Table 3.
Change over time in dietary intakes (loge-transformed) in the intervention group compared with the control group of the MEDIM (‘impact of Migration and Ethnicity on Diabetes in Malmö’) intervention study. The analysis included all individuals with data available for at least two time points
Crude model | Extended model* | |||||
---|---|---|---|---|---|---|
Outcome | β † | P value | 95 % CI | β | P value | 95 % CI |
Total energy intake (kcal) n CG 34, n IG 37 | −0·024 | 0·24 | −0·065, 0·016 | −0·022 | 0·30 | −0·065, 0·020 |
Carbohydrate (g) n CG 34, n IG 37 | −0·020 | 0·42 | −0·069, 0·029 | −0·019 | 0·45 | −0·070, 0·032 |
Carbohydrate (E%) n CG 34, n IG 37 | 0·006 | 0·66 | −0·023, 0·036 | 0·005 | 0·74 | −0·026, 0·036 |
Saturated fat (g) n CG 34, n IG 37 | −0·043 | 0·15 | −0·102, 0·016 | −0·036 | 0·24 | −0·097, 0·025 |
Saturated fat (E%) n CG 34, n IG 37 | −0·017 | 0·39 | −0·058, 0·023 | −0·014 | 0·50 | −0·055, 0·027 |
Total fat (g) n CG 34, n IG 37 | −0·040 | 0·16 | −0·091, 0·015 | −0·034 | 0·22 | −0·089, 0·021 |
Total fat (E%) n CG 34, n IG 37 | −0·014 | 0·29 | −0·041, 0·013 | −0·013 | 0·35 | −0·041, 0·015 |
Protein (g) n CG 34, n IG 37 | 0·002 | 0·92 | −0·038, 0·043 | 0·0002 | 0·99 | −0·042, 0·042 |
Protein (E%) n CG 34, n IG 37 | 0·025 | 0·07 | −0·002, 0·054 | 0·022 | 0·14 | −0·007, 0·050 |
Sucrose (g) n CG 34, n IG 37 | −0·094 | 0·08 | −0·201, 0·013 | −0·071 | 0·18 | −0·176, 0·034 |
Sucrose (E%) n CG 34, n IG 37 | −0·067 | 0·18 | −0·166, 0·031 | −0·047 | 0·32 | −0·142, 0·047 |
Fibre intake (g/4184 kJ (1000 kcal)) n CG 34, n IG 37 | 0·043 | 0·08 | −0·004, 0·091 | 0·040 | 0·11 | −0·009, 0·089 |
n CG, number in control group (n for extended models can vary slightly because of missing values); n IG, number in intervention group (n for extended models can vary slightly because of missing values); E%, percentage of energy intake.
Adjusted for age, sex, BMI, economic difficulties, self-efficacy, smoking, physical activity and education.
β is a parameter estimate for the interaction term ‘group status’בtime since baseline visit (months)’. The β coefficient indicates percentage change (calculated as 100×exp(β – 1)) in the outcomes per month in the intervention group compared with the control group.