Page 20 - Sample Moderate prematurity, socioeconomic status,
and neurodevelopment in early childhood
P. 20
Chapter 2
Statistical analyses
Firstly, we assessed child and family characteristics across children with low, intermediate, and high SES. Secondly, we examined prevalence rates of ASQ scores in the abnormal range for term-born and MP children with low, intermediate, and high SES. Thirdly, we performed univariate logistic regression analyses to assess the crude effect of SES and GA on developmental delay by using standardized scores for SES and GA, meaning that both have a mean of 0 and standard deviation of 1. Finally, we assessed the effects of SES, GA, and the interaction between SES and GA on developmental delay in three consecutive multivariable logistic regression models. In the first model we assessed separate effects of SES and GA, with mutual adjustment. Next, we assessed potentially synergistic effects of SES and GA by adding the SES*GA interaction (Model 2). In the final Model 3 we adjusted for the effect of potential confounders, which were identified on the basis of the literature and differences in background characteristics. We decided not to adjust for family composition and ethnicity of the mother to prevent over-adjustment for factors that significantly correlated with SES. We used SPSS for Windows 18.0 for all the statistical analyses. A P value of less than .05 was considered statistically significant.
RESULTS
In Table 1 we show the characteristics of all preterm and term children with low, intermediate, and high SES. Characteristics were statistically different between SES groups for family composition, maternal age, and maternal ethnicity. In analyses for the preterm and term-born children separately, we found no other characteristics that differed with statistical significance. The GA distribution of the 926 MP children was: 32 weeks, 113 (12.2%); 33 weeks, 177 (19.1%); 34 weeks, 257 (27.8%); 35 weeks, 379 (40.9%).
Figure 1 shows prevalence rates of ASQ scores in the abnormal range for term- born children (A) and MP children (B) with low, intermediate, and high SES. Except for domains gross motor and communication (MP children), the proportion of children with scores in the abnormal range increased as SES decreased. For the ASQ total score, prevalence rates were 7.2%, 4.0%, and 2.8% in term-born children with low, intermediate, and high SES, respectively, compared to 12.5%, 7.8%, and 5.6% in MP children, respectively.
30