Can task shifting improve efficiency of HIV self-testing kits distribution? A case study in Mali

Background: The ATLAS project introduced HIV self-testing (HIVST) in consultations of people living with HIV (PLHIV) at public health facilities in Côte d'Ivoire, Mali and Senegal for secondary distribution to their partners. Preliminary data from a qualitative study (observations of consultations, interviews with distributing agents) carried out in two clinics in Mali highlight implementation challenges associated with the counselling on self-testing and kit distribution currently done by the medical staff (doctor/nurse) and reported time-consuming. While implementation teams are considering the possibility of delegating certain tasks, it is important to consider the cost of alternative delivery models. Materials & Methods: We analysed preliminary economic costs data for the provision of rapid HIV testing services (HTS) (analysis period: October 2018 – September 2019) and HIVST services (August 2019 – October 2019) in these same two Malian clinics. Above service level costs are excluded. We then modelled the costs of provision using alternative cadres of medical and non-medical staff (psychosocial counsellors/peer educators) and the consumables used to simulate task shifting scenarios for the provision of HTS and HIVST services. The three scenarios correspond to 1. partial delegation: individual counselling done by non-medical staff and HIVST distribution by the medical staff ; 2. total delegation: individual counselling and distribution done by non-medical staff only; and 3. total delegation with group counselling: where group counselling and distribution are done by non-medical staff only. Results: Findings show that the unit costs per HIVST provided for the observed model are 58% higher than those of a conventional rapid test: $7,50 and $4.75, respectively. The costs are less high in scenarios of partial ($5.45, +15%) or total ($5.29, +11%) delegation but always higher than those of a rapid test due to the greater costs of consumables (HIVST kit). Finally, in the case where counselling on self-testing were carried out in a group, the costs per kit provided ($4.44, -6%) would become slightly lower than those of a rapid test, where counselling is always done individually. Conclusion: Task delegation from medical to non-medical staff can generate substantial cost savings. These preliminary results can guide the implementation strategy of HIVST in care consultations, to ensure sustainability from early introduction through scale-up.