Since the launch of WHO's '3 by 5' initiative in 2003, many countries in sub-Saharan Africa have established national antiretroviral treatment (ART) programmes. Although the WHO target of providing access to ART for 3 million people by 2005 was not achieved, by end-2005 an estimated 1.3 million people in low- and middle-income countries had access to treatment (about 20% of those estimated to be in need) (WHO and UNAIDS, 2006). By mid-2005, the WHO target had already been overtaken by an even more ambitious aim. In July 2005, the G8 group of industrialized countries committed to the goal of achieving 'as close as possible to universal access to treatment for all those who need it by 2010.' (UNAIDS, 2006, G8 Gleneagles Summit, 2005). Nonetheless, the challenges in the region remain great. Health systems are weak, and the target orientation of ART programmes risks an emphasis on initiating people on ART at the expense of ensuring effective use of medicines. As discussed in Chapter 2, extremely high levels of adherence (at least 95%) are needed to ensure positive treatment outcomes and prevent the development of drug-resistance (Paterson et al., 2000).

Up till now, only limited operational research has been carried out to identify adherence problems in resource-poor settings and to strengthen adherence support (Jaffar et al., 2005; Bennet, Boerma and Brugha, 2006; Kent et al., 2003; Akileswaran et al., 2005; Farmer et al., 2001). Previous studies on adherence to ART in Africa have provided quantitative estimates of adherence and data on clinical outcomes, mainly from experimental settings (Ivers, Kendrick and Doucette, 2005; Coetzee et al., 2004; Orrell et al., 2003; Koenig, Léandre and Farmer, 2004; Gill et al., 2005). A recent review of six of these studies reported that 68%-99% of patients took at least 95% of their medicines. The authors, Ivers et al., conclude that adherence levels in Africa are high, i.e. comparable to those in industrialized settings. However, Gill and colleagues (2005) and Laurent et al., 2002) stress that there is no room for complacency, noting that adherence rates tend to deteriorate over time.

There is little evidence as to why some ARV users do not achieve optimal adherence rates or about how to improve adherence support in resource-poor settings (Koenig, Léandre and Farmer, 2004; Gill et al., 2005). Reports on sub-optimal adherence to ART in developed countries indicate that the key factors are patient- and treatment-related, including substance and alcohol abuse, complexity of dosing regimen and 'pill burden', dietary restrictions and side-effects (DiMatteo, 2004; Chesney, 2000; American Public Health Association, 2004; WHO, 2004). The few studies conducted in Africa suggest that in resource-poor settings other factors may predominate (Hardon, Hodgkin and Fresle 2004). Weiser et al., in a study in Botswana in 2003, identified financial constraints as the major obstacle to adherence. Ivers and colleagues (2005) found in their meta-analysis of 10 studies conducted in resource-poor settings that providing medication free of charge to patients was associated with a 30% higher probability of having an undetectable viral load at months 6 and 12. In resource-poor settings, cost appears to be an important determinant of adherence.