Descriptive Epidemiology of Ambulatory Activity in Rural, Black South Africans


Medicine & Science in Sports & Exercise:
doi: 10.1249/MSS.0b013e3181ca787c
BASIC SCIENCES: Contrasting Perspectives

Purpose: We investigated the distribution of objectively measured ambulation levels and the association of ambulation levels to adiposity levels in a convenience sample of adolescent and adult, rural black South Africans.

Methods: We analyzed 7-d pedometry data, collected over a period of nine consecutive days, in 789 subjects (women, n = 516; men, n = 273). Adiposity measures included body mass index (BMI) and waist circumference (WC). Obesity was defined as BMI ≥ 30 kg·m−2 or WC ≥ 102 cm for men and WC ≥ 88 cm for women.

Results: The average age- and BMI-adjusted 7-d ambulation level was 12,471 steps per day (95% confidence interval (CI) = 12,107-12,834). Ambulation levels differed between sexes (P = 0.0012), and weekday ambulation differed from weekend ambulation (P = 0.0277). Prevalences, age adjusted to the world population, for sedentarism (SED; <5000 steps per day), low active-somewhat active (5000-9999 steps per day), and active-very active (ACT; ≥10,000 steps per day) were 8.0%, 25.5%, and 66.6%, respectively. In contrast, published self-reported national prevalences for physical inactivity, insufficient physical activity, and physically active have been estimated to be 43%-49%, 20%-27%, and 25%-37%, respectively. After adjusting for sex and age, adiposity measures remained significantly associated with steps per day (BMI, r = −0.08; WC, r = −0.12; P < 0.03). Adjusting for sex, age, village, and season, SED increased the risk of obesity by more than twofold compared with ACT (P < 0.05). Achieving <10,000 steps per day compared with ACT was associated with an increased multivariate-adjusted obesity risk of 86%-89% (P < 0.001).

Conclusions: Ambulation levels were high for this rural African sample, and prevalences for SED and ACT differed from published self-reported estimates.

Author Information

1Physical Activity Epidemiology Laboratory, University of Limpopo, Turfloop Campus, Polokwane, SOUTH AFRICA; and 2Department of Medical Sciences, University of Limpopo, Turfloop Campus, Polokwane, SOUTH AFRICA

Address for correspondence: Ian Cook, B.Sc. (Med.) Hons., Physical Activity Epidemiology Laboratory, University of Limpopo, Turfloop Campus, PO Box459, Fauna Park, 0787, Polokwane, South Africa; E-mail:

Submitted for publication June 2009.

Accepted for publication October 2009.

Article Outline

Globally, regionally, and nationally, physical inactivity is considered a leading risk factor for chronic diseases of lifestyle (19,23). Physical inactivity has been estimated to contribute 1.3% to global mortality, 22% to ischemic heart disease (IHD) burden, and 7% to stroke burden (19). In low- and high-mortality developing regions, physical inactivity contributes 5% and 21%-22% to stroke and IHD disease burden, respectively (19). Recently, the South African Comparative Risk Assessment Collaborating Group attributed 20%-30% of IHD, ischemic stroke, and type 2 diabetes mellitus to physical inactivity (23).

The South African prevalence of physical inactivity has been estimated to be 43%-49% (23). This estimate is based on data collected using the International Physical Activity Questionnaire (IPAQ) (15) during the 2003 World Health Study and provides the first nationally representative data set on physical inactivity in South Africa (21). However, caution has been raised (21) regarding the low validity and reliability coefficients for the IPAQ in rural areas of developing countries (15).

A recent large-scale, population-based survey of objectively measured physical activity (PA) in a highly urbanized, developed country has reported discrepancies between objective and subjective measures of PA (36). The possibility of discrepancies from a South African perspective has also been raised (11). Estimates for physical activity/inactivity using objective measures (accelerometers, pedometers) in rural, black South African women suggest higher PA volumes and lower physical inactivity prevalences than what is suggested by subjectively measured surveys (11). There is thus a need for objectively measured PA surveys to be conducted, especially in rural settings, to verify the self-reported prevalences of physical activity/inactivity for rural South African populations (15,21). Moreover, there is a dearth of objectively measured PA data from African settings (8,13,14). To date, only one objectively measured, cross-sectional survey of 7-d PA has been reported for a South African sample (13). The study reported crude PA and inactivity prevalences of 39.7% and 13.7%, respectively, and low to moderate associations between average daily step counts and adiposity measures for a sample of rural, black women (13). However, the sample size was small, limited to women, and the spring-levered pedometer used could not store daily step counts for later retrieval (13).

Therefore, the primary objective of this study was to investigate the distribution of pedometry-assessed ambulation in a larger, more representative sample of a rural black South African population, particularly concerning some objectively measured public health indices, using an accurate and robust accelerometer-based pedometer with a 7-d memory capacity. The secondary objective was to investigate the relationship between pedometry-assessed ambulatory activity and measures of adiposity.

Back to Top | Article Outline


Back to Top | Article Outline
Field site.

The Dikgale Health and Demographic Surveillance System (DHDSS) field site was established in 1995, and a yearly census has been conducted since 1996 with trained, local enumerators visiting households. The DHDSS consists of a relatively stable population of around 8000 residents. Inhabitants of this area are northern Sotho-speaking people, belonging to the Pedi group (2-4,12). The DHDSS is situated in the Limpopo Province, which constitutes 10.2% and 11.8% of the total land area and total population of South Africa, respectively (34). Nearly 90% of the Limpopo population live in nonurban areas, and 97.2% of the population are black Africans. Nationally, 42.8% live in nonurban areas, and 79% are black South Africans. The Limpopo provincial age-sex profile of the black African population yields a pattern typical for a developing population, triangular with a large base and concave sides. In addition, the contribution of the 0- to 4-yr-old age group for both sexes is contracted (34). The DHDSS has a population pyramid and age structure nearly identical with that of the Limpopo Province (12,34). Moreover, the prevalence of socioeconomic, infrastructural, and education variables shows similar trends in the DHDSS (see Results section) compared with provincial estimates (34). These data suggest that the DHDSS is representative of the Limpopo Province, specifically concerning the black, nonurban population.

Back to Top | Article Outline
Sample size.

Using data from a previous cross-sectional study (13) and assuming small effect sizes (Cohen's conventions) for correlations, independent t-tests, contingency tables, one-way ANOVA, multiple linear regression, and binary logistic regression, sample size estimations ranged from 485 to 1096 (median = 792) for α = 0.05 and 1 − β = 80% (20).

Back to Top | Article Outline

Subjects were conveniently recruited (n = 830) from households in the DHDSS between December 2005 and December 2007. Before the commencement of the survey, local community chiefs were visited to explain the purpose of the study, to request permission to recruit subjects from the communities, and to have the leaders inform the communities about the survey. A random sample of 1000 subjects was generated from the DHDSS relational database and distributed to trained fieldworkers. However, fieldworkers reported difficulty in contacting the subjects during house-to-house visits. Direct contact was necessary because provincial estimates for telephone access are 20.2% for cellular telephone only and 3.1% for landline telephone only within the dwelling (34). Because of time and financial constraints, it was decided that fieldworkers would recruit subjects house to house, at common meeting places, and through general mouth-to-mouth promotion of the survey. Signed informed consent was obtained from all participants. The study was approved by the ethics committee of the University of Limpopo (Turfloop Campus).

Back to Top | Article Outline
Study protocol.

The participants were contacted twice over a 9-d period. On the first occasion, subjects were recruited and completed the informed consent and the relevant sections of a health questionnaire and provided anthropometric data. Anthropometric measurements and interviews were performed by trained, local fieldworkers. Finally, subjects were instructed on the required procedures for wearing the pedometer. Nine days later, the pedometers were collected. Subjects received a small honorarium on completion of the study.

Back to Top | Article Outline
Anthropometric measurements.

Standard anthropometric measures of stature (nearest 1 cm), body mass (nearest 1 kg), and waist circumferences (WC; midway between the lower rib margin and the iliac crest, steel anthropometric tape, nearest 1 cm) were obtained. Obesity was defined as body mass index (BMI) ≥ 30 kg·m−2 or WC ≥ 88 cm for women WC ≥ 102 cm for men (31). We also categorized subjects using BMI categories (underweight, <18.5 kg·m−2; normal weight, 18.5-24.9 kg·m−2; overweight, 25-29.9 kg·m−2; obese, ≥30 kg·m−2) (29).

Back to Top | Article Outline
PA volume.

To objectively quantify PA volume of the subjects, the participants wore piezoelectric pedometers (NL-2000; New Lifestyles Inc., Kansas City, MO), not affected by pedometer tilt or adiposity level (16), for nine consecutive days so that when the pedometers were collected, the total number of steps during the seven complete days was stored and recalled (32). Data for day 1 and day 9 were omitted because these were incomplete days. The pedometer was worn on the right waist, securely attached to a nylon belt and sealed with surgical tape. The pedometers could be removed for sleeping and bathing purposes by unclipping the nylon belt. Ambulation PA volume was defined as the average steps per day (total steps recorded/ days pedometer worn). The average steps per day was calculated for a 7-d period (weekly: Monday to Sunday), a 5-d period (weekday: Monday to Friday), and a 2-d period (weekend: Saturday to Sunday). Public health indices (thresholds) for steps per day were defined as follows (37): sedentary, <5000 steps per day; low active, 5000-7499 steps per day; somewhat active, 7500-9999 steps per day; active, ≥10,000 steps per day; and very active, ≥12,500 steps per day. Subjects were also described as complying with public health pedometry guidelines (≥10,000 steps per day) or not (<10,000 steps per day) for each day. A summary variable was created indicating the number of days a subject was compliant (0-7 d), and from this continuous variable, a categorical compliance variable was also created (0-1, 2-3, 4-5, and 6-7 d).

Back to Top | Article Outline
Statistical analysis.

Data were expressed as mean (SD), mean (95% confidence interval (CI)), or n (%). Crude prevalences for pedometry thresholds were age standardized according to the world population (1). Where necessary, skewed data were transformed (natural logarithms). Independent t-tests and one-way ANOVA with post hoc multiple comparison analyses (Sidak's t-test) assessed group differences. To examine the ordered association between public health pedometry thresholds and categorical variables, we conducted Pearson's chi-square analyses. We performed multiple linear regression analyses with BMI and WC as dependent variables and reported the association between continuous variables with zero-order and partial correlation coefficients. Explanatory variables that were forced into the model were pedometry output, age, sex, village, and season. The mean number of days that complied with ≥10,000 steps, adjusted for age and BMI, was calculated using a univariate general linear model. Repeated-measures general linear models were used to determine steps per day across days, adjusting for age and BMI. Models were created for 7 d, five weekdays and two weekend days. Post hoc multiple comparison analyses (Sidak's t-test) assessed group differences. To investigate the risk of obesity in relation to pedometry-assessed ambulation, we derived odds ratios (OR ± 95% CI) from logistic regression models for BMI- and WC-defined obesity, adjusting for sex, age quartile, village, and season (13). Data were analyzed using the Statistical Package for the Social Sciences for Windows (Version 14.0; SPSS Inc., Chicago, IL). Significance for all inferential statistics was set at P < 0.05.

Back to Top | Article Outline


Of the 830 subjects on the main database, 792 had pedometry data. One subject was missing anthropometry data, and the pedometry data for two male subjects were excluded as outliers (36.4 yr = 52,010 steps per day, 45.2 yr = 53,418 steps per day). A complete data set for anthropometric, demographic, and pedometry data was provided by 789 subjects. Because <100 subjects per village were recruited from three villages (Ga Tjale, Moduane, and Ntsima), statistics are reported for these villages combined (n = 122). Moreover, because these villages are geographically closely linked, grouping these villages was valid. Unpublished results from a 2006 survey conducted in the DHDSS found that although the majority of subjects reported living in formal, bricked housing (96.0%) and had electricity supply in the house (72.2%), only 8.4% reported a tap inside the house, almost 80% still used wood for cooking, and 20.3% reported access to a motor vehicle within the household. Completion of secondary school (grade 12) was attained by 15.9% of the subjects. Subjects reported perceived overall health rating as good to very good (81.2%).

Descriptive statistics for anthropometric, demographic, and pedometry variables are reported in Table 1 as unadjusted means and prevalences. Between-subject effects for adjusted average ambulatory activity (full week) are reported in Table 2. The only adjusted mean, significantly less than 10,000 steps, was that for women in the Mantheding village (9084 steps per day, 95% CI = 8309-9859). Ambulation levels were significantly lower in women compared with men for the 2-, 5-, and 7-d periods (differences = −1483, −1154, and −1248 steps, respectively, P < 0.01). Overall, ambulation levels were significantly lower in the oldest age quartile (difference = −2956 to −3789 steps, P < 0.0001) and in the Mantheding village (difference = −2956 to −3789 steps, P < 0.0001) compared with the other age and village groupings. The ambulatory activity of the BMI obese was lower compared with BMI normal (difference = −1843 steps, P = 0.0785) (Table 2). Overall, subjects recorded the highest and the lowest ambulatory levels in spring and autumn, respectively (P = 0.0273, data not shown). Pedometer threshold prevalences by sex are reported in Table 3. Associations were significant across cross-tabular comparisons (P < 0.001).

Within-subject effects for adjusted ambulatory levels across 7 d for sex, age, BMI, village, season, and days ≥10,000 steps are shown in Figures 1A-F. Overall, steps per day were higher on Saturday compared with Wednesday (difference = 931 steps, P = 0.0286) (Fig. 1A). Younger age groups showed increasing activity over the week toward Saturday, whereas the ambulation levels in older age groups were more consistent over the week (Fig. 1B). Ambulatory activity did not significantly vary across days for BMI categories, although the obese tended to record lower steps per day for most days and the steps per day for the underweight and normal weight categories peaked at Saturday (Fig. 1C). Subjects in the Mantheding village recorded lower ambulation compared with all other villages (Fig. 1D). Ambulatory activity varied across days depending on the season, with levels higher in spring and lower in autumn (Fig. 1E). Step counts were relatively constant across the week for all compliance groups, except in particular the 4- to 5-d compliance group (Saturday > Wednesday) and the 6- to 7-d compliance group (Saturday > Friday) (Fig. 1F).

Regression analysis revealed that pedometry-assessed ambulation was inversely and significantly, albeit weakly, associated with adiposity measures, after adjusting for several covariates (BMI: zero order r = −0.27, partial r = −0.08, P = 0.0237; WC: zero order r = −0.34, partial r = −0.12, P = 0.0006). Compared with <5000 steps per day, achieving 10,000 steps per day significantly reduced the risk of obesity (BMI: 0.69, 95% CI = 0.52-0.92; WC: 0.71, 95% CI = 0.52-0.96). The reduction in obesity risk (BMI and WC) was more pronounced at moderate levels of ambulation such that risk changed 29%-31% from 5000 to 10,000 steps per day, 18%-20% from 10,000 to 15,000 steps per day, and 13%-14% from 15,000 to 20,000 steps per day. The risk of obesity (BMI and WC) was more than twofold higher in sedentary to low active activity zones compared with the active to very active zone (OR = 2.09-2.37, P < 0.04). Achieving <10,000 steps per day was associated with a nearly 90% greater risk of obesity compared with achieving 10,000 steps per day (OR = 1.86-1.89, P < 0.005). Because only 15 (1.9%) men were classified as obese (see Table 1), we ran the BMI and the WC logistic models again only for women and found very similar significance levels, patterns, and magnitudes of risk as reported above.

Back to Top | Article Outline


This is a novel study, presenting for the first time an analysis of objectively measured ambulatory patterns from a health and demographic surveillance site in South Africa's Limpopo Province and adding to the growing repository of PA data for rural South Africa (11,13,14,26,35). The major findings of this study are twofold. First, prevalences for activity and inactivity are substantially different from estimates gleaned from self-reported measures. Second, ambulatory levels in the adult, rural population are high because of increased nonleisure, subsistence demands.

Nationally, representative estimates of the prevalence of sedentarism in rural South Africans for men and women are 33.6% and 39.2%, respectively, using self-reported measures (40) that are substantially higher than the age-adjusted 8.0% for <5000 steps per day and 25.5% for 5000-9999 steps per day found in this study. Similarly, Alberts et al. (4) reported age-adjusted prevalences of 45%-48% for no exercise at home and 17%-27% for walking or being physically active at work in DHDSS adults ≥30 yr (n = 2106). However, age-adjusted prevalences for adults ≥30 yr in this study (n = 382) were 12%-22% for <7500 steps per day, 23%-25% for 7500-9999 steps per day, and 55%-63% for ≥10,000 steps per day. Because the questionnaire used by Alberts et al. (4) was not designed to probe PA performed at low to moderate intensities and since few adults especially women in the DHDSS participate in sport or exercise, there is the possibility that the PA questionnaire could have displayed a "floor" effect, such that below a certain threshold, important health-contributing PA behaviors were not probed (11). From the preceding discussion, it would seem that self-report measures are overreporting sedentarism and underreporting physically active behaviors within a rural population in a developing economy. In contrast, results from the 2003-2004 NHANES survey revealed that the adherence to PA public health guidelines for self-report measures was 25%-33% compared with <10% for accelerometer measurements (36). Taken together, these results suggest that reliance solely on self-report measures to assess PA levels might result in dramatic underestimates of the true variation in PA between mostly rural, developing economies and highly urbanized, industrialized economies.

These divergent findings between objective and subjective measures of PA have been suggested to be due to several factors (36). First, the overestimation of reported PA might be the result of misclassification of sedentary and light activities as moderate. Second, the duration of activities could be overestimated. Third, subjects may provide relative responses to questions of intensity despite the presentation of intensity examples or physiological cues. Guthold et al. (21) have voiced similar concerns about self-report measures, suggesting factors such as the interpretation of questions, the understanding of PA intensity, and the climate complicate the use of self-report questions across countries and cultures. Interestingly, one South African study (n = 402), using an interview-based questionnaire, has reported sedentary (8.0%) and moderately active (73.4%) prevalences very similar to our results in an adult, rural, black population (35). This study used a simple adaptation of the IPAQ questionnaire, which excluded questions about leisure because of the lack of leisure-time concept in that population. Instead, IPAQ was adapted around the understanding that walking and carrying a load constituted a major portion of the day (35). In contrast to our findings, women were more active than men and reported less sedentary behavior, although not significantly so (P = 0.275) (35).

Presently, much of the adult PA in the DHDSS revolves around subsistence-related activities (2,4,9), whereas vigorous leisure-time activities are more prevalent in the adolescent age group. Unlike industrialized countries, where leisure-time pursuits tend to dominate adult PA choices (24,33), investigations in traditional communities and developing economies have found a wide range of PA levels on the basis of occupation and gender (24,25,28,39). Our experience has been that subsistence tasks performed within the household are usually light to moderate in intensity, distributed throughout the day. This is in contrast to industrialized productivity that requires high bursts of energy over relatively shorter periods of time (30). Within the DHDSS, communal lands are available for planting of crops, such as maize and some vegetables, and cattle grazing (2). The lack of electricity in many dwellings would preclude the use of labor-saving devices (22,27). The use of wood for cooking is still common and requires the manual collection and transport of the raw material or purchasing of wood from sellers coming to the villages. Water is often not available in or around the dwelling and is collected in containers and transported manually. Journeys to the local store, clinic, communal fields, and bus or taxi stops are mostly completed on foot because few DHDSS residents have access to motor vehicle transport within the household (7).

Pedometry-assessed ambulation levels in an industrialized setting for adults (mean age ± 47 yr) were 44% for sedentarism (<5000 steps per day) and 13.9% for active behavior (≥10,000 steps per day) (38). In contrast, our results for the third and fourth age quartiles revealed prevalences for sedentarism and active behavior of 5.1%-70.6% and 19.3%-36.0%, respectively. A population-based sample of adults (n = 1102, ≥25 yr) from another industrialized country reported ambulation levels of 10,900 and 11,200 steps per day for men and women, respectively (difference, P = 0.34) (17). The ambulation levels for the same age group in our study were 12,134 and 10,721 steps per day for men and women, respectively (difference, P < 0.0008).

A population-based sample of adults (n = 704, ≥20 yr) in a developing country has found that 44.0% of adults achieved ≥10,000 steps per day (men = 55.2%, women = 39.1%) (6). In comparison, we found a higher prevalence (59.5%) for the same age group (men = 72.1%, women = 55.2%). Moreover, for the 20- to 30-yr age group, we found 20% and 35% higher prevalences for men and women, respectively, for ≥10,000 steps per day. However, for the ≥60 yr age group, prevalences were similar for ≥10,000 steps per day.

Between the ages of 13-18 yr, we found that approximately 35% of adolescents did not meet suggested minimum pedometry thresholds (18). The recent first South African National Youth Risk Behaviour Study reported that 34.4% of men and 43.0% of women participated in insufficient or no PA (5). The congruence, especially for men, between estimates for the prevalence of physical inactivity in adolescents from self-reported measures (5) and our objective measures is in contrast to the divergence of prevalences in adults. The congruence is likely because self-report of vigorous PA is more accurately recalled (33) and the greater participation of adolescent youth, particularly men, in vigorous leisure-time PA (5).

Because of time and financial constraints, the major limitation of the study was the use of convenience sampling. However, using unpublished data collected during a separate DHDSS socioeconomic status survey conducted in 2006, there were no significant differences between a socioeconomic score for households that participated and those that did not participate in the present survey (n = 830, P = 0.657). Socioeconomic and 5-d uniaxial accelerometry data collected in the DHDSS found that individuals from households with greater access to motor vehicles, readily available water and electricity supply, and less reliance on wood for cooking purposes are significantly less physically active than individuals from households without motor vehicle access, with restricted access to water, and with reliance on wood for cooking purposes (n = 138, P < 0.03) (9). Consequently, our sample does not appear to be overly biased on the basis of socioeconomic data. A minor aspect was our decision, due to financial constraints, to use a pedometer instead of an accelerometer, which might make comparison problematic with accelerometer-based studies (36). However, the NL-2000 is strictly speaking an accelerometer (piezoelectric), with a signal-processing algorithm that reduces the number of false-positive signals (10). Consequently, our results should be comparable to surveys using accelerometers, although a disadvantage was that the NL-2000 does not provide minute-by-minute data for download to a computer for further analysis (36).

In conclusion, this novel study reports pedometry-assessed ambulatory levels in a rural, black South African sample. In so doing, objective, unit-based prevalences of various ambulatory levels could be quantified and direct comparisons made with data from various settings.

This study was supported by the Research Development and Administration Division of the University of Limpopo (Turfloop Campus) and the Thuthuka Programme of the National Research Foundation.

The continuing and unwavering support of the individuals, families, and communities within the DHDSS is gratefully acknowledged.

The authors thank Vicki Lambert for her insightful comments on the manuscript. The authors do not have a professional relationship with companies or manufacturers who may benefit from the results of this study. The results of this study do not constitute endorsement of the products by the authors or the American College of Sports Medicine.

Back to Top | Article Outline


1. Ahmad O, Boschi-Pinto C, Lopez A, Murray C, Lozano R, Inhoue M. Age Standardization of Rates: A New WHO Standard. GPE Discussion Paper Series: No 31. EIP/GPE/EBD. Geneva: World Health Organization; 1999. p. 1-14.
2. Alberts M, Burger S. INDEPTH DSS site profiles: Dikgale DSS, South Africa. In: Sankoh OA, Kahn K, Mwageni E, Ngom P, Nyarko P, editors. Population and Health in Developing Countries. Volume 1. Population, Health, and Survival at INDEPTH Sites. Ottawa: IDRC; 2002. p. 207-11.
3. Alberts M, Burger S, Tollman SM. The Dikgale field site. S Afr Med J. 1999;89(8):851-2.
4. Alberts M, Urdal P, Steyn K, et al. Prevalence of cardiovascular diseases and associated risk factors in a rural black population of South Africa. Eur J Cardiovasc Prev Rehabil. 2005;12(4):347-54.
5. Amosun SL, Reddy PS, Kambaran N, Omardien R. Are students in public high schools in South Africa physically active? Outcome of the 1st South African National Youth Risk Behaviour Survey. Can J Public Health. 2007;98(4):254-8.
6. Anjos LA, Wahrlich V, Vasconcellos MTL. Distribution of pedometer count in a population-based sample of adults from Niterio, Rio de Janerio, Brazil. Med Sci Sports Exerc. 2005;37(5 suppl):S324.
7. Bell AC, Keyou G, Popkin BM. The road to obesity or the path to prevention: motorized transportation and obesity in China. Obes Res. 2002;10(4):277.
8. Benefice E, Cames C. Physical activity patterns of rural Senegalese adolescent girls during the dry and rainy seasons measured by movement registration and direct observation methods. Eur J Clin Nutr. 1999;53(8):636-43.
9. Cook I, Alberts M, Lambert EV. Development of a four-item physical activity index from information about subsistence living in rural African women. Int J Behav Nutr Phys Act. 2009;6:75.
10. Cook I. Pedometer step counting in South Africa: tools or trinkets? S Afr J Sports Med. 2006;18(3):67-78.
11. Cook I. Physical activity in rural South Africa-are current surveillance instruments yielding valid results? S Afr Med J. 2007;97(11):1072-3.
12. Cook I, Alberts M, Burger S, Byass P. All-cause mortality trends in Dikgale, rural South Africa, 1996-2003. Scand J Public Health. 2008;36(7):753-60.
13. Cook I, Alberts M, Lambert EV. Relationship between adiposity and pedometer-assessed ambulatory activity in adult, rural African women. Int J Obes. 2008;32(8):1327-30.
14. Cook I, Lambert EV. The sources of variance and reliability of objectively monitored physical activity in rural and urban Northern Sotho-speaking Africans. S Afr J Sports Med. 2008;20(1):21-7.
15. 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.
16. Crouter SE, Schneider PL, Bassett DR. Spring-levered versus piezo-electric pedometer accuracy in overweight and obese adults. Med Sci Sports Exerc. 2005;37(10):1673-9.
17. Dwyer T, Hosmer D, Hosmer T, et al. The inverse relationship between number of steps per day and obesity in a population-based sample-the AusDiab study. Int J Obes. 2007;31(5):797-804.
18. Eisenmann JC, Laurson KR, Wickel EE, Gentile D, Walsh D. Utility of pedometer step recommendations for predicting overweight in children. Int J Obes. 2007;31(7):1179-82.
19. Ezzati M, Hoorn SV, Rodgers A, Lopez AD, Mathers CD, Murray CJL. Estimates of global and regional potential health gains from reducing multiple major risk factors. Lancet. 2003;362(9380):271-80.
20. Faul F, Erdfelder E, Lang AG, Buchner A. G*Power 3: a flexible statistical power analysis program for the social, behavioral, and biomedical sciences. Behav Res Methods. 2007;39:175-91.
21. 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.
22. Herrin AN. Rural electrification and fertility change in the Southern Philippines. Pop Dev Rev. 1979;5:61-86.
23. Joubert J, Norman R, Lambert EV, et al. Estimating the burden of disease attributable to physical inactivity in South Africa in 2000. S Afr Med J. 2007;97(8):725-31.
24. Katzmarzyk PT, Mason C. The physical activity transition. J Phys Act Health. 2009;6(3):269-80.
25. Kruger HS, Venter CS, Vorster HH. Physical inactivity as a risk factor for cardiovascular disease in communities undergoing rural to urban transition: the THUSA study. Cardiovasc J S Afr. 2003;14(1):16-23.
26. Kruger HS, Venter CS, Vorster HH, Margetts BM. Physical inactivity is the major determinant of obesity in black women in the North West Province, South Africa: the THUSA study. Transition and Health during Urbanisation of South Africa. Nutrition. 2002;18(5):422-7.
27. Lanningham-Foster L, Nysse LJ, Levine JA. Labor saved, calories lost: the energetic impact of domestic labor-saving devices. Obes Res. 2003;11(10):1178-81.
28. Levine JA, Weisell R, Chevassus S, Martinez CD, Burlingame B, Coward WA. The work burden of women. Science. 2001;294(5543):812.
29. National Institutes of Health. Clinical guidelines on the identification, evaluation, and treatment of overweight and obesity in adults-the evidence report. Obes Res. 1998;6(suppl 2):51-209S.
30. Panter-Brick C. Issues of work intensity, pace, and sustainability in relation to work context and nutritional status. Am J Hum Biol. 2003;15(4):498-513.
31. Puoane T, Steyn K, Bradshaw D, et al. Obesity in South Africa: the South African Demographic and Health Survey. Obes Res. 2002;10(10):1038-48.
32. Schneider PL, Crouter SE, Bassett DR. Pedometer measures of free-living physical activity: comparison of 13 models. Med Sci Sports Exerc. 2004;36(2):331-5.
33. Shephard RJ. Limits to the measurement of habitual physical activity by questionnaires. Br J Sports Med. 2003;37(3):197-206.
34. Statistics South Africa. Provincial Profile 2004. Limpopo (Pretoria): Statistics South Africa; 2006. Report No.: 03-02-03(2001). p. 1-100.
35. Thorogood M, Connor M, Tollman S, et al. A cross-sectional study of vascular risk factors in a rural South African population: data from the Southern African Stroke Prevention Initiative (SASPI). BMC Public Health. 2007;7:326-36.
36. Troiano RP, Berrigan D, Dodd KW, Masse LC, Tilert T, McDowell M. Physical activity in the United States measured by accelerometer. Med Sci Sports Exerc. 2008;40(1):181-8.
37. Tudor-Locke C, Bassett DR. How many steps/day are enough? Preliminary pedometer indices for public health. Sports Med. 2004;34(1):1-8.
38. Tudor-Locke C, Ham SA, Macera CA, et al. Descriptive epidemiology of pedometer-determined physical activity. Med Sci Sports Exerc. 2004;36(9):1567-73.
39. Walker AR, Walker BF, Walker AJ, Vorster HH. Low frequency of adverse sequelae of obesity in South African rural black women. Int J Vitam Nutr Res. 1989;59(2):224-8.
40. World Health Organization Web site [Internet]. Geneva (Switzerland): World Health Organisation; [cited June 4, 2009]. Available from:


©2010The American College of Sports Medicine