Table 3. Best-fit model predicting household-level long lasting insecticidal net (LLIN) retention after a targeted distribution, controlling for repeated observations on households and community-level clustering among 10 Indigenous Batwa communities in Uganda.
| Variable | Odds ratio | 95% confidence interval | P-value |
|---|---|---|---|
| Socioeconomic factors | |||
| Household assets (tertiles) | |||
| Poorest | Ref | Ref | Ref |
| Poor | 1.59 | [0.79–2.77] | 0.17 |
| Least poor | 3.32 | [1.54–6.49] | <0.01 |
| Highest education of female ≥16 in household | |||
| No formal education | Ref | Ref | Ref |
| Some primary education and beyond | 1.09 | [0.70–1.68] | 0.69 |
| Number of people in household | 1.05 | [0.94–1.16] | 0.38 |
| Number of rooms in home | 1.25 | [0.98–1.62] | 0.07 |
| Roofing material on home | |||
| Iron sheets | Ref | Ref | Ref |
| Grass thatched or banana fibers | 0.34 | [0.07–1.54] | 0.16 |
| Time control | |||
| Survey month/year | 0.01* | ||
| January 2013 | Ref | Ref | Ref |
| July 2013 | 0.54 | [0.27–1.08] | 0.08 |
| November 2013 | 0.22 | [0.11–0.45] | <0.01 |
| January 2014 | 0.17 | [0.08–0.34] | <0.01 |
| April 2014 | 0.12 | [0.06–0.24] | <0.01 |
* Global significance of the variable in the model calculated from a Deviance test