Table 4. Bivariate and multivariate analysis of factors associated with attrition for patients who started antiretroviral therapy before and after the implementation of HIV “Treat All” in the 9 pilots districts in Zimbabwe (Multiple imputation, N = 3787).
| Variable | Categories | HR | p-value | (95%CI) | aHR | p-value | (95%CI) |
|---|---|---|---|---|---|---|---|
| Cohort | Before “Treat All” | 1 | |||||
| After “Treat All” | 1.40 | 0.009 | (1.09–1.80) | 1.73 | <0.001 | (1.30–2.31) | |
| Sex | Female | 1 | |||||
| Male | 1.12 | 0.336 | (0.89–1.41) | 1.45 | 0.005 | (1.12–1.87) | |
| Age group | Adults | 1 | |||||
| Children | 0.80 | 0.010* | (0.42–1.52) | ||||
| Adolescents and young adults | 1.30 | (0.91–1.85) | |||||
| Elderly | 0.55 | (0.37–0.84) | |||||
| HIV testing modality | Voluntary | 1 | |||||
| Antenatal | 1.52 | 0.002* | (1.10–2.11) | ||||
| Others | 0.84 | (0.64–1.12) | |||||
| Baseline tuberculosis status | Negative screening | 1 | |||||
| Presumptive tuberculosis | 2.05 | 0.233* | (0.83–5.06) | ||||
| On tuberculosis treatment | 1.23 | (0.70–2.16) | |||||
| WHO stage | I-III | 1 | |||||
| IV | 2.52 | 0.0450 | (1.02–6.21) | 2.89 | 0.0220 | (1.16–7.11) | |
| Functional status | Normal | 1 | |||||
| Impaired | 1.06 | 0.830 | (0.62–1.83) | ||||
| Level of care | District/provincial hospital | 1 | |||||
| Primary health facility | 0.81 | 0.448 | (0.47–1.39) | ||||
| Partner support | Not supported | 1 | |||||
| Supported | 1.31 | 0.432 | (0.66–2.61) | ||||
| Pregnant when starting ART | a No | 1 | |||||
| Yes | 2.08 | <0.001 | (1.58–2.75) | 3.47 | <0.001 | (2.36–5.11) | |
| Pregnancy##Cohort | Not confirmed#After “Treat All” | 1 | |||||
| Confirmed#After “Treat All” | 0.48 | 0.0120 | (0.27–0.85) |
*Overall p-value
a Includes non-pregnant women and men, HR: hazard ratio, aHR: adjusted hazard ratio, CI: confidence interval, WHO: World Health Organisation, ART: antiretroviral therapy, #: interaction, Children: 0–14 years, Adolescents and young adults: 15–24 years, Adults: 25–49 years, Elderly: +50 years.