villages. From capital and towns, random population samples were drawn from the central population register. From villages, all adults were invited to participate. Pregnant women, individuals not born in Greenland or Denmark, and individuals who had moved out of the study area at the time of the study were excluded from the population sample. We confined the study to Inuit as defined by the participant and the interviewer based on language and self-perceived ethnicity at the time of enrolment in the health examination. In total, 2874 adult Inuit age 18 yr and older participated in a clinical examination and were interviewed. A detailed description of the methods is available elsewhere (4). The study was approved by the ethical review committee for Greenland. Written informed consent was obtained from all participants.

Back to Top | Article Outline

Procedures of health examination.

At the day of the health examination, face-to-face interviews were conducted by trained native Greenlandic-speaking interviewers in the language chosen by the participant (Greenlandic or Danish). Information on sociodemographic factors and lifestyle, including PA in the form of IPAQ-L, was obtained during the interview. Height (nearest 0.1 cm) and weight (nearest 0.1 kg) were measured.

Back to Top | Article Outline

Self-reported PA.

Information on PA was collected using a modification of the interviewer-administered IPAQ-L. Participants indicated time spent on PA in the previous 7 d: how often (the number of days per week) and for how long (the average duration per day) separately for vigorous intensity, moderate intensity, and walking in the four domains (work, transportation, domestic, and leisure time). The original English version of the PA questionnaire was translated into Greenlandic and back-translated by two translators bilingual in Danish and Greenlandic and familiar with Greenlandic living conditions. The questions were adjusted to arctic living conditions by replacing some of the activity examples by culturally relevant examples. In the domestic domain, we combined the two questions concerning moderate intensity (outside and inside activity) into one; gardening is nonexistent in arctic living conditions, and common activities such as getting fishing equipment ready are done both inside and outside the house. We also did a brief interview with five of the main interviewers about their experience with interpretation of the questions.

Back to Top | Article Outline

Combined accelerometry and heart rate monitoring.

A combined accelerometer and heart rate monitor (ACC + HR) (Actiheart®; CamNtech Ltd, Cambridge, UK), described in technical detail elsewhere (8), was provided to a subgroup of the participants (n = 2055). The monitor was set up to measure acceleration and heart rate in 30-s intervals and attached to the participant’s chest by two standard ECG electrodes (MXC55; MediMax, Edison, NJ). The participant was instructed to wear the monitor for 24 h·d−1 for at least 2 d and preferably 4 d from the day of the health examination. Because of study logistics, only a limited time was available at each study location, especially for data collection in villages. Together with a finite stock of monitors, this explains why not all participants were given a monitor and why the length of recordings from some participants was shorter. A subgroup of participants (n = 135) conducted an individual calibration test (8-min step test) as described previously (10). The step tests were used to define a population-specific calibration equation of the heart rate–activity energy expenditure relationship.

Back to Top | Article Outline

IPAQ-L data processing.

Data on PA from the modified IPAQ-L were initially scored according to guidelines from the IPAQ group (16). All participants who reported days (frequency) but not time (duration) of PA or vice versa were treated as missing. Total daily PA of more than 960 min (16 h) was scaled linearly (sedentary time not included). In each domain, minutes spent on PA at moderate and vigorous intensity and walking for more than 180 min·d−1 (at each activity) were truncated to 180 min. Reported time spent on PA between 0 and 10 min was accepted even though the questionnaire was restricted to activity of at least 10-min duration. In a review of Murphy et al. (27), most of the studies did not find alterations between accumulated and continuous patterns of exercise, and it was concluded that further research was required to determine whether even shorter bouts (<10 min) could provide a health benefit. Time spent at moderate- and vigorous-intensity PA and walking and total time spent in each domain were calculated. PAEE on each activity was calculated by multiplying time reported (min·wk−1) by the net metabolic cost of each activity, which was expressed in METs. Net metabolic cost of each activity was assigned according to the physical activity compendium’s gross MET values (1), subtracted by 1 MET to account for resting metabolic rate (RMR). An estimate of total daily sedentary time was calculated from time spent sitting (including activities, such as TV and computer use and reading), to which we added 8 h as presumed time spent sleeping (not included in IPAQ-L). Time spent at moderate-intensity activity was analyzed with and without the inclusion of walking.

Back to Top | Article Outline

Accelerometry and heart rate data processing.

Data from ACC + HR monitoring were manually trimmed to indicate the end of each participant’s recording, after which heart rate data were preprocessed using robust Gaussian process regression for inference of latent heart rate trace as described elsewhere (31). The combination of prolonged periods of large heart rate uncertainty (disturbances in heart rate signal mostly related to the wear of the monitor and eventually interference by cloth or electrode detachment) accompanied by no acceleration was used to classify all measured time points as wear or nonwear. We included individuals with >48 h of monitor wear data. Caloric intensity of PA was estimated by combining the acceleration-based estimate of intensity (10) with the heart rate-based estimate from the population-specific equation (see previous discussions) in a branched equation modeling framework (9). Briefly, this method predominantly uses the accelerometer estimate during low levels of heart rate and movement and the heart rate estimate when both heart rate and acceleration levels are high, with equal weighting for other conditions. Resulting time series of activity intensity (J·min−1·kg−1) were summarized into total PAEE (kJ·kg−1·d−1) and time spent at different intensity levels (sedentary as <1.5 MET, moderate as 3–6 MET, and vigorous as >6 MET) while minimizing diurnal bias from potentially unbalanced data accumulated over the day. Intensity categories were defined using multiples of 1 MET as derived using the Oxford equations for resting metabolic rate (21).

Back to Top | Article Outline

Statistical methods.

Descriptive characteristics of the study sample are presented as means with SDs for normally distributed continuous variables and medians with interquartile ranges for nonnormal distributed variables. Results are stratified by sex, age groups, and place of residence. Differences between men and women were tested by t-test for continuous data and chi-square test for categorical data. The linear association between self-reported and objective activity estimates was examined by the nonparametric Spearman rank correlation coefficient (ρ). Level of agreement was examined by modified Bland–Altman plots (6). Median difference between the measurements (IPAQ-L minus ACC + HR) was plotted against the objective estimate, with lines indicating the median difference (median bias) and 95% limits of agreement (2.5 and 97.5 centiles) (nonparametric data). The differences of the medians were analyzed by Wilcoxon signed-rank test. Sensitivity analyses were performed including only participants with ACC + HR monitoring of ≥72 h. Gaussian process regression of heart rate was performed in JAVA using a MySQL database, and all other analyses were carried out in STATA version 11.

Back to Top | Article Outline


The study population consisted of 2874 Inuit adults. After data processing, valid data from the IPAQ-L were obtained from 2798 participants (97.4%), of whom 1999 (71.4%) had worn an ACC + HR monitor. After excluding recordings with insufficient valid PA data (<48 h), 1508 participants with complete data from both IPAQ-L and ACC + HR were available for analysis. The proportion of men and women that was monitored by ACC + HR for more than 48 h did not differ from those not monitored (men, 44% vs 45%; women, 56% vs 55%; P = 0.6); however, a smaller proportion of participants from villages and towns was monitored by ACC + HR (Nuuk, 21% vs 8%; towns, 61% vs 70%; villages, 18% vs 23%; P < 0.001). Moreover, a smaller proportion of participants age 70 yr or older was monitored by ACC + HR, compared with those not monitored (4% vs 11%, P < 0.001). The characteristics of the study population are displayed in Table 1. Men (range, 18–84 yr old) were slightly older than women (range, 18–85 yr old), and women had higher body mass index (26.5 kg·m−2) than men (25.8 kg·m−2) (P = 0.02). One fifth of the participants lived in the capital, Nuuk, and more than half in other towns.

The median PAEE estimated by the IPAQ-L was almost similar compared with the median estimate from ACC + HR for all subgroups, although significant differences were found for women, for the participants in villages, and in the age group of 45–54 yr (Table 2). Self-reported PAEE (kJ·d−1·kg−1) was moderately, although significantly, correlated with objectively measured PAEE in analyses stratified by sex, age groups, and place of residence (r = 0.20–0.35, P < 0.001) (Table 3). The level of agreement between the two methods for measuring PAEE stratified by sex and residence is illustrated in Figure 1. The median bias for PAEE was small and indicated a fair agreement between the two methods at population level; however, large individual differences in PAEE were found (Table 2). The same tendencies were found for all subgroups (data not shown).

Self-reported time spent on PA at moderate intensity (MPA) was significantly lower than objectively assessed MPA when walking was not included in the estimation but was significantly higher when walking was included (Table 2); correlations across strata of sex, age groups, and residence, however, were similar, with a tendency toward higher correlations when walking was included (Table 3). Figure 2 illustrates MPA including walking. The median bias was large, and the 95% limits of agreement indicated high individual variability for both men and women in the different subgroups. The asymmetry of the 95% limits of agreement around the median bias highlights the substantial overestimation of MPA by self-report. The same pattern was found for all subgroups (data not shown). In contrast, time spent at moderate-intensity PA without walking included was only slightly overestimated by IPAQ-L for women and for participants in villages and slightly underestimated for the rest of the subgroups. For all subgroups, a significant difference between median time spent on moderate-intensity PA measured by IPAQ-L and ACC + HR was found except for participants age 45 yr and older (Table 2). Light-intensity PA is not included in IPAQ-L, so estimates areonly available from ACC + HR data. More than 50% of the population spent 6 h daily or more on light-intensity PA (1.5–3 METs), which made up 54.4% (interquartile range, 44.9–62.8) of total PAEE (data not shown).

The median duration of vigorous-intensity PA measured by IPAQ-L and ACC + HR monitoring differed significantly for all subgroups except for women and for participants living in Nuuk and in towns (Table 2). Men reported more time spent at vigorous-intensity PA and women reported less vigorous PA as compared with ACC + HR measurements. Time spent on PA at vigorous intensity estimated from IPAQ-L was significantly correlated with the ACC + HR measurement in all subgroups, but the correlation was generally weak (r = 0.11–0.27, P < 0.05) (Table 3). Both IPAQ-L and ACC + HR showed that more than 50% of the population spent less than 10 min daily at vigorous-intensity PA. A poor and nonsignificant correlation was found for sedentary activity measured by self-report and ACC + HR for all subgroups, and the median differed significantly for all subgroups (Tables 2 and 3). Sedentary behavior was highly underestimated by the questions in IPAQ-L, even after adding 8 h of presumed time spent sleeping.

We repeated the analysis for the subgroup of participants who wore the ACC + HR monitor for ≥72 h and found similar results (data not shown).

Back to Top | Article Outline


We found moderate validity for questionnaire-based overall PAEE and weak to moderate validity for different intensities of PA and sedentary time compared with ACC + HR monitoring stratified by sex, age groups, and residence. The Bland–Altman plots showed relatively small median differences for all variables of PA; however, the individual variability in PA measures was high.

Studies testing the validity of the IPAQ-L questionnaire against different criterion measures have shown different levels of correlation. Craig et al. (11) validated IPAQ-L in 12 different cultural settings with accelerometry and found a rank correlation of around 0.30 for overall PA, although this varied greatly between study sites, from −0.27 to 0.61. Similarly, a study examining the Chinese version of IPAQ-L reported a correlation of 0.35 for overall PA against accelerometry (25). Only one study has validated the long form of IPAQ-L using DLW, which is considered as the gold standard for total energy expenditure during free living and found a correlation of 0.38 (26). Taken together, these previous validation results are more or less in line with our correlation coefficients for overall PA (0.20–0.35).

We found that self-reported PA at different intensities was more weakly correlated with ACC + HR estimates than overall PA. The findings from other studies are not clear (7,18); however, two studies among non-European populations (19,28) have shown comparably low correlations as in our study for moderate and vigorous-intensity PA. It is suggested that cultural differences may affect the interpretation of the intensity of the activity. Median self-reported time spent at vigorous-intensity PA was found to be substantially higher (more than fourfold) compared with objective estimates among participants living in villages. Accordingly, we stratified the analysis by job status and found that hunters and fishermen substantially overreported vigorous-intensity PA (median: IPAQ-L, 34 min·d−1, vs ACC + HR, 8 min·d−1). Traditional activities, such as hunting and fishing, may be more difficult to recall and classify into moderate or vigorous intensity because these activities do not have a regular time schedule and vary in intensity. According to the interviewers in our study, traditional activities such as hunting and fishing could be misinterpreted as a vigorous-intensity activity because of its psychological exhausting and time-consuming character and because of demanding climate conditions. Going hunting is generally considered to be physically demanding; however, hunting often includes periods with waiting time.

Despite the substantial overreporting of time spent at moderate-intensity PA (walking included) by IPAQ-L, the Bland–Altman plots showed only small median bias for overall PAEE measured by self-report and ACC + HR monitoring, which suggests that IPAQ-L is a valid measure for overall PAEE at the group level; however, it is less valid when measuring different intensities of PA. Light-intensity PA was only measured by ACC + HR monitoring but contributed substantially to daily PA. Our findings are contrary to other studies that have demonstrated that PAEE on the group level is overestimated by IPAQ-L, a bias for which social desirability has been suggested as a plausible explanation (7,18). The attention from the media on the positive health effect of PA might have been less marked in Greenland compared with more Westernized countries, and thus, the risk of social desirability bias may be somewhat lower.

Back to Top | Article Outline

Moderate intensity with and without walking.

We found that IPAQ-L substantially overestimated moderate-intensity PA when walking was included as a moderate-intensity PA. Ekelund et al. (15) found in a study of IPAQ-S that walking might be difficult to accurately quantify. In IPAQ-L, walking activity is asked for in all domains of the questionnaire, which might increase the risk of reporting the same walking activity twice. According to guidelines from the IPAQ group, walking is defined as moderate-intensity PA and assigned the MET value of 3.3 METs; however, in the compendium of Ainsworth et al. is listed various intensities of walking corresponding to different MET values (2.0 to 8.0 METs) (1). One could argue that a slow pace of walking corresponds to light intensity and not moderateintensity PA. Moreover, the IPAQ-L does not ask for light-intensity activities, which may result in participants classifying light-intensity PA as moderate-intensity PA. A qualitative interview with the interviewers in our study pointed out the difficulties in estimating total time spent on moderate-intensity PA and that standing activity, such as teaching or working in a shop, was sometimes misinterpreted as walking activity; this may add to explain the high amount of moderate-intensity activity reported in our study. Accordingly, we did the analyses of moderate-intensity activity with and without walking and walking considered as a light-intensity activity (data not shown) and found a substantially higher level of agreement between the two methods when walking was not included as MPA.

Back to Top | Article Outline

Sedentary time.

Knowledge about the health risks of sedentary behavior is increasing (20,35). The question about sitting in IPAQ-L has demonstrated acceptable validity and reliability and has been used to compare the prevalence of sitting time in an international study in 20 countries (3,11). We added 8 h of presumed sleeping time to sedentary time estimated from IPAQ-L and compared it with time spent on activity of less than 1.5 MET from the day-and-night ACC + HR recordings. We found a substantial underestimation of sedentary behavior by IPAQ-L, which could be explained by the fact that frequent activities such as standing and lying (and sleep) are estimated by ACC + HR monitoring as sedentary activity (<1.5 METs), whereas even though the question of sitting time in the IPAQ-L includes some aspects of lying, it may not capture all sedentary activities in daily life. Moreover, we estimated time spent sleeping to 8 h in the IPAQ-L processing, and individual variations in sleeping time could be another explanation.

Back to Top | Article Outline

Study population.

The present study is a population-based study including a representative sample of Inuit in Greenland. For logistic reasons, not all participants were given a monitor or fulfilled the criteria of ACC + HR monitoring for 48 h or more (65.4%). Nevertheless, we found only small differences in age, sex, and residence between our smaller study sample and the entire study population, which imply that the results of this study are applicaple to the population of Greenland. In the present study, we did not find significant differences in validity between rural and urban communities, defined by living in a village, a town, or in Nuuk. However, traditional rural activities such as hunting and fishing might influence the interpretation of the different intensities of PA.

Back to Top | Article Outline

Strengths and limitations.

The gold standard for measuring PAEE in free-living individuals is the DLW method, combined with a measure of RMR. However, this method is expensive and cannot provide information about the intensity, frequency, and patterns of PA. Studies in non-Western countries have shown that it is particularly important to monitor both HR and movement in the estimation of PAEE in rural populations because of a higher number of activities that cannot be fully measured by a classic uniaxial accelerometer, such as digging and heavy lifting, which are activities comparable with traditional activities in Greenland. We therefore consider the use of combined accelerometry and heart rate monitoring for estimating PAEE in this study as a strength (2,8,12), although the lack of dynamic individual calibration in everybody is a potential weakness (10). A status report on the assessment of PA by self-report finds the use of an interviewer-administrated questionnaire to increase the validity of the responses (30). According to the interviewers in our study, some participants found the interpretation of moderate and vigorous-intensity PA difficult. Furthermore, because of the wide differences in living conditions, climatic differences, and dialects across the country, the interviewers had to pay particular attention to the choice of words and the examples given of different activities. Therefore, the use of face-to-face interviews undertaken by Greenlandic interviewers bilingual in Danish and Greenlandic is a strength in this study. An important observation in the translation of the questionnaire into Greenlandic was that no word exists for PA. PA is translated to “use of the body” and that may, in a higher degree, refer to sports activities instead of activities of daily living. However, we did not find this misclassification likely because we would have assumed a substantial degree of underreporting.

Our study has some potential limitations. First, we did not conduct any repeated administrations of IPAQ-L because of logistical reasons in this comprehensive data collection process. Knowledge about reliability is an important metric of any instrument, and it is recommended that future research on measurements of PA in the Arctic include a test–retest element.

Second, the administration of the two instruments meant that they did not refer to the same period. The monitor was given to the participants on the day they were interviewed about their PA in the preceding 7 d. However, the short interval between the periods is unlikely to have introduced substantial bias in the results, and one may even consider the present results to reflect more truly the convergent validity of these instruments to assess habitual PA. Third, the estimates from the IPAQ-L were calculated as the average of the previous 7 d (no information on sleeping hours but 8 h were estimated for sleep), whereas estimates from the ACC + HR monitor were calculated as an average of 2–5 d including nights. ACC + HR monitoring from both week and weekend day were obtained from 965 (67%) of the participants. Rennie and Wareham (29) estimated that 3 d of recording yielded a validity coefficient at 0.85 for the assessment of energy expenditure. In our study, that was the case for 810 (56.3%) of the participants. Ideally, 7 d of objective recording would have been preferable to capture variations in PA during the week, but this was not feasible. However, our sensitivity analyses that included only participants with more than 72 h of recording showed very similiar results as the primary analyses, which suggests that the law of diminishing returns may govern these behavioral data.

Although objectively measured PA is considered a more valid measure of PA, the IPAQ-L has the advantage of measuring domains of daily life PA, which are important, particularly in non-European countries where being physically active at work and at home is more common compared with leisure time PA. The domains in the IPAQ-L provide the opportunity to track changes in PA patterns along with social changes, e.g., whether decreasing occupational PA is compensated by an increase in leisure time PA, not necessarily changing the total amount of PA; therefore, objective and subjective measures complement each other.

Back to Top | Article Outline


The long version of IPAQ modified to arctic living conditions is a valid measure for overall PAEE among adult Inuit in Greenland at population level; it is, however, less valid to measure different intensities of PA and sedentary activities. Time spent at moderate-intensity PA was substantially overreported by IPAQ-L when walking was included in this category. The validity did not differ significantly between rural and urban communities. Using IPAQ-L at the individual level will be subject to a high degree of uncertainty.

Studies on how culture, social norms, and language affect the interpretation of PA questions are important to improve the validity of the IPAQ-L in non-Western countries.

This study was funded by the Karen Elise Jensen Foundation, Denmark. The authors are grateful to the participants and the participating communities. The authors would also like to thank Kate Westgate and Stefanie Mayle at the MRC Epidemiology Unit, Cambridge, United Kingdom, for assistance in data processing.

The authors declare that there are no conflicts of interest.

The results of the present study do not constitute endorsement by the American College of Sports Medicine.

Back to Top | Article Outline


1. Ainsworth BE, Haskell WL, Herrmann SD, et al.. 2011 Compendium of Physical Activities: a second update of codes and MET values. Med Sci Sports Exerc. 2011; 43 (8): 1575–81.
2. Assah FK, Ekelund U, Brage S, Wright A, Mbanya JC, Wareham NJ. Accuracy and validity of a combined heart rate and motion sensor for the measurement of free-living physical activity energy expenditure in adults in Cameroon. Int J Epidemiol. 2011; 40 (1): 112–20.
3. Bauman A, Ainsworth BE, Sallis JF, et al.. The descriptive epidemiology of sitting a 20-country comparison using the International Physical Activity Questionnaire (IPAQ). Am J Prev Med. 2011; 41 (2): 228–35.
4. Bjerregaard P. Inuit health in transition. Greenland survey 2005–2009. Population sample and survey methods. In: National Institute of Public Health. 2009 (cited 19 October 2012):1–13 Available from:
5. Bjerregaard P, Young TK, Dewailly E, Ebbesson SO. Indigenous health in the Arctic: an overview of the circumpolar Inuit population. Scand J Public Health. 2004; 32 (5): 390–5.
6. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet. 1986; 327 (8476): 307–10.
7. Boon RM, Hamlin MJ, Steel GD, Ross JJ. Validation of the New Zealand Physical Activity Questionnaire (NZPAQ-LF) and the International Physical Activity Questionnaire (IPAQ-LF) with Accelerometry. Br J Sports Med. 2010; 44 (10): 741–6.
8. Brage S, Brage N, Franks PW, Ekelund U, Wareham NJ. Reliability and validity of the combined heart rate and movement sensor Actiheart. Eur J Clin Nutr. 2005; 59: 561–70.
9. Brage S, Brage N, Franks PW, et al.. Branched equation modeling of simultaneous accelerometry and heart rate monitoring improves estimate of directly measured physical activity energy expenditure. J Appl Physiol. 2004; 96: 343–51.
10. Brage S, Ekelund U, Brage N, et al.. Hierarchy of individual calibration levels for heart rate and accelerometry to measure physical activity. J Appl Physiol. 2007; 103: 682–92.
11. Craig CL, Marshall AL, Sjostrom M, et al.. International Physical Activity Questionnaire: 12-country reliability and validity. Med Sci Sports Exerc. 2003; 35 (8): 1381–95.
12. Crouter SE, Churilla JR, Bassett DRJ. Accuracy of the Actiheart for the assessment of energy expenditure in adults. Eur J Clin Nutr. 2008; 62 (6): 704–11.
13. Dahl-Petersen IK, Joergensen ME, Bjerregaard P. Physical activity patterns in Greenland: a country in transition. Scand J Public Health. 2011; 39: 678–86.
14. Egeland GM, Denomme D, Lejeune P, Pereg D. Concurrent validity of the International Physical Activity Questionnaire (IPAQ) in an liyiyiu Aschii (Cree) community. Can J Public Health. 2008; 99 (4): 307–10.
15. Ekelund U, Sepp H, Brage S, et al.. Criterion-related validity of the last 7-day, short form of the International Physical Activity Questionnaire in Swedish adults. Public Health Nutr. 2006; 9 (2): 258–65.
16. Guidelines for data processing and analysis of IPAQ—short and long forms. Web site (Internet); (cited 19 October 2012). Available from:
17. Guthold R, Ono T, Strong KL, Chatterji S, Morabia A. Worldwide variability in physical inactivity a 51-country survey. Am J Prev Med. 2008; 34 (6): 486–94.
18. Hagstromer M, Oja P, Sjostrom M. The International Physical Activity Questionnaire (IPAQ): a study of concurrent and construct validity. Public Health Nutr. 2006; 9 (6): 755–62.
19. Hallal PC, Simoes E, Reichert FF, et al.. Validity and reliability of the Telephone-Administered International Physical Activity Questionnaire in Brazil. J Phys Act Health. 2010; 7 (3): 402–9.
20. Healy GN, Wijndaele K, Dunstan DW, et al.. Objectively measured sedentary time, physical activity, and metabolic risk: the Australian Diabetes, Obesity and Lifestyle Study (AusDiab). Diabetes Care. 2008; 31: 369–71.
21. Henry CJK. Basal metabolic rate studies in humans: measurement and development of new equations. Public Health Nutr. 2005; 8 (7A): 1133–52.
22. Jørgensen ME, Bjerregaard P, Borch-Johnsen K. Diabetes and impaired glucose tolerance among the Inuit population of Greenland. Diabetes Care. 2002; 25 (10): 1766–71.
23. Lachat CK, Verstraeten R, Khanh IN, et al.. Validity of two physical activity questionnaires (IPAQ and PAQA) for Vietnamese adolescents in rural and urban areas. Int J Behav Nutr Phys Act. 2008; 5: 37.
24. Lee PH, Macfarlane DJ, Lam TH, Stewart SM. Validity of the International Physical Activity Questionnaire short form (IPAQ-SF): a systematic review. Int J Behav Nutr Phys Act. 2011; 8: 115.
25. Macfarlane D, Chan A, Cerin E. Examining the validity and reliability of the Chinese version of the International Physical Activity Questionnaire, long form (IPAQ-LC). Public Health Nutr. 2010; 13: 1–8.
26. Maddison R, Ni MC, Jiang Y, et al.. International Physical Activity Questionnaire (IPAQ) and New Zealand Physical Activity Questionnaire (NZPAQ): a doubly labelled water validation. Int J Behav Nutr Phys Act. 2007; 4: 62.
27. Murphy MH, Blair SN, Murtagh EM. Accumulated versus continuous exercise for health benefit a review of empirical studies. Sports Med. 2009; 39 (1): 29–43.
28. Nang EEK, Ngunjiri SAG, Wu Y, et al.. Validity of the International Physical Activity Questionnaire and the Singapore prospective study program physical activity questionnaire in a multiethnic urban Asian population. BMC Med Res Methodol. 2011; 11: 241.
29. Rennie KL, Wareham NJ. The validation of physical activity instruments for measuring energy expenditure: problems and pitfalls. Public Health Nutr. 1998; 1 (4): 265–71.
30. Sallis JF, Saelens BE. Assessment of physical activity by self-report: status, limitations, and future directions. Res Q Exerc Sport. 2000; 71 (2): 1–14.
31. Stegle O, Fallert SV, MacKay DJC, Brage S. Gaussian process robust regression for noisy heart rate data. IEEE Trans Biomed Eng. 2008; 55 (9): 2143–51.
32. Strath SJ, Brage S, Ekelund U. Integration of physiological and accelerometer data to improve physical activity assessment. Med Sci. Sports Exerc. 2005; 37 (11 Suppl): S563–71.
33. Thompson D, Batterham AM, Bock S, Robson C, Stokes K. Assessment of low-to-moderate intensity physical activity thermogenesis in young adults using synchronized heart rate and accelerometry with branched-equation modeling. J Nutr. 2006; 136: 1037–42.
34. van Poppel MN, Chinapaw MJ, Mokkink LB, Van Mechelen W, Terwee CB. Physical activity questionnaires for adults. A systematic review of measurement properties. Sports Med. 2010; 40 (7): 565–600.
35. Wijndaele K, Brage S, Besson H. Television viewing time independently predicts all-cause and cardiovascular mortality: the EPIC Norfolk Study. Int J Epidemiol. 2011; 40: 150–9.


©2013The American College of Sports Medicine