Skip to main content
. 2019 Feb 12;19:33. doi: 10.1186/s12905-019-0719-0

Table 3.

Multivariate logistic regression analyses of family factors for body weight gain in men

Model 1 (⊿BMI> = 2.5 kg/m2) Model 2 (⊿BMI> = 3.5 kg/m2)
ORsa 95% CIsa p for trendb ORsa 95% CIsa p for trendb
Marital status
 Unmarried 0.77 (0.52–1.14) 0.79 (0.51–1.22)
 Married 1.00 (Ref) 1.00 (Ref)
 Bereavement/Divorce 1.10 (0.74–1.65) 0.835 1.27 (0.82–1.98) 0.311
Family structure
 Single 1.03 (0.68–1.56) 1.37 (0.88–2.15)
 Couple 1.11 (0.86–1.43) 0.99 (0.74–1.32)
 Two generations 1.00 (Ref) 1.00 (Ref)
 Three generations 1.13 (0.90–1.41) 0.776 1.14 (0.89–1.48) 0.934
Family relationships
 Good 1.00 (Ref) 1.00 (Ref)
 Somewhat good 0.90 (0.75–1.09) 0.89 (0.72–1.10)
 ‘Not so good’/‘Not good’ 0.55 (0.19–1.54) 0.752 1.28 (0.45–3.66) 0.793

CIs confidence intervals, ORs odds ratios, Ref reference

aORs and CIs were adjusted for age, BMI and physical activity (as continuous variables for the three variables), smoking status (never, ex- and smokers = 0, 1 and 2), habitual drinking (never, ex- and drinkers = 0, 1 and 2), feeling stressed (many times, normal and rare = − 1, 0 and 1) and education level (< 12, 12 and ≥ 12 = 0, 1 and 2)

bTrend association was assessed by assigning ordinal numbers (− 1, 0 and 1) to unmarried, married and ‘bereavement/divorce’ for marital status, those (0, 1, 2 and 3) to single, couple, two generations and three generations for family structure, and then those (0, 1 and 2) to good, somewhat good and “not so good/not good” for family relationships, respectively