Fig. 2

Estimated adjusted incidence of preeclampsia by length of residence for various immigration reasons in Norway, 1990–2013. The incidences were estimated for primiparous and multiparous immigrant women by using generalized additive logistic regression models, adjusted for year of birth, maternal age at birth, parity, marital status at birth, chronic hypertension, pre-pregnancy diabetes, maternal income and education. The incidence trajectories for each immigration group are shown for secondary school and third income quartile and at the means of the other covariates (see Table 1). Due to small numbers, lengths of residence above the 95th percentile of the distributions were excluded. The distribution of length of residence is shown on the x-axis as frequency bars (highest and lowest frequencies are shown on right vertical axis)