The desk research identified 112 publications related to malaria control and elimination in Sri Lanka. Roughly a quarter (26) of these publications are studies on vector control in Sri Lanka. There were 72 grey documents identified and reviewed, 56 of which were reports from the Anti-Malaria Campaign, the Sri Lanka Ministry of Health, or reports written by consultants about a project implemented by either organization. There were 16 documents reviewed from the World Health Organization. A list of these documents is available in Appendix S1
The Anti-Malaria Campaign (AMC) Directorate in Colombo guides and coordinates all malaria control activities (see ). Under the purview of the AMC is formulation of national malaria control policy, monitoring national malaria trends, technical guidance to subnational malaria control programmes, inter-district coordination, and coordination of training and research activities. Entomological and parasitological surveillance is also undertaken by the AMC. Decentralization in 1989 shifted the administration of malaria control activities to the districts. Health services are managed by the Regional Director of Health Services (RDHS) and responsibility for malaria control activities rests with the Regional Malaria Officer (RMO) in each district. RMOs work jointly with the Medical Officers of Health (MOHs), whose offices provide varying levels of support for vector control activities.
Organizational diagram of the Sri Lanka Anti-Malaria Campaign.
Challenging and enabling factors
Sri Lanka's long-running civil conflict affected the whole country but was concentrated in an estimated eight districts 
. A ceasefire agreement was declared in 2002, and there was a decline in civilian casualties around this time: a decline in “battle-related deaths” was reported from 4,000 (2000) to 1,000 (2001), then down to zero in 2002 and 2004 
. The ceasefire also linked the Jaffna Region in the north to the rest of Sri Lanka through regular commercial passenger flights and the reopening of route A9, a major artery for transportation between Jaffna and Colombo 
. As a result, delivery of supplies to the northern areas may have increased. However, for the purposes of this case study, the number of districts considered to have conflict during these years remained at eight because reports indicate that the ceasefire was not respected by both sides at all times and sporadic fighting continued 
The ceasefire officially ended in 2006 when violence resumed in the northeast 
. The same eight districts are considered to have had active conflict from 2005 to 2007, decreasing to six in 2008 and to four in 2009. Over this period deaths related to the conflict rose, peaking in 2008 (11,144) 
. By May 2009 the war was declared over and by December route A9 was again open.
Socio-economics, health and environmental factors
National gross domestic product (GDP) per capita increased from $715 (current USD) in 1995 to $2,375 in 2010 
. Total health expenditure per capita (current USD) also rose from $26 in 1995 to $84 in 2009. The adult literacy rate in Sri Lanka was estimated to be 91% both in 2001 and in 2006. 76.6% of the population was reported to have access to electricity across the country in 2009 
Historically, transmission of malaria has increased in Sri Lanka when pooling occurs in rivers and streams, which is conducive for the breeding of the primary vector, Anopheles culicifacies.
Transmission increases with monsoon rain events in the dry zone. Transmission may also increase when the monsoon is weak or does not occur in the intermediate zone 
. The literature did not report any major droughts, flooding, or shifts in vector breeding during the period 1995–2011. However, the World Bank Health Services Project reported that a drought in 2001 may have contributed to project-area declines in malaria incidence 
. Studies have had reasonable success in linking rainfall to malaria incidence, when using a two to three month period for forecasting 
The tsunami of December 2004 led to a massive loss of life and displacement of 860,000 people. Hospitals and administrative buildings were destroyed 
. However, reports indicate that surveillance and prevention activities for malaria were maintained by local health authorities and NGOs and no malaria epidemic accompanied the tsunami (2004–2005) 
Epidemiology of malaria and vectors
From 1995 to 1999, the number of malaria infections confirmed by microscopy or rapid diagnostic tests (RDTs), or confirmed infections, rose from 142,294 to 264,549 (). Then, beginning in 2000, cases began to decline. From 1999 to 2007, cases were reduced from 264,549 to 198 (99.9%). There was a small uptick in total cases, combined indigenous and imported, to 670 in 2008 through to 2010 (736). In 2011, 175 cases were confirmed, of which 124 were indigenous, meaning that the infection originated in Sri Lanka.
Total confirmed infections from Active and Passive Case Detection, Sri Lanka, 1995 to 2011.
Malaria-related mortality in Sri Lanka has similarly declined, from a peak of 115 deaths in 1998 to zero indigenous deaths each year since 2008. In years 2009 and 2011 there was one death, each an imported case from Nigeria.
Annual parasite incidence (API), or the number of confirmed infections of all Plasmodium species divided by the estimated population at risk, was 11.9 per 1,000 in 1995, reached a peak of 22.1 in 1999, then declined to less than 1 by 2004 (0.9). In 2010 the estimated API of indigenous cases was 0.1. The slide positivity rate (SPR), or the proportion of slides found positive for indigenous cases for any Plasmodium parasites among the slides collected 
, was 13.0% in 1995, peaked in 1999 at 16.7%, then declined starting in 2000 (11.8%) to 0.2% in 2005. In 2010 the SPR was 0.1%. Studies employing polymerase chain reaction (PCR) assay have found no evidence of sub-microscopic parasitaemia in previously endemic portions of the country 
As national malaria morbidity declined, the profile of all people infected with malaria, indigenous and imported cases combined, became mostly adult males (ages 15–49) with P. vivax infections, as opposed to P. falciparum infections. The proportion of all confirmed cases occurring in persons over the age of 15 was 58.8% in 1999, 77.0% in 2006, and 95.4% in 2011 (). In 1999, 53.9% of all infections were in males. By 2006 this grew to 59.6% and in 2011 reached 92.6% of cases. The proportion of infections due to P. vivax in all cases, including indigenous and imported, grew from 75.9% in 1999 to 95.4% in 2006 before leveling off at 90.3% in 2011. One P. ovale infection was diagnosed in 2005, then one P. malariae in 2008, which was acquired outside of Sri Lanka.
Annual percentage of confirmed infections for 1999, 2002, 2006 and 2011.
Male gem miners and male military personnel became a major risk group for malaria infection. Other at-risk groups are considered to be people living along rivers and streams with high vector density and mobile populations, such as chena (slash and burn) cultivators.
Non-conflict districts (17 districts) had a similar API (33.0) to the eight conflict districts (29.9) in 1995, but by 2000 conflict districts accounted for the majority of infections (71.5 average API in conflict and 35.1 in non-conflict districts). In 2005, when national incidence was much lower, API across both areas was the same (0.4). shows the spatial distribution of API across districts during 2000, 2005, and 2010.
The SPR was higher in conflict districts throughout most of the study period. In 1995, conflict districts had an SPR of 17.0% compared to 11.7% for non-conflict districts. As seen with API, by 2005 the SPR was the same across both regions (0.2%) and has not changed in recent years.
The principal vector in Sri Lanka is of the Anopheles culicifacies
species E of the complex which has sibling species A, B, C, D, and E 
. Species B is considered a poor vector in Sri Lanka 
. The species complex is largely found in stream habitat 
. Species E can be found in a wider range of habitat, mainly in river margins in the rock and sand pools, agricultural sites, and in wells and irrigation channels 
. Throughout its history, Sri Lanka has had major irrigation projects, funneling its rivers into dams, reservoirs and tunnels 
. The vector is considered both endo- and exophagic (outdoor and indoor feeding behavior), primarily endophilic (indoor resting), an “intensely domestic species,” and has a dusk to night biting time 
. An. culicifacies
is considered to be zoophilic in nature, except Species E 
. In southeastern Sri Lanka in the 1990s, which was a period of higher endemicity, the entomological inoculation rate (EIR) of the ten Anopheles species studied was estimated at 0.0029 infectious bites per person per night for P. vivax
and 0.0109 for P. falciparum
. More recent EIR estimates are not available as the number of vectors have been extremely low and finding sporozoite-infected mosquitoes even more challenging.
Another important vector on the island is An. subpictus
, of which two sibling species are present in Sri Lanka, Species A and B. Species A is associated with inland areas and Species B with coastal areas 
. An. subpictus
is found in coastal and brackish water, and pools and rice fields 
. This vector is also considered to be zoophilic and is both endo- and exophagic 
. It is endophilic in nature and bites at dusk or night. A study in the eastern part of Sri Lanka in 1989 and 1990 estimated an EIR for An. subpictus
ranged from 0.00006 to 0.007 infectious bites per night for P. vivax
and from 0.0002 to 0.005 for P. falciparum
Two main surveillance measures are used in Sri Lanka – passive and active case detection. “Activated passive case detection,” or APCD, is a form of passive case detection used in Sri Lanka comprised of dedicated malaria-only screening facilities in public health facilities. APCD capacity increased in the late 1990s by a near doubling of the number of microscopists in district hospitals. Outside of APCD facilities, health facilities conduct passive case detection without malaria-specific screening centers. All of these facilities rely mostly on microscopy. RDTs were distributed starting in 2001, but are only for emergencies, such as in the months following the tsunami. 45,000 RDTs were procured in 2004 with Global Fund funding. If all RDTs were used, this figure would represent only 4% of the 1.2 million malaria tests conducted that year.
Limited clinical diagnosis still occurs but is discouraged by the Directorate, and clinicians are instructed to take a blood film two weeks later if possible. Quality control of microscopically-confirmed diagnosis occurs at the regional and national laboratories: regional laboratories perform quality control of all positive and negative blood smears while the national reference laboratory carries out quality control on all positive tests, including RDT-positive, and 10% of negative smears.
Active case detection (ACD) was introduced in 1997. Over the years, mobile malaria clinics have targeted mobile populations resulting from the conflict and remote, inaccessible populations in all areas. Today ACD is also part of the reactive case investigation procedures (focal screening). The aim is to detect asymptomatic and symptomatic parasite carriers, including relapsing P. vivax cases, who may contribute to post-monsoon epidemics. RDTs are occasionally used for these clinics, but the majority of tests are conducted by microscopy. The World Bank International Development Association (IDA) supported initial ACD activities and the Global Fund increased support for them starting in 2003.
The annual blood examination rate (ABER), or the number of blood slides collected out of the total national population, was 6.1% in 1995, 9.4% in 2000, and declined to 5.0% in 2005 with little change through 2010 (4.8%).
In all years, the majority of confirmed cases nationwide were identified through APCD. In 1995, 89.8% of cases were identified through this method, with no significant change in 2000 (89.4%). APCD identified 94.0% of all cases in 2005. ACD accounted for only a small percentage of positive cases (0.9%) in 1997 and 2000 (1.1%), rising to 13.1% in 2007. Passive case detection and other blood surveys found the remaining confirmed infections.
As the number of cases declined after 1999, district-level staff had more time to dedicate to case investigation. In 2009, the AMC Directorate developed standard operating procedures (SOPs) for every confirmed or suspected infection, which include follow-up of confirmed infections for 28 days post-treatment together with case investigation procedures and additional measures, such as household malaria screening, entomological surveillance within 24 hours, and focal IRS in a one-kilometer radius. Also in 2009, the programme instituted case investigation reviews, where each case and the follow up measures taken are reviewed in detail by AMC Directorate staff and Regional Malaria Officers, at a meeting in the capital. The information gathered in the case investigation and in the case reviews is used to detect any deviations in vector behaviour (see Entomological Surveillance section, below) as well as to monitor clinical manifestations and parasite clearance time. The results of these investigations, in combination with mapping with geographical information systems (GIS) technology, which commenced three years ago, are used for the purpose of epidemic forecasting, the cornerstone of the national malaria elimination strategy. The programme continually seeks to strengthen surveillance in order to quickly detect epidemics.
In 2008, the AMC Directorate introduced individual case reporting to the AMC and a year later a policy of 24-hour case reporting was implemented. Regional Malaria Officers report cases by email, by phone or through the hotline maintained by the AMC Directorate. An elimination surveillance database was developed to house this information and for rapid analysis. There is a national health information system and there is a national requirement to report malaria cases to this system. However, the AMC uses a separate, web-based system that enables the programme to conduct 24-hour reporting. The AMC expects to integrate the malaria reporting system with the national system and with other diseases once malaria is eliminated from the country.
Monitoring and evaluation is an important aspect of the surveillance system and the entire programme, and the systems in place have been greatly strengthened through implementation of the Global Fund grants. A framework and plan for monitoring and elimination was developed in 2010, based on the framework put forward in 2009 by the Roll Back Malaria (RBM) – Malaria Evaluation Reference Group (MERG) 
. Indicators for disease surveillance and management as well as vector surveillance and control are included in the plan. Monitoring and evaluation of malaria activities is coordinated by the Regional Malaria Officers, at the district level, and at the national level by the AMC Directorate. Data is collected at the periphery at the smallest administrative level (Grama Niladari Division), with processing carried out by the RMOs and Medical Officer of Health Area staff 
. Data is used either immediately for corrective action or is processed upwards to the district, provincial, and national levels. Feedback to the periphery occurs typically through the case review monthly meetings with districts teams, as mentioned above, and includes other stakeholders. However, when urgent this feedback will occur more rapidly.
Since 1995, average ABER was higher in conflict districts. Conflict districts had an average ABER of 9.9% in 1995, compared to 5.4% in non-conflict districts. Conflict districts nearly doubled their ABER to 18.5% by 2000 while non-conflict districts only had a minor increase to 7.9%. By 2009, ABER decreased to 10.4% in conflict districts and 4.2% in non-conflict districts in 2009.
Conflict districts in some years had a slightly higher average per capita rate of ACD testing because the mobile clinics targeted hard to reach, at-risk populations, a majority of them located in the conflict areas. The per capita rate of ACD testing in conflict districts peaked in 2008 at 6.3% when 107,629 blood films were taken by these clinics. That year, 0.4% of the population in non-conflict districts was tested through ACD clinics.
In conflict districts, RMOs and their staff remained in their posts and were provided with vehicles and RDTs to conduct mobile clinics whenever it was safe to do so. There were reports of LTTE members assisting with and as beneficiaries of ACD clinics. In addition, the Sri Lankan Red Cross, the International Committee of the Red Cross (ICRC), and Medecins Sans Frontieres (MSF) assisted in providing diagnosis and treatment services. In addition, a Sri Lankan private not-for-profit organization, Tropical and Environmental Diseases and Health Associates (TEDHA), trains and deploys microscopists to APCD facilities in previous conflict districts as part of the Global Fund Round 8 grant, contributing to the scale-up of surveillance since 2009.
An entomological surveillance system was created in Sri Lanka shortly after the 1934–1935 epidemic, aiming to forecast increases in seasonal transmission and potential epidemics through identification of changes in vector breeding 
. Trained mosquito collectors collected larvae and adult mosquitoes in dwellings on a monthly schedule. In 1940 the programme added mandatory inspections of rivers and streams for larvae by public health inspectors in each jurisdiction 
These activities continue today at both the central and district level. Routine pyrethrum spray collections in dwellings, cattle-baited net and hut trap collections, window trap collections, and larval mosquito surveys are conducted in malarious districts at pre-determined sentinel sites. Susceptibility tests and bio assays detect evidence of insecticide resistance. Data obtained from these tests are used in planning IRS and in ITN/LLIN distribution. Since 2008, TEDHA has also conducted entomological surveillance in its target districts.
Entomological surveillance in Sri Lanka serves two major purposes – it is part of the epidemic forecasting system and is also an essential component of the national integrated vector management (IVM) strategy 
. IVM is used as a management tool in Sri Lanka and has been successful in agricultural areas through a combination of IRS, ITNs/LLINs, and larviciding and has contributed to the reduction in incidence. IVM in Sri Lanka brings together relevant sectors, community engagement and vector surveillance research to inform the use of insecticides and to determine the most appropriate mix of vector control interventions, environmental management and larval control. IVM in Sri Lanka began in the 1970s, when the hydroelectric and irrigation development project of Mahaweli River led to increases in malaria transmission 
. Vector control and larval source management were used in response, with participation of communities and with involvement of the irrigation and agriculture sectors 
. In the late 2000s, Farmer Field Schools were used as a platform to make the connection between vector control for health and agriculture, educating farmers about the relationship between public health and agriculture, and involving them in vector management activities 
Indoor residual spraying
Indoor residual spraying (IRS) was introduced in 1945 and became the primary vector control tool in Sri Lanka, where perennial spraying targeted all households in malarious districts. Following WHO recommendations issued in the mid-1990s 
, the AMC Directorate initiated a targeted spraying programme, focusing on historical areas of transmission, higher proportion of P. falciparum
, chloroquine-resistant confirmed infections, and proximity to vector breeding sites 
. A spatial mosaic insecticide rotation was then implemented in 1998, using a combination of up to six insecticides of two classes, organophosphates and pyrethroids. For example, in 2004, one zone of Kurunegala District applied Fenitrothion (organophosphate), while a neighboring sub-district used Cyfluthrin (pyrethroid). This spatial insecticide rotation has continued to today, unless there are delays in delivery of IRS supplies. The AMC instituted case-based and focal outbreak spraying in 2008, as a result of declining incidence.
In 1975 DDT was replaced by Malathion as reports of resistance to DDT increased. The first synthetic pyrethroid, Lambda-cyhalothrin, was introduced in 1994 and other new insecticides followed. The pyrethroid introduction may have increased community acceptance, which was already considered high (90% found in one study area for Malathion), as they emit less odor and do not leave visible residue on house walls 
. In 2002, Malathion was taken out of use because of mounting evidence of resistance.
National IRS coverage (estimated coverage of the population at risk) fluctuated over the 15-year period, from 64.8% (1995) to 46.5% (2000), then back down to 22.5% in 2005. In 2008, with the declining API there was a shift to case-based and focal outbreak IRS. By 2010, national coverage was down to 5.9% of the population at risk.
The AMC Directorate conducted IRS in conflict and LTTE-controlled districts, notwithstanding the challenges, including risk of landmines. LTTE personnel assured the AMC Directorate that support would be given to malaria control measures in their zones – partly because their combatants were severely affected by malaria. RMOs in neighboring stable districts report that they assisted conflict districts throughout all years by coordinating IRS along and at times over the border. The government sent supplies, including insecticides, to conflict districts by requesting permission from the Ministry of Defense to send shipments via the sole accessible road to the northeast or, alternatively, by passenger ship. LTTE and AMC Directorate communication increased during the ceasefire period, from 2002 to 2006. It is likely that supply delivery became easier during this period of relative calm.
This communication and collaboration allowed for the continuation of IRS in the conflict zone. In 1995, the population at risk protected by IRS reached 23.5% in the conflict districts, while higher coverage of 79.6% was estimated for non-conflict districts. However, the population at risk protected in conflict districts increased to 52.2% (2000) and 45.9% (2005). This was a higher level of coverage than in non-conflict districts in 2000 (43.7%) and 2005 (10.9%).
The second primary vector control tool in Sri Lanka after IRS is the distribution of ITNs and LLINs. ITNs were distributed since 1999 and LLINs were introduced in 2004 with support from the Global Fund. Non-conflict districts were prioritized for ITN/LLIN distribution according to several factors: P. falciparum percentage in the past three years, mortality, number of pregnant women and children affected by malaria, proximity to a mosquito breeding site, and presence of internally displaced persons (IDPs) or migrant populations. Because security conditions changed frequently in conflict zones with associated displacement of populations, there was no formal stratification process for ITN/LLIN distribution.
In 2005, 14.8% of the population at risk was estimated to be covered by a LLIN, rising to 22.7% by 2009 and to 34.6% in 2010. Coverage estimates are based on an average three-year lifespan of an effective LLIN and assumes appropriate use. A study conducted in 2008 on use of LLINs found that a range of 89.6% to 90.9% of respondents slept under a LLIN 
ITNs/LLINs were a key tool in conflict districts because of their higher caseloads, IDPs, and logistical challenges in conducting IRS. The Global Fund Round 1 grant supported the distribution of LLINs in conflict districts. The Ministry of Health, as part of the Global Fund grants, collaborated with a Sri Lankan NGO, Lanka Jatika Sarvodaya Shramadana Sangamaya (Sarvodaya), in the distribution of LLINs in northern conflict districts. The United Nations Children's Fund (UNICEF) and WHO also distributed LLINs. Through this network, 38.1% of the population at risk in conflict districts was covered by an LLIN in 2005, and 3.3% were covered in non-conflict districts. By 2009 coverage was similar in conflict districts (40.9%) and had increased in non-conflict districts (19.1%).
Treatment and prophylaxis
Since the mid-1990s, it was recommended that all fever patients were to be tested for malaria. Since 2007, testing is recommended only for fever cases with malaria-related history and symptoms - body aches, joint pain, headache, nausea, vomiting, or diarrhea.
Sri Lanka has a national health service, and consultation and treatment are provided free at all public hospitals. Global Fund support allowed for the scale-up of diagnosis and treatment of malaria. In addition, travelers to endemic countries receive free chemoprophylaxis for up to three months, based on destination of travel.
From the mid-1990s until 2006, chloroquine and primaquine (0.25 mg/kg/day for adults), with a five-day regimen in endemic areas and 14 days in low transmission areas, was used for P. vivax.
To ensure radical cure, or parasite clearance from both the blood and liver stages, a mandatory 14-day primaquine course was extended nationwide in 2006 
. Prevalence of Glucose-6-phosphate-dehydrogenase (G6PD) deficiency is relatively low (range of 1%–3%) 
. Patients are not routinely screened for G6PD deficiency before treatment.
In 1999, it was estimated that 51% of P. falciparum
infections were resistant to chloroquine, and by 2004 several cases of resistance to sulphadoxine-pyremethamine were detected 
. As part of the malaria elimination strategy and as a result of an increased number of imported P. falciparum
infections, artemisinin-based combination therapy (ACT), artemether-lumefantrin, was introduced in 2008. Primaquine for treatment of the gametocyte stage of the parasite has been documented to have been used in Sri Lanka since 1956 or earlier 
. The National Treatment Guidelines recommend that all P. falciparum
patients are admitted for three days, and P. vivax
patients should receive follow-up visits to ensure compliance with the primaquine regimen 
While reports in the early 1990s indicated that self-treatment was common 
, more recent studies describe a low level of self-treatment for malaria and patient preference for confirmed diagnosis 
Cost of malaria control and elimination
The Sri Lankan government and the Global Fund were and currently are the main sources of funding for malaria control in Sri Lanka. Funding for malaria control at the district level, based on risk and available resources, is allocated by the Ministries of Finance and Health. Sri Lanka successfully applied for and received funding for its malaria programme from the Global Fund in Rounds 1, 4, and 8. The approved grant amount was US $7.3 million in Round 1, $3.7 million in Round 4, and $21.6 million in Phase 1 of Round 8. The AMC Directorate, in collaboration with the Ministry of Health and the Global Fund Country Coordinating Mechanism Sri Lanka, determines which districts to include in grant proposals. Of the two districts selected for detailed costing studies, Anuradhapura received Global Fund support in Rounds 1 and 8, and Kurunegala in Rounds 4 and 8. WHO and World Bank/IDA provided additional support, which is only partially represented in the expenditure data. Both of these districts are located outside of the previous conflict zone, where major investments are targeted to scale up surveillance under the Global Fund Round 8 grant.
From 2004 to 2009, Anuradhapura District reported a decline of 48% in malaria programme expenditure per person at risk (). Expenditures in Kurunegala District did not significantly change.
Costs per person at risk in 2004 and 2009 by intervention category, in $USD, two districts.
From 2004 to 2009, there were also differences between Anuradhapura and Kurunegala in the proportion of total expenditure allocated to programme components. The proportion allocated for prevention declined in Anuradhapura from 2004 to 2009 (43.6% to 29.1%), while the percentage for surveillance slightly increased. In contrast, the proportion of malaria expenditures in Kurunegala on prevention and surveillance measures stayed consistent over these years.
The proportion of total expenditure distributed across cost categories (e.g., consumables) shifted from 2004 to 2009 for both districts. Anuradhapura reported a slight decrease in proportion of expenditure on personnel from 2004 to 2009 (80.8% to 74.0%) while Kurunegala reported an increase in the proportion for personnel (48.3% to 67.5%). Nationally, malaria full-time employees decreased by 29% during this period, from 2,991 to 2,113 
Initial cost estimates for elimination of P. falciparum
and P. vivax
in Sri Lanka, according to the five-year Global Fund Round 8 proposal budget projections, is $1 USD per person and $5 per person at risk