Table 4.
Multivariable logistic regression models for the association between total DDS and variety in food intake and prevalence of depression in the PREDIMED-Plus study participants. Odds ratios (95 % confidence intervals) (total population = 6625)
| Total DDS | Q1 (n 1657) | Q2 (n 1656) | Q3 (n 1656) | Q4 (n 1656) | P for trend * | ||||
|---|---|---|---|---|---|---|---|---|---|
| OR | 95 % CI | OR | 95 % CI | OR | 95 % CI | OR | 95 % CI | ||
| Model 1 | 1 | Ref. | 0·89 | 0·75, 1·04 | 0·87 | 0·74,1·02 | 0·73 | 0·62, 0·87 | <0·001 |
| Model 2 | 1 | Ref. | 0·92 | 0·78, 1·08 | 0·88 | 0·75, 1·04 | 0·76 | 0·64, 0·90 | 0·001 |
| Vegetable group | C1 (n 551) | C2 (n 1319) | C3 (n 2492) | C4 (n 2263) | |||||
| Model 1 | 1 | Ref. | 0·82 | 0·65, 1·04 | 0·76 | 0·61, 0·94 | 0·72 | 0·58, 0·90 | 0·004 |
| Model 2 | 1 | Ref. | 0·83 | 0·65, 1·05 | 0·78 | 0·63, 0·97 | 0·75 | 0·60, 0·94 | 0·017 |
| Fruit group | C1 (n 848) | C2 (n 4529) | C3 (n 779) | C4 (n 469) | |||||
| Model 1 | 1 | Ref. | 0·89 | 0·75, 1·06 | 0·80 | 0·63, 1·00 | 0·81 | 0·62, 1·06 | 0·051 |
| Model 2 | 1 | Ref. | 0·91 | 0·76, 1·08 | 0·81 | 0·64, 1·03 | 0·79 | 0·60, 1·04 | 0·043 |
| Cereal group | C1 (n 353) | C2 (n 4791) | C3 (n 1396) | C4 (n 85) | |||||
| Model 1 | 1 | Ref. | 0·69 | 0·54, 0·87 | 0·71 | 0·55, 0·92 | 0·72 | 0·41, 1·25 | 0·197 |
| Model 2 | 1 | Ref. | 0·69 | 0·54, 0·89 | 0·71 | 0·54, 0·94 | 0·81 | 0·46, 1·45 | 0·320 |
| Proteins group | C1 (n 25) | C2 (n 1258) | C3 (n 2787) | C4 (n 2555) | |||||
| Model 1 | 1 | Ref. | 0·31 | 0·13, 0·71 | 0·26 | 0·11, 0·59 | 0·23 | 0·10, 0·52 | <0·001 |
| Model 2 | 1 | Ref. | 0·31 | 0·13, 0·72 | 0·26 | 0·11, 0·61 | 0·24 | 0·10, 0·56 | <0·001 |
| Dairy group | C1 (n 690) | C2 (n 2454) | C3 (n 2622) | C4 (n 859) | |||||
| Model 1 | 1 | Ref. | 0·97 | 0·79, 1·19 | 0·89 | 0·73, 1·08 | 0·88 | 0·69, 1·11 | 0·105 |
| Model 2 | 1 | Ref. | 0·99 | 0·81, 1·21 | 0·93 | 0·76, 1·15 | 0·91 | 0·72, 1·17 | 0·292 |
C, category; DDS, dietary diversity score; Q, quartile (Q1, less diversity; Q4, more diversity).
DDS/food group measure as continuous variables in order to estimate P for trend.
Values are presented as OR and 95 % CI for prevalence of depression (≥18 p at Beck Depression Inventory II and/or a current depression diagnosis) as categorical variable according to total DDS and variety in food intake.
Model 1: Adjusted for sex and age.
Model 2: Additionally adjusted for energy intake, smoking habits, physical activity, educational level, BMI, living alone, civil status, sleep duration and presence of chronic diseases.
Values presented in bald showed a statistically significant association (P < 0·05).
DDS cut-off points for each quartile: (Q1 = 0·8–3·9, Q2 = 4·0–4·6, Q3 = 4·7–5·4, and Q4 = 5·5–8·0).
The variety in each food group was classified into four categories (C): (C1 = 0 points), (C2 => 0–≤0·5 points), (C3 => 0·5–<1 points) and (C4 ≥ 1 point).