Table 3.
Multivariable Poisson regression analyses of the relationship of total diabetes with educational attainment and BMI, overall and by World Bank income classification
| Model (variable adjusted for) | Overall (N = 562,445) | P | LIC (N = 26,629) | P | LMI (N = 510,377) | P | UMI (N = 25,439) | P |
|---|---|---|---|---|---|---|---|---|
| Model 1 | ||||||||
| No formal education | Ref. | Ref. | Ref. | Ref. | ||||
| Primary school | 1.30 (1.15–1.46) | <0.001 | 1.53 (1.29–1.82) | <0.001 | 1.29 (1.18–1.41) | <0.001 | 1.21 (0.96–1.52) | 0.107 |
| ≥Secondary education | 1.58 (1.43–1.75) | <0.001 | 2.15 (1.77–2.61) | <0.001 | 1.65 (1.54–1.77) | <0.001 | 1.42 (1.12–1.79) | 0.003 |
| Model 2 | ||||||||
| No formal education | Ref. | Ref. | Ref. | Ref. | ||||
| Primary school | 1.24 (1.10–1.40) | <0.001 | 1.44 (1.21–1.71) | <0.001 | 1.16 (1.07–1.27) | <0.001 | 1.20 (0.95–1.51) | 0.119 |
| ≥Secondary education | 1.36 (1.22–1.52) | <0.001 | 1.66 (1.37–2.02) | <0.001 | 1.24 (1.15–1.34) | <0.001 | 1.34 (1.07–1.69) | 0.012 |
| Model 3 | ||||||||
| No formal education | Ref. | Ref. | Ref. | Ref. | ||||
| Primary school | 1.17 (1.04–1.31) | 0.009 | 1.36 (1.15–1.62) | <0.001 | 1.10 (1.01–1.20) | 0.025 | 1.17 (0.93–1.46) | 0.177 |
| ≥Secondary education | 1.24 (1.11–1.38) | <0.001 | 1.47 (1.22–1.78) | <0.001 | 1.14 (1.06–1.23) | 0.001 | 1.28 (1.02–1.61) | 0.033 |
| Model 4 | ||||||||
| Underweight BMI | 0.65 (0.57–0.75) | <0.001 | 0.70 (0.58–0.85) | <0.001 | 0.71 (0.64–0.79) | <0.001 | 0.56 (0.35–0.89) | 0.015 |
| Normal weight BMI | Ref. | Ref. | Ref. | Ref. | ||||
| Overweight BMI | 1.89 (1.71–2.07) | <0.001 | 1.70 (1.43–2.02) | <0.001 | 1.80 (1.68–1.92) | <0.001 | 1.95 (1.54–2.14) | <0.001 |
| Obese BMI | 2.89 (2.56–3.26) | <0.001 | 1.54 (1.24–1.91) | <0.001 | 2.64 (2.37–2.95) | <0.001 | 3.45 (2.77–4.30) | <0.001 |
Data are RR (95% CI).
Model 1 was adjusted for age and sex. Model 2 was adjusted for age, sex, and wealth. Model 3 was adjusted for age, sex, wealth, and BMI. Model 4 was adjusted for age, sex, education, and wealth.
SEs were adjusted for clustering at the primary sampling unit level, and each country was weighted proportionate to its population size.
Burkina Faso, Chile, Costa Rica, and Fiji were dropped from multivariable prediction models due to a lack of data on household wealth quintile. Seychelles was also excluded, since a harmonized wealth quintile was not available at the time of this analysis.
Multivariable Poisson regression for the relationship between total diabetes and wealth quintile is provided in Supplementary Data.