Indicator-focussed technical assistance in South Africa's HIV programme: A stepped-wedge evaluation

There is a lack of research on technical assistance (TA) interventions in low- and middle-income countries. Variation in local contexts requires tailor-made approaches to TA that are structured and replicable across intervention sites whilst retaining the flexibility to adapt to local contexts. We developed a systematic process of TA using multidisciplinary roving teams to provide support across the various elements comprising local HIV services.

To examine the effectiveness of targeting specific HIV and TB programme indicators for improvement using roving teams.

We conducted a cluster-randomised stepped-wedge evaluation of a TA support package focussing on clinical, managerial and pharmacy services in the Mopani district of the Limpopo province, South Africa (SA). Three roving teams delivered the intervention. Seventeen primary and community healthcare centres that had 400-600 patients on antiretroviral therapy (ART) were selected for inclusion. The TA package was implemented for six consecutive months across facilities until all had received the same level of support. Data were collected from the relevant health management information systems for 11 routine indicators.

The mean proportion of PLWH screened for tuberculosis (TB) at ART initiation increased from 85.2% to 87.2% (P = 0.65). Rates of retention in care improved, with the mean proportion of patients retained in care at three months post-ART initiation increasing from 79.9% to 87.4% (P < 0.001) and from 70.3% to 77.7% (P < 0.01) after six months. Finally, the mean proportion of patients with TB who completed their treatment increased from 80.6% to 82.1% (P = 0.75).

Tailored TA interventions in SA using a standardised structure and process led to a significant improvement in retention-in-care rates and to non-significant improvements in the proportion of PLWH screened for TB and of those who completed their treatment.