Introduction

The hepatitis B virus (HBV) is a blood-borne pathogen that infects the liver and is a cause of global morbidity and mortality with an estimated 257 million people living with chronic HBV infection in 2015 [1]. Infection can cause acute hepatitis or lead to chronic infection resulting in liver cirrhosis and hepatocellular carcinoma. Within the World Health Organization’s (WHO) Southeast Asian Region, Myanmar has the highest rates of perinatal transmission [2]. There is an estimated 16 perinatal chronic HBV infections per 1000 live births in Myanmar and an estimated 6.5% of the population in Myanmar has chronic HBV [2, 3]. There is also an estimated 3.8% of 5 year old children in Myanmar who are hepatitis B surface antigen (HBsAg) positive [2]. Further data from the Thai-Myanmar border established a HBV seroprevalence of 8.5% among pregnant women whom were primarily ethnic Karenni [4]. Given the high prevalence of HBV infection, high population health burden is likely. Cirrhosis of the liver, a long-term consequence of HBV infection was estimated to be the fifth leading cause of death in Myanmar in 2017 [5].

The prevention of mother to child transmission (PMTCT) of HBV is important as the likelihood of developing chronic hepatitis B is estimated to be as high as 90% among those infected at birth, and between 15 and 40% of those with chronic infections will progress to develop cirrhosis or hepatocellular carcinoma [6, 7]. Rates of vertical transmission vary significantly depending on the characteristics of maternal infection, ranging from 10–40% with positive maternal HBsAg and negative maternal hepatitis B e antigen (HBeAg) to as high as 70–90% if both HBsAg and HBeAg are positive [6, 7]. Although maternal screening programs and universal vaccination of newborns (birth dose) have reduced HBV vertical transmission rates worldwide, at present such programs are not consistently available in Myanmar. HBV vaccination, HBV birth dose, as well as potential use of antiviral therapy from 28 weeks gestation to 3 months postpartum is part of Myanmar’s central governmental health policy, but implementation is not complete in remote regions, and this situation is likely to worsen with the February 2021 military coup [2, 3, 8]. This may be due to the lack of governmental health facilities in rural regions and limited accessibility. Travel times to governmental facilities are on average longer than to ethnic and community based health service providers [9]. In Myanmar, birth dose HBV vaccination is mainly provided for institutional deliveries in hospitals, and was introduced in 2016 [2]. Although Myanmar has an estimated three dose HBV vaccine coverage of 90%, only 17% of the population receive a HBV vaccination series at birth in 2019, up from 7% in 2018 [10]. Additionally, only 36% of births in Myanmar occur in health facilities, creating a gap in the delivery of HBV birth dose vaccination [2].

The combination of hepatitis B immunoglobulin (HBIG), birth dose vaccine, and three dose vaccine series is estimated to reduce the vertical transmission rate to 3% [6]. Within the last few years, there are also recommendations for the use of peripartum antivirals in addition to vaccine and HBIG for decreasing vertical transmission and this is the formal recommendation of the WHO as of 2020 [11]. However, the access to WHO recommended therapy is not always possible in resource limited settings [11]. Therefore, further studies assessing feasible regimens in resource limited settings where testing and treatment are a challenge may be required. A 2015 meta-analysis showed no significant difference in HBV infection or in sero-protection markers between birth dose vaccine alone versus vaccine plus HBIG for HBsAg+/HBeAg- mothers [12]. A 2017 Chinese study showed no differences in perinatal transmission between birth dose vaccine alone versus vaccine plus HBIG [13]. However, both of these studies were conducted among women who were HBsAg + / HBeAg-. A third study in rural India, which screened only for HBsAg positivity, found that vaccine alone had an effectiveness of only 64% in a non-study setting [14].

Little is documented about the factors that may influence HBV birth dose immunization program implementation in a rural, resource limited, and culturally variable Myanmar setting. Several studies illustrate HBV immunization programs in rural locales within surrounding Asian countries, but no studies have been conducted in Myanmar. A study on HBV immunization in rural South India described implementation and immunization rates [14], while another study in rural China explore vaccination rates, vaccine awareness, and serology to measure out-of-cold-chain vaccine success [15]. Additional studies in rural Laos and Indonesia also investigated out-of-cold-chain HBV vaccination and barriers to providing the birth dose of HBV vaccine respectively [16, 17]. The present study adds to the existing literature on HBV immunization implementation in a rural setting within Asia.

Myanmar presents a unique opportunity to proactively integrate context and implementation science into program design to create culturally aware and specific sustainable interventions that may grow with the health system over time and engage all stakeholders. Using the Myanmar context and existing definitions for evaluating program implementation, this study describes a targeted HBV vaccination program for PMTCT in a rural region and assesses the feasibility of such a program in a resource limited setting.

Methods

A mixed methods analysis was completed. Based on the framework of Palinkas et al. (2011), a mixed method design in an implementation study [18] allows for evaluation of both process and outcomes. Additionally, mixed methods analysis allows exploration of the content as well as the context of the intervention [18]. As this study appears to be the first describing a HBV immunization program in a conflict-affected rural region in Myanmar, combining quantitative and qualitative data provides a more complete picture of implementation and increases the rigor of the evidence through triangulation [19].

Program description

In 2016, a pragmatic field test case study in Karenni and Shan states, Myanmar implemented a targeted HBV vaccination program for PMTCT in rural areas. The HBV vaccination program was a collaboration between three partners: the Civil Health and Development Network (CHDN) of Karenni State, Myanmar, B.K.Kee Foundation, an American non-governmental organization operating out of Yangon, Myanmar, and Queen’s University in Kingston, Canada. CHDN is a network bringing together six ethnic health organizations to provide access to quality health care services for the communities in Karenni State. Its mission is to deliver healthcare services for populations where government services are limited or nonexistent. The vaccination program was funded by B.K.Kee Foundation, and technical and financial support for the monitoring and evaluation components was provided by Queen’s University.

From January 2016 to February 2017, one thousand pregnant women were recruited sequentially and screened for HBV with the ABON Hepatitis B Surface Antigen Rapid Test Device. The relative sensitivity and specificity of these devices are 99.8% and 99.2% respectively [20]. Verbal informed consent was given by the women for both HBV screening and HBV vaccination for their newborns. Recruitment occurred in ten villages under the jurisdiction of CHDN in Karenni (nine villages) and Shan (one village) states, with a catchment population of 144,000 (Fig 1). Within this catchment area, there were an estimated 9400 women of reproductive age [21]. These villages are small, and under normal circumstances, the majority of women in these villages would be followed by Community Health Workers (CHWs) from CHDN during their pregnancies, and therefore recruitment occurred at routine antenatal visits. CHWs provide routine prenatal care, as well as screening and treatment for other illnesses, and child wellness checks among others. In many settings they are the only healthcare providers available within a reasonable travel distance. Women who screened positive for HBsAg continued to be followed by CHWs at clinic check ups and home visits until the birth of their children as would normally occur throughout their pregnancy. A three-dose vaccination schedule was implemented for newborns of HBsAg positive mothers, with the first dose scheduled to be given within the first 24 hours of birth, the second at one month, and the final dose at six months. Individualized systems were put in place to notify CHDN workers of the need for the birth dose, depending on where the woman delivered (at home, at a CHDN clinic, or in a government or private hospital). Specific recommendations around location of delivery did not depend on HBsAg status. Generally, contact was made by cellphone by a family member if the birth was unattended, or by a health worker if the birth was attended. This occasionally required travel by the person making the call due to unreliable cellphone service in certain villages. ABON HBsAg rapid tests were completed at 9 months of age to assess for HBV infection in the children.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Map of CHDN main office and clinic locations.

Map courtesy of U.S. Geological Survey.

https://doi.org/10.1371/journal.pone.0261470.g001

Vaccines were stored in and delivered from the CHDN main office due to a lack of reliable electricity and/or refrigerators in rural clinics and the presence of a backup generator in the main office. Vaccine refrigerators were monitored with twice daily temperature checks for optimal temperatures between 2 to 8°C. CHWs at the clinics contacted the main office when vaccine was required and these vaccines were then delivered to each clinic. The cold chain was maintained during transport through cold boxes adapted following WHO recommendations [22].

Training of clinic staff was completed in a two-day pre-implementation workshop, and monitoring of study progress and data collection was undertaken through field visits to each clinic one month prior to the start of recruitment and nine months into the program. Training included information about the theory behind the program (information on HBV, why vaccination would be important, etc.), standardized program rollout, and effective maintenance of data logbooks. The follow-up nine-month field visit assessed the implementation of the program through the collection of quantitative and qualitative data. All clinics were supported by the CHDN central office which also provided the day-to-day and field supervision during implementation.

Results

Sociodemographic data and participant outcomes

Demographic information for the 1000 pregnant women screened are summarized in Table 1. The median age was 27 years old with a median of 4 people living in the same household. The majority of the women were married, identified subjectively as middle income, and were farmers.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Demographics of pregnant women screened for HBsAg.

https://doi.org/10.1371/journal.pone.0261470.t001

Out of 1000 women screened for HBV, 46 tested positive for HBsAg providing an estimated HBV prevalence of 4.6% among women of childbearing age in the catchment area. Of the 46 HBV positive women, 40 had live births, 3 had stillbirths, and 3 had spontaneous abortions. Of the 40 live births, 37 (92.5%) received their first dose of HBV vaccine within the first 24 hours of birth. Two newborns (5%) received the first dose, but not within the recommended first 24 hours. One (2.5%) newborn dropped out of the program before vaccination. The remaining 39 children (97.5%) received their second and third dose of HBV vaccine on schedule. All 39 children who remained in the program tested HBsAg negative at 9 months of age.

Of the 30 births with a recorded location, 17 (56.7%) were homebirths while 12 (40%) were at a hospital and 1 (3.3%) was at a clinic. Out of the homebirths, 2 (11.8%) did not receive their first dose within 24 hours. All hospital and clinic births had timely first doses.

Characteristics of the birth setting for children born to HBV positive mothers (n = 46) are provided in Table 2.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Characteristics of birth settings.

https://doi.org/10.1371/journal.pone.0261470.t002

Pre-implementation training and community education

CHDN’s CHWs were responsible for providing opportunities for face-to-face contact with participants for patient education, counselling, testing, and vaccination. Administration of the program was based at the CHDN central office in the provincial capital of Karenni State, Loikaw. A train-the-trainer model was used with B.K.Kee and Queen’s University partners (physicians and epidemiologists) training CHWs from each of the 10 regional clinics and four members of CHDN’s central administrative team on the objectives, implementation, and monitoring of the project. These CHWs, with support from CHDN, were then responsible for program implementation and for training of the other clinic staff in their respective clinics.

In total, forty CHDN staff were trained within the first two months of the program with 39 of those being CHWs. CHWs held information and education sessions regarding HBV in their communities 78 times over the study period with a total of 1921 participants. These information sessions were held in a variety of settings including in clinics, homes, and community spaces. The median number of people attendees per session was 24 with a range of 1–99 people per session.

CFIR mapping (Table 3).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Consolidated framework for implementation research theme mapping.

https://doi.org/10.1371/journal.pone.0261470.t003

Researchers mapped implementation themes from interviews, observations, and examination of monitoring materials onto the CFIR constructs, as illustrated in Table 3. The most salient findings and themes about barriers of implementation are discussed in more depth below [25].

With respect to the initial training sessions, analysis of participant observation field notes resulted in three unique themes related to implementation barriers. The first related to an inner setting or culture CFIR domain. The idea of informed consent appeared to be a new, unclear, and unfamiliar concept to participants in the context of clinical care and as part of the research project. Asking for explicit consent was not part of normal clinic process. Participants felt that the process of obtaining informed consent would be intimidating for clients and health care providers. Physical consent sheets were provided as a tool to help guide the process, but these were also described as intimidating and CHWs preferred verbal consent to minimize patient concerns. Based on this feedback, shorter verbal consent was obtained instead of written consent.

The second theme from the training activities related to implementation barriers was in the process or planning domain. The data collection tools that had been developed prior to the training were not adequate. Due to project timing constraints, local CHDN staff had not had sufficient opportunity to inform tool development prior to training. To be functional, data collection tools required more collaborative development and a better understanding of local clinic contexts ideally with input from front-line CHWs. Tools were translated into Burmese, but the translation proved to be in language that was too complicated for participants (mostly whose first language was Kayah instead of Burmese). Tools were extensively edited during the training process with participant involvement and feedback to improve usability, specifically around a preference for close-ended categorical variables in data collection rather than open-ended free text.

A third kind of implementation barrier that emerged as a theme in participant observation field notes from the training activities was also in the process or planning domain. There were limitations in the training structure and ease of its implementation. The initial training session was mostly didactic in nature with the larger group, then breaking up into smaller groups to practice. However, it became evident that there were socio-cultural and hierarchal groups and separations between trainees impeding some participants from participating actively in training. Hierarchy was present at multiple levels with CHWs (of varying ages and levels of experience), CHDN administration staff, B.K.Kee funders, and researchers from Queen’s University all occupying potentially different places on that hierarchy. There was hesitancy in asking clarification questions by the CHWs because of the complex hierarchies involved. Language provided another layer of stratification. Training was partially in delivered in English (with simultaneous translations to Burmese) and partially delivered in Burmese. However, the local CHWs operate in a variety of dialects of Kayah, the local languages spoken in Karenni state. Although Burmese is the assumed lingua franca, comprehension of Burmese was a barrier in the training. It was unclear during the training session itself if all concepts were transferred effectively and this likely contributed to the need for ongoing support and retraining by CHDN staff during the early phases of implementation.

Discussion

Our results indicate the feasibility of a vaccination program for remote, ethnic minority communities in a resource limited setting in Myanmar. The program was successful in preventing HBV vertical transmission in HBV positive mothers with no vaccinated infants testing positive for HBsAg. This is consistent with modelling that suggests that controlled temperature chain outreach strategy for delivery of HBV birth dose vaccine can reduce the burden of HBV associated with perinatal infection [27]. This is particularly important in the context of the February 2021 military coup which has led to a disintegration of the country’s healthcare system and is likely to further marginalize ethnic minority groups in Eastern Myanmar and elsewhere.

It is interesting that the HBV prevalence in the current study was lower compared to other studies. On the Thai-Burmese border, 8.5% of women were HBsAg positive whereas in this study only 4.6% were positive [4]. Our prevalence is also lower than the 6.5% estimated for Myanmar as a whole [3]. Further research is warranted to determine if there are any systematic differences between subgroups to explain different HBV prevalence or whether the differences are due to sampling bias.

It is noteworthy that no vaccinated infants tested positive for HBV. Without vaccination, we expect 4–36 newborns to be positive for HBsAg, given a vertical transmission rate of 10–90% depending on relative HBsAg and HBeAg positivity in the mothers [6]. Importantly, this was only assessed through a highly sensitive and specific point of care HBsAg test, and not confirmed through gold standard serology, allowing for the possibility of false negatives and misestimates of initial HBV prevalence among pregnant women. The study was also completed prior to the formal recommendations for third trimester antivirals [11] therefore further study is required regarding the implementation and effectiveness of antiviral provision in this setting. Nevertheless, this study adds to the body of evidence supporting the effectiveness of HBV vaccination alone in preventing vertical transmission. The current estimates of PMTCT effectiveness are also higher than those reported for a HBV vaccination program in rural India [14].

This study’s qualitative results around implementation are a unique strength. To our knowledge, no current published studies describe the implementation factors and considerations in such a resource limited rural setting or in the context of Myanmar. A similar study describes barriers to immunization among Burmese migrant workers living in Thailand, but it does not discuss the in-country Myanmar context nor does it use an implementation framework [28]. This study adds to the literature in describing potential barriers and considerations for successful program implementation both in this specific setting as well as resource-poor settings and in conflict affected settings more generally.

CFIR has been used in the past to analyze implementation of immunization programs both in developed and developing countries [29, 30]. Similar to other immunization programs, we identified barriers in multiple domains. However, unlike other studies, the identified concerns in this setting were not consistently spread among all implementation domains. Instead, most concerns were focused on the outer setting and within the process group of constructs. There were few perceived barriers concerning the innovation itself such as the feasibility or validity of the program or an individual’s ability to implement. Interestingly, participants did not identify challenges in recruiting women to participate in the program, nor challenges in immunizing or testing neonates. While stigma around hepatitis B infection might have been a limitation, this was not identified by participants, rather they were unable to meet the demand for hepatitis B testing due to limited resources. Although a major barrier was adaptability of the innovation to the resource limited setting, the program was well accepted by all participants. The concern about logistical issues was readily overcome by CHDN. The program may have been so well accepted due to a local understanding that immunization is well within the scope of healthcare delivery for CHDN. Alternatively, acceptance may have been higher due to the positionality of the researchers in comparison to the participants. Because of the power differential, certain themes may be hidden during the research process [31].

Identified barriers emphasized the importance of process and understanding the outer setting. There were various barriers associated with the planning and execution phases of the process such as issues with adapting training materials to local needs and the monitoring and evaluation requirements of the program. These barriers could be easily mitigated through longer term development of the tools as well as early involvement of stakeholders and end users to ensure that the tools are appropriate and useful. The overall program was designed and implemented quickly over a few months and would have benefitted from a longer planning phase and training period to overcome these process barriers. Unforeseen issues such as consent not being part of normal health care could then have been identified and addressed in advance. In future partnerships, a longer planning phase and earlier engagement with the local stakeholders is recommended.

The outer setting barriers echo the complex environment of Karenni state and the importance of adapting programs to the local context. The friction between governmental and non-governmental stakeholders was not fully anticipated by project leads. Additionally, further pre-planning would have been useful to assess the needs of the community and to address unintended consequences such as increasing stress in the community about HBV without the ability to offer broader testing and treatment. A needs assessment would have been beneficial to both mitigate the unintended consequences and to set priorities in such a resource limited setting [32, 33].

The program in Karenni State provides a model for collaborative efforts in vaccination program implementation in a resource limited setting. The incorporation of the local ethnic health organization, CHDN, likely explains the success of this program. Mapping data onto the CFIR, it became apparent that the inner setting domains were sources of strength that moderated implementation barriers. There were positive structural characteristics with a combination of stability and resources in the partnership between CHDN and B.K.Kee. CHDN also had effective internal processes that fostered positive communications and readiness for implementation. It appeared that most of the barriers identified were able to be overcome due to the commitment and capacity of CHDN. Despite the complex environment, CHDN was also able to adapt to the changing circumstances and to support their staff throughout implementation. Issues such as the lack of electricity to power vaccine refrigerators and complicated relationship issues between CHDN and the government were overcome. Having local stakeholder involvement allowed for a community based participatory approach to implementation, which is associated with more effective and sustainable interventions [34].

The successful implementation of this program is also a hopeful step in this conflict-affected setting. CHDN was able to navigate the relationship between Burmese government officials and overcome this barrier to improve overall health in the region. At the time of this study, it was hoped that the ability to work parallel to the Myanmar government might provide a roadmap of peace through health in the region. Obviously, the landscape has changed dramatically since the military coup of February 2021. Nevertheless, it may be that further, potentially collaborative, efforts centred around health and led more by Burman and ethnic health workers rather than governmental agents or policy could enhance the relationship leading to beneficial effects to the local population. Reaching an understanding about superordinate health goals can be a mechanism that is further developed to achieve lasting peace [35].

Strengths of this study include the multiple modalities of capturing the implementation experience and contextual factors. The mixed methodology incorporating multiple data collection and analytical approaches allowed for increased credibility of the results and increased depth of analysis into barriers. This approach enabled analysis on both the process and the outcome, which are key to implementation success.

Limitations of this study include the lack of a control group to reliably assess vaccine effectiveness. The before-after nature of the study only allowed estimation of vaccine effectiveness based on assumptions of baseline HBV positivity. Furthermore, exact maternal HBV status with viral loads and HBeAg testing was unable to be completed due to the resource limited setting therefore clouding understanding of possible reasons of why this regimen was effective. Additionally, monitoring was only done at discrete time points by the study team without continuous embedment during the study period. With already existing issues in data management, data collection inaccuracies could have affected the results. Language and cultural barriers were a significant issue in interpretation of the findings. However, the incorporation of multiple perspectives mitigated potential problems through triangulation of the themes.

Conclusion

Taken together, the results show that a complicated vaccination program is feasible and effective in a resource limited rural region in Myanmar without government health system involvement. Despite the complexities of the setting, which have only increased recently barriers were able to be overcome. Ethnic health organizations have demonstrated their innovation and creativity for decades in the context of waxing and waning conflict with the central government⁹. This vaccination campaign was successful in preventing HBV vertical transmission, with no infants testing positive for HBsAg. Further longitudinal research is needed to assess the sustainability of this intervention and its potential expansion to other immunizing agents.

Acknowledgments

We wish to acknowledge the courageous workers of the Civil Health and Development Network who implemented this project, and who continue to work tirelessly under very difficult circumstances to provide healthcare to the people of Karenni State. Likewise, we wish to acknowledge B.K.Kee Foundation, without whose funding and support the original plan for Hepatitis B screening in pregnancy and immunization of neonates would not have been possible.