Background: To reach zero AIDS deaths and zero new HIV infections, antiretroviral therapy (ART) must be delivered to all people living with HIV. As HIV testing is decentralized and community testing and self-testing become more widely available, identifying and implementing strategies that facilitate linkage to HIV care (LTC) from the community to the clinic to start ART is crucial.
Methods: In seven urban communities in Zambia and South Africa, randomised to receive the universal-testing-and-treatment intervention (UTT) within the HPTN071(PopART) trial, Community HIV-care Providers (CHiPs) offer home-based HIV testing, referral to government clinic services for HIV-positive individuals, support for LTC, and ART initiation. CHiPs deliver the intervention in “rounds”, during which they (re‑)visit all households in their community. Round 1 (R1) was November 2013-June 2015, R2 June 2015-September 2016, and R3 was September 2016-December 2017; 19 and 15 months respectively. CHiPs record data on electronic registers, including dates of referral, LTC and ART initiation, and conduct follow-up visits to support LTC and retention. From R2, with intensified efforts in R3, various strategies were implemented to facilitate LTC, including: targetted follow-up of individuals not yet LTC, extra counselling, more supervisory support to CHiPs, and increased coordination with the clinic. For each round, we estimated the time to ART initiation after first referral to care among individuals aged ≥15 years, using the Kaplan-Meier method and Cox regression for “time-to-event” analysis.
Results: In Zambia 6,197, 3,435, and 2,295, and in South Africa 1,375, 1,262 and 754 individuals were referred by CHiPs to HIV care in R1, R2, and R3 respectively (Table 1). The median time to ART initiation after CHiP referral to care was ~10 months, ~6 months, and ~3 months in R1, R2, and R3 respectively (Figure 1,p< 0.001), with a reduction overall and for both men and women (Table 1).
Conclusions: More rapid ART initiation for HIV-positive individuals after referral to HIV care was achieved over three rounds of community-wide intervention. We believe that improvements came from a combination of better targeted community and clinic activities, expanded national HIV treatment guidelines, increasing understanding of universal ART, and enhanced community acceptance of UTT.

CountryReferred to HIV care, and self-reported they were not on ART on the date of referral (N)ART initiated, by months after referral (%) 3 monthsART initiated, by months after referral (%) 6 monthsART initiated, by months after referral (%) 12 months Estimates for % initiated ART by 12 months after referral are not given for individuals referred to care in R3, because as yet (as of Sep 30 2017) there is insufficient follow-up time during Round 3 to report reliable estimates for this time point; it will be possible to report such estimates (or at least up to 9 months after referral) after Dec 31 2017Hazard ratios adjusted from Cox regression, adjusted for gender, age group, and community of residence
Zambia OverallR1 6,197 R2 3,435 R3 2,29632 % 39% 50%42% 50% 64%55% 67%1 (Ref) (p<0.001) 1.37 [1.29-1.46] 2.13 [1.97-2.29]
Zambia MenR1 2,053 R2 1,117 R3 69833% 44% 51%43% 56% 66%57% 71%1 (Ref) (p<0.001) 1.51 [1.35-1.68] 2.18 [1.90-2.50]
Zambia WomenR1 4,144 R2 2,318 R3 1,59831% 37% 50%42% 48% 64%54% 65%1 (Ref) (p<0.001) 1.32 [1.22-1.43] 2.14 [1.95-2.34]
South Africa OverallR1 1,375 R2 1,262 R3 75425% 36% 54%36% 47% 68%54% 66%1 (Ref) (p<0.001) 1.46 [1.28-1.66] 2.88 [2.50-3.32]
South Africa MenR1 457 R2 436 R3 27620% 31% 46%28% 42% 62%47% 58%1 (Ref) (p<0.001) 1.57 [1.24-2.00] 2.82 [2.16-3.68]
South Africa WomenR1 918 R2 826 R3 47827% 39% 58%39% 50% 71%57% 71%1 (Ref) (p<0.001) 1.42 [1.22-1.65] 2.93 [2.48-3.46]
[Table 1 ? Time to ART initiation after CHiP referral in Rounds 1-3, by country and gender]

Figure 1 time to ART start by country and by round
[Figure 1 time to ART start by country and by round ]