#PHTwitJC 7 on Sunday 27th November (from 8pm GMT) will discuss A.Timperio et al (2010) Neighbourhood physical activity environments and adiposity in children and mothers: a three-year longitudinal study, IJBNPA 2010, 7:18 .
All are welcome to follow or contribute to the discussion. If you’ve never joined in a Twitter ‘chat’ before, see the ‘About’ page, above.
Environmental factors appear to be behind the growth in obesity in developing nations, including Australia (the setting of this study). Environmental factors are mediated by behavioural ones, which include the amount of physical activity, the focus of this study.
There has been increasing interest in place and health in the academic literature, much of it suggestive of complex interactions between individual perceptions, behaviour, and features of place (e.g. Cummins et al 2007 ).
Children from two age groups (5-6 years and 10-12 years) and their mothers/female carers were recruited from schools in Melbourne, Australia. Their height and weight were measured to produce a BMI score (converted into z index for children: more information about z-scores is here). The women’s height and weight calculations were self-reported. BMI was calculated again at three-year follow up. In addition, family pairs’ neighbourhoods were assessed using GIS data within 800m and 2km of the home. Factors included in this assessment included road and traffic density, availability of open spaces, walking/cycling paths and leisure opportunities.
This study was part of a wider longitudinal study called Children Living in Active Neighbourhoods (CLAN). The paper points the reader to related publications for details about recruitment. This is from another paper from the same study:
Children (aged 5–6 and 10–12 y) and their families were recruited from 19 state primary schools in high (n=10) and low (n=9) socioeconomic areas in metropolitan Melbourne, Australia. Stratified random sampling was used to select schools within the sampling frame. Five schools declined to participate and were replaced with randomly selected schools. Although schools were selected randomly, all schools in high socio-demographic areas were from the east of Melbourne, while all schools in low socio-demographic areas were from the west. This reflects the socioeconomic distribution within Melbourne.19 All students within Grades prep (5–6 y of age) and Grades five and six (10–12 y of age) in participating schools were invited to participate. However, only families who provided active consent by returning a signed consent form by the required date were eligible. In total, 291 families of children aged 5–6 y (27% of those invited) and 919 families of children aged 10–12 y (44% of those invited) took part in the study.
690 children (+ female carers) participated in the three-year follow up measurement. After excluding those with incomplete data and families who had moved neighbourhoods within the three year period, results from 409 children and 369 female carers are included in the analysis presented in the paper.
The other paper from the study cited above, indicates that there was also a questionnaire survey for children and adults about their perceptions of their neighbourhoods ; however this is not discussed in the paper we are considering, which reports associations between objective calculations about neighbourhood, and adiposity.
There was little association between the objectively measured neighbourhood characteristics and BMI/changes in BMI. Those that were apparent did not hold consistently across the age and sex groups (the children were analysed according to age group at baseline, and sex). Most features that demonstrated any significance in the statistical analysis were within the 800m rather than the 2km distance band. Associations that did emerge include:
- Length of access paths within 800m of home inversely associated with z-score for both age cohorts of children
- Number of sport/recreation public open spaces inversely related to z-scores for the older age cohort of children.
- Length of walking/cycling tracks within 800m positively associated with BMI for female carers
- Length of busy roads within 800m negatively associated with BMI for female carers
- Number of options for aerobics/fitness/swimming within 2km associated with greater relative decreases in BMI for the female carers.
The discussion highlights inconsistencies in the findings, and some surprising findings such as the third point, above. The authors present possible explanations for some of the findings and the differences between age groups. For example, access paths may be significant for facilitating children’s independent outdoor play. Number of busy roads in the neighbourhood appeared to be associated with overweight for the older children but not the younger –older children are less likely to be accompanied by parents and therefore more likely to be exposed to heavy traffic (or choose not to expose themselves to the heavy traffic). The authors suggest that slowing traffic through residential areas may be part of an obesity prevention strategy.
The study’s strength is in the attempt to analyse features of ‘natural’ neighbourhoods based on distance from actual home, rather than of administrative divisions. The authors conclude based on the equivocal findings, that the smaller distance radius (800m) is more significant and reflects real sensations of the local area better, especially for children (who are unlikely to walk much further than that). They suggest that including perceptions of neighbourhood environments, and further features including the food environment, in future research, would be beneficial.
Questions for discussion:
- Is the study design appropriate for the aims?
- Was the sampling technique clear and appropriate? What about follow-up?
- Were the measurements used clear, relevant and valid (BMI/x score)?
- Were the objective neighbourhood criteria chosen clear, relevant and valid?
- The results were inconclusive:
- are there any that are particularly noteworthy?
- Are the researchers’ explanations valid?
6. Are there implications for further research?
7. Are there implications for public health policy & practice?