Provision of HIV services by community health workers should be strengthened to achieve full programme potential: a cross-sectional analysis in rural South Africa

Abstract

Objective:

South Africa’s community health workers (CHWs) provide a bridge between the primary healthcare (PHC) facility and its community. We conducted a cross-sectional analysis to determine the contribution of the community-based HIV programme (CBHP) to the overall HIV programme.

Methods:

We collected service provision data from the daily activity register of CHWs attached to 12 PHC facilities in rural Mopani District, South Africa. Personal identifiers of individuals referred to the facility for HIV services were recorded and verified against facility routine patient registers to determine effectiveness of referral.

Results:

HIV services were provided on 18 927 occasions; 30% of the total activities performed by CHWs during the study period. CHWs assessed 12 159 individuals for HIV risk (13% coverage of the study population); only 290 (2%) were referred for HIV testing services. Referral was effective in 213 (73%) individuals; evidence of an HIV-positive status was found for 38 (18%) individuals. However, 30 (79%) of these individuals were referred by CHWs despite being on ART. Adherence support was provided during 5 657 visits; only one individual was referred for complications. Finally, of 864 individuals lost to the ART programme, CHWs managed to find 452 (52%) for referral back to the facility; only 241 (53%) of these were (re)initiated on ART.

Conclusions:

Provision of HIV services by CHWs should be strengthened to fully deliver on the programme’s potential. Human resource investment, home-based HIV testing and improved tracing models constitute potential strategies to enhance CHWs impact on the HIV programme.

INTRODUCTION

South Africa has the largest Human immunodeficiency virus (HIV) programme in the world, with currently more than 3.9 million individuals on antiretroviral therapy (ART) [1]. HIV services are primarily delivered at primary healthcare (PHC) facilities [2]. Early HIV diagnosis, efficient linkage to care and a high rate of retention in care are important determinants of programme success [3]. Despite scale-up of access to HIV services and ART across the country, various barriers to programme success remain, such as operational and socio-cultural factors and improper integration into the health system [4–6]. Ward-based outreach teams (WBOTs), run by community health workers (CHWs), are positioned at the community-facility interface and constitute an important component of the healthcare system to strengthen HIV programme implementation [7–10]. HIV services provided by CHWs include: (1) identification and referral of at-risk individuals for HIV testing, (2) treatment adherence support to individuals on ART and early identification of individuals with complications or side-effects to be referred to the facility for assessment, and (3) tracing and referral of HIV-infected individuals that have been lost to the ART programme. Tracing of individuals lost to the ART programme is done using reports generated by administrative clerks from the Three Interlinked Electronic Register (TIER.Net), South Africa’s electronic ART register [11].

Several studies from sub-Saharan Africa have shown that community health programmes can be used to strengthen HIV service delivery at primary healthcare facilities, especially through mobilising of individuals in need of care [12–15]. For example, a systematic review of 34 studies from high, middle, and low income countries shows that community based programmes have the potential to improve the uptake of HIV services, and can contribute to more equitable uptake of referrals from communities to healthcare facilities [15]. There are only limited data available on the volume, type and effectiveness of HIV services routinely delivered by CHWs. It is imperative to understand the roles of CHWs in the CBHP and to assess the impact of HIV services provided by CHWs in order to ensure that the CBHP achieves maximum impact and delivers on its intended goals. We therefore undertook a cross-sectional study to characterise HIV services delivered by the CHWs in the CBHP and to assess the impact of these services on the overall HIV programme in a rural district of South Africa.

METHODS

Study setting and design

This study was conducted in rural Mopani District, South Africa, where the community health programme was initiated in 2012; CHWs have been providing HIV services from its inception. A cross-sectional study to assess the contribution of the CBHP to the overall HIV programme was conducted from January through September 2017 in the sub-districts of Greater Giyani and Greater Letaba. We purposively selected the 12 WBOTs (comprised of 90 CHWs) attached to 12/49 PHC facilities in these two sub-districts (six each); these facilities were chosen for their similar size of catchment population (approximately 9 000 individuals each). Ethical approval was obtained from the Human Research Ethics Committee at the University of Witwatersrand, Johannesburg, South Africa (Reference number: M1611111) as well as the Limpopo Provincial Health Research Committee.

Study procedures

We obtained demographic data for the 12 wards from the Statistics South Africa (StatsSA) community survey 2016 [16]. The estimated HIV-infected population in the 12 wards was calculated using the Thembisa model [17]. We used the activity registers completed by CHWs daily to collect information on the services provided to their community. In this register, each CHW records the particulars of all individuals attended to as well as which services were provided (using a code book). The services that are provided include screening and adherence support for HIV, TB, malnutrition, pregnancy, and non-communicable diseases. With regard to provision of HIV services, the following information is routinely recorded: (a) at-risk individuals referred to the nearest facility for HIV Testing Services (HTS), (b) HIV-infected individuals on ART to whom adherence support is provided, and (c) tracing and referral back to the facility of individuals lost to the ART programme.

We used the patient identifiers (name, surname, and birth date) as recorded in the daily activity register to determine whether individuals who were referred to the PHC facility by CHWs actually visited the facility for services. The patient identifiers were triangulated with several data sources at each facility: the paper-based headcount register in which all facility visitors should be registered, the paper-based HTS and pre-ART register in which individuals with a positive HIV result are recorded, and TIER.Net, the electronic ART patient register of South Africa [11]. If an individual was found in at least one register within two weeks of CHW referral, we considered the referral as successful. In the case where an individual was found on TIER.Net, we looked at the ART initiation date to determine whether the individual was already on ART prior to CHW referral for HIV screening or that the individual was actually newly diagnosed as result of HIV screening, and subsequently initiated on ART. The CHW referral and back-referral forms were not used properly as stipulated by the provincial guidelines due to operational constraints [18]. Individuals referred for HTS were considered HIV-uninfected if evidence of a facility visit was found in either the headcount or HTS register, but no record in the pre-ART or ART register. Facility staff identified individuals lost to the ART programme and the CHWs were asked to trace them.

Statistical analysis

Data were captured into a Research Electronic Data Capture (REDCap) electronic database hosted by University of the Witwatersrand [19]. Validity checks were done to ensure completeness and to identify and correct consistency and logic errors. Descriptive statistics are provided as number with proportion and median with range. Logistic regression was used to compare demographic variables between individuals with and without successful facility referral. Variables with a p value ≤ 0.1 on univariate logistic regression were included in the multivariate logistic regression model. A p value <0.05 was regarded as statistically significant.

RESULTS

Daily activities recorded by the CHWs

The CHWs within the participating wards serve a population of 91,484 individuals with 6,917 individuals (8%) registered on TIER.Net to be on ART at the beginning of the study (Table 1). 58% of the population in this area are within the working age-group (15–60 years of age) and 56% are female. CHWs recorded 18 927 household visits in the study period during which the following services were provided: HIV services (18 688 times), screening for TB (16 445 times), diabetes mellitus (11 143 times), hypertension (14 115 times) and antenatal care and pregnancy (1 553 times). Altogether HIV services constitute 30% (18 688/61 944) of the services provided. The breakdown of HIV services provided was as follows: 12 159 (65%) individuals were assessed for HIV risk and, when appropriate, referred for HTS, adherence support was provided to 5 657 individuals on ART (30%), and 864 individuals (5%) were traced because they were lost to the ART programme.

Table 1:

Demographics of the study population and HIV services provided in the twelve study wards in Mopani District, South Africa.

		Participation wards in Greater Giyani sub-district	Participation wards in Greater Letaba sub-district	Total
Population		46 959	44 525	91 484
Gender	Male	20 577 (44%)	20 079 (45%)	40 656
	Female	26 382 (56%)	24 446 (55%)	50 828
Age	0–14 years	16 873 (36%)	14 374 (32%)	31 247
	15–24 years	10 980 (23%)	10 594 (24%)	21 574
	25–44 years	11 063 (24%)	11 147 (25%)	22 210
	45–60 years	4 426 (9%)	4 655 (11%)	9 081
	>60 years	3 617 (8%)	3 755 (8%)	7 372
Estimated HIV-infected adult population		6 035 (13%)	6 118 (14%)	12 153
Gender	Male	2 107 (35%)	2 129 (35%)	4 236
	Female	3 928 (65%)	3 989 (66%)	7 917
Age	15–24 years	859 (14%)	876 (14%)	1 734
	25–44 years	4 011 (67%)	4 006 (66%)	8 017
	45–60 years	1 038 (17%)	1 105 (18%)	2 144
	>60 years	127 (2%)	131(2%)	258
Number of PLHIV on ART		3 050 (51%)	3 867 (63%)	6 917
Services provided by CHWs during the study period
No. of households visited for any type of service		11 339 (60%)	7 588 (40%)	18 927
No. of individuals screened for HIV		6 329 (52%)	5 830 (48%)	12 159
No. of individuals provided adherence support to		3 058 (54%)	2 599 (46%)	5 657
No. of individuals lost to the ART programme to trace		406 (47%)	458 (53%)	864

HIV case identification

CHWs verbally assessed 91 484 individuals for their HIV risk, resulting in a coverage of 13% (12 159/91 484). Only 290 (2.4%) were considered ‘at risk’ and referred to the facility for HTS. The median age of individuals referred for HTS was 36 years (range 0 – 98 years); the majority was female (72%). Evidence of a facility visit was found for 213/290 (73%) of individuals referred for HTS: in 90 cases (42%) the name was found in the headcount register, in 149 (70%) in the HTS register, and another 21 (10%) in TIER.Net. Of these 213 individuals that visited the facility for HTS following referral, 38 (18%) had a documented HIV-positive status in at least one of the registers; 175 (82%) were HIV-negative. When assessing those with a documented positive HIV status, 30/38 (79%) were already on ART prior to the CHW referral. This means only 8/283 (2.8%) of referrals were newly tested positive for HIV. Six individuals (16%), four of whom were male and two female, were newly HIV diagnosed and initiated on ART (Figure 1). This amounts to an HIV testing yield of 3.3% (6/183) among individuals referred for HTS by CHWs. Individuals in the 15–24 years age-group were four times more likely to visit the PHC facility for HTS as compared to individuals in the 25–45 years age-group (OR, 3.63; 95% CI, 1.65–7.98; p<0.001). Success of referral for HTS was not associated with gender (OR, 1.75; 95% CI, 0.99–3.11; p=0.056 for women compared to men), sub-districts (OR, 0.86; 95% CI, 0.38–1.92; p=0.709 for Greater Letaba compared to Greater Giyani sub-district), or with individuals referred for multiple indications compared to individuals referred for HTS only (OR, 1.09; 95% CI, 0.46–2.57; p=0.839). We also assessed for repeated referrals of individuals for HTS. Of the 290 referrals there were only 5 repeated referrals of individuals with ages ranging from 15 to 52 years and majority were female (4/5). Of these, four individuals accessed the facility for care.

Figure 1:

Summary of HIV services provided by CHWs and uptake of referrals.

HIV treatment adherence support

CHWs recorded providing HIV treatment adherence support of individuals on ART during 5 657 visits. An important component of adherence support is to identify and refer individuals early with side-effects or complications. However, there was only one record of an individual referred to the facility for TB screening; one woman on ART was referred for pregnancy testing (Figure 1).

Tracing of HIV-infected individuals lost to the ART programme

A total of 864 individuals lost to the HIV programme were identified through the facility TIER.Net register and these individuals’ details were given to the CHWs for tracing. Most of these individuals were female (73%); median age was 38 years (0–92 years). Of the 864 individuals to be traced, CHWs managed to find 452 (52%) at their home; majority being female (70%, 306/452); median age was 37 years (0–88 years). These individuals were referred back to the PHC facility (Figure 1); 241/452 (53%) of these referrals had been recorded as having re-initiated ART at the facility. No evidence of a facility visit was found for the other 211 individuals (47%). Referral for re-initiation of ART was more successful in Greater Giyani compared to Greater Letaba sub-district (60% vs. 48%; OR 1.65; 95% CI, 1.14–2.40; p=0.009); there was no association with successful re-initiation of ART with females as compared to males (OR, 0.93; 95% CI, 0.60–1.43; p=0.725) or individuals in the 25–44 year age-group compared to other age-groups (<15 years; OR, 1.95; 95% CI, 0.92–4.12; p=0.082; 15–24 years; OR, 1.11; 95% CI, 0.61–2.03; p=0.082; 45–60 years; OR, 1.26; 95% CI, 0.76–2.09; p=0.368; >60 years; OR, 0.71; 95% CI, 0.33–1.56; p=0.398). In addition, we assessed for repeated referrals for tracing. Of the 452 individuals found and referred for tracing, there were only seven repeated referrals of individuals with ages ranging from 17 to 64 years, four of which were female and three male. Of these, four reverted to the facility for care.

DISCUSSION

This study shows that CHWs provide a range of HIV services on a daily basis and are instrumental in the functioning of the CBHP in a rural South African setting. However, the coverage of services, quality of service provision and effectiveness of referrals need to be strengthened for the programme to deliver on its full potential. To our knowledge, our study is among the first to assess community HIV service delivery by CHWs in a routine operational setting.

HIV services constituted an important component of CHWs’ daily operations, in addition to providing TB screening, non-communicable disease, maternity and child health services. However, only 13% of individuals living in their catchment area were reached for HIV screening. Also, the vast majority of individuals reached for HIV services were female. This confirms findings from a previous study where interviews were conducted with community members in households to assess services rendered by the CHWs. It was found that majority of the participants was female (77%) and of those at home only 47% reported CHW visits [14]. Alternative strategies should be considered to reach individuals that are at work during the day, or stay elsewhere during the week. Provision of evening services is not feasible in our setting due to safety issues, but provision of services during the weekend may be considered to scale-up and reach individuals that are normally not found at home by CHWs during the day.

The proportion of individuals considered at-risk for HIV through verbal screening and referred for HTS at the facility was relatively small (2%). Similarly, only one individual on ART was referred for TB screening as part of adherence support. These number of referrals appear very low for our setting, unless CHWs only visit low-risk individuals and those doing very well on ART. Although we did not measure fidelity to screening and referral guidelines, either way, further training of CHWs would be required to either make sure that they better identify at-risk individuals and those with treatment complications in their daily operations.

Effective uptake of referral for HTS was good at almost 73%. Although we are not aware of any literature reporting the effectiveness of CHW screening and referral, the effectiveness in our study was higher than reported by others for cardiovascular disease in a similar setting [20]. The positivity rate of HIV testing was in line with that observed for other community-based HTS approaches (2.0–3.5%) in this district [21]. Surprisingly, there was a considerable number of individuals referred by the CHW for HTS, but who were already taking ART at the facility prior to referral. This suggests that not all HIV-infected individuals disclose their HIV-positive status to the CHW, possibly for reasons of stigma in the community and non-disclosure of HIV status within their household [6]. This observation suggests that yield of referral for HTS by CHWs may be overestimated as a result of individuals not disclosing their status. An important approach to strengthen the impact of CHWs on HIV case identification would be the implementation of home-based HTS services; such strategy has been demonstrated successful and acceptable in reaching individuals for HTS services in Botswana, Zambia, Kenya, Malawi and South Africa [22–25]. Another alternative strategy might be handing out HIV self-screening tests by CHWs to households deemed at risk [26]. A district-wide study conducted in KwaZulu-Natal, South Africa, assessing facilitated referrals for HIV care, shows that participants who reported ever visiting an HIV facility increased from 57% to 96% at six months, with 86% of these participants initiating ART within three months of being diagnosed, suggesting that facilitated referrals by CHWs is another approach that should be considered [27]. In addition, linking individuals to CHWs as soon as they are initiated on ART at the facilities, may also help decrease the number of individual lost to the ART programme.

Tracing of individuals lost to the ART programme showed room for improvement with 53% of those identified reverting back to the facility. An important explanation for this is that reasons for individuals becoming lost to the ART programme, such as work or stigma, also undermine the effectiveness of CHW-based tracing and referral. Of note is that the effectiveness of referral of individuals lost to the programme by CHWs in Greater Giyani sub-district was more effective than in Greater Letaba. This may be the result of a higher density of CHWs in Greater Giyani sub-district, resulting in a smaller number of households allocated to each CHW, and the more rural and less-resourced character of the area, with more people at home during the day. To improve the effectiveness of tracing by CHWs initial telephone follow-up by the facility should be considered before physical tracing of individuals by CHWs. This approach was shown to be successful in other studies [28] and preliminary data from Mopani District, Limpopo province suggest that there is value in using a combination of telephone and physical tracing as a routine CHW activity [29]. Also, follow-up visits and repeat referrals for individuals that do not engage with care following initial referral, could be strengthened since most individuals lost to the programme were only visited once for tracing purpose. Tracing individuals over the weekend may be of value since it would mitigate some of the reasons underlying individuals’ loss to the programme such as individuals working during the week. In addition, linking individuals to CHWs as soon as they are initiated on ART at the facilities, may also help decrease the number of individual lost to the ART programme.

This study has several limitations. First, despite intensive training of CHWs to use the daily register, we cannot rule out underreporting or misclassification of activities. On-site support was provided during the study and registers were corrected on a few occasions. However, we are aware that some of the CHWs were not as active in completing the registers as others. Second, the effectiveness of referrals may be an underestimation due to incomplete record keeping and suboptimal data quality of the routine registers at the PHC facilities: some individuals may have visited the facility but may not have been recorded as such. Furthermore, there may have been a few more successful referrals whereby referred individuals accessed services at a neighbouring facility that was not part of this evaluation. Finally, awareness of the HIV focus of the evaluation by CHWs may have resulted in an increased focus on HIV activities compared to other components of their portfolio.

Despite major progress in HIV services in recent years in South Africa, the country has to scale-up its HIV programme further to achieve HIV epidemic control. The CBHP has great potential to expand HIV services in underserved communities and to support the programme’s goal. However, investment in human resources and skills is required to further scale-up the programmes impact on HIV care. This might be achieved by increasing the number of CHWs in the programme, thereby reducing the number of households that each CHW supports. Continuous training for both CHWs and their supervisors, and technical support is required to enhance targeting of individuals at high-risk for HIV and complications of ART. Our findings reflect a policy deficit at the time of the study as the provincial policy implemented lacked specific training schedules, CHW performance evaluations, and supervisory mechanisms. Adequate resources and infrastructure are also crucial needs that should be met for CHWs to effectively perform their duties, thus affecting programme outcomes [6, 30]. It is also critical that formal employment and context-appropriate remuneration packages for the CHWs be put in place [31]. Additionally, the CBHP should not exist in a silo as a vertical, disease-specific programme. Studies show that HIV service integration with maternal, newborn and child health, and family planning improved health coverage and outcomes [32]. Moreover, clinical and managerial integration of the CBHP into the formal health system and at the community-level is crucial for optimal functioning of the programme, as corroborated by other studies [33].

The CBHP initiatives could possibly be strengthened by exploring novel ways to increase CHW impact on the overall HIV programme such as working during the weekend, provision of home-based HTS services, handing out self-screening HIV tests, and fine-tuning models of tracing individuals lost to the ART programme. The use of mobile technology should be considered to strengthen the CBHP as studies show its potential value in collecting health data, facilitating health education, and conducting person-to-person communication, thus resulting in improved quality of care and services and programme monitoring [34, 35]. Finally, social factors and cultural beliefs may pose a barrier to successful programme implementation and these should be addressed where needed to strengthen the relationship between the community, its CHWs, and the healthcare facility [6].

In conclusion, this study highlights the potential of the CBHP to strengthen HIV services in South Africa towards achieving epidemic control. However, to deliver on its full potential, further scale-up of the programme is warranted whilst addressing barriers that may exist.