Table 2.
Predictors of self-reported chronic conditions
|
Predictor variables |
Overall chronic conditions |
Diabetes |
Hypertension |
|||
|---|---|---|---|---|---|---|
| Adjusted odds ratio* (95% CI) | p value | Adjusted odds ratio* (95% CI) | p value | Adjusted odds ratio* (95% CI) | p value | |
|
Sex |
|
|
|
|
|
|
| Male |
- |
- |
- |
- |
- |
- |
| Female |
3.2 (2.6, 4.0) |
<0.001 |
2.5 (1.8, 3.5) |
<0.001 |
4.6 (3.6, 5.8) |
<0.001 |
|
Age groups (years) |
|
|
|
|
|
|
| ≤19 |
- |
- |
- |
- |
- |
- |
| 20-39 |
6.7 (4.8, 9.5) |
<0.001 |
10.9 (4.9, 24.0) |
<0.001 |
12.2 (7.3, 20.3) |
<0.001 |
| ≥40 |
58.8 (36.3, 95.2) |
<0.001 |
106.8 (40.7, 280.2) |
<0.001 |
116.1 (59.5, 226.4) |
<0.001 |
|
Monthly per capita income | ||||||
| First quintile |
- |
- |
- |
- |
- |
- |
| Second quintile |
0.8 (0.6, 1.0) |
0.047 |
0.8 (0.5, 1.2) |
0.226 |
0.8 (0.6, 1.1) |
0.211 |
| Third quintile |
0.5 (0.3, 0.8) |
0.002 |
0.5 (0.2, 1.1) |
0.097 |
0.6 (0.4, 1.0) |
0.056 |
| Fourth quintile |
0.4 (0.2, 0.7) |
0.001 |
0.3 (0.1, 1.1) |
0.072 |
0.4 (0.2, 0.9) |
0.023 |
| Fifth quintile |
0.2 (0.1, 0.5) |
<0.001 |
0.2 (0.1, 1.1) |
0.072 |
0.3 (0.1, 0.9) |
0.026 |
|
Household poverty status |
|
|
|
|
|
|
| Above the poverty line |
- |
- |
- |
- |
- |
- |
| Below the poverty line |
3.0 (1.5, 5.8) |
0.002 |
0.6 (0.5, 0.7) |
<0.001 |
1.9 (0.7, 4.9) |
0.196 |
|
Religion |
|
|
|
|
|
|
| Islam |
- |
- |
- |
- |
- |
- |
| Hinduism |
0.9 (0.8, 1.1) |
0.227 |
1.0 (0.8, 1.1) |
0.527 |
0.9 (0.8, 1.1) |
0.177 |
| Christianity |
1.2 (1.0, 1.5) |
0.078 |
1.0 (0.8, 1.2) |
0.665 |
1.2 (0.9, 1.5) |
0.175 |
|
Interaction terms |
|
|
|
|
|
|
| Sex*Religion |
0.7 (0.6, 0.8) |
<0.001 |
|
|
0.7 (0.6, 0.8) |
<0.001 |
| Sex* Monthly per capita income |
|
|
0.8 (0.8, 0.9) |
<0.001 |
|
|
| Age group*Monthly per capita income |
1.1 (1.1, 1.2) |
<0.001 |
1.2 (1.0, 1.4) |
0.007 |
1.1 (1.0, 1.2) |
0.019 |
| Age group*Household poverty status | 0.6 (0.5, 0.8) | <0.001 | 0.7 (0.5,1.0) | 0.039 | ||
- Referent category. *Adjusted odds ratio as obtained from multivariable logistic regression models. All the predictor variables were included in the initial model, including two-way interaction terms that were significant at p < 0.05 during binominal logistic regression. Similar to a backward elimination technique, the predictors that were not significant at p < 0.05 were then dropped individually, and the resultant models were compared for goodness of fit (using a likelihood-ratio test) until no further improvement was possible.