(WITHDRAWN) Mixture Modeling As a Person-Centered Approach to Understanding Infant Health Risks

Background: Rates of preterm birth (PTB) and low birth weight (LBW) have decreased since their peak rates in the early 2000s, but profound disparities exist in the rates at which Black and White families are impacted by both PTB and LBW. Although it is clear that Black families, as a whole, are at a greater likelihood of experiencing PTB and LBW, there are many other factors that must be accounted for in order to better understand the cycle of reproductive disadvantage.  

Evidence suggests that social factors play an important role in understanding and explaining poor infant health outcomes. Given the complex nature of social determinants of health, institutions such as the CDC and the NICHD are now prioritizing research related to the etiology of PTB and LBW. Notably, social workers are uniquely positioned to highlight the ways that social factors contribute to reproductive disadvantage, given our strong understanding of intersectionality and the historical roots of power and oppression.

Purpose: First, this presentation will highlight the ways that social workers are uniquely capable of promoting a more thorough understanding the social factors which aid in determining risk for PTB and LBW. Second, this presentation will present the results of a latent-class mixture model (LCMM), which will illustrate the ways that more complex approaches to quantitative data analysis can help to further inform knowledge regarding the many ways that social factors can contribute to poor infant health outcomes. 

Methods: The sample for this study consists of 4,336 newly postpartum women who reported on their recent perinatal experiences. A LCMM was first conducted to identify whether homogeneous subgroups exist within our sample. Next, results from the LCMM were used to determine whether class membership (e.g., unique subgroups of mothers) predicted risk for PTB and LBW.

Results: Three latent classes were identified: (1) low risk for PTB/LBW (44%); (2) moderate risk (19%); and (3) high risk (36%). Compared to class 1, the odds of PTB for class 3 increased by 51% (OR = 1.51, p < .001; 95% CI 1.30, 1.75) and the odds of LBW by 26% (OR 1.26, p = .01; 95% CI 1.07, 1,48).  The odds of PTB were lower for class 2 compared to class 3 (OR .667, p < .001, 95% 535, .830). There were no statistically significant differences between class 1 and 2. Odds of LBW were lower for class 2 compared to class 3 (OR .619, p < .001,95% CI .510, .751).

Conclusions: Results from this analysis illustrate that examining risk for PTB and LBW using latent classes allows for a more in-depth understanding of risk profiles for at-risk mothers than the simplistic understanding allowed through a race-based analysis. Theoretical and practical implications will be discussed in relation to these findings. Specifically, we will discuss the utility of an intersectional approach to analyses, the importance of the social work perspective in understanding risks for PTB and LBW, and ways that this perspective can be used to positively impact existing policy and practice standards.