TABLE III. Adjusted odds ratio (ORadj) and 95% confidence interval (CI) obtained from the multiple Firth penalized logistic regressions for domestic and intra-domiciliary infestation. Using a multi-model inference approach, one model (the full model) was identified as the best model for domestic infestation, while three competing models (delta AICc < 2) were averaged for intra-domiciliary infestation. Relative importance of a variable refers to the relative presence of the variable in the competing models (1 = present in all three models; 0 = absent in all three models).
| Domestic infestation* | Intra-domiciliary infestation** | ||
| Variable | ORadj (CI) | Relative importance | ORadj (CI) |
| Number of householders | 0 | ||
| 1-2 | 1 | - | |
| 3-5 | 0.68 (0.42 - 1.1) | - | |
| ≥ 6 | 0.89 (0.47 - 1.67) | - | |
| No. of dogs | 0.7 | ||
| 0 | 1 | 1 | |
| 1 | 1.14 (0.5 - 2.86) | 1.14 (0.34 - 3.81) | |
| 2 | 1.65 (0.73 - 4.15) | 1.72 (0.52 - 5.7) | |
| ≥ 3 | 2.68 (1.22 - 6.59)*** | 1.36 (0.4 - 4.66) | |
| Cats | 1 | ||
| No | 1 | 1 | |
| Yes | 1.87 (1.19 - 2.9)*** | 2.41 (1.32 - 4.4)*** | |
| Chickens | 0.2 | ||
| No | 1 | 1 | |
| Yes | 8.77 (4.28 - 17.79)*** | 1.16 (0.29 - 4.55) | |
| Wall and roof with materials that could offer refuge for triatomines | |||
| Wall | 1 | ||
| No | - | 1 | |
| Yes | - | 0.24 (0.01 - 6.22) | |
| Roof | 1 | ||
| No | - | 1 | |
| Yes | - | 1.06 (0.04 - 26.56) | |
*: domestic infestation refers to the occurrence of a live T. infestans in any domicile or peridomestic structure; **: intra-domiciliary infestation refers only to the presence of the vector inside the domicile; ***: p < 0.05.