Social Determinants of Childhood Overweight and Obesity in Elementary School Children Richard Lazenby, Sarah Angus, Tracey Galloway and Hanh Huynh University of Northern British Columbia 1. INTRODUCTION Overweight and obesity (OW/OB) rates in children and adolescents have increased noticeably in the past 25 years (Shields, 2004; Tremblay et al., 2002). This rapid increase is attributed to environmental factors (Crooks, 1999) – those social factors that impact the health of the home / work / school environment and ultimately the health of the individual. This study examines the pattern of OW/OB measured in elementary schools in Prince George, BC, Canada within four neighbourhoods differing across a range of socioeconomic (SES) variables (Figures 1 and 2). Our working hypothesis is that OW/OB rates by neighbourhood will vary inversely with SES (e.g., Moffat et al., 2005). 2. SETTING Prince George is located in the central interior of British Columbia, Canada with a population of 77,148 (2005), of which approximately 7,100 are Aboriginal. The regional economy is dependant on cyclical resource extractive industries (forestry, oil/gas and mining). School District 57 includes 37 elementary schools with approximately 9100 students. The four schools included in this study are located within Prince George city limits.

3. MATERIALS AND METHODS Sample: • 283 children; grades 1 – 6; ages 6 – 14 yrs • Four catchments (C1 – C4); 25 dissemination areas (DA) (2001 Census) Anthropometric Variables: • Height, weight, waist and hip circumference • Height measured to the nearest mm using a Seca 220 portable stadiometer (Seca Corp., Hamburg, GDR) • Weight measured in lbs and converted to kgs, using a LifeSource ProFIT UC-321 digital scale (A&D Medical, Milpitas, CA) • Ht-for-Age, Wt-for-Ht, BMI-for-Age, WHR derived • Triplicate measures, means used; two observers

5. RESULTS: OVERWEIGHT AND OBESITY • C4 has the highest rate of obesity among the four schools (> 95th percentile) – nearly double that of the CCHS outcomes (Shields 2004) (Table 2; Figure 4). It also has the highest mean and widest range for WHR. • C3 has the highest rate of overweight (85th - 95th centile), approximately 50% higher than the CCHS findings. • C1 and C2, the two more affluent catchments, were closest to the CCHS findings, though clearly OW/OB is still a concern for these schools. Several children were underweight ( < 5th centile), most frequently in the lowest SES school. Table 2: OW-OB Outcomes By School

Social Determinants (SD) Variables (2001 Census): • Average Family Income • Income Composition (% Public Transfer Payments) • % Lone Parent Families • % Aboriginal (vs. ‘Other’) • % < LICO (Low Income Cut-Off, ca $30,000 for family of 4) • Educational Attainment (% University)

School

n

% UW

% OW

% OB

Mean WHR (range)

C1

105 (49 F, 56 M)

0.95

18.10

6.67

0.8077 (0.689-0.927)

C2

64 (28 F, 36 M)

0

12.50

7.81

0.8263 (0.721-1.048)

C3

51 (27 F, 24 M)

0

27.45

7.84

0.8218 (0.710-1.024)

C4

63 (25 F, 38 M)

4.76

7.94

14.29

0.8356 (0.716-1.097)

Analysis: • EpiInfo 3.3.2 based on CDC (2000) growth reference (Kuczmarski, RJ et al. (2002). CCHS reference (Shields, 2004) • Kruskal-Wallis ANOVA and Spearman’s Rho for SD variables; cross-tabs, Χ2 and t-test for anthropometric variables (Systat v 14)

4. RESULTS: SOCIAL DETERMINANTS (SDs) • The four catchments, C1 – C4, are clearly distinguished across a range of social determinants (Figure 3; K-W results not shown). • C4 is the lowest income catchment area with more lone parent families, a higher aboriginal population, a higher income portion of government transfer payments, and a greater population living below the LICO levels (Figure 3). • There are strong and predictable associations among the SDs, (Table 1); for example, the likelihood of living below the LICO is highly and positively correlated with being Aboriginal and/or a lone parent.

40 35 30 25

% OB % OW

20 15 10 5

Figure 1: SD57 Catchment areas based on 2001 Census dissemination areas.

100 90 80 70 60 50 40 30 20 10 0

0 C4

C4 C3 C2 C1

Average Family Income (000s)

% Transfer Payments

% Aboriginal

% Lone Parent

% LICO

%PostSecondary

% University

Figure 3. Distribution of SDs by Catchment

Table 1. Spearman’s Rho correlation among SD variables.

Figure 2: Average family income, Prince George 2001.

% Lone Parent % Aboriginal % University Average Income % Transfer % LICO

% Lone Parent % Aboriginal % University Average Income % Transfer 1 0.784 1 -0.84 -0.779 1 -0.885 -0.796 0.843 1 0.865 0.827 -0.822 -0.917 1 0.843 0.848 -0.802 -0.848 0.821

Presented at the Human Biology Association meetings, March 2007, Philadelphia, PA

7. SUMMARY (1) Significant differences in social determinants of health exist among the four school catchments; (2) Prevalence of OW children (> 85th centile) in two of four neighborhoods exceeds age- and sex-matched standards while OB (> 95th percentile) rates in all schools match standards, though the lowest SES neighborhood has OB prevalence almost twice as high; (3)OW-OB rates are greater for boys than girls relative to the CDC (2000) standards; this is true even for the catchment (C1) with the highest SES profile. 8. CONCLUSION The trend to OW-OB in elementary school age children transcends a simple social determinants analysis. Children in Prince George across the spectrum of social strata are progressing towards higher OW-OB values relative to national standards. However, the trend is more marked among the relatively disadvantaged (defined by poverty, ethnicity, parental status and education).

C3

C2

C1

CCHS, 2004

Figure 4. OW and OB Among Catchments Relative to CCHS (2004)

6. RESULTS: Z-SCORE ANALYSIS (CDC, 2000) • Although elevated relative to the CDC (2000) standards, cross-tabs analysis found that, in the same overall, neither boys nor girls are significantly more OW (85th - 94.99 percentile) nor more obese (> 95th centile); • However, when examined by individual school catchment, significant difference do appear, distinguished by sex - girls demonstrate fewer significant differences with respect to CDC standards than boys (Table 3). This result mirrors that of others (e.g., Moffat et al., 2005), and points to socially determinant interactions of SES, food security and gender.. Table 3. Mean z-scores for HT, WT and BMI relative to DCD (2000) standards; one sample t-test against z = 0.0. Male

Female

HAZ

WAZ

BMIZ

HAZ

WAZ

BMIZ

C1

0.292*

0.448*

0.452*

0.243

0.247

0.251

C2

0.331*

0.312*

0.226

0.497*

0.510*

0.396

C3

0.529*

0.615*

0.521*

0.418*

0.659*

0.643*

C4

-0.017

0.411*

0.512*

0.118

0.007

0.046

9. REFERENCES CITED • Canada Census (2001) Statistics Canada, Ottawa, On. • Crooks, DL (1999) Child growth and nutritional status in a high-poverty community in eastern Kentucky. Amer. J. Phys. Anthropol. 109 (1): 129-142. • Kuczmarski, RJ et al. (2002) 2000 CDC growth charts for the US: Methods and Development. Vital and Health Statistics 11(246). • Moffat, T et al. (2005) Stature and adiposity among children in contrasting neighborhoods in the city of Hamilton, Ontario, Canada. Amer. J. Hum. Biol. 17: 355-367. • Shields M (2004) Measured obesity: overweight Canadian children and adolescents. Nutrition: Findings from the Canadian Community Health Survey. • Systat (2006) Systat Software, Inc. Richmond, California • Tremblay, MS et al. (2002) Temporal trends in overweight and obesity in Canada, 1981 – 1996. International Journal of Obesity 26: 538-543. 10. ACKNOWLEDGEMENTS • Children and parents, School District 57 Board of Trustees, and teachers and staff at participating schools; • Prince George Community Foundation; Vancouver Foundation; Heart and Stroke Foundation Prince George BC • Canadian Tire Prince George BC; Save-On-Foods, Prince George BC; Costco, Prince George BC