Optimizing health systems to improve delivery of decentralized care for patients with drug resistant tuberculosis

Tuberculosis is a disease of the poor and remains a significant cause of disease and death globally. Our failure to effectively control the TB epidemic has resulted in the emergence of TB bacteria that have become resistant to the main drugs we use in treatment. These drug-resistant bacteria (DR-TB) can then be transmitted to others. In some settings DR-TB has now reached epidemic levels. While treatment is available, it's lengthy, complicated, expensive, and results in poor patient outcomes.

South Africa has a high burden of DR-TB, with more than 26,000 cases reported in 2013. Data suggest that a third of these patients do not have access to appropriate treatment, with only 40% cure for the rest. Fortunately, there are some recent advances that could assist in combatting the DR-TB epidemic. A rapid diagnostic test for DR-TB has been marketed and has now been rolled out across South Africa. This test (Xpert) has increased the number of patients identified and reduced the delay in receiving the diagnostic result. In addition, a new drug for treating DR-TB has just been registered in South Africa. Bedaquiline has the potential to dramatically improve the effectiveness of DR-TB treatment.

However, DR-TB diagnosis and treatment needs to be delivered in the context of the existing health system, the characteristics of which are likely to influence the impact of new interventions. In order to increase access to DR-TB treatment, South Africa moved to a policy of decentralized treatment provision in 2011, i.e., providing DR-TB treatment at lower levels of the health system, with less reliance on hospital treatment. This policy has been variably implemented across different provinces, health districts and settings (i.e., rural versus urban) in South Africa.

In order to maximize the benefit of new interventions, we aim to assess health system factors that enhance or undermine the delivery of treatment for DR-TB, specifically in regard to decentralization of care. We aim to determine what works and why it works, across different settings. Health system factors include such things as: the allocation of appropriate (number and training) human resources, financing at appropriate levels, organizational management, efficient communication and referral systems, sufficient inpatient capacity and access to other services needed for the care of DR-TB patients. These health system factors determine whether individuals with DR-TB disease access care, how long this takes, the quality of care they access, whether they transmit their infection to others and finally whether they have a successful treatment outcome or not. Ultimately, our findings will be used to identify feasible and effective strategies to improve decentralized care for patients with DR-TB in South Africa and other similar settings.

Diagnosis of drug-resistant tuberculosis (DR-TB, specifically rifampicin-resistance) in South Africa has increased dramatically in recent years with the roll-out of the Xpert MTB/RIF rapid PCR diagnostic test. While more than 26,000 cases were reported in 2013, only 42% of these are reported to have started on second-line treatment and less than 50% of those starting treatment are reported to be successfully treated. In 2011, South Africa instituted a policy of decentralization for DR-TB treatment, in an attempt to move away from lengthy and expensive inpatient treatment and the consequent waiting lists for bed availability and treatment initiation. Given the range of variation in disease burden, population density and health system capacity across South Africa, the policy allowed for different interpretations of decentralized care, with greater or lesser reliance on hospitalization. Preliminary data from an ongoing study suggests that there is a large variation in outcomes among diagnosed DR-TB patients across health districts in South Africa. Considerable diversity in the extent and nature of decentralization was also observed.

In this study we aim to utilize data on patient outcomes (treatment initiation, time to treatment and treatment success) from existing patient cohorts to purposively select a sample of health districts across three provinces in South Africa. Starting from the national level, through to the provinces and subsequently districts, we then aim to conduct a realist study to identify features of decentralized care and health system factors that are associated with positive patient pathways to care and outcomes. We aim to understand the mechanisms underlying what works and why it works in terms of decentralized DR-TB care, in particular settings and contexts. In collaboration with policy makers and health authorities, we will then identify strategies for short and medium-term improvements in the delivery of decentralized care for DR-TB.

We recognize that gaps exist between recommendation, policy change and successful implementation of revised policies. We will therefore focus on developing very specific recommendations that target short, medium and longer-term change. Our experience from our previous work, described below, leads us to believe that there are 'low hanging fruit', which are amenable to intervention, e.g. simple improvements in the prompt delivery of laboratory results to clinicians at primary care level are likely to increase access to treatment in the short term. Medium term improvements might result from improved access to specific DR-TB training for health care workers at primary care level, while longer term improvements might result from the implementation of more complex referral pathways for patients requiring more specialised care than can be provided at the primary care level.

This research aims to improve the ability of health systems (and in particular the South African health system) to effectively treat diagnosed DR-TB cases. Therefore, the beneficiaries of this research will be:
1) Rural and urban patients with DR-TB with varied access to appropriate and timely services;
2) Local managers and health workers tasked with implementing DR-TB decentralized care.
3) Policy makers (both provincial and national level) within the National Department of Health and in similar settings globally.

DR-TB Patients will benefit from improved health systems that more effectively deliver care for DR-TB. Specifically, we will develop recommendations to improve linkage to appropriate care following diagnosis, to support the roll out of decentralized care (closer to the patient's home) and to retain patients in care. Ultimately, we hope that this will lead to improved quality of life and reduced morbidity and mortality amongst DR-TB patients.

Local managers and health workers tasked with implementing decentralized care for DR-TB cases will benefit through the identification of interventions (e.g., training, supportive supervision, performance management, etc.) that can support and enhance their capacity to deliver decentralized care, by maximising the use of resources, evidence and skills available to them.

Policy makers within the South African health services will benefit from evidence-based recommendations on how to best organize the health services for DR-TB. This will include a clear description of optimal interventions for decentralization in different contexts, identification of strategies that represent best practice, and guidelines on how to support local health systems to maximize the potential benefit of new interventions for DR-TB (such as new diagnostics and models of care). The recommendations will also address the need and mechanisms to improve health information systems for DR-TB. Whilst these recommendations will be based within the South African context, it is anticipated that they will also be useful for policy makers in other low and middle income countries with a high burden of DR-TB.
Additionally, research staff trained in this approach as part of this study will be able to contribute more broadly to the development of an evidence-based approach to health policy and implementation in South Africa.

We also aim to utilize data from the patient cohorts and findings from this study to inform and populate mathematical models, aimed at addressing key questions in the health system response to DR-TB. For example, while access to centralised, specialist care might result in improved outcomes for the individuals able to access such care, more widespread access to treatment through decentralization might result in somewhat reduced quality of care for individuals, but treatment success for a greater proportion of diagnosed patients. Policy decisions need to weigh up these key trade-offs; mathematical models assessing future impacts on DR-TB transmission of different approaches could inform such decisions.