Why we must decentralise
While South Africa has improved case detection for drug-resistant tuberculosis (DR-TB) with the introduction of the Xpert TB test, the capacity to treat patients diagnosed with Rifampicin-resistant TB has not kept pace. More individuals are being diagnosed than are treated, and many of those who don’t receive treatment die or infect others around them, or both.
Decentralisation of DR-TB treatment offers the opportunity for more patients to access appropriate treatment for their DR-TB disease. Historically, DR-TB treatment has been provided through small, individualised programmes, with specialist clinical support and often long periods of hospitalisation. Such an approach might work where patient numbers are small, although the burden this places on individual patients is large. Instead, decentralisation aims to treat patients at much lower levels of the healthcare system, with much less reliance on hospital admission.
Prior to 2011, national policy in South Africa mandated hospitalisation of all DR-TB patients for treatment initiation. This resulted in long delays, ongoing transmission, and undoubtedly high mortality. Recognising the deficiencies in the centralised system, decentralised care was piloted at a few sites, including some districts in KwaZulu-Natal and Khayelitsha in the Western Cape. As these were quite different settings, with Khayelitsha being a more urban and densely populated when compared to the rural districts of KZN covering large areas, decentralisation was implemented differently.
In Khayelitsha, decentralisation meant that patients who were well enough to attend their local primary care clinic could both start and continue their treatment at the clinic. Only patients who required admission based on their clinical condition were admitted, regardless of sputum smear positivity. Decentralisation included building capacity at local primary care clinics in order to manage DR-TB patients, with patients attending their local clinic daily to receive treatment.
At the predominantly rural pilot sites in KZN four decentralised DR-TB sites based at district hospitals were established across the province in DR-TB ‘hot spots’. As three- quarters of the patients were co-infected with HIV and very ill at the time of DR-TB diagnosis, most patients were initially hospitalised. On discharge, mobile injection teams administered the daily injectable and provided support for patients in the community. Some patients who lived close to primary health care (PHC) clinics accessed this care at their clinic.
Based on evidence of long waiting lists in most provinces and, to a lesser extent, the relative success of these pilots, the National Department of Heath (NDOH) revised policy in 2011 to support decentralisation. The new policy outlined the establishment of decentralised and satellite DR-TB units as additional levels between the centralised specialist DR-TB hospitals (that were relied upon before 2011) and primary healthcare facilities (see figure 1 on next page). These 40- bed decentralised DR-TB units are designated as the health system level for treatment initiation, while satellite DR-TB units ensure treatment follow up. The policy also outlines the provision of mobile teams to provide injections for patients unable to attend their local clinic on a daily basis. Admission for treatment initiation is required for all patients who are sputum smear positive or are diagnosed with XDR- TB. However, patients who refuse admission, but still desire treatment can receive treatment on an ambulatory basis.
Hospital admission has often been described as a means of ‘isolation’ and protecting the community from further transmission. However, TB transmission predominantly takes place before patients are diagnosed and started on appropriate treatment. Infectiousness, i.e. the risk to others, reduces dramatically with treatment, even with second-line treatment for DR-TB. As a result, ‘isolating’ patients after treatment initiation is unlikely to reduce community transmission. Rather, early diagnosis and treatment initiation is the best infection control.
The extent to which this new policy has been implemented across the provinces is unclear. Some areas have decentralised care substantially, for example the Cape Metro region and some rural areas in the Western Cape, along with most districts in KZN. However, in most provinces hospitalisation is still relied upon. Large patient numbers and limited beds results in decentralised care on an ad hoc basis, and not always in accordance with the national guidelines. We fear that healthcare workers at primary healthcare clinics may have been inadequately prepared, trained or resourced to manage DR-TB at a community level.
In addition to improving access to treatment, decentralised care often results in more rapid treatment initiation. Given that early mortality is significant, particularly among those co-infected with HIV, rapid treatment has the benefits of reducing deaths and ongoing transmission. Decentralised care is also less disruptive to patients’ lives, allowing patients who are often primary caregivers to remain at home and continue family responsibilities. Hospital admission, for long periods, often at considerable distances from patients’ homes can be extremely isolating. From a health system perspective, decentralisation offers better use of scarce healthcare resources, both through reduced reliance on expensive hospital care and integration with existing healthcare services, i.e. the routine drug-susceptible TB programme.
Effective decentralisation in South Africa needs to be undertaken with care as the health system is often weak and overburdened. Due to the lack of trained staff and constant staff movement, many primary healthcare facilities are fragile, so careful planning and ongoing monitoring are needed to ensure guidelines are correctly implemented and patients are managed appropriately. Although different models will be appropriate in different settings, ongoing provincial support and monitoring are necessary to ensure that districts own the DR-TB programme and integrate DR-TB services into the existing district primary healthcare and TB services. By integrating the DR-TB programme, with its extra resources, into existing district services, these services will be strengthened as opposed to further fragmentation by the introduction of yet another vertical programme.
In turn, district management has to ensure that facilities are supportive of the DR-TB service, and practices such as the rotation of key clinical posts, are minimised. Ongoing training, mentoring and support, clear referral pathways and clinical audits are necessary to ensure clinical staff implement the treatment guidelines in their entirety. From the nurse- initiated and managed antiretroviral treatment (NIMART) programme we learnt that training is insufficient and primary care staff will not implement as they have been trained in the absence of supportive supervision.
Given that the management of DR-TB in conjunction with HIV is sometimes complicated, simple guidelines and algorithms have to be developed to assist clinicians who have limited experience of DR-TB. Guidelines to manage the following issues are needed:
- Patients with discordant GeneXpert, culture and drug sensitivity testing results.
- Otoxicity (hearing loss due to the DR-TB medications).
- Patients who are not responding to treatment.
- Those who need palliative care.
Additional resources invested in counselling and community support are needed, given the difficulties many patients experience in adhering to DR-TB treatment due to side effects.
As the decentralised programmes from Khayelitsha and KZN demonstrate, decentralisation doesn’t require the same approach in different settings. Mobile teams might be key to reaching patients in more rural areas, while less necessary in others. Short term admission in the order of a week or two, for treatment initiation for a greater proportion of patients might be suitable in areas with limited health system capacity, while treatment initiation at primary care will be preferable in others. Whatever approach is chosen, well-defined, written policy guidelines, outlining the pathways that patients follow after diagnosis, is required for individual districts and provinces.
A system of monitoring how well the DR-TB programme is working is also required. Currently, South Africa relies on laboratory records to assess case detection, and a centralised electronic database to assess treatment. Under this system, case detection is most likely overestimated, while treatment initiation underestimated. Decentralising data recording to the levels of the health system where patients are diagnosed and treated enables a more accurate reflection of where we are, and will enable ongoing evaluation for programme improvement as we move forward.
The South African Constitution enshrines the right to health care and thus the responsibility to provide universal access to appropriate, second-line treatment for DR-TB. To date this is not a reality. The current decentralisation policy needs to be adequately resourced, implemented fully and expanded to allow for innovations such as nurse-initiated treatment. With the prospect of new, simpler and shorter DR-TB treatment regimens, the vision is that all primary care facilities in South Africa will be able to provide DR-TB treatment routinely.
Loveday is a senior scientist at the Medical Research Council and engaged with her doctoral studies in the Public Health Department at the University of KwaZulu-Natal. Her doctoral work evaluates different models of care for patients with MDR- TB in KwaZulu-Natal.
Dr Cox is a senior researcher at the Division of Medical Microbiology, University of Cape Town.