With the realization of the value of ART as a means of preventing HIV transmission, the question of how best to retain HIV-diagnosed in care becomes all the more pressing. Recent STI blogs have covered such topics as the potential role of computer-generated reminders in retaining patients (sti blogs03/05/13), as well as the re-engagement of patients lost to care in a developed world context (sti blogs17/06/13). But what about the retention of the HIV diagnosed in limited-resource settings such as sub-Saharan Africa? Here, the stakes seem even higher – given the severity of the epidemic, and likely obstacles to full engagement such the relative inaccessibility of medical facilities to poor people living in rural settings (sti-Fried & Eyles). What is the evidence that decentralizing HIV care from hospitals to local health care centres or the community could help to improve retention in care, and thereby contribute to controlling the epidemic?
A recent systematic review (Kredo & Gardner) produced by the Cochrane Collaboration surveys the few studies that address this question, and its results have fed into the recent WHO guidelines (WHO 2013 ART guidelines).
The paper itself considers care options within the three general categories of “partial decentralization” (ART initiated by hospital, maintained by local health centre), “full decentralization” (ART both initiated and maintained by health centre), and “decentralization to the community” (ART initiated by hospital or health centre and maintained by the community). The options are examined in regard to their impact on loss to care, mortality, and attrition (a combination of loss to care + mortality).
Most of the evidence surveyed (i.e. 12/16 studies) consists in retrospective cohort studies susceptible to various kinds of bias. Consequently the evidence is largely graded low or very low quality. However, the review flags up some moderate quality evidence that:
– partial decentralization is associated with reduced attrition at 12mnths. (RR=0.46)
– full decentralization is associated with reduced loss to care at 12 mnths. (RR=0.3)
– decentralization to health centres and decentralization to the community have a similar impact on loss to care, mortality and attrition.
The low and very low quality evidence regarding other associations points almost entirely in the direction of decentralization having a positive impact on retention in care.
On this basis the WHO guidance on the operational aspects of HIV management proposes the de-centralization of ART delivery and its integration with maternal and child health clinics, along with other strategies to improve retention in HIV care and task-shifting to close human resource gaps.
The studies reviewed demonstrate that that decentralization is at least feasible. However, the authors also remind us in their conclusion that the decentralizing measures, to which they refer, “were in the context of a range of support structures and investments to ensuring delivery, including training, supervision and additional devices such as computer-aided or checklist-based decision aids”. The implication seems to be that, without such support structures and investments – which may be difficult to achieve in certain contexts (for an interesting parallel see: sti blog25/6/12) the benefits of decentralization cannot be counted on.