Fidelity of HIV programme implementation by community health workers in rural Mopani district, South Africa: a community survey

Community health programmes have strong potential to strengthen primary health services in low- and middle income countries (LMICs) [1]. The value of these programmes has been demonstrated for maternal, child, and mental healthcare in various settings in LMICs. For example, studies from South Africa show that CHWs have been successful in improving maternal and child health outcomes [2], and providing a social support system [3]. Reviews of quantitative and qualitative studies show that CHWs can provide an important human resource capacity and have the potential to contribute to HIV services in sub-Saharan Africa [4, 5]. There is growing evidence on successful scale up and integration of community-based programmes into national health systems [6‐8]. However, despite its strong potential, only limited data are available with regard to provision of HIV services at community level [9, 10].

South Africa, the country with the largest HIV programme in the world, has implemented a community health programme since 2012 with the aim to improve access to and provision of high-quality primary healthcare (PHC) services [11]. South Africa’s community health programme is delivered through ward-based outreach teams (WBOTs), each comprising a professional nurse that serves as team leader and up to five community health workers (CHWs) [11]. The local guidelines stipulate that the team of five CHWs should support a population of 7660 individuals in total [11]. The services provided by the CHWs include general health education, health status monitoring, and referral of individuals in need of care to the PHC facility [11].

There is a strong programmatic focus on vulnerable populations, including individuals with chronic illnesses; HIV or tuberculosis (TB); pregnant women; as well as screening for malnutrition and gastroenteritis; and checking immunisation, vitamin A and deworming status in children.

Provision of HIV services constitutes an important component in the community health programme. The CHWs contribute to HIV programme delivery through providing health education to prevent HIV infection, identifying individuals that need to test for HIV, referring HIV-infected individuals not yet in care to start antiretroviral therapy (ART), providing adherence support to those on ART, tracing and referral of HIV-infected individuals that have been lost to the ART programme, and identifying individuals who have clinically failed on ART and require further assessment [11]. Visits to HIV-infected individuals should happen on a monthly basis while assessment of HIV risk and potential referral for testing should happen at least annually [11].

Regular assessment and monitoring of implementation performance of any health programme is essential to obtain maximum impact of its implementation and to ensure that the programme meets its intended goals [12]. The same holds true for the HIV programme as provided by CHWs. However, to the best of our knowledge there is no record of a comprehensive assessment of implementation of this programme in South Africa [13]. Studies from other countries, not addressing HIV, show that lack of such information could negatively affect programme effectiveness [2‐4]. Knowledge of how and why a community health programme works, or does not work, in terms of implementation is required to optimise programme implementation before successful scale-up [1, 6, 14].

An important method to assess implementation of health programmes is to measure fidelity to the intended programme specifics. Fidelity to programme implementation is generally measured on four aspects: ‘coverage’ (services provided to everyone that is supposed to receive these services), ‘frequency’ (services provided by the prescribed frequency), ‘duration’ (service delivery that is uninterrupted) and ‘content’ (the correct services provided) [15]. In our context, this means measurement of HIV programme implementation by CHWs of these four components against the programme specification in the local government guidelines [11]. In this study we aim to measure fidelity to HIV programme implementation as provided by CHWs from the community’s perspective in rural Mopani District, South Africa. We conducted a community survey to obtain insight into the current status of programme implementation and to identify areas that warrant strengthening to improve effectiveness of HIV service delivery by CHWs.

This is the first study to measure fidelity of implementation of the HIV programme as provided by CHWs in rural South Africa. We are not aware of any other HIV programme with fidelity data from similar settings to provide a comparison to our observations. However, when compared to other South African and global studies measuring CHW chronic disease services other than HIV, we observed similar findings with regard to duration and content, and similarly found that coverage and frequency should be improved [18‐20].

In terms of coverage and frequency, we measured a 47% coverage of households from which the participant reported a CHW visit at any point in time, despite the policy that all households ought to be visited on an annual basis to assess vulnerability status. This is lower than reported by studies of chronic diseases that showed coverage in the range of 57% to 74% [19, 20], but good compared to a systematic review of 38 studies that reported 18% coverage for community health programmes overall [18]. In addition, frequency of visits was relatively low at 39%, but this was similar to other studies [18]. Increasing coverage and frequency of CHW visits could directly impact on the HIV programme as participants who had been visited by CHWs were more likely to have discussed HIV and know their HIV status than those that had not been seen by a CHW. Various factors may have impacted on coverage and frequency in our area. First, operational factors may play a role: coverage was higher in Greater Giyani sub-district where in our opinion the programme stronger managed and where there are on average more CHWs employed in each community than in Greater Letaba (resulting in lower number of households covered by each CHW). Second, an important factor to consider is the capacity of an individual CHW to visit all allocated households in the section within the set timeframe. It is possible that in our rural area, the target number of 270 households, which in practice becomes 250–400 households, is too large to cover within the specific programme deliverables. Reduction of the number of households allocated to each CHW by increasing the number of CHWs per ward should be considered to improve coverage [11]. Third, geographic factors such as houses located in rocky ad mountainous areas may impact on accessibility of houses to CHWs. In addition, challenging terrain would likely take the CHW more time to reach a household, limiting the number of households that can be visited during a day. Although an effect cannot be ruled out, we did not observe any obvious patterns or clusters when looking at households reporting visits compared to those not visited. Distance between the household and nearest PHC facility was not associated with the report of a CHW visit as most houses were within walking distance to the facility. Despite the need to improve coverage and frequency, we observed high appreciation of CHW visits by the community; most participants were happy with the CHW visits and services provided, whereas most of those that had not been visited by the CHW would really appreciate a visit. We did not observe a difference between male and female respondents with regards to interest in and appreciation of CHW visits. This contradicts findings from another report from Pakistan that suggests that men are unwilling to interact with the female CHWs [21].

In terms of duration and content, we measured high duration (100%) of programme implementation among those visited by the CHWs as compared 39% reported the aforementioned systematic review of 38 studies [18]. This suggests that services are relatively continuous and uninterrupted once there is a structural relationship between household members and CHWs. In addition, the content of service provision for HIV was 67%) and no major gaps were identified. This is good when compared to a systematic review of 38 studies that reported content provision of chronic diseases services in the range of 40% to 66% [18, 19]. This indirectly shows the impact of the intensive CHW training programme (phase 1 and phase 2 of the national curriculum) that was conducted in Mopani District [11]. Participants with whom HIV was discussed were more likely to know their HIV status suggesting a direct link between CHW activities and HIV programme performance. We did not distinguish between HIV-infected and HIV-uninfected participants. However, HIV should be discussed with all household members.

This study shows that despite limited resources, there is a reasonably good implementation of HIV programme activities by CHWs in our rural South African district. Although coverage and frequency should be improved, there is good content and duration of service provision to those that are reached by the CHWs. This supports the high promise of CHW implementation of the HIV programme on a structural basis. Resource investment in the CHW programme, through increasing the number of CHWs and strengthening operational structures, may have to be expanded to achieve the programme’s full potential and effectiveness, the sub-district with the lower resources also scored lower in implementation fidelity. These findings are consistent with other studies showing that adequate staffing, resource provision, and a human resource management approach are imperative for successful implementation of community-based programmes [22, 23]. In light of this, the use of human resource strategies such as clarification of roles, job satisfaction, and adequate remuneration should be considered to improve the South African CBHP. Although operational determinants may play a role, other factors that contribute to the programme success or provide a barrier to successful coverage and frequency need to be determined within a local context, in order to further improve implementation fidelity. Moreover, alternate strategies should be considered to reach the large proportion of households where we found no one at home, such as worksite wellness and local media initiatives [24].

This study has several limitations. Firstly, several types of bias may have occurred including recall and desirability bias. Selection bias may have also occurred as there was nobody at home at a substantial proportion of households. Possibly, these represent higher socio-economic group as they may be at work during the day. Although we recruited more women than men, there was no difference in their response. The response rate could have possibly been increased by visiting households multiple times instead of once. However, this was not feasible for operational reasons. Secondly, we have measured implementation fidelity from the community’s (recipient) perspective. Triangulation of our findings with similar measurement from the CHW’s perspective could have strengthened our findings, but resources did not permit this. We did not allocate a health vulnerability status to the households included in this study, which would have allowed for a more precise assessment of some components of implementation fidelity. This was due to the fact that household registration data, including vulnerability status, were not available through the CHWs at the time of study and could not be reliably collected as part of the questionnaire. For operational purpose, we collected data on CHW visits within the last 6 months as opposed to the recommended 12 months, which may have resulted in an underestimation of frequency. Finally, a time component was not included in the question where individuals were asked if they knew their HIV status.

In conclusion, this study shows that implementation fidelity of the HIV programme as provided by CHWs in a rural South Africa district is similar to other CHW programmes that report on chronic illnesses, other than HIV. There is room for improvement, in particular of coverage and resource investment is required to increase frequency of household visits. While this study informs policy makers on specific programme areas to improve this CBHP, further quantitative and qualitative research in similar contexts is required to further gauge barriers and facilitators to the fidelity of CHBP implementation.