Table 5. Ordinary least squares regressions of the natural logarithm of cardiovascular risk on sociodemographic covariates and a district-level fixed effect.
| Female (n = 419,478) | Male (n = 375,642) | |||||||
|---|---|---|---|---|---|---|---|---|
| Adjusted for age group only1 | Adjusted for all covariates2 | Adjusted for age group only1 | Adjusted for all covariates2 | |||||
| Coefficient3 (95% CI) |
p-Value | Coefficient3 (95% CI) |
p-Value | Coefficient3 (95% CI) |
p-Value | Coefficient3 (95% CI) |
p-Value | |
| Household wealth quintile | ||||||||
| 1 (poorest) | Ref. | Ref. | Ref. | Ref. | ||||
| 2 | 2.22 (1.75–2.69) | <0.001 | 2.44 (1.97 to 2.91) | <0.001 | 0.64 (0.22–1.05) | 0.003 | 0.83 (0.42–1.25) | <0.001 |
| 3 | 4.68 (4.18–5.17) | <0.001 | 5.32 (4.82 to 5.82) | <0.001 | 1.97 (1.53–2.41) | <0.001 | 2.45 (2.00–2.90) | <0.001 |
| 4 | 7.47 (6.96–7.98) | <0.001 | 8.31 (7.79 to 8.84) | <0.001 | 3.91 (3.46–4.36) | <0.001 | 4.55 (4.08–5.02) | <0.001 |
| 5 (richest) | 13.11 (12.57–13.66) | <0.001 | 14.14 (13.55 to 14.74) | <0.001 | 7.71 (7.23–8.19) | <0.001 | 8.44 (7.91–8.97) | <0.001 |
| Educational attainment | ||||||||
| <Primary school | Ref. | Ref. | Ref. | Ref. | ||||
| Primary school | 7.20 (6.72–7.67) | <0.001 | 4.03 (3.54 to 4.51) | <0.001 | 1.99 (1.57–2.41) | <0.001 | 0.86 (0.44–1.28) | <0.001 |
| Middle school | 8.80 (8.31–9.29) | <0.001 | 4.09 (3.59 to 4.59) | <0.001 | 2.69 (2.29–3.10) | <0.001 | 0.83 (0.42–1.24) | <0.001 |
| Secondary school | 11.61 (11.07–12.15) | <0.001 | 4.59 (4.02 to 5.16) | <0.001 | 5.35 (4.95–5.76) | <0.001 | 2.23 (1.80–2.66) | <0.001 |
| High school | 10.71 (9.98–11.44) | <0.001 | 2.12 (1.36 to 2.88) | <0.001 | 4.59 (4.09–5.10) | <0.001 | 0.54 (0.01–1.07) | 0.047 |
| >High school | 10.69 (9.99–11.39) | <0.001 | −0.66 (−1.42 to 0.10) | 0.087 | 7.06 (6.60–7.52) | <0.001 | 1.09 (0.57–1.61 | <0.001 |
| Location | ||||||||
| Rural | Ref. | Ref. | Ref. | Ref. | ||||
| Urban | 11.96 (11.62–12.30) | <0.001 | 12.51 (12.14 to 12.88) | <0.001 | 6.92 (6.62–7.23) | <0.001 | 7.35 (7.02–7.68) | <0.001 |
Standard errors were adjusted for clustering at the level of the primary sampling unit.
1These models included 1 sociodemographic characteristic, age group, and a binary indicator variable for each district as explanatory variables.
2These models included all variables listed in the table, age group, and a binary indicator for each district as explanatory variables.
3Coefficients were multiplied by 100 so that they can be interpreted as an approximation of the percentage change in cardiovascular risk associated with a 1-unit change in the explanatory variable.
CI, confidence interval; Ref., reference category.