Table 2. Results from six models of the association between active trachoma in children aged 1–9 years and community sanitation usage, Amhara, Ethiopia, 2011–2014.
| Characteristics | Modela |
|||||
|---|---|---|---|---|---|---|
| 1 | 2 | 3 | 4 | 5 | 6 | |
| No. of individuals | 62 869 | 62 806 | 62 037 | 61 351 | 61 351 | 61 351 |
| No. of households | 35 977 | 35 963 | 35 477 | 35 061 | 35 061 | 35 061 |
| No. of clusters of households | 1 510 | 1 510 | 1 510 | 1 510 | 1 510 | 1 510 |
| aOR (95% CI) with community sanitation usage of:b | ||||||
| ≥ 80% | 0.51 (0.35–0.73) | 0.50 (0.33–0.75) | 0.54 (0.36–0.81) | 0.62 (0.42–0.92) | 0.74 (0.49–1.11) | 0.68 (0.45–1.04) |
| 60 to < 80% | 0.67 (0.47–0.97) | 0.68 (0.45–1.03) | 0.73 (0.49–1.10) | 0.78 (0.52–1.17) | 0.88 (0.58–1.32) | 0.84 (0.55–1.26) |
| 40 to < 60% | 0.91 (0.66–1.26) | 0.95 (0.66–1.35) | 1.00 (0.70–1.42) | 1.02 (0.72–1.46) | 1.11 (0.78–1.58) | 1.06 (0.74–1.52) |
| 20 to < 40% | 1.03 (0.76–1.39) | 1.07 (0.77–1.49) | 1.10 (0.79–1.53) | 1.11 (0.80–1.54) | 1.16 (0.83–1.60) | 1.12 (0.81–1.55) |
| < 20% | Ref. | Ref. | Ref. | Ref. | Ref. | Ref. |
| Random effects | ||||||
| Intraclass correlation coefficient | ||||||
| Same cluster | 0.34 | 0.37 | 0.37 | 0.36 | 0.36 | 0.35 |
| Same household | 0.45 | 0.52 | 0.52 | 0.51 | 0.51 | 0.51 |
| Median OR | ||||||
| Cluster | 4.80 | 6.07 | 6.02 | 5.86 | 5.83 | 5.82 |
| Households within cluster | 2.19 | 2.66 | 2.65 | 2.63 | 2.63 | 2.63 |
| Variance between clusters (SE) | 2.03 (0.18) | 2.52 (0.23) | 2.50 (0.22) | 2.41 (0.22) | 2.39 (0.22) | 2.38 (0.22) |
| Variance between households (SE) | 0.68 (0.08) | 1.05 (0.10) | 1.04 (0.10) | 1.03 (0.10) | 1.03 (0.10) | 1.03 (0.10) |
aOR: adjusted odds ratio; CI: confidence interval; OR: odds ratio; Ref.: reference group; SE: standard error.
a Results were weighted to account for the unequal probabilities of selection and down-weight survey participants excluded from the assessment of trachoma prevalence because they were aged < 1 year or > 9 years. Model 1 was adjusted for survey round. Model 2 was like model 1 but also adjusted for sex and age of the child– with age centred at 5 years. Model 3 was like model 2 but also adjusted for household access to bathing water within a journey of < 30 minutes and household latrine with evidence of use. Model 4 was like model 3 but also adjusted for household access to electricity, household education, household ownership of mobile phone, radio and/or television and iron roof. Model 5 was like model 4 but also adjusted for community mean household wealth. Model 6 was like model 5 but also adjusted for the median number of times surveyed residents in a community had reportedly received a mass drug administration against trachoma.
b Usage was measured as the proportion of households with a latrine with evidence of use.