Although tuberculosis (TB) mortality has fallen by 47% between 1990 and 2015, TB remains one of the world’s deadliest infectious diseases. In 2016 there were still approximately 1.3 million deaths due to TB (among HIV-negative), surpassing an estimated 1.0 million due to HIV / AIDS. Moreover, in recent years, public health experts have become increasingly concerned about the emergence and spread of drug-resistant strains of TB, particularly those that are not controlled by one or both most effective anti-TB drugs, rifampicin and isoniazid (referred to as DR-TB going forward). Growing drug-resistance is a major hurdle to global TB control, threatening to stall or even reverse the declines in TB incidence and mortality [1].

Recent WHO figures indicate that 19% of previously-treated TB cases and 4.1% of new TB cases globally were DR-TB. This burden is unequally distributed, both spatially and economically: of the 600,000 DR-TB cases emerging in 2016, 91% originated in just 30 countries, 22 of which are classified as low or lower-middle income [2]. In most of these countries, government spending on health per capita is relatively low, and out-of-pocket expenditures exceed 15% of total healthcare costs (well over the suggested benchmark threshold for universal health coverage) [1, 3]. These and other systemic barriers - such as high costs of formal diagnosis and treatment; limited accessibility; poor management and infrastructure; and weak regulatory oversight - result in health systems that are already overstrained yet faced with a growing and increasingly complex burden of diseases, creating ideal conditions for the continued rise of drug-resistance [4]. In addition to these macro-level factors, however, the role of patients and their behavior is also critical. Individual beliefs and preferences, as well as household social and financial constraints, can lead to suboptimal risk-taking, healthcare-seeking and compliance-related behaviors, the consequences of which may be further exacerbated by weaknesses in the underlying health system; for instance, poor health literacy coupled with limited access and high prices can lead to delays in diagnosis and treatment [5–7]. Alternatively, lack of patient motivation to comply with treatment, combined with low levels of monitoring will lead to treatment failure and contribute to the emergence of DR-TB [8, 9].

Myanmar exemplifies the case of a lower-income country with a high DR-TB burden. The TB incidence rate is 361 per 100,000 population, approximately 2.5 times higher than the global average. DR-TB incidence is approximately 13,000 per year, placing Myanmar among the highest DR-TB burden countries in the world. Overall, 27% of previously treated TB cases and 5.1% of new TB cases were estimated to be DR-TB, higher than both the global and regional average [1]. Despite progress towards eliminating infectious diseases per the “Good Health and Wellbeing” portion of the Sustainable Development Goals, Myanmar’s health system is still developing. Although international attention and funding has turned towards Myanmar in recent years, low levels of human capital and government spending in healthcare (3.4% of general government spending in 2014–2015) [10] continue to constrain access, coverage and service quality. A recent systematic review in Myanmar identified a critical lack of evidence on TB in the local context to inform policy decisions. Specifically, the study highlighted a need for funders and policymakers to have better information about drivers of drug-resistance that could be the target of preventative efforts and health systems strengthening [11]. To address this gap, our present study investigates the healthcare-seeking behavior and preferences of recently diagnosed DR-TB patients.

As expected, most participants are of the Bamar ethnic group (84%), the dominant ethnic group in Myanmar, and are Buddhist (92%), the dominant religion in the country. Of the 202 participants, 41% reported living in informal housing (without official address and usually built of impermanent materials rather than permanent structures).

Drug-resistant TB is a challenge for resource-constrained settings where healthcare is provided by a range of healthcare providers in the public and private sectors. In countries such as Myanmar, in addition to low government funding [10], fragmented healthcare systems may contribute to the emergence of drug-resistance and reduce capacity to manage and diagnose drug-resistant cases [13–15]. To our knowledge, ours is the first study to quantitatively evaluate the health-seeking behavior of DR-TB patients in Myanmar, focusing on behavior with respect to the use of private healthcare providers.

We found that the vast majority of participants prefer to utilize the private sector and take prescribed medications from private providers when they start feeling unwell. Similar findings are reported in other Southeast and South Asian countries with dominant private sectors [16, 17]. It is important to highlight the heterogeneity of the provider landscape for allopathic health services in Myanmar and other regional countries; the private healthcare sector includes independently operating, usually unregulated GPs, pharmacies, and hospitals as well as international and local non-government organizations (NGOs) such as Population Services International (PSI). Private GPs also undertake dual practice which means that they are employed in public sector and see private patients outside of office hours as well [18]. Thus, the diverse entities that constitute the private sector add to its complexity. In contrast to the high use of allopathic private providers, we found limited use of traditional or complementary medicine practitioners in Yangon. Finally, regardless of which provider in the private sector participants initially choose, after a diagnosis of DR-TB in the public health sector, we found that the majority switch to using the public sector and do not continue to be treated by private healthcare providers.

A highly-fragmented pathway to care with a large de-facto role for the private sector, as found in our study, can facilitate the development of DR-TB. The quality of care received at unregulated private providers is poorly studied or documented through the routine monitoring systems that are applied in the public sector. There is some evidence indicating that diagnosis and treatment in the private sector can deviate from guidelines for appropriate patient management [19] and that poor treatment compliance in the private sector contributes to drug-resistant TB [20–22]. We do not know the length of delay to diagnosis of DR-TB in Myanmar; a long delay to initiating treatment may result in spread of infection and this should be monitored among newly diagnosed DR-TB patients.

From the perspective of patient choice, use of the private or public sector is determined by both push and pull factors. Few studies in Myanmar have looked at the motivations for initially using private as opposed to public healthcare. One study found most people went to a private GP nearby since it was more convenient and was encouraged by former patients who had been cured of the disease. Long waiting times and perceived prolonged risk to exposure in the waiting rooms were also reported as deterrents to the use of the public sector [18].

In our study, we found that women were more likely to use private providers, consistent with other research that also found that male TB patients outnumbered females in Yangon’s public health sector [23]. Evidence from other settings indicates that women may experience greater stigma from TB or challenges in travelling long distances to reach healthcare, and private providers may be perceived to allow greater patient privacy and be more conveniently located [24–26]. Further research is needed to explain other demographic differences, such as age-related disparities in seeking care from private providers.

The most likely main determinant of switching to the public sector after diagnosis with DR-TB is the high cost of future visits in the private sector, because DR-TB requires expensive medication [27, 28]. A study in Cambodia also found that patients would initially visit private healthcare providers and switch to the public sector once they knew they had TB, because they wanted to access free TB treatment which would not be available in the private sector [29]. Proximity and access to health public facilities did not appear to be a major barrier in our study in an urban setting, with 78% of the participants able to reach NTPs within 30 min.

A major strength of this study is the collection of data via detailed, face-to-face interviews with all newly-diagnosed DR-TB participants in the studied townships, which has not previously been collected in Myanmar. However, our study is limited in that we only focus on selected townships within Yangon, and so our findings could more representative of urban settings. For example, the finding that few DR-TB patients used traditional healers may differ in rural settings. Costs of seeking care are also missing from our dataset as we did not receive ethical approval to collect this information. Likewise, our study lacks data on delays to diagnosis of DR-TB, as well as the time period between first seeking care in the private healthcare sector and shifting to the public sector. Lastly, there is considerable diversity in private sector providers that we have not investigated. Differences in quality of care, referral practices and costs may differ between registered/regulated private provider and informal or untrained providers.

As funding increases for infectious disease control in Myanmar, this research provides initial evidence for policy making and resource allocation decisions [10, 11]. Specifically, our study indicates that private allopathic healthcare providers are the first point of contact when patients are unwell, and they can therefore play a major positive or negative role in healthcare provision in urban settings. Investments that can improve the quality of care, delays to diagnosis, and continuity of care within the private sector, or coordination between the public and private sector, have the potential to impact a large proportion of patients and reduce the growth in drug-resistance in Myanmar. The current study adds to the limited evidence on healthcare-seeking behavior among DR-TB patients in Asian settings with pluralistic health systems. Future studies should examine what motivates patients to use the private sector as the first point on contact when ill, and subsequently understand the quality of care for different types of providers within this sector. Longitudinal studies can help elucidate factors associated with timing of diagnosis and treatment uptake, and treatment adherence, which will support the design and funding of effective interventions in the future.