Table 4. Results of the logistic mixed model regressions of ALRI with main cooking fuel.
| Variables | Categories | Univariable | Multivariable Best-Fitting Models | ||||
|---|---|---|---|---|---|---|---|
| Model development dataset (N = 48815) a | Model development dataset (N = 48815) a | Final dataset(N = 56437) b | |||||
| Exposure to risks | |||||||
| Main cooking fuel | OR (95% CI) | p-value | OR (95% CI) | p-value | OR (95% CI) | p-value | |
| Clean fuels | - | - | - | - | - | - | |
| Kerosene | 1.60 (1.12, 2.28) | 0.0099 | 1.30 (0.81, 2.10) | 0.2755 | 1.23 (0.77, 1.95) | 0.3875 | |
| Coal, charcoal | 1.68 (1.39, 2.03) | <0.0001 | 1.46 (1.12, 1.90) | 0.0053 | 1.35 (1.06, 1.72) | 0.0139 | |
| Wood | 1.65 (1.38, 1.96) | <0.0001 | 1.38 (1.07, 1.79) | 0.0143 | 1.32 (1.04, 1.66) | 0.0209 | |
| Lower-grade biomass | 1.51 (1.07, 2.13) | 0.0186 | 1.12 (0.70, 1.79) | 0.6429 | 1.07 (0.69, 1.66) | 0.7743 | |
| Maternal smoking | Yes | 1.20 (1.03, 1.40) | 0.0221 | 1.21 (1.04, 1.42) | 0.0147 | 1.22 (1.06, 1.41) | 0.0070 |
| pLR = 0.0080 | No | - | - | - | - | - | - |
| Crowding | Not crowded | - | - | Factor was not selected for the best-fitting model | |||
| Crowded | 1.04 (0.98, 1.10) | 0.2048 | - | ||||
| Time to nearest water source pLR = 0.0013 | [coded in 10 min intervals] | 1.02 (1.01, 1.03) | <0.0001 | 1.01 (1.00, 1.02) | <0.0001 | 1.01 (1.01, 1.02) | 0.0010 |
| Non-modifiable risk factors | |||||||
| Child sex | Male | 1.10 (1.03, 1.16) | 0.0017 | 1.10 (1.04, 1.17) | <0.0001 | 1.13 (1.07, 1.19) | <0.0001 |
| pLR < 0.0001 | Female | - | - | - | - | - | - |
| Child age | [years] | 0.91 (0.89, 0.94) | <0.0001 | 0.91 (0.87, 0.94) | <0.0001 | 0.90 (0.87, 0.93) | <0.0001 |
| Birth order | 1 | ||||||
| pLR = 0.0118 | 2 | 0.89 (0.81, 0.98) | 0.0144 | 0.88 (0.80, 0.97) | 0.0080 | 0.89 (0.81, 0.97) | 0.0104 |
| ≥3 | 0.90 (0.84, 0.97) | 0.0064 | 0.88 (0.82, 0.96) | 0.0023 | 0.90 (0.83, 0.97) | 0.0048 | |
| Household socio-economic status | |||||||
| Shelter index pLR = 0.0004 | Low | - | - | - | - | - | - |
| Intermediate | 0.99 (0.91, 1.07) | 0.8175 | 1.00 (0.93, 1.09) | 0.9257 | 1.02 (0.94, 1.10) | 0.6996 | |
| High | 0.78 (0.72, 0.85) | <0.0001 | 0.83 (0.75, 0.92) | 0.0002 | 0.86 (0.79, 0.94) | <0.0001 | |
| Wealth index | Poorest | - | - | Factor was not selected for the best-fitting model instead shelter index was selected | |||
| Poorer | 0.98 (0.90, 1.07) | 0.6410 | - | ||||
| Middle | 0.96 (0.87, 1.04) | 0.3156 | - | ||||
| Richer | 0.87 (0.79, 0.95) | 0.0033 | - | ||||
| Richest | 0.75 (0.68, 0.84) | <0.0001 | - | ||||
| Maternal education | None | 1.30 (1.00, 1.70) | 0.0533 | 0.87 (0.64, 1.17) | 0.3474 | 0.91 (0.68, 1.22) | 0.5265 |
| pLR = 0.0002 | Primary | 1.46 (1.13, 1.90) | 0.0043 | 1.02 (0.77, 1.37) | 0.8693 | 1.08 (0.82, 1.44) | 0.5782 |
| Secondary | 1.26 (0.97, 1.64) | 0.0880 | 0.96 (0.72, 1.27) | 0.7530 | 1.01 (0.77, 1.34) | 0.9313 | |
| Higher | - | - | - | - | - | - | |
| Paternal education | None | 1.53 (1.27, 1.84) | <0.0001 | 1.36 (1.11, 1.68) | 0.0037 | 1.29 (1.06, 1.57) | 0.0102 |
| pLR = 0.1148 | Primary | 1.44 (1.20, 1.72) | <0.0001 | 1.23 (1.01, 1.50) | 0.0400 | 1.22 (1.01, 1.48) | 0.0363 |
| Secondary | 1.38 (1.15, 1.65) | <0.0001 | 1.23 (1.02, 1.50) | 0.0346 | 1.20 (1.00, 1.45) | 0.0481 | |
| Higher | - | - | - | - | - | - | |
| Never married d | 1.41 (1.14, 1.75) | 0.0018 | 1.15 (0.92, 1.46) | 0.2250 | 1.16 (0.93, 1.45) | 0.1833 | |
| Possession of health card | Yes | 0.99 (0.91, 1.08) | 0.7887 | Factor was not selected for the best-fitting model | |||
| No | - | - | - | ||||
| Vulnerability | |||||||
| Breastfeeding duration | Currently breastfeeding | 1.23 (0.97, 1.55) | 0.0891 | Factor was not selected for the best-fitting model | |||
| <12 month | 1.04 (0.80, 1.35) | 0.7492 | - | ||||
| 12–24 month | 1.01 (0.80, 1.29) | 0.9092 | - | ||||
| >24 month | 1.10 (0.84, 1.43) | 0.4845 | - | ||||
| Never | - | - | - | ||||
| Stunting | Not stunted | - | - | - | - | - | - |
| Stunted | 1.02 (0.96, 1.09) | 0.4513 | 1.12 (1.01, 1.23) | 0.0256 | 1.11 (1.01, 1.22) | 0.0290 | |
| Vaccination index | Low | 0.97 (0.90, 1.05) | 0.4787 | 0.83 (0.75, 0.92) | <0.0001 | 0.85 (0.77, 0.93) | <0.0001 |
| Intermediate | 1.10 (1.02, 1.19) | 0.0181 | 0.97 (0.87, 1.07) | 0.5179 | 0.98 (0.89, 1.09) | 0.7296 | |
| High | - | - | - | - | - | - | |
| Contextual factors | |||||||
| Rainy season | Yes | 0.74 (0.68, 0.80) | <0.0001 | 0.76 (0.55, 1.04) | 0.0907 | 0.79 (0.59, 1.05) | 0.1079 |
| No | - | - | - | - | - | - | |
| Geographic location | Countryside | 1.06 (0.85, 1.32) | 0.6137 | 0.90 (0.71, 1.14) | 0.3941 | 0.87 (0.70, 1.09) | 0.2228 |
| pLR = 0.0212 | Small city | 0.89 (0.70, 1.14) | 0.3529 | 0.79 (0.62, 1.01) | 0.0625 | 0.75 (0.60, 0.95) | 0.0147 |
| Town | 0.92 (0.72, 1.16) | 0.4673 | 0.88 (0.69, 1.13) | 0.3158 | 0.86 (0.68, 1.07) | 0.1717 | |
| Capital, large city | - | - | - | - | - | - | |
| Religion | Christian | - | - | - | - | - | - |
| pLR = 0.0088 | Muslim | 0.91 (0.81, 1.02) | 0.0908 | 0.89 (0.80, 1.00) | 0.0549 | 0.88 (0.80, 0.97) | 0.0132 |
| No or other religion | 1.03 (0.93, 1.15) | 0.5420 | 0.99 (0.88, 1.10) | 0.8088 | 0.98 (0.88, 1.09) | 0.7190 | |
| Unknown religion | 0.78 (0.64, 0.95) | 0.0147 | 0.76 (0.62, 0.92) | 0.0060 | 0.78 (0.64, 0.95) | 0.0131 | |
| Sex of household head | Male | 0.97 (0.90, 1.04) | 0.3472 | Factor was not selected for the best-fitting model | |||
| Female | - | - | - | ||||
| Age of household head | [years] | 1.00 (1.00, 1.00) | 0.1424 | Factor was not selected for the best-fitting model | |||
| Interactions | |||||||
| Main cooking fuel * rainy season | Clean fuels | - | - | - | - | - | - |
| c pLR<0.0001, <0.0341 | Kerosene | - | - | 1.33 (0.67, 2.63) | 0.4111 | 1.34 (0.69, 2.59) | 0.3884 |
| Coal, charcoal | - | - | 1.16 (0.81, 1.65) | 0.4149 | 1.14 (0.83, 1.58) | 0.4129 | |
| Wood | - | - | 0.94 (0.68, 1.30) | 0.7142 | 0.91 (0.68, 1.23) | 0.5443 | |
| Lower-grade biomass | - | - | 1.30 (0.67, 2.54) | 0.4373 | 1.39 (0.75, 2.57) | 0.3008 | |
| Stunting * child age | Not stunted | - | - | - | - | - | - |
| c pLR = 0.0792, 0.0411 | Stunted | - | - | 0.94 (0.89, 0.99) | 0.0300 | 0.95 (0.90, 1.00) | 0.0413 |
| Vaccination index * child age | Low | - | - | 1.06 (0.99, 1.12) | 0.0850 | 1.05 (0.99, 1.11) | 0.1118 |
| c pLR< 0.0014, 0.1495 | Intermediate | - | - | 1.06 (0.99, 1.13) | 0.0761 | 1.05 (0.99, 1.11) | 0.1259 |
| High | - | - | - | - | - | - | |
| Main cooking fuel * child age | - | - | Factor was not selected for the best-fitting model | ||||
* refers to an interaction between two variables. Please note, the presented ORs are coming from the estimate of the corresponding beta and are of pure statistical interest. Interpretable ORs for variables with interaction terms can be found in the text.
pLR is the p-value of the LR-Test comparing the final model with the final model without the corresponding covariate.
a Dataset without missing values in any explanatory variable (maternal smoking, crowding, time to nearest water source, shelter, wealth, and vaccination index, paternal and maternal education, possession of health card, child sex and age, birth order, breastfeeding duration, stunting, geographic location, sex and age of household head).
b Dataset without missing values in any explanatory variable of the final model (maternal smoking, time to nearest water source, shelter, and vaccination index, paternal and maternal education, child sex and age, birth order, stunting, geographic location).
c Both p-values relate to LR-Tests comparing (i) the final model with a model without main or interaction effects, and (ii) the final model with a model with main effects but without interaction effects.
d Women who were never married are per design of the questionnaires not asked about the education of their husband/partner.