Page 48 - Sample Moderate prematurity, socioeconomic status,
and neurodevelopment in early childhood
P. 48
Chapter 4
problem items and one open-ended item for recording any problems not listed on the form. Each item can be rated as either: 0 = not true; 1 = somewhat or sometimes true; 2 = very true or often true. By summing the ratings for sets of items, we computed internalizing and externalizing problems, and a total problems score. We used the American cut-offs based on the guidelines for the CBCL regarding cross-cultural comparisons. In the guidelines is written that for Dutch children the same cut-offs are advised as in American children.22
Covariates
Data on background characteristics were collected using a general parental questionnaire that was sent to the parents simultaneously with the CBCL. The questionnaire consisted of questions about pregnancy, delivery, developmental and medical conditions of the child, family composition, and socioeconomic status. Furthermore, retrospective medical files for all children were available from PCH centres, paediatricians, midwives, and obstetricians. The most important variables (e.g. gestational age) from the parental questionnaires were crosschecked with information from the medical records. In more than 95% of cases, gestational age was calculated by using the last date of menstruation, and confirmed by early ultrasound measurements. When inconsistencies were found, these were checked against information in discharge letters.
Statistical analyses
First, we assessed child and family characteristics of the study participants categorized by low, intermediate, and high SES. Next, we examined prevalence rates of clinically relevant CBCL scores in MP children according to SES categories. Third, we performed logistic regression models to examine independent and joint effects of MP birth and SES on behavioural and emotional problems. In these analyses we used standardised measures for gestational age and SES, meaning that both risk factors had a mean of 0 and a SD of 1. In the first logistic regression model, we included MP birth and SES, mutually adjusting for their effects on behavioural and emotional problems. In the second model we assessed whether MP birth and low SES had joint effects on CBCL outcomes by adding the MP birth * SES interaction. In the case of joint effects, the combination of MP birth and low SES would be less meaningful because of shared effects (adversities do not multiply in that case). In the third and final model, we adjusted for the effect of potential confounders which we had identified in relevant literature and
64