Different challenges, different approaches and related expenditures of community-based tuberculosis activities by international non-governmental organizations in Myanmar

Background International non-governmental organizations (INGOs) have been implementing community-based tuberculosis (TB) care (CBTBC) in Myanmar since 2011. Although the National TB Programme (NTP) ultimately plans to take over CBTBC, there have been no evaluations of the models of care or of the costs of providing CBTBC in Myanmar by INGOs. Methods This was a descriptive study using routinely-collected programmatic and financial data from four INGOs during 2013 and 2014, adjusted for inflation. Data analysis was performed from the provider perspective. Costs for sputum examination were not included as it was provided free of charge by NTP. We calculated the average cost per year of each programme and cost per patient completing treatment. Results Four INGOs assisted the NTP by providing CBTBC in areas where access to TB services was challenging. Each INGO faced different issues in their contexts and responded with a diversity of strategies. The total costs ranged from US$ 140 754 to US$ 550 221 during the study period. The cost per patient completing treatment ranged from US$ 215 to US$ 1 076 for new cases and US$ 354 to US$ 1 215 for retreatment cases, depending on the targeted area and the package of services offered. One INGO appeared less costly, more sustainable and patient oriented than others. Conclusions This study revealed a wide variety of models of care and associated costs for implementing CBTBC in diverse and challenging populations and contexts in Myanmar. Consequently, we recommend a more comprehensive evaluation, including development of a cost model, to estimate the costs of scaling up CBTBC country-wide, and cost-effectiveness studies, to best inform the NTP as it prepares to takeover CBTBC activities from INGOs. While awaiting evidence from these studies, model of CBTBC that have higher sustainability potential and allocate more resources to patient-centered care should be given priority support. Electronic supplementary material The online version of this article (doi:10.1186/s40249-017-0263-9) contains supplementary material, which is available to authorized users.


Multilingual abstract
Please see Additional file 1 for translations of the abstract into the six official working languages of the United Nations.

Background
Although the global mortality rate of tuberculosis (TB) in 2015 was 47% lower than in 1990, TB now ranks alongside the human immune deficiency virus (HIV) as a leading cause of death worldwide [1]. In 2014, an estimated 9.6 million people developed TB and 1.5 million died from the disease (including 0.4 million who were HIV co-infected) [1]. Myanmar is classified by the World Health Organization (WHO) as one of the 30 TB, TB/HIV and MDR-TB high-burden countries with a TB incidence of 373/100 000 population [2]. Consequently, TB control has been one of the priorities in Myanmar's national health plan.
Additionally, some population groups are of particular concern as they may be contributing to the ongoing infection rate. These include Internally Displaced Persons (IDPs), urban slum dwellers and hard-to-reach populations, especially those living in rural areas, hilly regions and border areas [3]. These populations maintain reservoirs of infection and continue to spread the disease in the community. Therefore, in 2011 the NTP in collaboration with international non-governmental organizations (INGOs), started to implement a community-based TB care (CBTBC) programme for active case finding (ACF) in two regions and three states [3]. The aim of ACF is to identify TB infected patients, to initiate treatment and ensure follow-up until completion [3].
The groups of concern range widely in their needs and the strategies required to perform active case finding. Urban slum dwellers are easier to reach physically but may be difficult to locate in the chaos of the slums. People living in remote mountainous areas pose a physical challenge in locating and maintaining contact over time. Thus, although active case finding is the goal, how to achieve this in some contexts can be challenging. The four INGOs in Myanmar each approach their target population using different strategies, adapted to the population's unique circumstances. As the NTP ultimately hopes to take over CBTBC, it is important to document the strategies employed and the costs associated with delivering the care. To date there have been no evaluations of the care nor of the costs of providing ACF in Myanmar by INGOs, although recent studies in Cambodia pointed out that community-based active case finding and ACF targeting household and neighborhood contacts are highly cost-effective, with the additional benefit of early case finding of patients from vulnerable age groups, i.e., younger and older [4,5].
Thus, the aim of this study is to describe the differences in provision of CBTBC and associated costs by four INGOS in Myanmar over the period of 2013 and 2014.

Design
This is a descriptive study using routinely-collected programmatic and financial data from four INGOs.

General setting
Myanmar is a multi-ethnic country located in South East Asia with the population of 51 million. It is bordered by India, Bangladesh, China, Laos and Thailand. There are over 100 languages and dialects spoken in Myanmar, which contribute to language barriers for health service delivery. In addition, the geography of the country, with a number of rivers and mountains, make many areas of the country hard to access [6], creating further barriers to delivering health care. The country is administratively divided into the Nay Pyi Taw Council Territory and 14 States and Regions. Administratively, the States and Regions are the same, but geographically, the States are situated along the border areas and have more ethnic populations while the Regions are situated in the center of the country. Generally, the Regions are more urbanized than the States. Seventy five percent of the population live in rural areas and only 25% are urban inhabitants [7].
The Department of Public Health and Department of Medical Services are the service providers and regulators in protecting the health of the people [8]. However, INGOs have been playing an increasingly important role within the evolving political and administrative context. The government used to be the main source of financing, with provision of services virtually free until user charges were introduced in the form of cost sharing in 1993; since then, household out-of-pocket payment has become the main source of finance of the health system. Health care is now highly fragmented both in provision and financing [8].

Specific setting
The NTP is the main provider of TB control and care services. It runs TB centers in 15 Regions and States, with 101 TB teams at district and township level. However, there are no TB centers or teams at the ward, village tract and village level [2].

Community based TB care
A number of implementing partners, both local and international NGOs, are currently implementing CBTBC on a large scale across the country, with the financial support from international donors. The core elements of CBTBC being implemented by the partners are the same. Given different contexts and circumstances, however, their approaches vary widely. The elements of CBTBC include: 1) community mobilization, 2) recruitment of community volunteers, 3) training of community volunteers, 4) TB awareness-raising within communities through volunteers, 5) detection of TB suspects by volunteers, 6) attending DOT by the volunteers until finishing the course of TB treatment, 7) counseling TB patients for treatment adherence, and 8) support for TB patients (transportation, nutritional support, incentives) [3].

Sources of data, data variables and data collection
All four INGOs provided TB programme monitoring and evaluation reports, activity reports and financial reports that became the data sources. Programme costs included human resources, training, communication materials and community awareness-raising, health products and equipment, TB patient supports, monitoring and evaluation activities, planning and administration and overhead. The components of human resource costs included salaries for TB staff of each INGO, both local and international, and incentives or payment for the community volunteers. Training costs were comprised of fees for trainers, food, travel and accommodation costs for the trainers and trainees, costs for learning aids and stationery, and fees for renting training venues. Communication materials involved expenses for production and delivery of IEC materials, costs for community awareness-raising activities and social mobilization. The costs for health products and medical equipment included procurement of medicines to treat side effects and multivitamins and establishment and maintenance of microscopic laboratories. Costs incorporated into TB patient supports were transportation and meal allowances for patients and volunteers during outpatient visits. Other costs included chest X-rays, supplementary food support, plus transportation and meal allowances for the patients and attendants during inpatient hospital stays. Costs for monitoring and evaluation included travel and accommodation for TB staff of INGOs and NTP officials for supervision visits, as well as costs for monthly volunteer meetings, quarterly experience sharing meetings, mid-term and annual evaluation meetings. Costs for planning and administration included the expenses for office supplies, stationery and printing costs. Overhead costs included cost for office rental, vehicle rental, fuel and maintenance costs, and telecommunications. The cost for anti-TB medicines were obtained from NTP's programme report [9].
In this study, we did not include cost for sputum examination as the procedure is identical in all models of care.

Analysis
A descriptive analysis was performed adopting a health service provider perspective. A data extraction sheet was prepared to collect the relevant data from the reports of INGOs. We calculated average cost per year of each INGO and cost per patient completing treatment. As this study includes costs for the years 2013 and 2014, cost data from 2013 were adjusted for inflation and expressed in 2014 constant prices. For staff and community volunteers jointly using other services, their salaries or incentives were allocated proportionally, based on the time spent for community based TB care activities.

Results
Four INGOs provided community based TB care in 22 townships in Myanmar during the study period (Fig. 1). The respective models of care, populations served and types of services of the INGOs are shown in Table 1. Table 2   The number of TB cases detected, the number of patients completing treatment and the cost per patient completing treatment through community based TB    INGO (C) worked in one of the conflict-affected areas in Myanmar and therefore it mainly provided care to internally displaced persons. Similar to INGO (A) it carried out active case detection via mobile teams and community volunteers. On top of that, it established its own diagnostic facilities. It had the second highest total cost and cost per patient treated, again probably due to the context in which it operated.
INGO (D) provided care in settings similar to INGO (A) and the population served was also comparable. In contrast to other INGOs, it delivered community-based TB care by establishing self-help groups (see Table 1). It was able to detect a large number of cases and achieved the lowest cost per patient treated among the four INGOs.
As there are no standard tools to assess the performance of CBTBC in Myanmar, we were unable to evaluate the overall performance of each INGO based on standard indicators. Neither were we able to assess individual elements of the interventions that were carried out. Instead, we used existing data from qualitative and mixed-methods studies that have been conducted in Myanmar to appraise the work of the INGOs in terms of their potential sustainability [11,12].
Three of four INGOs, (A, B and D) supported the NTP within its definition of the "role and responsibilities of implementers in delivering CBTBC" (4). However, INGO (C) operated in parallel to NTP by providing CBTBC with their own diagnostic facilities. The number of presumptive TB cases detected and the number of TB patients completing treatment were relatively higher in INGOs (A) and (D) while INGO (B) reported the highest costs per patient treated, which may be explained by the fact that it targeted communities in some of the most hard-to-reach areas of the country.
The study revealed several interesting findings regarding sustainability, mainly related to INGO infrastructure and the involvement of community members. First, infrastructure created by external partners, such as the diagnostic facility run by INGO (C), would be hard to sustain without funding support from the INGO. Handing over such facilities to NTP would result in unnecessary duplication, given that the NTP already has its own network of facilities in these areas.
Second, as sustainability is an important issue in CBTBC, some INGOs performed various measures to ensure that the delivery of CBTBC could be continued after their withdrawal. INGO (B) trained and utilized existing local community health workers to continue delivery of CBTBC after their withdrawal. However, the community health workers were also required to perform other tasks such as environmental sanitation, health education, nutrition surveillance, malaria case finding and treatment, case management of pneumonia and diarrhoea in children under five, and identification of pregnancy at risk for referral according to community's need [6]. This workload may have diluted the effectiveness of CBTBC by that INGO. On the other hand, using existing community health workers could result in economies of scale. However, INGO (B) spent more than other INGOs for a TB patient to be successfully treated. INGO (A) and (D) formed local groups to help TB patients in preparation of their future withdrawal from the country. INGO (A) formed village mobility working groups (VMWGs), which were expected to recruit and support volunteers who can provide CBTBC in their respective areas with the groups' own funds. Similarly, INGO (D) formed self-help groups (SHGs) to look after TB patients and take over CBTBC when the supporting INGO phased out. Consequently, these SHGs performed fund raising activities alongside CBTBC. A study assessing the cost of implementing SHGs showed that half of them were able to mobilize funding to cover up to 40% of the costs for TB patient support [13]. Furthermore, a previous study has shown that the majority of volunteers of INGO (D) stayed with the organization longer than those of INGO (A) (four vs. three years) [12]. The study on effectiveness of SHGs showed that 92% of the members remained in SHGs after completion of TB treatment [11]. These findings suggest that the CBTBC model used by INGO (D) may be more sustainable.
A study conducted in Zambia and Rwanda, which assessed the cost of community-based programmes for HIV control, indicated that the highest cost was for direct project services such as health education, referrals, primary health care services and provision of nutrition (7). In our study, only INGO (D) allocated more of their budget to patient support whereas INGOs (A) and (C) spent mostly on staff salaries.
Total costs were found to be the highest for INGO (A) and the lowest for INGO (B). However, the cost per patient completing treatment was found to be relatively higher for INGOs (B) and (C) than INGOs (A) and (D). Based on the study assessing involvement of community volunteers in TB control, the socio-demographic characteristics of the target population and population size of coverage areas of INGO (A) and INGO (D) are quite similar [12]. But, INGO (D) spent less than INGO (A) per patient to complete treatment. Thus, based on the available evidence, INGO (D) appeared to show the allocation of their financial resources more patient-oriented.
Our study has several strengths and limitations. First, this was a comprehensive study that mapped and described all INGOs providing community-based TB care in Myanmar. Second, we were able to obtain financial data from all INGOs using a standardized data collection tool, which allowed for broad cost comparisons. A limitation of this study was that we were unable to ascertain or validate the cost data obtained from the INGOs. However, as the INGOs in the country have their own internal auditing mechanisms, we are confident that the integrity and validity of the data used in this study is acceptable. There are a number of international studies evaluating the cost and cost effectiveness of communitybased TB care, however, the ability to compare our data to international literature is limited due to the differences in models of care and contexts. Studies across Africa found the cost per patient of completing treatment under community -based TB care were quite varied, from US$ 60.7 in Ethiopia, US$ 128 in Tanzania and US$ 726 for new cases and US$ 1 419 for retreatment cases in South Africa [14][15][16]. Studies done in Bangladesh and Brazil showed that it cost US$ 64 and US$ 548 respectively for a patient to complete treatment by community-based TB care [17,18]. The cost for community-based TB care in Myanmar seemed relatively higher than most of the previous international studies. However, these other studies were conducted among populations residing in urban and densely populated rural area which might reduce the transportation and other management costs. Again, the overhead and general operational expenses of INGOs included in our study made the CBTBC more expensive.

Conclusion
In conclusion, the findings from this study may inform the NTP about the models of community-based TB care and their associated costs. We recommend that standardized tools to evaluate the CBTBC performance should be developed and comprehensive evaluations, including the development of a model to estimate the cost of scaling up CBTBC country-wide and costeffectiveness studies, should be carried out, especially as CBTBC activities are eventually to be taken over by the government. We suggest that while awaiting evidence from cost-effectiveness studies, models of CBTBC that have a higher sustainability potential and that allocate the largest share of their resources to patient-centered care should be supported.