Changes in experiences with discrimination across pregnancy and postpartum: age differences and consequences for mental health

Abstract

Objectives: We aimed to contribute to growing research and theory suggesting the importance of examining patterns of change over time and critical life periods to fully understand the effects of discrimination on health, with a focus on the period of pregnancy and postpartum and mental health outcomes.

Methods: We used hierarchical linear modeling to examine changes across pregnancy and postpartum in everyday discrimination and the resulting consequences for mental health among predominantly Black and Latina, socioeconomically disadvantaged young women who were receiving prenatal care in New York City.

Results: Patterns of change in experiences with discrimination varied according to age. Among the youngest participants, discrimination increased from the second to third trimesters and then decreased to lower than the baseline level by 1 year postpartum; among the oldest participants, discrimination decreased from the second trimester to 6 months postpartum and then returned to the baseline level by 1 year postpartum. Within-subjects changes in discrimination over time predicted changes in depressive and anxiety symptoms at subsequent points. Discrimination more strongly predicted anxiety symptoms among participants reporting food insecurity.

Conclusions: Our results support a life course approach to understanding the impact of experiences with discrimination on health and when to intervene.

Figures

FIGURE 1—

Estimated interactions between age at the start of the study and the curvilinear pattern of change over time in experiences with discrimination among young women: New York City, 2008–2012. Note. Time 1: mean gestation = 18/81 weeks. Time 2: mean gestation = 30.34 weeks. Time 3: mean postpartum = 26.47 weeks. Time 4: mean postpartum = 58.29 weeks. The discrimination scale used had a range of 1–4. The scale is reduced here to make the interactions and the distinct patterns for each age group easier to visualize. The time variable in the analysis was coded to reflect the mean interval in weeks between the completion of each time point. The sample size was n = 549.