Hierarchical Semi-Bayes Methods for Misclassification in Perinatal Epidemiology

BACKGROUND

Validation data are used to estimate the extent of misclassification in epidemiologic studies. In the Penn MOMS cohort, prepregnancy body mass index is subject to misclassification, and validation data are available to estimate the extent of misclassification. We use these data to estimate the association between maternal prepregnancy body mass index and early preterm (<32 weeks) birth using a semi-Bayes hierarchical model, allowing for more flexible adjustment for misclassification.

METHODS

We propose a two-stage model that first fits a Bayesian hierarchical model for the bias parameters in the validation study. This model shrinks bias parameters in different groups toward one another in an effort to gain precision and improve mean squared error. In the second stage, we draw random samples from the posterior distribution of the bias parameters to implement a probabilistic bias analysis adjusting for exposure misclassification in a frequentist outcome model.

RESULTS

Bias parameters from the hierarchical model were often more substantively reasonable and often had smaller variance. Adjusting results for misclassification generally attenuated the strength of the unadjusted associations. After adjusting for misclassification, underweight mothers were not at increased risk of early preterm birth relative to normal weight mothers. Severely obese mothers had an increased risk of early preterm birth relative to normal weight mothers.

CONCLUSIONS

The two-stage semi-Bayesian hierarchical model borrowed strength between group-specific bias parameters to adjust for exposure misclassification. Model results support evidence of an increased risk of early preterm birth among severely obese mothers, relative to normal weight mothers.

Relationship of Early Pregnancy Waist-to-Hip Ratio versus Body Mass Index with Gestational Diabetes Mellitus and Insulin Resistance

OBJECTIVE

To determine the risk of gestational diabetes mellitus (GDM) and insulin resistance (IR) in obesity defined by body mass index (BMI), waist-to-hip ratio (WHR), or both combined.

METHODS

Secondary analysis of a randomized multicenter trial of antioxidant supplementation versus placebo in nulliparous low-risk women to prevent pregnancy associated hypertension. Women between 9 and 16 weeks with data for WHR and BMI were analyzed for GDM (n = 2,300). Those with fasting glucose and insulin between 22 and 26 weeks (n = 717) were analyzed for IR by homeostatic model assessment of IR (normal, ≤ 75th percentile). WHR and BMI were categorized as normal (WHR, < 0.80; BMI, < 25 kg/m(2)); overweight (WHR, 0.8-0.84; BMI, 25-29.9 kg/m(2)); and obese (WHR, ≥ 0.85; BMI ≥ 30 kg/m(2)). Receiver operating characteristic curves and logistic regression models were used.

RESULTS

Compared with normal, the risks of GDM or IR were higher in obese by BMI or WHR. The subgroup with obesity by WHR but not by BMI had no increased risk of GDM. BMI was a better predictor of IR (area under the curve [AUC]: 0.71 [BMI], 0.65 [WHR], p = 0.03) but similar to WHR for GDM (AUC: 0.68 [BMI], 0.63 [WHR], p = 0.18).

CONCLUSION

Increased WHR and BMI in early pregnancy are associated with IR and GDM. BMI is a better predictor of IR compared with WHR. Adding WHR to BMI does not improve its ability to detect GDM or IR.

Eighty-year trends in infant weight and length growth: the Fels Longitudinal Study

OBJECTIVES

To investigate secular trends in weight and length growth from birth to 3 years of age in infants born from 1930 to 2008, and to assess whether these trends were associated with concurrent trends in pace of infant skeletal maturation and maternal body mass index.

STUDY DESIGN

Longitudinal weight and length data from 620 infants (302 girls) were analyzed with mixed effects modeling to produce growth curves and predicted anthropometry for infants born from 1930 to 1949, 1950 to 1969, 1970 to 1989, and 1990 to 2008.

RESULTS

The most pronounced differences in growth occurred in the first year of life. Infants born after 1970 were approximately 450 g heavier and 1.4 cm longer at birth, but demonstrated slower growth to 1 year of age than infants born before 1970. Growth trajectories converged after 1 year of age. There was no evidence that relative skeletal age, maternal body mass index, or maternal age together mediated associations between cohort and growth.

CONCLUSIONS

Recent birth cohorts may be characterized not only by greater birth size, but also by subsequent catch-down growth. Trends over time in human growth do not increase monotonically, and growth velocity in the first year may have declined compared with preceding generations.