Table 3. Main effects.
| Attribute | Level | 1 | 2 | ||||
|---|---|---|---|---|---|---|---|
| Coefficient | 95% CI | p-value | Coefficient | 95% CI | p-value | ||
| Food Insecurity | Skipped 0 meals | 1.303 | 1.114–1.491 | <0.001 | 6.692 | 5.437–8.236 | <0.001 |
| Skipped 1 meal | 0.875 | 0.736–1.013 | <0.001 | 3.414 | 2.831–4.117 | <0.001 | |
| Sleep | 3–5 hours of sleep | 0.0968 | 0.00774–0.186 | <0.05 | 1.136 | 1.010–1.278 | <0.05 |
| 8 hours of sleep | 0.335 | 0.211–0.459 | <0.001 | 1.673 | 1.420–1.971 | <0.001 | |
| Income | Earned income of Ush 100,000 in last month | 0.807 | 0.719–0.896 | <0.001 | 3.314 | 3.034–3.619 | <0.001 |
| Earned income of Ush 200,000 in last month | 0.993 | 0.882–1.105 | <0.001 | 4.527 | 3.950–5.187 | <0.001 | |
| Pain | No pain that makes it difficult to perform daily activities | 0.390 | 0.274–0.505 | <0.001 | 1.785 | 1.557–2.045 | <0.001 |
| R-squared | 0.259 | ||||||
| Observations | 2,560 | ||||||
| Number of participants | 320 | ||||||
| Number of blocks | 8 | ||||||
Note: Column 1 shows results from the CA sample analyzed using an OLS regression specification, while column 2 shows results from the CA sample analyzed using an ordered logit specification. All coefficients represent improvements in expected ART adherence (coefficients in ordered logit column represent odds ratios) and are compared to the reference categories (Food Insecurity: Skipped 2 meals, Sleep: 0 hours of sleep; Income: No income; Pain: pain that makes it difficult to perform daily activities). 95% confidence intervals and p-values are shown adjacent to coefficients. Standard errors were clustered at the individual level.