Group based prenatal care in a low-and high risk population in the Netherlands: a study protocol for a stepped wedge cluster randomized controlled trial

Within the control period, all participants will be part of the control group and will receive usual individual prenatal care. See Fig. 1.

Within the intervention period, pregnant women are offered perinatal care in the CP model. In the CP model traditional prenatal care with one-to-one visits with a perinatal care provider is replaced by the use of a group model [35]. Care is provided by a midwife or an obstetrician (facilitator) and a co-facilitator to groups of eight to twelve women of similar gestational age. Groups meet eight to ten times during pregnancy at the usual scheduled visits, with sessions running for 90 to 120 min. The usual prenatal health assessment is integrated with information, education and peer support. All involved care providers will be trained by licensed trainers to perform CP. The intervention period begins when the center starts to recruit for their first CP group, after the randomized time point. All women included in the study during this period will be part of the intervention group. In addition to the study information, participants will be informed about CP at their first prenatal consultation and will be given the choice to participate in CP. Women that do not participate in CP will receive usual individual care, resulting in two sub-groups within the intervention group. See Fig. 1.

We will use data from the Perined database. The Perined database contains linked and validated routine care information concerning pregnancy, delivery, (re)admissions and pregnancy outcomes [36]. Data are routinely and separately registered by midwives, obstetricians, general practitioners and paediatricians/neonatologists. For this study, the Perined data will be complemented with data from the questionnaires to be completed by participating women at four time points: Table 1 provides an overview of the timing of the different items. Reminders will be sent after 1, 2 and 3 weeks and participants will be reminded by their care provider. Finally, we will use implementation data from CP-group-evaluations and qualitative data from semi-structured interviews with care providers. Outcome measurements are described below. More details on instruments used in the questionnaires are provided in Table 2.

The basic characteristics that will be collected are age, (parental) country of birth, religion, educational level, marital status, employment status, and parity.

The primary neonatal outcomes are perinatal mortality (defined as death per 1000 still- and life births from a GA above 22 weeks to 7 days postpartum) and perinatal morbidity. Perinatal morbidity (composite outcome) is defined as presence of congenital abnormalities, small for gestational age (birth weight below 10^th percentile), preterm birth (<37 weeks of GA), Apgar score after 5 min < 7 and/or admission to a neonatal intensive-care unit (immediately after birth). These outcomes will be retrieved from the Perined database, as well as gestational age at birth and birth weight. Other neonatal outcomes that will be addressed in the self-administered questionnaires are child’s crying- and sleeping behavior per 24 h (and to what extent women’s expectations on crying behavior are met), consultations with a general practitioner and/or pediatrician and admission to the hospital within 6 weeks postpartum.

We will collect data on maternal mortality and morbidity (defined as intensive care admission, eclampsia/HELLP, and/or postpartum hemorrhage ≥ 1000 ml), and prenatal and perinatal referrals to specialized obstetrical care.

Birth outcomes that will be collected are: intended and actual place of delivery, type of delivery (spontaneous or induced), mode of delivery (non-operative vaginal delivery, operative vaginal delivery, and planned and unplanned cesarean section), augmentation (yes/no), and need for analgesia (epidural, remifentanil, pethidine). These outcomes will be retrieved from the Perined database and complemented by data from the questionnaires regarding expectations on labor pain, intentional use of analgesia, perceived duration of delivery (Table 1). We will measure women’s readiness for labor using an item designed by Ickovics et al. [17]. Women’s delivery experience will be measured with the Labor and Delivery Index (LADY-X) [37] and women’s sense of control during delivery with the Labour Agentry Scale (LAS-10) [38].

Health behavior outcomes and health literacy will be collected using the questionnaires. Health behavior outcomes are: physical activity (number of 30-min exercise per week and self-judgment physical activity), nutritious behavior (breakfast-, vegetable-, fruit-, juice- and snack-consumption, and self-judgment eating behavior), substance use (tobacco-, alcohol-, and soft or hard drug consumption, and intentions towards smoking cessation), medication use (folic acid use and use of (prescribed) drugs), intention of infant feeding and attitude towards breastfeeding, and infant feeding on day one, at 1 week postpartum and at 6 weeks postpartum. Women’s prenatal and postnatal care knowledge will be measured with a scale developed by Ickovics et al. [17].

Psychosocial outcomes include perceived stress, coping, social support, and depression. Stress will be measured using the Prenatal Distress Questionnaire (NuPDQ), which is a revised version of the original 12-item scale developed by Yali and Lobel in 1999 [39‐41]. Socio-economic and relational stress will be measured by four items of the Cambridge Worry Scale (CWS), complementing the NuPDQ [42‐47]. We have added a concluding item on the amount of experienced stress on a scale from 0.0 to 10.0. To measure coping we will use a self-developed short instrument measuring three constructs based on the coping strategies as described by De Ridder & Schreurs (1994), and Savelkoul et al. (2000): problem focused active coping, emotion focused active coping and emotional passive coping [48‐50]. Each coping strategy is measured on a subscale and contains three items, which are based on the items of the revised Prenatal Coping Inventory (NuPCI), using the same answering scale (5 point Likert like scale, 0 = never tot 4 = very often) [51]. For example: “How often did you take a walk or performed other physical exercise to feel better?” (emotion focused active coping), “How often did you try not to think about it?” (emotional passive coping), and “How often did you talk to others in the same situation?” (problem focused active coping). Social support will be measured using the social support list-12 Interaction [52, 53]. Depression will be measured using the Edinburgh Postnatal Depression Scale (EPDS) [54].

We will measure women’s readiness for baby care using an item designed by Ickovics et al. [17]. Women’s self-efficacy in child care will be evaluated using the Parental Expectations Survey (PES) [55].

Women’s experience and satisfaction with prenatal care will be evaluated using the Patient Participation and Satisfaction Questionnaire (PPSQ) [56, 57].

We will assess the implementation of CP by monitoring the percentage of women that start CP, the addressed content within the group sessions, the involvement of women and their partners (adherence), and model fidelity. These data will be collected from the CP-group-evaluations, filled out by the group facilitators at the end of each session. These evaluations also contain data on group cohesion. Women’s experience with CP and their inhibiting and facilitating factors to participate will be addressed in the self-administered questionnaires: see Table 1. Inhibiting and facilitating factors in the implementation process and the care providers satisfaction with prenatal care will be addressed in individual semi-structured in-depth interviews. Health care utilization of women will be measured in the self-administered questionnaires, addressing preconceptional and prenatal utilization of general health care, and health care utilization provided by a perinatal care center.

A minimal sample of 600 pregnant women in both the intervention and control condition is needed to be able to accept with 95% confidence and an upper confidence limit less than 1.85, that after the intervention has been implemented, there is at least no significant difference in infant morbidity (using 14% of infant morbidity as outcome) between CP and individual care. This sample size is also largely sufficient to find amongst others a difference in prenatal care satisfaction or in proportion of breastfeeding comparable to Ickovics et al, with an α of 0.05 and a power of 0.90 [17]. We will account for a 20% loss to follow-up and aim to include in total 1600 women (800 in both conditions).

Data entry will be automatic by using a secured online survey system or (in case of hard copy) scanned and checked using the Teleform software. Data will be stored in a digital data base only accessible to the researchers. Before analysis, all data will be cleaned improving data quality.

We will conduct intention-to-treat analyses to compare the primary and secondary outcomes of the control group with those of the complete intervention group. Also a per protocol analysis will be performed comparing the outcomes of the three subgroups (see Fig. 1): control group (1), CP-participants (2) and non-CP-participants (3). Analyses will be descriptive and inferential (univariate and multivariate), and we plan to perform multilevel techniques to account for the clustering effect among participants in centers and CP-participants in CP-groups. The analysis of the implementation data and health care use will be descriptive. The relation between the degree of implementation and primary and secondary outcomes will be analyzed using multi-level analysis. Furthermore, the interviews with care providers will be conducted using a topic list based on the results of the analysis of the implementation data. They will be transcribed and qualitatively analyzed using a framework approach.