Violence Victimization Predicts Body Mass Index One Decade Later among an Urban Sample of African American Young Adults: Sex as a Moderator and Dehydroepiandrosterone as a Mediator

Abstract

Psychological stressors such as violence victimization are known contributors to obesity. However, moderators and mediators of this association have not been studied, although they might offer pathways for intervention or prevention. Using a sample of African American young adults, this study tested: (1) the moderating effect of sex on the effect of violence victimization on trajectories of body mass index (BMI), and (2) the mediating effect of dehydroepiandrosterone (DHEA) on this association. This 13-year longitudinal study followed 73 male and 80 female African American young adults who lived in an urban area from 1999 to 2012 when the youth were 20-32 years old. The independent variable was violence victimization measured in 1999 and 2000. The dependent variable was BMI measured in 2002 and 2012. The mediator was DHEA measured in 2001 and 2002. Multilevel path analysis was used to test if males and females differed in violence victimization predicting change in BMI (Model I) and the mediating effect of DHEA change on the above association (Model II). The results of Model I suggested that the change in violence victimization from 1999 to 2000 predicted change in BMI from 2002 to 2012 for females, but not males. Based on Model II, the DHEA change from 2000 to 2001 for females fully mediated the association between violence victimization from 1999 to 2000 and increases in BMI from 2002 to 2012. Our findings suggest that violence victimization in urban areas contributes to the development of obesity among African American female young adults and change in DHEA mediates this link. Violence prevention may have important implications for obesity prevention of African American young women who live in unsafe urban areas. This study also suggests that DHEA may be involved in the violence victimization-obesity link for African American women.

Keywords: African Americans; Dehydroepiandrosterone; Obesity; Sex; Violence.

Fig. 1

Summary of Model 1; multi-group path model to test sex differences in the effect of violence victimization and BMI change among African Americans. a The model among men where change in violence victimization from 1999 to 2000 is not associated with change in BMI from 2002 to 2012. b The same model among women where all paths are significant, suggesting that change in victimization from 1999 to 2000 predicts change in BMI from 2002 to 2012. Chi square = 18.03, degrees of freedom = 9, probability level = < .001, CFI = .91, CMIN/DF = 2.001, RMSEA = .048. DHEA, dehydroepiandrosterone; BMI, Body Mass Index

Fig. 2

Summary of Model I1; multi-group path model shows sex differences in the mediating effect of change in Dehydroepiandrosterone (DHEA) from 2000 to 2001 on the association between change in violence victimization from 1999 to 2000 and BMI change from 2002 to 2012 among African Americans. a The model among men where all horizontal paths from pre to post victimization, DHEA, and BMI were positive and significant; however, all vertical paths between victimization, DHEA, and BMI were non-significant. b The same model among women where in addition to the significant horizontal paths between victimization, DHEA, and BMI (with the exception of the path from DHEA 1 to DHEA 2 which was only marginally significant), the three following vertical paths were also significant: DHEA –pre ➔ BMI –pre, victimization post ➔ DHEA –post, and DHEA –post ➔ BMI –post. Chi square = 76.658, degrees of freedom = 33, probability level = < .001, CFI = .90, CMIN/DF = 2.323, RMSEA = .044 (90% CI = .031–.057. DHEA, dehydroepiandrosterone; BMI, body mass index