Study design
In this national population-based study, we used individual record data from the Medical Birth Registry of Norway (MBRN) (12, 13) and Statistics Norway (14). The data were merged using each woman’s unique national identity number. The MBRN is the repository for mandatory notification of all births in Norway, and includes information on women’s obstetric background, maternal health before and during pregnancy, current pregnancy, labour and birth, and maternal and infant outcomes. The MBRN data are collected from medical records and women’s self-reported obstetric history. SSB provides information on migration and socioeconomic factors.
Setting
In Norway, the health care system is considered of high quality with low maternal and child mortality rates (15). All women are entitled to free maternity care in Norway, and the vast majority of women give birth in public hospitals (99%) (16). Unless there are medical complications necessitating specialist obstetric care, women may choose antenatal care provided by either a general practitioner, a midwife, or a combination of the two (17). However, inequalities in health care have been reported and migrant women in Norway appear more likely to receive suboptimal care compared to non-migrant women (18). In 2018, 29% of children born in Norway were born to a migrant mother (19).
Study population
The main goal of this study was to compare subsequent birth outcomes in migrant women who already had a child before arriving in Norway (defined as the exposure group) with the same outcomes in migrant women with a first birth in Norway (defined as the comparison group). In order to control for possible parity-related differences between exposure and comparison groups, we restricted the exposure group to include women with only one birth before arriving in Norway (Figure 1).
Initially, MBRN comprised 1,620,532 births during the period 1990-2016. Births to second generation migrant women, those with unknown or mixed background, such as adoptees or women with one Norwegian-born and one foreign-born parent, were excluded (n=87,696). The final sample included the second and any subsequent singleton births to foreign-born women with two foreign-born parents (n=96,068 births to migrant women), and Norwegian-born women with two Norwegian-born parents (n=521,004 births to Norwegian-born women) giving birth in Norway between the years 1990 and 2016 (Figure 1).
Country of woman’s first birth
To derive information on whether a woman had a first child before or after immigration to Norway, we used the following algorithms:
Migrant women: The country of a woman’s first birth was determined by the woman’s first parity registered in the MBRN dataset. If a parous woman’s first birth was in the dataset, the birthplace of her firstborn baby was classified as Norway. If the woman’s first birth was not in the dataset, the birthplace was classified as other than Norway. Women with permission to stay in Norway prior to 1990 may or may not have given birth in Norway before 1990 (the study period commencement) and were therefore excluded (n=35,929).
Norwegian-born women: To identify country of first birth we excluded births to any woman 13 years or older in 1990 whose first birth was not available in the MBRN dataset (n=193,243) and therefore could in theory have had previous babies before 1990. The women’s first parity registered in the dataset was then used to identify country of first birth in the Norwegian-born women. The age limit was chosen based on the fact that the youngest mothers in our dataset were 13 years of age.
Adverse neonatal outcomes
Gestational age was based on ultrasound estimation or, when such information was lacking, calculated from the first day of the last menstrual period. Very preterm birth, moderately preterm and post-term birth were defined as births in gestational week 22-31, 32-36 and >42, respectively. In the analyses of very preterm birth, moderately preterm and post-term birth, we excluded births with unknown gestational age (migrant women n=1,512; Norwegian-born women n=12,677) and term births were used as comparison group. In the analyses of small for gestational age (SGA) and large for gestational age (LGA) we also excluded births with unknown birthweight (migrant women n=63; Norwegian-born women n=403). For calculating SGA and LGA, we used Norwegian standards combining information on gestational age, birthweight and gender (20). Low Apgar Score was defined as <7 at 5 minutes. Stillbirth was defined as a pregnancy loss at ≥22 weeks of gestation or birthweight ≥500 grams if data on gestational age were missing. Neonatal death was defined as a live born infant at ≥22 weeks of gestation (or with a birthweight ≥500 grams if data on gestational age was missing) who died within 28 days after the birth.
Other variables
From the MBRN, we also obtained data on year of birth, maternal age (<25, 25-34, ≥35 years), single status (yes, no), parity (1, 2, 3, ≥4), smoking in early pregnancy (yes/no) and previous stillbirth (yes, no).
For each birth year, SSB provided data on maternal level of education (no education, primary school, secondary school, university/college, missing), mother’s gross income (categorized into quartiles, missing), reason for immigration (Nordic migrants, work/education, family reunion or establishment, refugee, missing), and paternal origin (Norwegian-born, foreign-born, missing). Maternal country of birth from SSB was used to classify women according to seven Global Burden of Disease super regions (GBD) (21): High income countries; Central Europe, Eastern Europe, and Central Asia; Sub-Saharan Africa; North Africa and Middle East; South Asia; Southeast Asia, East Asia, and Oceania; Latin America and Caribbean. Maternal length of residence was calculated as the difference between the year of birth and the year a woman officially received her permission to stay in Norway (<2 years, 2-5 years, 6-9 years, ≥10 years). Maternal age at immigration was calculated as the difference between maternal age at birth and her length of residence (<18 years, ≥18 years).
Statistics
Neonatal outcomes were compared between births to: 1) migrant women with a first birth before immigration to Norway versus those with a first birth after immigration, and 2) Norwegian-born women with a first birth outside Norway versus those with a first birth in Norway. We also compared births to migrant women with a first birth before immigration to Norway versus Norwegian-born women with a first birth outside Norway.
Logistic regression analyses were used to investigate possible associations between country of first childbirth (Norway/Other than Norway) and adverse neonatal outcomes in subsequent births. Associations were reported as odds ratios with 95% confidence intervals. Adjustment variables were year of birth, maternal age, parity, marital status, maternal education and mother’s gross income. To account for dependency between births by the same mother, we used robust standard errors that allowed for within-mother clustering.
To avoid list-wise deletion and potential bias due to missing data in covariates in the adjusted regression models, we used a multiple imputation technique to replace missing values in covariates. Ten imputed datasets were created using the multivariate normal model (22). Separate imputation models were created for each neonatal outcome and included the respective outcome (very preterm birth, moderately preterm birth, post-term birth, SGA, LGA, low Apgar score, stillbirth or neonatal death), as well as country of first childbirth and adjustment variables.
Analyses were performed using Stata IC version 16 (Stata Statistical Software, College Station, TX, USA) for Windows.