Latent Class Analysis of Low Birth Weight and Preterm Delivery among Australian Women

Abstract

To identify patterns of health, socioeconomic, behavioral, and psychosocial indicators that may be associated with low birth weight delivery or preterm birth. Data were analyzed from the Australian Longitudinal Study on Women's Health. A total of 9075 live singleton births among 3801 women were linked to state perinatal records with birth outcome data. Survey data were used to identify 11 indicators for latent class analysis. Latent classes were tested for association to birth outcomes. These indicators also were used along with covariates in main effect multiple logistic regression analyses of birth outcomes. Latent class analysis revealed 5 classes, including those characterized by low education, recent drug use, stress/anxiety/depression, smoking/drinking/low education/multi-risk, and a low risk referent group. The stress/anxiety/depression class was associated with preterm delivery (OR 1.87, 95% CI 1.20-2.92), and the smoking/drinking/low education/multirisk class was associated with low birth weight (OR 1.54, 95% CI 1.02-2.30). Traditional logistic regression analyses for main effects identified some measures not captured by the latent classes, and the latent classes identified variable combinations not captured by the main effect analysis. Unique latent classes were associated with preterm delivery vs low birth weight. Both latent class analysis and main effects analyses may be combined to improve understanding of birth outcome risks. Clinical and programmatic interventions to reduce risks of low birth weight and preterm delivery may benefit from risk profiles that women experience.