The adoption of a sedentary lifestyle in low to middle income countries is an important problem needing to be addressed (Popkin, 2009). A country such as South Africa is already battling the double burden of communicable disease (tuberculosis and HIV epidemics), and this coupled with problems arising from diseases of lifestyle associated with poor physical activity levels may cripple an already overburdened health care system. A child who is overweight is more likely to be overweight or obese as an adult (Biro and Wien, 2010) and engaging in exercise has numerous benefits in terms of weight loss and general health.
Over one third of South African public high school pupils (Grades 8-12) do not engage in sufficient physical activity (Amosun et al., 2007). Mamabolo et al (2007), studied physical activity levels in a population of 14 year old South African adolescents from a rural high school and showed a reduction of physical activity levels with advancing pubertal status. In addition, the authors found that weekend physical activity levels were negatively correlated with Body Mass Index (BMI) (Mamabolo et al., 2007). More recently Pienaar et al (2012) have shown sustained increased levels of physical activity in boys (but not girls) after an exercise intervention. Of concern however, is the observed decline in physical activity levels in both boy and girl South African adolescents over a three year period (Pienaar et al., 2012). Other studies conducted in South African children report on older children (Engelbrecht et al., 2004; Lennox et al., 2008; Mciza et al., 2007) but none have examined physical activity levels in an urban South African setting across 12 years of formal schooling.
In 2001, the American Academy of Paediatrics recommended that children should view less than two hours of television each day (American Academy of Pediatrics, 2001). High income countries such as Australia and the USA have implemented these guidelines into national recommendations and other countries such as Norway have adopted other initiatives to decrease screen time (Overby et al., 2013). No such national guidelines currently exist for South African children. The South African Youth Risk Behaviour Survey (SAYRBS) provided nationally representative information regarding time spent watching television and being physical active in South African high school children (Reddy et al., 2008). While the (SAYRBS) data are comprehensive and provide useful insight into physical activity behaviour, no data are provided on primary school aged children and relationships with body mass status have not been evaluated.
Ethnic disparities in physical activity levels have been reported in South African children (McVeigh et al., 2004). While lower levels of physical activity in Black children compared to White children may be related to socio economic differences (McVeigh et al., 2004; Kriska, 2000), there may also be different leisure time preferences between different ethnic groups. Little is known about how South African Indian children spend their leisure time and since this population forms a large part of the broader South African population, it is important that this group is not neglected in studies of physical activity and health especially in children. The association between physical activity, inactivity and body mass underpins the importance of understanding how youth use their leisure time. In addition, insight into after school behaviour has implications for policy development and caregivers (Olds et al., 2009).
Few studies conducted in low to middle income countries have used sedentary time and physical activity guidelines as "barometers" to assess whether inhabitants of those countries are meeting international recommendations. Narrow age ranges, not relating body mass across the school years to race, physical activity or sedentary behaviour data on Indian children remain limitations of current studies. In the present study, we aimed to examine the levels of physical activity over a period of change from childhood to adolescence in Black, White and Indian children. The aim of our study was to describe gender and race related patterns of physical and sedentary activity levels in a large sample of South African children aged between 5 and 18 years and to determine whether there were associations between these variables and body mass status.
Subjects and study design
We contacted 14 (primary and/or secondary) schools in the Johannesburg Metropolitan area. The schools included in the study offered physical education lessons as part of their curriculum as well as after school sports. Permission to conduct the study in the schools was granted by the principal, primary caregiver as well as the children themselves. In total, 10 schools agreed to participate in the project. Two thousand children from grade 1 (first year of primary school) to 12 (final year of secondary school) enrolled in the schools, were invited to take part in a physical activity and inactivity questionnaire survey. The municipality from which our participants were selected encompasses inhabitants of diverse ethnic backgrounds. We thus chose to collect data from "semi-private" schools within the greater Johannesburg region. The schools were matched for a similar fee structure and were able to offer the students at the schools similar access to opportunities for being physically active. The specific area from which we chose our schools was in a 30 km radius of Central Business District (CBD) of Johannesburg. The primary language of instruction at all schools we recruited was English. Questionnaires, consent forms and instructions were sent home with the children to the primary caregiver, who completed the questionnaire in conjunction with and on behalf of the child. Detailed instructions and examples on how to complete the questionnaire were provided. One week after distributing the questionnaires, we returned to the school to collect the questionnaire and to record the height and body mass of the individual. Only children who returned accurately completed, eligible questionnaires and who consented to anthropometric measures being made were included in the final analysis. The study was approved by the Human Research Ethics Committee of the University of the Witwatersrand (protocol number: M10828).
Physical Activity Questionnaire (PAQ)
We used a PAQ which has been previously validated with actigraphy in an ethnically diverse and similar cohort of South African children (McVeigh & Norris 2012). The questionnaire was used to gather information pertaining to the child's participation in physical (in)activity across six domains. These domains included 1) school physical education classes; 2) informal activities (e.g. playing with friends outside etc); 3) sedentary after school activities (television, computer, computer and video games); 4) transport to and from school; 5) extra mural/formal activities and 6) sleep. The PAQ was scored based on a rating of each activity's metabolic expenditure equivalent (MET) (Ainsworth et al. 2000). The MET physical activity score (METPA) was calculated by multiplying the intensity (multiples of basal metabolic rate (metabolic equivalents)) by the duration of...