Effect of universal test and treat on retention and mortality among people living with HIV-infection in Uganda: An interrupted time series analysis.

Background: Few studies have analysed the effect of HIV universal test and treat (UTT) on retention and mortality among people living with HIV (PLHIV) in routine care. We examined six-month retention and mortality at 11 health facilities (HFs) run by a large NGO, The AIDS Support Organisation (TASO), before and after UTT.

Methods: We used a quasi-experimental study using patient data extracted from 11 TASO HFs. Two periods, one before UTT (2015-2016) and the other during UTT (2017-2018) were compared. The primary outcome was six-month retention defined as the proportion of PLHIV who were alive and in care at six months from enrolment. The secondary outcome was six-month mortality defined as the proportion of PLHIV who died within six months from enrolment. We performed an interrupted time series analysis using graphical aids to study trends in six-month retention and mortality and a segmented regression to evaluate the effect of UTT. We used a generalized linear mixed model (GLMM) and generalized estimating equations (GEE) to account for facility-level clustering.

Results: Of the 20,171 PLHIV registered between 2015 and 2018 and included in the analysis, 12,757 (63.2%) were enrolled during the UTT period. 5256/7414 (70.9%) of the pre-UTT period compared to 12239/12757 (95.9%) of the UTT were initiated on ART treatment with 6 months from enrolment. The median time from enrolment to initiating ART was 14 (interquartile range (IQR): 0-31) days for the pre-UTT compared to 0 (IQR: 0-0) days for the UTT period. The median age at enrolment was 32.5 years for the pre-UTT and 35.0 years for the UTT period. Overall, the six-month retention just after scale-up of UTT, increased by 9.2 percentage points (p = 0.002) from the baseline value of 82.6% (95% CI: 77.6%-87.5%) but it eventually decreased at a rate 1.0 percentage point (p = 0.014) for cohorts recruited each month after UTT. The baseline six-month mortality was 3.3% (95% CI: 2.4%-4.2%) and this decreased by 1.6 percentage points (p = 0.003) immediately after UTT. The six-month mortality continued decreasing at a rate of 0.1 percentage points (p = 0.002) for cohorts enrolled each month after UTT. Retention differed between some health facilities with Rukungiri HF having the highest and Soroti the lowest retention. Retention was slightly higher among males and younger people. Mortality was highest among people aged 50 years and more. The effect of UTT on retention and mortality was similar across sex and age groups.

Conclusion: Overall, UTT significantly led to an immediate increase in retention and decrease in mortality among PLHIV enrolled in HIV care from 11 HFs run by TASO in Uganda. However, retention (and mortality) significantly decreased for cohorts enrolled each month after UTT. Retention was highest in Rukungiri and lowest in Soroti HFs and slightly higher in males and younger people. Mortality was highest in older patients and lowest in adolescents. We recommend for innovative interventions to improve the overall retention particularly in facilities reporting low retention in order to achieve the UNAIDS 2030 target of 95-95-95.