Table 2

Associations of mean gestational age at birth and mean birth weight with pregnancy, parental and environmental factors from adjusted linear mixed-effects model (model 3)

Gestational age (days)
absolute differences (95% CI)
Birth weight (g) *
relative differences (95% CI)
Eligible populationComplete case populationEligible populationComplete case population
Intercept 277.3 (277.2 to 277.5)277.9 (277.7 to 278.2)3278 (3218 to 3339)†3298 (3180 to 3420)†
Sex
Female0011
Male−0.56 (-0.65 to -0.48)−0.63 (-0.79 to -0.47)1.045 (1.044 to 1.046)1.048 (1.046, 1.049)
Birth rank
1‡0011
2−0.39 (-0.49 to -0.29)−0.34 (-0.52 to -0.16)1.038 (1.037 to 1.039)1.039 (1.037, 1.041)
3−0.37 (-0.52 to -0.22)−0.16 (-0.44 to 0.11)1.050 (1.048 to 1.051)1.054 (1.051, 1.057)
≥4−0.24 (-0.50 to 0.02)0.24 (-0.25 to 0.72)1.058 (1.056 to 1.061)1.065 (1.059, 1.070)
Age mother (years)§
<20 (per 5 years decr.)−4.10 (-5.59 to -2.61)1.002 (0.987 to 1.017)
≥20–30‡0011
≥30–40 (per 5 years)−0.99 (-1.06 to -0.91)−0.93 (-1.07 to -0.78)1.000 (1.000 to 1.001)0.998 (0.997, 1.000)
≥40 (per 5 years)−2.93 (-3.36 to -2.50)−3.46 (-4.29 to -2.63)0.998 (0.994 to 1.003)0.998 (0.990, 1.006)
Civil status¶
Married0011
Not married−0.01 (-0.13 to 0.10)0.15 (-0.08 to 0.38)0.990 (0.989 to 0.991)0.993 (0.99, 0.995)
Nationality mother
Switzerland‡0011
S Europe0.20 (-0.01 to 0.40)0.39 (00 to 0.78)0.994 (0.992 to 0.996)0.995 (0.991, 0.999)
W Europe0.20 (0.02 to 0.38)−0.08 (-0.43 to 0.26)1.008 (1.006 to 1.010)1.007 (1.004, 1.011)
N Europe0.37 (-0.07 to 0.81)0.30 (-0.57 to 1.17)1.025 (1.020 to 1.029)1.022 (1.013, 1.031)
E Europe0.21 (0.04 to 0.38)0.33 (-0.01 to 0.68)1.013 (1.011 to 1.014)1.017 (1.014, 1.021)
Other−0.32 (-0.49 to -0.14)−0.67 (-1.05 to -0.30)1.007 (1.005 to 1.008)1.012 (1.008, 1.016)
Nationality father
Switzerland‡0011
S Europe−0.46 (-0.64 to -0.28)−0.28 (-0.62 to 0.06)0.991 (0.990 to 0.993)0.993 (0.989, 0.996)
W Europe0.07 (-0.11 to 0.25)0.30 (-0.04 to 0.63)1.008 (1.006 to 1.009)1.006 (1.003, 1.010)
N Europe0.51 (0.08 to 0.94)−0.24 (-1.09 to 0.62)1.013 (1.009 to 1.017)1.011 (1.003, 1.020)
E Europe−0.46 (-0.64 to -0.28)−0.01 (-0.38 to 0.36)1.009 (1.007 to 1.010)1.011 (1.008, 1.015)
Other−0.02 (-0.22 to 0.18)0.48 (0.05 to 0.90)0.992 (0.991 to 0.994)0.992 (0.987, 0.996)
Missing−3.87 (-4.24 to -3.50)0.989 (0.985 to 0.992)
Education mother
Tertiary‡01
Secondary−0.55 (-0.74 to -0.36)0.996 (0.995, 0.998)
Compulsory−0.90 (-1.22 to -0.58)0.993 (0.990, 0.996)
Education father
Tertiary‡01
Secondary−0.16 (-0.35 to 0.03)0.996 (0.994, 0.998)
Compulsory−0.25 (-0.58 to 0.07)0.997 (0.994, 1.000)
Altitude (m)
500‡0011
per 500 m increase0.07 (-0.09 to 0.23)0.03 (-0.24 to 0.30)0.989 (0.988 to 0.991)0.989 (0.987, 0.992)
Urbanisation
Urban‡0011
Peri-urban−0.43 (-0.57 to -0.28)−0.59 (-0.82 to -0.36)1.001 (1.000 to 1.002)1.003 (1.000, 1.005)
Rural−0.15 (-0.32 to 0.02)−0.29 (-0.55 to -0.02)1.000 (0.998 to 1.001)1.003 (1.001, 1.006)
Language region
German‡0011
French−0.62 (-0.77 to -0.47)−0.66 (-0.88 to -0.44)0.989 (0.987 to 0.990)0.988 (0.985, 0.990)
Italian−0.94 (-1.26 to -0.63)−1.11 (-1.55 to -0.68)0.982 (0.980 to 0.985)0.983 (0.979, 0.987)
Percent of spatial variance explained** 31%39%87%88%
  • *Birth weight was modelled on a log scale, which results in multiplicative effects. The model for birth weight was additionally adjusted for gestational age by a cubic spline function with knots at weeks 25, 30 and 35.

  • †In the model for BW, the intercept corresponds to an estimated mean birth weight (g) for a singleton girl born at gestational age 40 weeks as the first child (rank 1) in a German-speaking, urban region of elevation 500m, whose mother is 20-30 years old at birth and married, and both parents have Swiss nationality and tertiary education.

  • ‡Reference category.

  • §Age modelled by a piece-wise linear function: constant at reference range ≥20-30, and separate slopes for age <20, ≥30-40, and ≥40.

  • ¶Married or in registered partnership / Not married: Single, widow, divorced or in dissolved partnership.

  • **Percentage of regional variance explained by model predictors, i.e. percent reduction in variance of random effects (σ2) when compared to model with no predictors (model 0).