**Background:** Childhood obesity rates have risen dramatically worldwide, with rates in developed countries increasing from 16.6% to 23.2% between 1980 and 2013. Parental influence is critical for childhood obesity prevention, yet the mechanism by which parents' social support affects children's health practices remains unclear. The authors hypothesized that health knowledge (BMI knowledge and nutrition knowledge) mediates the relationship between parents' social support and children's obesity-related health practices.
**Methods:** A nationally representative cross-sectional survey was conducted in Singapore between June and November 2018. The Singapore Department of Statistics randomly drew 2,116 household addresses; 1,488 valid responses were returned (response rate 70.6%). All participants had at least one child aged 14 years or younger. The sample was 66.1% female, 30.1% younger than 36 years, 30.7% aged 36–40, 33.0% aged 41–49, and 6.3% aged 50+. Ethnic composition was 61.2% Chinese, 20.2% Malay, and 18.6% Indian. Social support was measured using five dichotomous items (e.g., availability of family, friends, doctors, other healthcare providers for childcare advice). Health knowledge was assessed via two constructs: BMI knowledge (3 items: knowledge of obesity measurement, BMI awareness, ability to assess child's weight status) and nutrition knowledge (6 Likert-scale items covering understanding of healthy diet, nutrition labels, and weight management). Health practice was measured by two items: frequency of children's exercise and frequency of eating at fast food restaurants (reversed), rated on 5-point scales. Structural equation modeling (SEM) via Stata 15.0 was used. Two models were built: Model 1 (predictor, mediators, outcome) and Model 2 (adding control variables: gender, age, education, employment status, residence type, mother/father involvement).
**Key Results:** Model 2 showed satisfactory fit: χ²/df = 154.08/80 < 3, RMSEA = 0.038, CFI = 0.904, SRMR = 0.037, explaining 20.20% of variance in health practice. Key findings: (1) Parents' social support showed a positive relationship with BMI knowledge (Coef. = 0.17, p < 0.001) and nutrition knowledge (Coef. = 0.18, p < 0.001), supporting H2. (2) BMI knowledge (coefficient = 0.10, p < 0.01) and nutrition knowledge (coefficient = 0.31, p < 0.001) were positively associated with children's health practice, supporting H3. (3) The total effect of social support on health practice was 0.081 (p = 0.071), and the direct effect was not significant (p > 0.1), while the indirect (mediated) effect was 0.081 (p = 0.007), indicating 100% full mediation by health knowledge, supporting H4. H1 (direct positive association between social support and health practice) was supported only when mediators were not included. Most control variables were not significant except age (β = –0.081, p = 0.06), father involvement (β = 0.088, p = 0.07), and residence type (β = –0.074, p = 0.09). Nutrition knowledge played a stronger mediating role than BMI knowledge.
**Clinical Implications:** The study demonstrates that social support from parents' networks does not automatically improve children's health practices—only social support that conveys adequate health knowledge is effective. This explains previously inconsistent findings in the literature. Public health interventions should leverage parents' social networks for disseminating pro-health information (e.g., 'share this with a parent of young children'). Developing community health champions as support resources for parents may be effective. The study also highlights a gap in public understanding of BMI derivation and nutrition labels, suggesting that health education must focus on making information practically relevant. Limitations include cross-sectional design (limiting causal inference), reliance on parent-reported data only, and limited measurement instruments for health practice and knowledge.