The MAMAACT Intervention

The intervention is a complex intervention and designed as a cluster randomised trial. 19 maternity wards participate and by simple randomization within thee groups of maternity wards according to the proportion of non-Western migrant women giving birth, 10 clusters are in the intervention group and 9 in the control group.

The effectiveness of the intervention will be analysed in a difference-in-difference design, where the change in outcomes from before to after intervention period at the intervention sites are compared to the change at the control sites. Cross sectional surveys will be conducted before and after the implementation of the intervention.

Participants for the pre-implementation survey will be enrolled from May to July 2018. Participants for the post-implementation survey will be enrolled from May to July 2019. Only participants from the post-implementation survey at the intervention sites will be exposed to the intervention. The start date for the enrollment of participants exposed to the intervention will therefore be May 1st 2019.

The primary trial outcome is studied using telephone surveys with pregnant women recruited through the antenatal care services. The questionnaire will include validated items on the health literacy dimension actively engagement with health care providers (5 items) and health system navigation (5 items). Women are recruited when they attend for the week 20 ultra sound scan and will be interviewed between pregnancy week 30+0 and 38+7. The questionnaires are managed in six languages (Danish, Arabic, English, Turkish, Somali and Urdu, and bilingual and trained interviewers conduct the interviews.

Using unpublished data from pregnant women at Hvidovre Hospital in 2016 (n=407), Danish born women are found to have a mean level of the relevant health literacy dimension on 4.16 and non-Western migrant women have a significantly lower mean at 3.97. The goal is that the MAMAACT intervention will to improve the mean level of the non-Western migrant women to the level of the Danish born women at baseline, thus an effect on 0.2 on the mean of the dimension are to be shown. A recent Danish cluster randomized trial aiming at improved breast feeding by improving the guidance of new mothers at maternity wards has reported intra-cluster correlation coefficients (ICC) under 0.005.

For the sample size calculation an estimation with 10 cluster in the control group and 10 clusters in the intervention group were used, however after recruitment and randomization 9 wards were in the control group and 10 in the intervention group. With an ICC on 0.005 with 10 clusters in the intervention arm and 10 clusters in the control arm an effect on 0.1 with a strength of 80 % can be detected if data from 2580 of the total population before and after the intervention (1290 from the intervention before and after and 1290 from the control group before and after) is collected.

With an ICC on 0.005 with 10 clusters in the intervention arm and 10 clusters in the control arm an effect on 0.2 with a strength of 80 % can be detected if data is collected from 500 non-Western women before and after the intervention (250 from the intervention both before and after and 250 from the control group both before and after).

Based on telephone survey data, effects on the skills in health system navigation among the women will also be analysed.

The overall effects of the MAMAACT intervention on stillbirth and neonatal health will be analysed by the medical birth registry and obstetric databases. Stillbirth is defined as death of a child without any signs of life after 22 completed weeks of gestation. Neonatal death is defined as death of a child after live birth before 1 months of age. No significant effects on these outcomes are expected due to low numbers in the short intervention period. Secondary trial outcomes are 5-minute Apgar score, umbilical-cord blood gas analysis, transfer to a neonatal intensive care unit, gestational age at birth, fetal birthweight according to gestational age, ICD-10 code for sign of asphyxia, and pregnancy complications (preeclampsia).

All effectiveness analysis will be using an intention-to-treat approach and will account for the clustering of individuals and potential confounding in a controlled mixed effect regression.

The implementation of the intervention will be analyzed using qualitative data. The resarch questions are: How do midwives and non-Western immigrant women react to the MAMAACT intervention across Denmark? Which barriers exist in the organisation of antenatal care, and how do they affect the national implementation of the MAMAACT intervention? What are non-Western immigrant women's everyday life conditions in Denmark, and how do they affect women's responses to potential pregnancy complications?