Table 3.
Results of generalized additive model.
| Independent variables | Respiratory infection/illness |
|||
|---|---|---|---|---|
| Estimate | Standard error | Odds ratio | p-value | |
| Cluster level variables | ||||
| Altitude (m) | 0·010 | 0·03 | 1·010 | 0·746 |
| Annual PM2·5 Levels (2019, μg/m3) | Smooth term–see graph | |||
| Annual Rainfall (2020, mm per year) | 0·115 | 0·023 | 1·122 | <0·001 |
| Malaria Prevalence (2020, Plasmodium falciparum parasite rate) | −0·133 | 0·026 | 0·875 | <0·001 |
| Urban designation (vs. rural) | −0·059 | 0·051 | 0·943 | 0·247 |
| Household variables | ||||
| Average temperature (month of interview) | Smooth term–see graph | |||
| Cooking fuel toxicity (Ref: Moderate to high toxicity) | ||||
| Non-toxic or Low-toxicity | −0·017 | 0·042 | 0·983 | 0·69 |
| Other | 0·001 | 0·079 | 1·001 | 0·989 |
| High quality housing material (vs. low quality) | −0·073 | 0·04 | 0·930 | 0·071 |
| Sanitation or toilet facilities (Ref: Improved sanitation) | ||||
| Open defecation | 0·047 | 0·043 | 1·048 | 0·271 |
| Unimproved sanitation | −0·079 | 0·1 | 0·924 | 0·429 |
| Season of Interview (Ref: Monsoon) | ||||
| Post-Monsoon | −0·162 | 0·057 | 0·850 | 0·005 |
| Summer | −0·515 | 0·072 | 0·598 | <0·001 |
| Winter | −0·346 | 0·066 | 0·708 | <0·001 |
| Smoking of tobacco in house: yes (vs. no) | 0·022 | 0·034 | 1·022 | 0·521 |
| Water source: piped into house | −0·063 | 0·043 | 0·939 | 0·138 |
| Wealth index poor (vs. not poor) | 0·037 | 0·048 | 1·038 | 0·443 |
| Child variables | ||||
| Age of child (months) | −0·189 | 0·025 | 0·828 | <0·001 |
| Female (vs. male) | −0·141 | 0·032 | 0·868 | <0·001 |
| Weight of child currently (kg) | −0·040 | 0·027 | 0·961 | 0·135 |
| Weight of child at birth (kg) | −0·010 | 0·016 | 0·990 | 0·525 |
| Total number of observations | 145,675 | |||
Respiratory Infection/Illness: Demographic Health Survey (DHS), has child suffered from cough AND fever AND short rapid breaths/difficulty breathing in last two weeks (yes vs. no)?
Cooking fuel toxicity: electricity, liquid petroleum gas, natural gas, and biogas are classified as non-toxic or low-toxicity, all else is moderate to highly toxic.
High quality housing material includes fabricated material, such as ceramic tiles, brick, cement, polished stone, marble. All else is low quality.
Sanitation/toilet facilities: As classified by DHS, improved is flush toilet or modified pit latrine, unimproved is open pit latrine or bucket, open is non toilet facility or location.
Urban designation: As classified by DHS, including large cities, small cities, and towns (other urban areas); all else rural.
Wealth Index: Composite measure of household's cumulative living standard, divided in quintiles by DHS, and dichotomized into poor (bottom two quintiles) vs. not poor.
Estimates and p-values obtained via generalized additive multivariable logistic regression, dependent variable was respiratory infection. p-values <0·05 were considered as statistically significant (in bold).
Models include random effects for cluster.