As highlighted in the Sustainable Development Goals (SDGs) [1], the world continues to face high rates of poverty and associated population health challenges, including the joint epidemics of tuberculosis (TB) and human immunodeficiency virus (HIV) [2]. High rates of TB-HIV co-infection necessitate HIV testing of all TB patients, yet, only 55% of notified TB patients in South Africa had a recorded HIV test result in 2015 [3]. With only about half of notified TB patients knowing their HIV status, the shortfalls and deficiencies in integration between TB and HIV services in the country become clear. Lack of progress in successfully integrating traditionally separate TB and HIV services into parallel programmes, staffing and approaches negates patients’ preference for combined services and results in wastage of scarce health resources and patients’ time and finances [4].

Both the link between TB and poverty [5–7] and the link between HIV and poverty [8–10] are well established. The Gini index for income inequality – an indicator for achieving SDG 10 to reduce inequality within and among countries – for South Africa was 63.4 in 2011 [11]. In the same year, 16.6% of the country’s population lived below the poverty headcount threshold of ZAR 13.81/US$ 1.90 (2011 exchange rate) per day [12]. The Free State is one of the poorest of the country’s nine provinces. Just more than 5% of the national public health-sector dependent population – of whom more than 80% are African and thus historically and socioeconomically disadvantaged – live in the Free State [3]. At ZAR 91994.00/US$ 12653.92 (2011 exchange rate), average annual household income in the province in 2010/2011 was substantially lower than that for the country as a whole at ZAR 119542.00/US$ 16443.19 (2011 exchange rate) [13]. In 2016, the Free State also had the lowest life expectancy for both males and females of all the provinces [14].

South Africa is extraordinarily challenged by the TB-HIV co-epidemic. The World Health Organization’s (WHO) global TB report shows that in 2016 the country had the highest total TB incidence rate in the world at 781 cases per 100,000 population compared to 140 cases per 100,000 population globally [15]. Further to the report, the HIV-positive TB mortality rate in South Africa was 181 per 100,000 population compared to five cases per 100,000 population globally and HIV prevalence in incident TB cases was 59% compared to 10% globally. In 2015, the Free State reported 15,833 cases of TB and that, in 2014, 70.3% of TB patients were HIV co-infected [3]. In 2014, the TB death rate in the province was substantially higher than that recorded for the country as a whole. A review of cases in the electronic TB register in the province from 2003 to 2012 revealed that both HIV co-infection and unknown HIV status were independently associated with increased likelihood of mortality while on TB treatment [16].

Directly observed support (DOT) continues to be commended as an efficacious approach to overcome interruption in both drug-susceptible TB (DS-TB) [17] and multidrug-resistant TB (MDR-TB) [18] treatment. Uncertainty however exists in the global and national guidelines about how and by whom DOT support for TB can be delivered most effectively and how DOT can simultaneously be used to strengthen uptake of HIV testing and treatment among TB patients. In 2017, the WHO conditionally recommended that the following treatment administration options may be offered to patients on DS-TB treatment: firstly, community- or home-based DOT was recommended over clinic-based DOT or unsupervised treatment; secondly, DOT administered by trained lay providers was recommended over DOT administered by family members or unsupervised treatment; and, thirdly, video observed treatment (VOT) could replace DOT where the technology was available [17].

The South African National Tuberculosis Management Guidelines released in 2014 state that the TB treatment supporter may be a healthcare worker, or a trained workplace or community health worker (CHW) or whoever patients choose to watch them swallowing the tablets “in a way that is sensitive and supportive to the patient’s needs” [19 p. 56]. Further to the guidelines, DOT services must be organised to suit the patients’ circumstances and must be provided as close to their homes as possible, unless patients live close to a clinic and it is convenient for them to take their treatment at the clinic. The guidelines prioritise the following groups for treatment support: children, elderly or infirm patients; those with a history of interrupting treatment, substance or alcohol abuse or mental illness; those who request a treatment supporter; and those who are homeless or living under poor social conditions. There is however considerable variation in how DOT is implemented, how HIV-related services have been integrated into DOT support, and the extent of cross-training to deliver both programmes’ services in South Africa.

Due to the wide use of an intensive service (directly observing patients taking their pills on a daily basis), DOT has been extensively studied for its impact on retention in treatment and treatment outcomes. Results have been mixed, and controversy remains about whether DOT is effective. Individual studies have shown that integrating DOT with an HIV home care programme improved outcomes for TB patients in Zambia [19]. A study in the Western Cape province of South Africa found community-based DOT to be more cost-effective than clinic-based DOT [20]. A study in Namibia showed that TB patients on the community-based TB treatment option had better cure rates than those on clinic-based DOT or self-administered TB treatment [21]. Another South African study in the Northern Cape province found the effects of DOT to be concentrated among patients undergoing retreatment for TB [22]. Qualitative studies in Ethiopia [23] and Pakistan [24] showed some of the potential advantages for clinics and patients with community-based DOT such as patients not having to travel to a clinic for DOT and alleviation of clinic-based healthcare workers’ time constraints by engaging CHWs.

However, a Cochrane review of 11 clinical trials concluded that TB cure and treatment completion rates were not improved by DOT [25]. A similar conclusion was drawn from a meta-analysis of ten studies (including five trials) which found that DOT was not significantly better than self-administered treatment in preventing microbiologic failure, relapse or adverse drug reactions [26]. Thereupon, a systematic review and meta-analysis of eight studies (including one trial) that compared the effectiveness of community-based versus clinic-based DOT concluded that community-based DOT had higher treatment success, but conceded that the evidence for this was not strong [27]. A systematic review of ten trials and eight quasi-experimental studies found evidence of beneficial effects from DOT with regard to treatment adherence and in terms of cure and treatment success rates, but no beneficial effect in terms of increasing the treatment completion rate [28]. Another systematic review of 23 studies (eight trials) concluded that while there was no convincing evidence that clinic-based DOT was more effective than self-administered treatment, there was convincing evidence that community-based DOT was more effective than self-administered DOT and that community-based DOT was as or more effective than clinic-based DOT [29]. This was corroborated by another systematic review and meta-analysis of eight trials and nine cohort studies finding that community-based DOT did improve TB treatment outcomes [30].

A systematic review and meta-analysis of 31 studies in 22 countries focusing specifically on MDR-TB patients showed that providing DOT for a full course of treatment was associated with a significantly higher treatment success rate [31]. However, this study reported no significant differences in the treatment success rates for DOT variously provided to MDR-TB patients by healthcare providers, family members and private DOT providers or between patients having clinic-based DOT and home-based DOT. Perhaps the most compelling argument in favour of maintaining DOT for both DS-TB and MDR-TB is that no large-scale TB programme without DOT has achieved the global TB targets, while most programmes using DOT achieve or nearly achieve these targets [32]. Nevertheless, it is argued that the effectiveness of DOT still needs to be more rigorously evaluated through a pragmatic experimental trial conducted in real-world programme settings [33].

Research in the Free State province showed relative patient satisfaction with HIV counselling provided to TB patients by both lay counsellors and nurses and recommended that expanded use of lay counsellors in TB/HIV programmes could help mitigate the human resource crisis that had resulted primarily from shortages of nurses [34]. However, despite undergoing a standard 59-day training programme, knowledge deficits among DOT supporters concerning TB, HIV and the link between TB and HIV have been identified [35–37]. Stigmatisation against people living with HIV/AIDS (PLWHA) has also been observed in a survey of CHWs (over half of whom were DOT supporters) in the province observing that about 15% of respondents agreed with the statement “Most people with HIV/AIDS only have themselves to blame” [38].

This study describes how DOT support by CHWs was used in one province of South Africa, the Free State, to provide HIV outreach, referrals, and education for TB patients in 2012. The study addressed the following specific research questions: 1) what type of support did patients receive for their TB treatment (home-based DOT, clinic-based DOT, or support by family/friends/employer?; 2) what were the characteristics of the TB patients who did and did not receive DOT support – in particular, did the most vulnerable patients receive support?; 3) what type of HIV-related services – in particular referrals to HCT – did patients receive from their supporters?; 4) how satisfied were patients with the HIV-related information they received, and how did that vary by type of supporter?; and 5) what effect did DOT support – in particular home-based DOT support – have on receipt of HCT?

The WHO End TB Strategy envisions a TB-free world. The first of the three pillars of the strategy is “integrated, patient-centred care and prevention”; the second is “bold policies and supportive systems”; and the third is “intensified research and innovation” ([41] p.2). “Social protection, poverty alleviation and actions on other determinants of TB” is a component of the second pillar ([41] p.2). One of the ways to do this is by providing effective community-based DOT and HIV services to TB patients at greatest socio-economic risk.

Guidelines for HIV providers in South Africa at the time of the survey (2012) suggested that lay counsellors should play a role in supporting patients [42]. For TB patients, lay counsellors were generally DOT supporters, many of whom had been cross-trained to provide HIV services. While much, but not all, prior research suggests that DOT can be effective for improving TB [22, 28–30] and MDR-TB [31] patient outcomes, research on the effect of DOT specifically on receipt of HIV services is limited. Our research from four municipalities in the Free State province suggests that DOT support – when it was received - facilitated access to a range of HIV counselling services. For example, DOT supporters were encouraging TB patients to take up HCT and educating them about the relationship between TB and HIV. However, at the time of the study the large majority of TB patients in these municipalities were not receiving any formal DOT support. Only 14.1% were receiving home-based DOT, while another 10.4% were receiving clinic-based DOT. This is in spite of the fact that DOT support for all TB patients was the goal of South African health policy at the time.

Prior research [43] and South African guidelines [44] emphasise that DOT support should especially be targeted to the most vulnerable patients. We examined specifically the effect of home-based DOT support and whether it was being targeted to the most vulnerable TB patients. With the exception of older patients being more likely to receive DOT at home, other vulnerable groups were not apparently targeted. Indeed, those in the poorest quality housing were less likely to have home-based DOT support than other TB patients. This finding is similar to that of a recent study in India showing that the most vulnerable TB patients faced the most difficulties in accessing DOT and completing treatment [45].

The advantage of age in receipt of home-based DOT did however not translate into higher rates of receipt of HIV counselling for older patients. It could be that DOT supporters were unassuming or demure about providing HIV-related advice to an older person or may have thought that an older person was unlikely to be in need of HIV counselling, especially when they were a friend or family member.

The study provides important information that is relevant to development of new policies concerning use of CHWs in addressing the TB/HIV co-epidemic. In particular, if cross-training in HIV services can become a standard part of DOT training, CHWs who have frequent contact with TB patients could become an important cadre of front-line HIV prevention personnel in non-clinical community-based settings. Further, in choosing who might receive DOT, it is critical that such a programme target those at the highest socio-economic and clinical risk.

The study is limited to four municipalities in Free State, South Africa at one point in time. In addition, data are self-reported. This particularly may affect how TB patients reports their DOT support. While the term “DOT supporter” implies Directly Observed Treatment, it is possible that the support that patients reported was not necessarily someone observing them taking medication regularly, but was rather someone they saw more occasionally who provided counselling and support. The data do not indicate some important aspects of the HIV services that the DOT supporters, or other supporters, were providing. For example, it is likely that some supporters provided HIV counselling (a direct service), and others encouraged patients to receive counselling at a clinic (a referral). The data do not distinguish these service differences. Another limitation of this analysis is that we did not compare patient treatment outcomes (for example mortality) between patients on DOT and those who were not, since all patients were alive and on treatment at the time of data collection.

Recent policy developments in South Africa provide some opportunities to improve services for TB patients, through a new “PHC re-engineering strategy,” which includes the deployment of facility-linked outreach and family health teams [46]. Currently, the scope of work of the CHWs serving on these teams, other than in exceptional cases, does not include DOT support. In addition, the teams do not visit households on a daily basis. It is not clear to what extent workers will be cross-trained in both TB and HIV prevention and treatment. Consequently, some of the benefits that co-infected patients might receive from dedicated daily DOT support may no longer be available. It will be important for the national and provincial governments to assure that the benefits of DOT for HIV counselling and prevention – as documented in this study – are not lost with the implementation of the new PHC strategy.