Greatly intensified efforts to control and prevent malaria on a global scale have resulted in substantial reductions in malaria-associated morbidity and mortality, with global malaria deaths decreasing by 47%, and incidence by 30% from 2000 to 2013, according to World Health Organization (WHO) estimates (1, 2). Aggressive control efforts have permitted many countries to begin the transition to elimination activities (1, 3), which specifically seek to interrupt local transmission. With increasingly finer focus, programs concentrate their attentions from the national scale to the residual foci of active transmission, and finally to individual cases (4, 5).

The WHO has suggested potential epidemiologic milestones for these transitions (3, 6). The move from a control program to a pre-elimination program involves the demonstration of a blood slide positivity rate (SPR) of less than 5% for suspected malaria cases. Subsequent transition to the elimination phase requires reaching an annual parasite index (API) of less than one case per thousand population at risk per year (4). Finally, WHO certification of malaria elimination requires a country to prove the absence of any locally acquired infections for a minimum of three consecutive years (4, 6). Once this is achieved, continued vigilance is required to prevent malaria resurgence (4, 7).

The small Himalayan nation of Bhutan (population 720,629 in 2012) has had malaria control programs in place since 1964 (8). Success of these programs has been measurable, as evidenced by dramatic decreases in malaria cases from nearly 40,000 cases in 1994 to 45 cases in 2013 (9). Distribution of malaria transmission in Bhutan has been limited by its topography and climate (10). Four northern districts experience essentially no malaria transmission as a result of their elevation, and any locally identified malaria cases are likely imported from outside districts or neighboring countries. Nine districts experience seasonal transmission during the summer months, while perennial transmission occurs in the seven southern districts bordering India (11). In 2011–2013, retrospective case mapping using Geographic Information System (GIS) was carried out, and districts have been reclassified as moderate, low, potential, and no transmission zones (8). Nationally, most cases occur in August and September, following the rainy season, with a secondary peak in April.

Two-thirds of all cases occur among males, and farmers are definitively the occupational group most at risk, followed by students and laborers (8). Bhutan’s proximity to India remains a major challenge to control as porous borders, and a large number of foreign workers are constant sources for new infections. Bhutan is surrounded by states with some of the highest SPRs for P. falciparum in India, and many of the foreign laborers come from the highly prevalent state of Orissa (1, 12).

The Plasmodium species endemic to Bhutan are P. falciparum and P. vivax (11, 13). Hypnozoites of P. vivax represent a major challenge to elimination efforts, functioning as an asymptomatic reservoir of infection, which require additional specific therapies, namely primaquine, to eradicate these forms (14). Primaquine also targets the gametocyte stages of P. falciparum, making it a critical tool for reducing transmission, but its use is challenged by hemolytic anemia occurring in individuals with glucose-6-phosphate dehydrogenase (G6PD) enzyme deficiency (14, 15).

All vector-borne disease-related activities in Bhutan – including surveillance, prevention, and outbreak control – are consolidated under the Vector-borne Disease Control Program (VDCP), a division of the Department of Public Health, Ministry of Health of Bhutan. Since surveillance was initiated in 1965, reported cases reached a peak in 1994 of 39,852 malaria cases (13). Since then, the VDCP launched an aggressive malaria control strategy, strengthening and expanding earlier efforts. Following those efforts, malaria incidence in Bhutan has drastically reduced. Malaria dropped from the 16th leading cause of years of life lost in 1990 to the 46th in 2010 (16). Bhutan is in the elimination phase and is progressing well over the coming years to be certified by the WHO as malaria free (2).

Elimination rests on a country’s ability to interrupt transmission of malaria within its borders. Surveillance is lauded as a keystone for malaria elimination, and it must be tailored to the specific epidemiology, public health capacity, and disease control strategy of a given country. In the case of Bhutan, the clinical protocols and vector control activities of the VDCP have been reported previously (11, 13). The Centers for Disease Control and Prevention (CDC) has published guidelines on how to evaluate public health surveillance systems and includes assessments of data quality, timeliness in reporting, system simplicity, and data quality, among other measures (17). Using this framework, we sought to evaluate the Bhutanese surveillance system as it strives to meet the objectives of the VDCP’s 2012–2016 malaria strategic plan: (1) to reduce the number of malaria-related deaths to 0 by 2016 (shifted to 2018 in the 2015–2012 National Strategic Plan), (2) to achieve 0 local malaria transmission by 2016, and (3) to obtain WHO malaria-free certification by 2020 (17, 18).

An evaluation of the Bhutan’s national malaria surveillance system from 2006 to 2012 was conducted using the CDC guidelines for evaluating public health surveillance systems (17). The guidelines focus on demonstrating that a surveillance system provides information that enhances public health decisions and describing system usefulness and attributes, including simplicity, flexibility, data quality, sensitivity, representativeness, timeliness, and stability. This report is structured based on these guidelines. Data were collected and analyzed in 2012 and 2013.

National malaria surveillance data from 2006 to 2012 were chosen for evaluation because these years align with the transition to the use of LLITNs and electronic record-keeping on the part of the VDCP and to achieve a finer focus on the pre-elimination years. Data were analyzed to determine numbers of blood slides collected and numbers of malaria cases by Plasmodium sp. and by district for each year during the study period. The evaluation included a search of PubMed, Google, and Google Scholar to identify studies pertaining to malaria in Bhutan. VDCP program data including surveillance reports, standard operating procedures, national documents were reviewed, and interviews with health workers, malaria technicians, district health officials, and VDCP personnel were conducted.

Finally, blood slide cross-check data for 2011 were analyzed including calculating the proportions meeting standards of preparation and percent agreement based on the results of the initial health center slide reading and the results from the VDCP cross-check. All data were analyzed using SAS 9.3 (SAS Institute Inc., NC, USA).

We aimed to assess the malaria surveillance system in Bhutan from 2006 to 2012 using CDC guidelines for the evaluation of public health surveillance systems. It should be noted that since our assessment, many of the recommendations described herein have been either enacted or planned as described in Bhutan’s updated 2015 to 2020 National Strategic Plan (8). Notably, a recent external review was conducted in collaboration with the WHO, and recommendations from this review are outlined in the strategic plan. Overall, from our assessment, the malaria surveillance system of Bhutan appeared strong and produced data that was useful and of good quality. The performance of the system has been critical to the major reductions to malaria morbidity and mortality in Bhutan. However, elimination will require the system to function at an even higher level, and measures are underway to meet this challenge (Table 3).

Our assessment of blood smear cross-check data from 2011 revealed a deficiency in diagnostic accuracy. This will require the addition of measures of reader agreement, additional training for local health workers, more rigorous attention to individual worker performance, and remedial measures when required. The predictive power of negative results is of particular significance because of consequences to an individual and as a source for future infections. The existence of a quality assurance component is a great strength of the structure of this surveillance system, but it is only of value if the results are utilized. Health worker accountability should be improved. While health centers that have repeated errors identified through the cross-check process are invited to send health workers to refresher trainings at the VDCP, workers are often not sent. This process should be formalized and made to be a requirement with set expectations for performance (29). The WHO and others recommend a system of on-going training and assessment of microscopists for competency in slide preparation and reading of both parasite presence and species, coupled with a clear protocol for remedial measures if expectations are not met (29, 34). District health centers must be encouraged to continue to send in the required number of blood slides as the blood slide cross-checking process is the primary means by which the VDCP can assess the sensitivity and specificity of diagnoses. Plans for the establishment of a National Quality Assurance system for malaria diagnosis should assist in making these adjustments (8). In addition, as new private laboratories are emerging in Bhutan, the VDCP should ensure that similar proficiency testing occurs in these centers as well.

Second, more sophisticated data management and analyses were required. The malaria surveillance system in Bhutan collected more information than was utilized, and harnessing this information more effectively would have aided in pro-elimination efforts. At a national level, training of personnel in statistics and epidemiology represented a significant challenge. As this is being addressed, data such as slide-reader agreement should be added to standard analyses. In addition, SPR calculated monthly at BHUs (instead of only at the year-end) may help to better understand the local transmission as compared to the raw case counts. These calculations can provide a way to compare districts and the national average, to assess progress over time, and to meaningfully communicate with outside entities concerned with public health.

Third, timeliness of identifying and treating cases is an area for improvement. The WHO has suggested using the proportion of P. falciparum infections with gametocytes present as a timeliness indicator, as treatment of P. falciparum infections within the first 6 days from the onset of symptoms is felt to be sufficient to prevent the development of gametocytes (6). The VDCP therefore may consider incorporating this indicator into its protocols. Ideally, the presence of gametocytes and resulting proportion should be recorded and calculated at the point of care.

In addition, to improve timeliness of reporting, significant enhancements will be needed to meet the WHO goal of “immediate” reporting of cases to national malaria programs (32). In the short-term, phone-based reporting, which is already regularly practiced with success in the southern districts, could be expanded to the seasonal transmission districts, potentially as a transient feature activated during the high case season. In the mid- to long-term, electronic reporting will enhance timeliness, but improved funding and training to expand internet access, acquire computers, and attain necessary technical expertise levels is required. With the advent of GIS case mapping, alongside the growth of mobile and internet capacity, significant improvements in timeliness is expected. Several countries have established SMS reporting systems with much success, including Sri Lanka in 2009 (35–38).

Fourth, from 2006 to 2012, the program functioned primarily in a passive detection capacity, but programmatic changes toward both reactive and active case detection are now underway (8). In several other countries, reactive case detection has successfully been used to focally increase sensitivity by requiring active blood screens of household members and other residents within a defined radius of any newly identified malaria case (35, 39). In addition, since 2012, case definitions have been modified to appropriately designate cases as imported or locally acquired, data which are critical to achieve and sustain elimination (26, 40). The addition of genetic analysis of parasites to confirm whether infections are locally acquired or imported would not seem to substantially challenge the structure of the current system (41). With such changes, the surveillance system will have to tighten its coverage, as reintroductions from porous borders and migrant workers will be an ongoing major threat to both achieving and sustaining elimination. The results of these case investigations then must be mined at the program level to attack foci of transmission and to implement targeted measures addressing populations identified as having the greatest risk, due to geographic location or occupational hazard (1, 27). Importantly, the use of spatial decision support systems, including GIS, is now underway, and will aid the in identifying hotspots and hotpops, and the deployment of targeted intervention strategies. In addition, targeted active screenings of workers at hydro-power project sites is underway. While this is an important step, care must be taken to ensure that workers do not become averse to seeking care for fear of loss of work privileges. In addition, innovative means of exposure reduction will have to be developed to prevent infection in other at-risk groups such as farmers, who are not protected by the primary vector control activities of IRS and LLIN use. In this case, the malaria surveillance system has been successful in identifying an at-risk population, but this information has not been translated into public health action.

The Roll Back Malaria initiative suggests that countries with highly malarious borders may choose to postpone elimination efforts until neighboring countries are able to control their high incidence levels (3) Delaying elimination efforts until control is achieved in India would mean a substantial delay. However, in the absence of effective surveillance and health service delivery in Assam, this state and its neighbors will continue to act as a source for re-introduction of infections in southern Bhutan (42). Cross-border collaborations have been limited, but they will have to be a priority for both countries to see case reductions in this area. Available funding for such collaborations, however, is limited (43). The Asia Pacific Malaria Elimination Network raises awareness of the collective progress of and challenges for elimination in the region acts as a catalyst to achieve consensus on important elimination issues, such as mobility and migration (9).

Finally, and more broadly, the successes of the VDCP and the malaria surveillance system can be leveraged in the control of emerging vector-borne diseases in Bhutan such as dengue, leishmaniasis, and chikungunya. Integrated vector management strategies can be used to progress toward elimination even as the spread of these other diseases is controlled, and the surveillance apparatus for malaria could be expanded to fill the present void of systematic surveillance of other vector-borne diseases.

Our evaluation had several limitations. The analysis of blood slide quality data could have been strengthened by an expert review of a random sample of all blood slides (both those sent in to the VDCP and those collected at BHUs but not sent in) as well as a direct assessment slide reading proficiency of local health workers and VDCP officials. Nevertheless, we believe that our assessment of the design and function of the surveillance system is well-founded despite these potential limitations.

Bhutan has seen major reductions in malaria morbidity and mortality since the 1994 peak of cases, and a substantial dwindling from 2006 to 2012, and major advances evident in 2013 data. Surveillance has played an important role in the ability of the VDCP to provide high quality, targeted control services to the foci of transmission during this period. Bhutan is continuing to develop and expand its surveillance system, and the addition of targeted active and reactive case detection, along with GIS data will be beneficial. The malaria elimination community must ensure that Bhutan does not become a victim of its own success, and that donors who have provided substantial support in the past do not turn their attention to other countries. Studies of malaria resurgence have found that nearly all instances of resurgence are due to the slackening of control efforts, which are themselves primarily due to insufficient resources (7, 40). In the absence of any external reduction in regional transmission potential, the activities used to achieve malaria control or elimination must be maintained (44). Despite such challenges and the potential of re-introductions of infections, elimination in Bhutan in the next few years appears possible with the continued responsiveness and adaptability of its surveillance system that has been demonstrated to date.

NW, SG, SD, ST, and KK conceived of the study. NW, SG, SD, AD, and SP obtained and analyzed the data. NW, KK, SD, and AD wrote the manuscript.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.