Original Research
Adolescent Nutrition

Sociodemographic Determinants of Obesity Among 12-Year-Old School Adolescents in Kuala Terengganu and Besut Districts, Malaysia

Background and Context

Childhood and adolescent obesity has emerged as one of the most pressing public health challenges globally, with prevalence rates increasing substantially over recent decades. In Malaysia, the National Health and Morbidity Survey has documented a concerning upward trend in childhood overweight and obesity, with combined rates exceeding 30% in several states. The downstream health consequences are severe: obese children face significantly elevated risks of developing type 2 diabetes, cardiovascular disease, non-alcoholic fatty liver disease, and psychosocial difficulties that persist into adulthood.

While environmental factors such as poor dietary habits and sedentary lifestyles are widely recognised as primary drivers of obesity, sociodemographic factors also play a pivotal role — particularly in developing nations. Parental education, household income, family size, urban versus rural residence, and cultural practices all shape the obesogenic environment in which children develop. However, evidence from sub-urban states in Malaysia’s East Coast region, including Terengganu, has been comparatively sparse.

Terengganu, situated on the northeast coast of Peninsular Malaysia, has a predominantly Malay population and includes both urban centres and extensive rural areas. The state’s socioeconomic profile differs notably from West Coast states such as Selangor and Penang, making locally generated evidence essential for tailoring public health interventions. This study from the Health Assessment of Terengganu Adolescents (HAT) project represents one of the largest population-based assessments of childhood obesity in the state.

Study Design and Methods

The researchers employed a cross-sectional design, recruiting all 12-year-old students from 136 government primary schools across the districts of Kuala Terengganu and Besut. These two districts were selected based on demographic and logistic considerations, as approved by the Malaysian Ministry of Education and the Terengganu State Education Department. The districts encompassed both urban and rural school locations, allowing for comparative analysis.

Body weight and height data were obtained from the 2015 National Physical Fitness Standard (SEGAK) assessment, a mandatory biannual fitness test conducted in all Malaysian government schools. Data were uploaded into a purpose-built Health Monitoring System (HEMS), an automated web-based platform developed by the research team for standardised data collection. BMI was calculated and classified according to the WHO 2007 BMI-for-age z-score system, with obesity defined as z-score exceeding +2 standard deviations.

Sociodemographic information — including parents’ educational background, family income, occupation, and household characteristics — was collected through self-reported questionnaires completed by parents. School locations were sub-classified as urban or rural based on the Terengganu State Education Department’s designations. Statistical analysis employed Pearson’s chi-square tests for bivariate associations and Forward Likelihood Ratio multiple logistic regression to identify independent determinants of obesity, with the Hosmer-Lemeshow goodness-of-fit test applied to evaluate model adequacy.

Principal Results

The final analytical sample comprised 3,798 adolescents aged 12 years, of whom 44% were boys and 56% were girls. The distribution of participants across the two districts reflected the population density patterns of each area. When students were classified by school location, a meaningful proportion attended rural schools, enabling robust urban–rural comparisons.

The logistic regression analysis revealed that several sociodemographic variables were independently associated with obesity. Gender emerged as a significant factor, consistent with prior national surveys showing differential patterns of adiposity between Malaysian boys and girls at this age. School location was also significant, reflecting the well-documented urban–rural gradient in obesity risk that has been observed across many developing nations.

Related publications from the same research programme, the HAT study, have provided complementary findings. A subsequent analysis published in PLOS ONE (2018) found significant associations between socioeconomic status and obesity within rural areas specifically, with household income and parental education showing differential effects. A further analysis in BMC Pediatrics (2021) identified that BMI z-score was positively associated with parental BMI (P<0.001) and birth weight (P=0.003), while inversely associated with household size (P=0.022), underscoring the importance of family-level factors.

The Malaysian Context: Childhood Obesity as a National Concern

Malaysia is experiencing a rapid nutritional transition, with childhood obesity rates among the highest in Southeast Asia. The 2019 National Health and Morbidity Survey indicated that roughly one in three Malaysian children aged 5–17 was either overweight or obese. This dual burden of malnutrition — where undernutrition and overnutrition coexist in the same communities — poses distinctive challenges for public health programming.

In Terengganu, the obesity landscape reflects both national trends and local specificities. A population baseline study from the same research group found that the prevalence of obesity was highest in Kuala Terengganu (13.0%), followed by Kemaman (12.5%) and other districts, with significant variations by gender across different age groups. These findings suggest that Terengganu’s adolescents are following the national trajectory of increasing adiposity, though rates remain somewhat lower than in more urbanised West Coast states.

The Malaysian government has responded to the obesity epidemic with several policy initiatives, including the National Plan of Action for Nutrition of Malaysia (NPANM), school-based nutrition education programmes, and restrictions on unhealthy food marketing to children. However, the effectiveness of these interventions depends on understanding the specific determinants operating within different communities — precisely the kind of evidence this study provides.

Implications for Policy and Intervention

The identification of sociodemographic determinants of obesity among Terengganu’s adolescents has direct relevance for intervention design. The significance of urban–rural differences suggests that obesity prevention strategies need to be tailored to local contexts rather than applied uniformly. Urban adolescents may benefit from interventions targeting access to fast food, screen time, and reduced physical activity, while rural programmes might need to address the increasing availability of processed foods and changing agricultural practices that affect dietary patterns.

The role of parental factors — including parental BMI, education, and household income — underscores the need for family-centred approaches. School-based interventions alone may be insufficient if the home environment continues to promote obesogenic behaviours. Community health workers, primary care clinics, and mosques could serve as important channels for reaching parents and caregivers with nutrition education and lifestyle counselling.

The SEGAK-HEMS data infrastructure developed through this research represents a significant methodological contribution. By linking mandatory physical fitness assessment data with sociodemographic information through an automated web-based system, the researchers created a model for ongoing obesity surveillance that could be replicated in other Malaysian states.

Limitations

Several limitations should be noted. The cross-sectional design precludes causal inference — while associations between sociodemographic factors and obesity were identified, the direction and mechanisms of these relationships cannot be definitively established. The reliance on parental self-reported information for sociodemographic variables introduces potential measurement bias, particularly for sensitive information such as household income. The SEGAK data, while standardised, were collected by physical education teachers in schools whose measurement precision may vary. Additionally, the study did not include dietary intake or physical activity data, limiting the ability to identify behavioural mediators of the observed associations.

← Back to MJPHM Homepage