The burden of Japanese encephalitis, the catch-up vaccination campaign, and health service providers’ perceptions in Myanmar: 2012–2017

Background Myanmar is endemic for Japanese encephalitis (JE) and has experienced several outbreaks in recent years. The vector-borne disease control (VBDC) program has collected hospital-based surveillance data since 1974. There is an urgent need to collate, analyze, and interpret the most recent information. The study aimed to describe (i) hospital-based JE cases and deaths between 2012 and 2017, (ii) a catch-up vaccination campaign in children in 2017, and (iii) health service provider perceptions about JE in one township in 2018. Methods This was a cross-sectional study of cases, deaths, and catch-up childhood vaccinations using secondary data from program records and a survey database of health service provider perceptions. Results Between 2012 and 2017, there were 872 JE cases and 79 deaths with a case fatality rate of 91 per 1000; 2016 was the year with most cases and deaths. Most cases (n = 324) and deaths (n = 37) occurred in children aged 5–9 years. Large case numbers were reported in delta and lowland regions (n = 550) and during the wet season (n = 580). The highest case fatality rates were observed in the hills and coastal regions (120 and 112 per 1000, respectively). Nationwide coverage of the catch-up JE vaccination campaign among 13.7 million eligible children was 92%, with coverage lower in the hills and coastal regions (84%) compared with delta and lowland regions and plains (94%). More vaccinations (65%) occurred through school-based campaigns with the remainder (35%) vaccinated through community-based campaigns. Structured interviews in one township showed that service providers (n = 47) had good perceptions about various aspects of JE, although perceived benefits of specific vector control measures were poor: spraying/fumigation (38%), garbage removal (36%), larvicide use (36%), and drainage of standing/stagnant water (32%). Conclusion The catch-up vaccination campaign was a successful response to high JE case numbers and deaths in children. However, ongoing surveillance for JE needs to continue and be strengthened to ensure comprehensive reporting of all cases, more knowledge is needed on disability in JE survivors, and all attempts must be made to ensure high percentage coverage of vaccination through routine and catch-up campaigns.

reported 404 cases of JE from 14 states/regions and Nay Pyi Taw Council Territory. Despite the growing importance of JE in Myanmar, the attention of service providers to provide an adequate coordinated response in a "one health approach to reduce zoonotic epidemic threats" has been limited. There have been gaps in public-private partnerships and poor coordination between health authorities, veterinarians, and the public in terms of education campaigns, vector control measures, vaccination of at-risk populations, and epidemic preparedness and response.
To address these gaps, in 2017, a nationwide JE vaccination catch-up campaign was started for schools and the community. Data were also collected from selected service providers about their knowledge, attitudes, and perceptions towards the disease and its prevention. The published and reported data on JE in Myanmar is local and fragmented [8,9,11], and there is a great need to have a comprehensive situational analysis of countrywide hospital-based cases over a 6-year period. This data, combined with information about the recent 2017 catch-up vaccination campaign and perspectives from service providers, will be valuable for the vector-borne disease control (VBDC) program and the Expanded Programme on Immunization (EPI) under the Department of Public Health, Ministry of Health and Sports, so that they can develop a new coordinated strategic plan to successfully reduce JE transmission in the country.
The aim of this study, therefore, was to describe the burden of hospital-based Japanese encephalitis, the vaccination campaign, and health service provider perceptions about the disease and its prevention in Myanmar. Specific objectives were to describe (i) the hospital-based cases with JE in Myanmar, associated characteristics, and deaths over a 6-year period from 2012 to 2017; (ii) the number of children vaccinated in the JE vaccination catch-up campaign in 2017; and (iii) health service provider perceptions about JE, its transmission, and prevention in one township, Bago Region in 2018.

Study design
This was a cross-sectional descriptive study using secondary data from program records and one survey database. Records of JE cases are kept at the hospital, and a focal person from the VBDC team collects these on a weekly basis through fax, electronic mail, and phone calls. During JE outbreaks, the collection of reports is daily rather than weekly. These case records are collated by focal persons of the VBDC team who transfer the data upwards from township to district to state/region and finally to the Central VBDC Programme at the Department of Public Health. The Central VBDC Programme reports weekly to the Ministry of Health and Sports, and from there, reports are sent to WHO SEARO through the WHO Country Office of Myanmar.

JE vaccination campaign
In 2017, the EPI undertook a nationwide JE vaccination catch-up campaign supported by funds from the Global Alliance for Vaccines and Immunizations (GAVI). The purpose of this was to provide high immunity to JE before JE vaccines were incorporated onto the national routine immunization schedule in 2018. Technical support was provided by WHO, the United Nations Children's Fund (UNICEF), and the Programme for Appropriate Technology in Health (PATH). Nearly 14 million children (aged 9 months to 15 years) were targeted using the WHO prequalified live attenuated (SA 14-14-2) JE vaccine, which is the same one to be used in the routine immunization schedule. The campaign was launched in two phases: a school-based phase targeting 5-15-year-old students between 15 and 23 November 2017 and a community-based phase targeting children aged 9 months to 5 years (out-ofschool) between 11 and 20 December 2017 [19].

Service provider perceptions in Letpadan Township, Bago Region
Bago Region includes 28 townships and about 4.8 million inhabitants. There are three distinct seasons in the region that are hot (March to May), wet (May to October), and cool (November to February). Between 2012 and 2017, there were 121 reported cases of JE and eight associated deaths. Letpadan Township was one of the five townships in Bago Region with a high number of cases and deaths [20]. In January 2018, a team from the Department of Medical Research conducted structured interviews in Letpadan Township following the JE vaccination catch-up campaigns. These interviews were carried out among service providers which included VBDC staff, health assistants, lady health visitors, public health supervisors, and midwives. The selection of participants for interviews was purposive so as to have comprehensive representation at the township level. Ethics approval for this particular survey was obtained from the Ethics Review Committee of the Department of Medical Research (Ethics/DMR/2018/008).

Study population
The study included (i) all patients diagnosed with JE in hospitals in Myanmar between 2012 and 2017, (ii) children vaccinated during the catch-up campaign in 2017 in both the school-based phase and community-based phase, and (iii) health service providers who were interviewed in the public sector in Letpadan Township, Bago Region.

Variables, data sources, and data collection
Data variables for the study included the following: for hospital-recorded cases of JE and deaths-year, month, age group, gender, urban/rural residence, season, and ecological area of the country for each case. The definition of ecological areas was adapted from Oo et al. [21] and is shown in Table 1. The source of data was the VBDC electronic EXCEL database for national aggregate data; for vaccinated children-ecological region, children eligible for vaccination, children vaccinated, children vaccinated during the school-based campaigns, and children vaccinated during the community-based campaigns. The source of data was the electronic EXCEL database for aggregate data obtained from the 2017 EPI records and donor reports; for health service provider perceptions-these were structured wide-ranging questions related to perceptions, attitudes, and opinions about JE and its prevention. The data source was secondary data from the survey database of the DMR that focused on Letpadan Township, Bago Region.

Analysis and statistics
Data were extracted and cleaned in an EXCEL file and exported to EpiData software (version 2.2.2.182 for analysis, EpiData Association, Odense, Denmark). A descriptive analysis was performed on JE cases, JE deaths, and results of the vaccination campaign using frequencies and proportions. Hospital-based JE case fatality rates were calculated by the number of deaths per 1000 JE cases per year. Cases and deaths were analyzed in relation to certain characteristics such as age group, gender, urban/rural residence, season, and ecological region. Maps of JE cases and deaths at state and regional levels were generated through a geographical information system using QGIS [22]. Perceptions, attitudes, and opinions of health service providers were analyzed by frequencies and percentages.

Hospital-recorded cases of JE and JE deaths
The  Table 2. Approximately, 85% of all cases were in children, with those aged 5-9 years being the predominant age group affected. The highest case fatality rates were also observed in the age group 5-9 years. Males were more likely to have JE compared to females, but case fatality was higher in females. Those residing in rural settings were at higher risk of JE and death compared with urban settings. The majority of cases and deaths, and the highest case fatality rates, occurred in the wet season from May to September. Most cases (63%) and deaths (53%) occurred in persons living in the delta and lowland areas although case fatality was highest in those residing in the hills. Figure 3 is a map that shows JE cases and deaths in relation to the four ecological regions of the country. The five main hot spots for JE cases were Yangon, Ayeyawady, and Bago regions followed by Rakhine and Shan states. The main hot spots for JE deaths were Rakhine and Shan states followed by Yangon, Ayeyawady, and Bago regions.

Childhood vaccination campaign for 2017
Numbers and proportions of children vaccinated against JE during school-based and community-based catch-up campaigns in 2017 for the whole country and by ecological region are shown in Table 3. Nationwide coverage of JE vaccination among eligible children was 92%. Vaccination coverage was lower in the hills and coastal regions (84%) compared to delta and lowland regions and the plains (94%). Nearly two thirds of the children were vaccinated through school-based campaigns with the remainder vaccinated through community-based campaigns. These ratios of school-based to community-based campaigns were similar in the delta and lowlands, plains, and coastal regions, while in the hilly regions, school-based campaigns were fewer and constituted 58% of vaccination coverage.

Perceptions, attitudes, and opinions of health service providers
Perceptions, attitudes, and opinions of 47 health service providers about JE infection in Letpadan Township, Bago Region, in January 2018, are shown in Table 4. Although 94% of health service providers accepted JE as a serious illness, their perceived risk about contracting JE and their increased risk of being a resident in the particular locality were just over 50% (Table 5). Over 90% of the health service providers showed a positive attitude towards the prevention of JE through compliance with VBDC guidelines. However, less than 40% believed that specific measures to prevent JE would be useful such as spraying to reduce the adult mosquito population, removing garbage dumps, using larvicides, and improving drainage. Over 90% of health service providers believed that epidemic surveillance of JE in humans and pigs and removal of the JE vectors (mosquitoes) would reduce JE transmission. Three quarters believed that JE vaccination should be given to children aged 9 months to 15 years. While less than 40% stated the need for stakeholders to be updated on JE, there was strong support for villagers to be educated about the disease through either community or school-based education sessions.

Discussion
This is the first study from Myanmar assessing the national burden and characteristics of hospital-reported JE cases and deaths over a 6-year period, documenting the results of the 2017 catch-up vaccination campaign in children and describing health service providers' perceptions of JE and its prevention. There were a number of important findings.
There were nearly 900 reported cases of JE with a case fatality rate of 91 per 1000. The majority of cases and deaths occurred in children, particularly those aged 5-9 years. The year 2016 was the worst for absolute numbers of cases and deaths, which led to the catch-up vaccination campaign in 2017. This was an ecological study, and we therefore did not capture any information about morbidity after recovery from the infection. Previous reviews point to significant neurological sequelae being present in up to 50% of survivors [3,23], so while this disease is less common in the country than dengue [21], it may nevertheless be very demanding on health service care and resources. While care is currently supportive, there is an ongoing work exploring the prospects for direct-acting antiviral treatment, especially with the use of nucleoside and nucleotide reverse transcriptase inhibitors [24,25]. Two studies in India have evaluated the use of minocycline, and while there was some evidence of reduced symptom duration and hospital stay, there were no survival benefits [26,27]. Finally, there has been some preliminary work on the use of intravenous immunoglobulin in Nepal [28], again without proof yet of survival benefit. The finding that JE cases were more frequent in the wet season and in delta and lowland regions (with paddy fields and areas prone to flooding) aligns with current knowledge about transmission of JE in Asian countries [29,30]. There are seasonal correlations in rural areas between mosquito abundance, JE virus seroconversion in pigs, and concurrent human JE outbreaks [29]. In the wet season, the widespread and timely implementation of mosquito larval control measures and space spraying of townships and wards with insecticides may help to contain mosquito proliferation and reduce human outbreaks. However, a recent report of increasing resistance of two important JE vectors (Culex tritaeniorhymchus and Culex vishnui) to DDT, malathion, and deltamethrin in India is of concern to the region [31]. Case fatality was higher in the hills, similar to what was found 2 years ago with dengue [21]. This may be because health facilities are not as well-resourced in these areas, health care workers are less experienced in providing supportive care for encephalitis, and there is poor immunity of the population to the JE virus.
It was encouraging to see the high uptake of JE vaccination in the 2017 catch-up campaign, especially in the delta and lowland areas where the burden of JE is highest. The SA 14-14-2 live attenuated JE vaccine, whether used on routine or campaign basis, has shown excellent and durable efficacy in China, Nepal, and South Korea [32], although results have been less impressive in India maybe as a result of cross-contamination from nonvaccinated to vaccinated districts [33]. Good surveillance will be needed in Myanmar to determine whether morbidity and mortality from JE are reduced through this strategy of a one-off catch-up vaccination campaign followed by routine immunization or whether further catch-up campaigns may be necessary, as is the case in Nepal [34].
Finally, the structured interviews in the township showed that service providers had a reasonable degree of knowledge about the serious nature of JE, the need for epidemic surveillance in humans and pigs, the importance of JE vaccination of children and about the value of regularly updated education of rural populations about JE. Knowledge related to specific vector control measures was poor, and this needs to be improved. There  have been no recent published studies on knowledge, attitude, or practice of service providers in relation to JE in Myanmar with which to compare these findings. The strengths of the study were the hospital-recorded cases and deaths of JE from all health facilities in all states and regions of the country over a 6-year period, making the findings representative of what is happening in the country. The conduct and reporting of this observational study were also in line with internationally recommended Strengthening the Reporting of Observational Studies in Epidemiology (STROBE guidelines) [35].
However, there were a number of limitations. First, we do not know how well or comprehensively all children in the country with acute encephalitis syndrome (AES) were investigated for JEV. In Mandalay hospital, four of 123 children presenting with acute encephalitis had JEV diagnosed in their CSF by neutralization tests and/or virus isolation [11], but if the tests are not performed on all children presenting with AES, the diagnosis will be missed. While it is likely that children with AES are admitted to the hospital given the serious nature of this condition, cases may develop quickly and die in the community. Our hospitalbased recording system may therefore have underestimated the true burden of JE in the country. Second, we also had no information about the clinical profile of JE patients, the time that they spent in hospitals and among the survivors the degree of disability and dysfunction, so the burden of disease on the health care services is not known. Third, while the vaccination uptake among eligible children was high, we do not know the reasons why some children, particularly in the hills and coastal areas, did not receive the vaccination. This is important because widespread and complete vaccination coverage is needed to control this disease [23].
Despite these limitations, there are a number of programmatic implications. First, ongoing surveillance of hospital-based cases and deaths must continue and be strengthened with oversight to ensure that all AES cases are fully investigated for JE. Second, Myanmar should undertake a prospective evaluation of the neurological Table 5 Health service providers' perceptions, attitudes, and opinions about JE infection and its prevention in Letpadan Township, Bago Region, Myanmar, in January 2018 sequelae and disability resulting from JE and use this locally generated evidence to better inform the health care services and the general public about the seriousness of the disease and the need for its prevention. In this regard, the country could consider whether to use a new outcome score based on a 15-item questionnaire which has been successfully developed and tested by the WHO to assess the severity of disability resulting from JE in children and whether the child is likely to be dependent [36].
Third, every attempt must be made to ensure that the SA 14-14-2 live attenuated JE vaccine is fully incorporated into childhood routine immunization schedules, with ongoing surveillance needed to determine whether additional catch-up vaccination campaigns are needed, as is the case in Nepal [34]. Ultimately, JE is a vaccinepreventable disease, and high percentage coverage with routine as well as catch-up campaigns is the intervention most likely to control and even eliminate the disease. Unfortunately, there has been a gradual shift in JEV genotypes in Asia over the years, especially from genotype III to genotype I [37]. Almost all currently available vaccines are developed against the genotype III strains of JE, so this genotypic shift raises concerns about vaccine efficacy. Studies have shown that these vaccines are effective against genotype I [38], but less so against genotype V [39]. In light of these findings, Myanmar needs to decide how much effort and investment is also needed to vaccinate domestic pigs and undertake vector control. This would be in line with a coordinated "one health approach." JEV in adult pigs does not result in symptomatic disease, but pigs can suffer from significant reproductive problems that include abortion, still-birth, and birth defects [40]. Protective vaccines are available for both pigs and horses [40].

Conclusion
This study has described the burden of hospital-recorded JE in Myanmar between 2012 and 2017. The disease was concentrated in children and seen particularly in the delta and lowland regions and during the wet season. Overall case fatality was 91 per 1000, with high rates observed in children aged 5-9 years and people residing in hills and coastal regions. The catch-up JE vaccination campaign through school-based and community-based phases was successful with an overall 92% coverage although this was lower in the hills and coastal regions. Structured interviews in one township showed that service providers had a reasonable degree of knowledge about various aspects of JE, although this was poor with respect to specific vector control measures. For the future, ongoing surveillance for JE needs to continue and be strengthened, more knowledge is needed on disability in JE survivors, and all attempts must be made to ensure high percentage coverage of vaccination through routine and catch-up campaigns.
Abbreviations CFR: Case fatality rate; CSF: Cerebrospinal fluid; EPI: Expanded Program on Immunization; JE: Japanese encephalitis; VBDC: Vector-borne disease control; WHO: World Health Organization