Does Neighborhood Crime Mediate the Relationship Between Neighborhood Socioeconomic Status and Birth Outcomes? An Application of the Mediational G-Formula

Abstract

While the link between living in low socioeconomic status (SES) neighbourhoods and a higher risk of adverse birth outcomes has been well established, the underlying mechanisms remain poorly understood. Using the parametric g-formula, we assess the role of neighbourhood crime as potential mediator for the relationship between neighbourhood SES and birth outcomes using data on singleton births occurring in the Netherlands between 2010 and 2017 (n = 1,219,470). We estimated total and mediated effects of neighbourhood SES on small for gestational age (SGA), low birthweight (LBW), and preterm birth (PTB) via three types of crime (violent crimes, crimes against property and crimes against public order). The g-formula intervention settings correspond to a hypothetical improvement in neighbourhood SES. The hypothetical improvement in neighbourhood SES resulted in a 6.6% (95%CI=5.6,7.5) reduction in the proportion of SGA, a 9.1% (95%CI=7.6,10.6) reduction in LBW, and a 5.8% (95%CI=5.7,6.2) decrease in PTB. Neighbourhood crime jointly accounted for 28.1% and 8.6% of the total effect on SGA and LBW, respectively. For PTB, we found no evidence of mediation. The most relevant pathways were crimes against property and crimes against public order. The results indicate that neighbourhood crime mediates a meaningful share of the relationship between neighbourhood SES and birth outcomes.

Does household income predict health and educational outcomes in childhood better than neighbourhood deprivation?

“…However, the multi-dimensionality of IMD may capture neighbourhood effects on young families regardless of income. [26] What this study adds We found that household income in early childhood is a stronger predictor of adverse health outcomes in adolescence than neighbourhood IMD. However, combining income and IMD adds useful information for understanding subgroups at highest risk.…”

“…[10] Research on health inequalities from the Nordic countries, the Netherlands and North America, reports use of income alongside area-based deprivation indices. [26,[28][29][30][31][32][33][34][35] A recent US report describes a comprehensive income dataset that incorporates nearly all taxable income which is used to address under-reporting in surveys and for tax administration, forecasting and research. [28,29] In the UK, improving information on costs and income is a priority for population studies on health and deprivation, as identified in a review and poverty experts.…”

Neighbourhood‐related socioeconomic perinatal health inequalities: An illustration of the mediational g‐formula and considerations for the big data context

“…However, we cannot rule out that in this case social environment may also influence SES. To avoid this issue, ideally, one must use a measure of the exposure that temporally precedes the measure of the mediator as done in Burgos Ochoa et al 44 However, in our real‐life example, this was not feasible as only two reporting years for the mediator (2014 and 2017) were available.…”