Discussion

In this setting, we explored the extent to which a referral system could be used to promote access to an HIV treatment clinic among individuals diagnosed at VCT, and to monitor rates of referral uptake. Similar mechanisms for monitoring referral rates could be implemented in any sites linking HIV-infected individuals with prevention, care, treatment and support services. Although it may not be appropriate for all linked HIV services to analyse referral uptake data, special studies can be conducted using this system to monitor the effectiveness of different HIV testing services in promoting access to ART, and can provide important insights into the degree of equity in access that is being achieved.

Furthermore, in selected sites, such as Kisesa, where community-level HIV data are collected through regular surveys, these analyses can be extended to monitor the entire process of accessing and initiating HIV treatment at a population level, enabling much-needed local estimates of ART coverage relative to treatment need, disaggregated by sex, residence and age to be derived [24].

Additional qualitative research in this setting has shown that even when economic barriers, such as the cost of transport, are addressed, knowledge and psychosocial issues remain important barriers to accessing HIV services [22,23,25-27]; these include HIV-related stigma, lack of family support, and denial of illness, as well as use of alternative healers. The referral uptake data generated through this method enabled us to compile a list of persons who did not register for treatment services following referral. HBC volunteers then provided additional support to these patients in the form of further information about HIV infection and associated prevention and treatment options. These patients had given prior consent to such contacts during the VCT session and lived locally, thus helping to overcome some of the barriers to attending the HIV clinic for patients who delayed their initial clinic appointment.

Furthermore, by monitoring appointment uptake, we were able to observe variations in referral uptake in relation to the level of support services that were being provided. Initial increases in the proportions taking up their referral appointment within a week correspond with the introduction of the community escort and transportation allowances at the beginning of 2006.

The lower proportion accessing the HIV clinic following a diagnosis made during the national campaign suggests that increasing opportunities to learn one’s status may not necessarily translate into effective access to HIV care and treatment, unless adequate resources for supportive counselling are also made available. In particular, the surge in the number of persons diagnosed during the last six months of the study period put pressure on the community escort scheme, such that it became difficult to offer this service to all referred patients during this period. It is also likely that the HIV testing campaign attracted individuals who were, on average, at an earlier stage of HIV infection compared with the population who actively sought VCT at the health centre, of whom a high proportion reported poor health as their reason for testing. This may have contributed to lower levels of motivation or readiness to attend the HIV treatment clinic among some persons who were diagnosed during the HIV testing campaign, partially explaining the lower overall referral uptake rates during this period.

The provision of a transportation allowance has been proposed as a strategy for improving access to HIV services, as well as to promote ART adherence and retention in care in several settings [11,28], and emerged as an important intervention in Kisesa in terms of facilitating regular attendance at the HIV treatment clinics. The cost of covering the return fare to the HIV treatment clinic in this setting was in the region of US$25-30 per patient per year, corresponding to a fraction of the total costs of providing medical treatment to HIV-infected patients, and is considered a sustainable use of programme funds by donor agencies, including the Global Fund for AIDS, TB and Malaria.

As such, the provision of transport fees should be viewed as an investment in terms of promoting timely registration for ART screening, which could result in earlier initiation of treatment and reduce the high mortality rates that have been widely observed among patients starting ART with very low CD4 counts [29,30]. Transportation allowances are also likely to facilitate adherence to treatment by delaying the need for second-line treatment, which cost around 10 times more than current generic first-line regimens.

Alternative strategies to donor-provided transportation fares should draw on lessons learned from programmes that have reported successes in using community cost-sharing or insurance schemes to cover transportation costs for referrals between primary and secondary level facilities [13,31]. Nevertheless, the longer-term solution needs to focus on bringing treatment services closer to local populations if barriers relating to the cost of reaching clinics are to be successfully addressed. Emerging evidence suggests that decentralization increases uptake of HIV treatment services and results in higher rates of retention in care [32,33]. This process of decentralization needs to be accompanied by interventions that address wider structural and social barriers that influence HIV clinic attendance, including poverty and stigma [22,23,26].

The involvement of key stakeholders throughout the design and implementation process led to high acceptance levels and satisfaction with referral monitoring procedures. Following this experience, other referral agencies linked to the same HIV clinic have adopted the same forms and are currently monitoring HIV treatment access [8]; similar systems are being piloted in other African countries. We have also used the same method to monitor referrals between VCT and the local HBC group. The next challenge is to encourage its adoption by other HIV testing services, including provider-initiated testing and counselling and those where subsequent referral uptake may be particularly low, such as mobile, door-to-door or other services outside a clinic environment, to ensure that national and international recommendations regarding strengthened linkages between testing and treatment sites are met, and that access to ART is improved.

The potential limitations of this study include the fact that referred individuals may have attended the HIV treatment clinic without their referral slip, or may have attended private providers, thus leading to an underestimate in the proportion of referred persons who subsequently accessed HIV care. In order to assess this, we cross checked registration books at the HIV treatment clinic to see if residents from the study area had enrolled for treatment having lost their slip, and found this to be the case for very few persons, whom we subsequently included in our uptake calculation.

Furthermore, in this setting, where the average income is approximately US$120 per year [34] and the average cost of a month of antiretroviral therapy drugs is in the region of US$25, it is unlikely that residents in the area were seeking private care. In terms of the qualitative data, we limited the potential for respondent bias by using trained fieldworkers, who established a rapport with study participants and explained the non-judgemental purpose of the study prior to commencing interviews and group activities.