Page 74 - Sample Moderate prematurity, socioeconomic status, and neurodevelopment in early childhood

Chapter 7
METHODOLOGICAL CONSIDERATIONS
The LOLLIPOP study
LOLLIPOP is one of the largest community-based cohort studies on MP and late preterm-born children. Moreover, unlike many other cohorts, LOLLIPOP comprises a control group of full-term children who were sampled from the same preventive child healthcare (PCH) centres and who were in the same age-range as the preterm children. Children born between 360 to 366 weeks’ gestation were not included since the main focus of the LOLLIPOP study was on outcomes of children born at 320-356 weeks’ gestation. The effect of not including these children is likely to be small because many MP children were born at 34 and 35 weeks’ of gestation, relative to 32 and 33 weeks’ of gestation. Numbers of MP children increased for every week of gestation, reflecting the community-based character of the cohort. The sample size of LOLLIPOP was based on numbers in order to compile growth curves for Dutch preterm children, which asked for a large number of children.
Parents could choose between joining the full study, which included follow-up measurements, or only the ‘growth curves’ part of the study. As a consequence, the number of participants joining the follow-up part was lower. Among the non-participating children low SES was more common (P<0.001), which may have influenced our results. However, we think that the effects of SES would then have been an underestimation of the real effects. Gender and SGA did not differ significantly between participating and non-participating children.
For the purpose of this thesis only data from the first measurement were used, i.e. when the children were nearly four years old. This measurement was planned as part of the scheduled PCH visit of the children, resulting in high responses (up to 95%) on the study questionnaires. Since the PCH doctors and nurses were very much involved in the study, they actively reminded the parents to fill in the questionnaires. Moreover, population coverage was high because in the Netherlands around 95% to 97% of children visit PCH centres regularly from birth up to four years of age.
Measurement of SES
We chose to use a composite measure of SES rather than single SES indicators, in order to fully account for effects that can be attributed to socioeconomic conditions.72, 73 We computed the composite score on the basis of five indicators: educational level of the mother, educational level of the father, the family income,
124