Table 3.
Effect of sector on adherence controlling for potential confounders 1
| Verified completed treatment 2 | Verified timely completion 3 | |||
|---|---|---|---|---|
| Adjusted odds ratio | p-value | Adjusted odds ratio | p-value | |
| Attended ADDO vs. public health facility | 0.65 (0.43, 1.00) | 0.048 | 0.69 (0.47, 1.01) | 0.056 |
| Age 4 | ||||
| Under 3 years (ref) | --- | --- | --- | --- |
| 3 years to under 8 years | 1.01 (0.67, 1.52) | 0.9 | 0.88 (0.61, 1.28) | 0.5 |
| 8 years to under 12 years | 1.07 (0.57, 2.09) | 0.8 | 1.12 (0.62, 2.01) | 0.7 |
| 12 years and above | 1.02 (0.56, 2.05) | 0.8 | 0.87 (0.60, 1.27) | 0.5 |
| Patient (or caregiver if patient below age 12) completed primary school | 1.68 (1.20, 2.36) | 0.003 | 1.06 (0.77, 1.45) | 0.9 |
| Socio-economic status 5 | ||||
| 1st quintile (most poor, ref) | --- | --- | ||
| 2nd quintile | 0.98 (0.62, 1.57) | 0.9 | 1.04 (0.66, 1.64) | 0.9 |
| 3rd quintile | 1.17 (0.73, 1.88) | 0.5 | 1.10 (0.70, 1.75) | 0.7 |
| 4th quintile | 2.25 (1.33, 3.81) | 0.003 | 1.64 (1.03, 2.65) | 0.039 |
| 5th quintile (least poor) | 2.24 (1.28, 3.81) | 0.005 | 2.34 (1.40, 3.93) | 0.001 |
| Distance from home to outlet within 2.5 km by GPS | 1.30 (0.92, 1.85) | 0.2 | 1.20 (0.87, 1.66) | 0.3 |
| Time of day drug was obtained | ||||
| Morning (ref) | --- | --- | --- | --- |
| Afternoon | 0.96 (0.62, 1.47) | 0.8 | 0.70 (0.48, 1.03) | 0.070 |
| Evening | 0.93 (0.50, 1.70) | 0.8 | 0.35 (0.19, 0.64) | 0.001 |
| Time between obtaining AL and interview (hours) | ||||
| 60-67 (ref) | --- | --- | --- | --- |
| 68-72 | 2.43 (1.39, 4.23) | 0.002 | 1.46 (0.81, 2.63) | 0.2 |
| 73-84 | 2.83 (1.49, 5.37) | 0.001 | 1.92 (1.00, 3.66) | 0.049 |
| 85 or more | 6.44 (3.19, 13.01) | <0.001 | 2.61(1.37, 4.95) | 0.003 |
1Number of observations = 912 (110 patients excluded from model due to missing data) and number of outlets = 77. Covariates are those presented in Table.
2Patient completed all doses, verified by pill count when available. Data missing for 5 patients.
3For each dose, patients took the correct number of pills at the correct time of day, verified by pill count when available. Data missing for 23 patients.
4Age categories based on recommended age breakdown for AL blister packs in Tanzania.
5Wealth quintiles pooled for public health facilities and ADDOs using principal component analysis of sampled patients based on standard Demographic and Health Survey variables.