289 - Interactions: Associations between Parent Social Networks, Family Health Behaviors and Child BMI

Publication Number: 289.5209

Michelle White, Duke University School of Medicine, Durham, NC, United States; Javier Rodriguez, Duke University School of Medicine, Durham, NC, United States; Janna Howard, Duke University School of Medicine, Durham, NC, United States; Katelyn M. Holliday, Duke University, Durham, NC, United States; Eliana M. Perrin, Johns Hopkins University Schools of Medicine, Nursing, and Public Health, Baltimore, MD, United States; James Moody, Duke University, Durham, NC, United States; Madelynn Wellons, Duke University, Durham, NC, United States; Sabina B. Gesell, Wake Forest School of Medicine of Wake Forest Baptist Medical Center, Winston-Salem, NC, United States

Background: Black and Hispanic children experience disproportionate risk of developing obesity. Social networks are powerful drivers of health behaviors and outcomes. While prior studies have largely focused on the role of parent social support, additional parent social network characteristics (i.e. network size, health behaviors of network members) may influence family health behaviors and child obesity risk.

Objective: To determine the relationship between parent social network characteristics, child body mass index percentile (BMIP) and family health behaviors.

Design/Methods: This was a cross-sectional study of parents of Black and Hispanic children 2-4 years recruited from 22 NC primary care clinics. Race/ethnicity was collected via parent report. Our social network survey was adapted from The Important and Health Matters Social Network Battery found in the PhenX Toolkit. Changes were made to the original survey with the assistance of experts in survey methodology and social network analysis. Caregivers were asked to name individuals (alters) within specific categories and describe characteristics of the alter. Family health behavior questions were derived from NHANES and measured via survey. All surveys were administered via phone in English or Spanish. Child BMIP was calculated based on measured height and weight at the child’s most recent well-child visit. Linear regression models were used to identify associations between parent network characteristics, BMIP and family health behaviors.

Results: There were 118 parents enrolled (48 parents of Hispanic children, 70 parents of Black children). Parents of Hispanic children named fewer supportive alters than parents of Black children (3.9 vs 4.9 p< 0.0.1). Forty-one percent of parents reported having alters who are both supportive and cause them stress (role strain). Among parents of Black children, a higher proportion of relationships with role strain was associated with higher child BMI percentile (B=0.30 p< 0.01). Also, for parents of Black children, a higher number of supportive alters was associated with more hours of child weekday sleep (B=0.3), less child weekday screen time (B= -0.2), and more meals consumed outside the home (B=0.6) (all p <.05).

Conclusion(s): There were important differences between parents of Black and Hispanic children, indicating that child obesity interventions which leverage parent social networks may benefit from strategies tailored to specific populations. Augmenting supportive parent relationships and developing coping strategies to minimize parent role strain may be opportunities to reduce obesity among Black children.

Table 1: Parent Demographic Information (N=118)1,2
Int PAS Tab 1 F.pdf1 Values represent mean (SD) or % (N), as appropriate
2 P values represent comparisons between parents of Black and Hispanic children using t tests where p<.1 = * p<.05 = ** p<.01 = ***
3 Child race/ethnicity defined by the parent. If they identified their child as both Black and Hispanic, they were coded as Hispanic. If they were identified as Black or Hispanic and another race then they were coded as Black or Hispanic for analysis.
4 If present, up to 2 parents/caregivers per child could be enrolled. The second parent/caregiver provided only demographic and social network information.

Table 2: Parent Social Network Characteristics (N=118)1,2
Int PAS Tab 2 F.pdf1 Values represent mean (SD) or % (N), as appropriate
2 P values represent comparisons between Parents of Black and Hispanic Children using
p <.05 = ** p<.01 = ***
3 Child race/ethnicity defined by the parent. If they identified their child as both Black and Hispanic, they were coded as Hispanic. If they were identified as Black or Hispanic and another race then they were coded as Black or Hispanic for analysis.
4 Role strain: alters parents identified as both supportive and causing stress or difficulties
5 Unhealthy is defined by the parent

Table 3: Association of Parent Social Network Characteristics and Child BMI Percentile (N=118)1
Int PAS Tab 3F.pdf1 Numbers represent beta coefficient of association between parent network characteristics and child BMI percentile using a linear regression model. Model was adjusted for the presence of two parents of the same child enrolled in the study: p<.1 = * p<.05 = ** p<.01 = ***
2 Child race/ethnicity defined by the parent. If they identified their child as both Black and Hispanic, they were coded as Hispanic. If they were identified as Black or Hispanic and another race then they were coded as Black or Hispanic for analysis.
3 Role strain: alters parents identified as both supportive and causing stress or difficulties
4 Unhealthy is defined by the parent