“Expectant Parents”: Study protocol of a longitudinal study concerning prenatal (risk) factors and postnatal infant development, parenting, and parent-infant relationships

Between November 2008 and July 2009, 835 pregnant women were invited by their midwives to participate in this study. Four midwifery practices in Eindhoven, the 5^th largest city of the Netherlands, agreed to participate in the study. At the first routine visit (between 9–15 weeks gestational age), midwives gave mothers information about the purpose of the study and invited them to participate. The oral information was accompanied by an information brochure with specific information about the study, which each mother received. If mothers were interested in participation, one of the researchers contacted them by phone to provide additional information and asked whether mothers wanted to enroll in the study. Partners were not directly approached by the researchers but the mothers were informed about the importance of involvement of their partners in the study. After parents received oral and written information about the protocol, both parents were asked for written consent. The informed consent form consisted of three different options. Parents could consent to (1) active participation in the complete research protocol, including two home visits, (2) active participation by filling in questionnaires but not by participating in home visits, or (3) passive participation by allowing the researchers to gather information from the midwife and National Health Care Centres, but without home visits or filling in questionnaires. Separate informed consent forms were sent to mothers and their partners. Once parents returned the signed forms, enrollment in the study was complete.

The “Expectant Parents” [“In Verwachting”] study protocol has been financed and approved by the Netherlands Organization for Health Research and Development (ZonMW, Grant 80-82405-98-074/157001020). It was also approved by the Medical Ethics Committee of St. Elisabeth Hospital Tilburg (date: 13-08-2008, register number: NL 23376.008.08).

Of the 835 invited women, women with a poor understanding of the Dutch and English language, those expecting multiple births, and women who were over 20 weeks of gestation at enrollment, were excluded from participation. Reasons for not giving written consent were withdrawal by the mother, miscarriage, and non-responding mothers. As Figure 1 demonstrates, this resulted in 466 completed informed consent forms of expectant mothers and 319 informed consents of their partners. All parents hereby gave permission to the researchers to retrieve information about the pregnancy and delivery from their midwives and information about the development of the child in the first year of life from National Health Care Centres. Of these parents, 409 mothers and 319 fathers agreed to complete questionnaires, of which 311 mothers and 243 fathers also agreed to participate during home visits (full participation).

The present study is a prospective longitudinal cohort study, in which pregnant women and their partners were followed from 15 weeks gestation until their child was 24 months old. As can be seen in Figure 2, pregnant women completed questionnaires at 15, 26, and 36 weeks gestational age. At 26 weeks of pregnancy, their partners also completed a questionnaire. At the same time a home visit took place during which a standardized interview concerning the prenatal representations of the unborn child was conducted with both parents separately.

Postnatally, there were five more measurement waves (at birth and at 4, 6, 12, and 24 months postpartum). Obstetric information about the birth of the child, including birth weight, Apgar score, and possible complications was registered by the midwives in line with their general practice guidelines. Additionally, midwives provided information about the presence of possible prenatal risk factors within families by completing an adapted Dutch version of the Dunedin Family Services Indicator (DFSI) [31]. At 4 and 6 months postnatally, both parents received questionnaires. At the child’s age of 6 months, an interview about the representations of their child was administered with both parents at their home, and the mother-child interaction was observed in several contexts within the home. Interviews generally lasted between 30 and 60 minutes and the observation of mother-infant interactions lasted approximately 20 minutes. All home visits were video-recorded. When children were 12 months old, mothers completed questionnaires concerning the child’s (social-emotional) development and information about the child’s physical development and well-being during the first year of life was retrieved from National Health Care Centres. At the child’s age of 24 months, the last measures concerning parental characteristics, the parent–child relationship, and the child’s development were completed by both parents.

Figure 3 shows which variables were investigated at different time points during the study. Generally, the study measures can be classified according to whether they concern parental characteristics, infant characteristics, or the parent-infant relationship. Therefore, the selected instruments are described below according to these categories.

The logistics of this study were carried out by three researchers (AM, CV, EdC) in close collaboration with the midwives participating in this study. Before starting data collection, a protocol was set up and discussed with participating midwives to ensure that a uniform protocol was followed by all midwifery practices. Participating midwives were instructed on how to recruit pregnant women for participation in the study and how to register the presence of possible prenatal risk factors.

Questionnaires were sent to parents one or two weeks before the time point they should be completed or before the home-visits. All questionnaires were available in Dutch and English. Reminders were sent when parents failed to return the questionnaires. Table 1 shows the number of parents that participated at each measurement wave.

The researchers (AM, CV, CR) and several research assistants were trained to administer and code the WMCI, concerning parents’ representations of their (unborn) children and to code observations of mother-child interactions. All interviews and mother-infant interactions were video-recorded and coded afterwards. A random subgroup of the interviews and observations was coded by more than one coder to determine inter-rater reliability.

Collected data were entered into an electronic database. Random samples of all manually processed questionnaires were double checked by the researchers to monitor the quality of the manual data entry. All measurements were checked by examination of the data, including their ranges, distributions, means, standard deviations, outliers, and logical errors.

Databases needed for answering specific research questions were centrally built from databases concerning different time points of the study. All information enabling identification of participants was erased from these databases, except identification numbers of each participant. Video-recordings of participants were stored on the computers of the researchers, which are only accessible with a password and not on web-based directories.

To answer the various research questions we will use structural equation modeling, regression analyses (HMR analyses), logistic regression analyses, and odds-ratio’s. Mediation and moderation analyses will follow Baron and Kenny’s requirements [60]. The power calculation is based on one of the main questions that will be addressed about the effects of prenatal (risk) factors on infant development. Assuming a moderate effect size of .30 or .40, a power of .80 (i.e., the minimal power for a similar study by Cohen [61]), an alpha of .05, and 11 parameters/predictors, we need a sample size of 220 participants (the power will be .83 with p = 11, r² = .09 or the power will be .99 with p = 11, r² = .16). Abovementioned power calculations are exact calculations, based on results of Gatsonis and Sampson [62]. Allowing for loss to follow-up by 24 months postpartum, we estimated that a sample of at least 240–260 women would be sufficient to test our hypotheses.