Background
Malaria has been decreasing in many parts of Laos and Vietnam, but it remains particularly endemic in remote, forest and forest fringe areas [
1‐
3], which often occur along the border [
4]. Malaria is often more difficult to control in border areas due to the more heavily forested, mountainous and inaccessible terrain, and because of unknown population movements across the border. In addition, these areas are most inhabited by ethnic minorities [
5,
6] with limited formal education [
7] and, therefore, less accessible for health education efforts. Furthermore, malaria control strategies and policies as well as the quality and management of the health care systems and conventions in data collection may differ across national borders, making cross-border collaboration difficult. One such area, with a clear trend towards greater infection and malaria morbidity near the international border, is on the border between Savannakhet province of Laos and Quang Tri province of Vietnam.
The malaria control strategy on the Lao side is based on long-lasting insecticide-treated nets (LLITN) for vector control and reducing human-vector contact [
8,
9]. However, in this particular area the strategy on the Vietnamese side relies mainly on indoor residual spraying [
10,
11]. There are also differences in practice on malaria diagnosis and treatment between the countries. Laos relies mostly on rapid diagnostic tests (RDT) while Vietnam uses microscopy as a rule. In both Vietnam and Laos the policy is that malaria patients do not need to pay for anti-malarial medicines when seeking care from the public health sector. However, there are differences in the access to public health facilities, the role of the private sector and auxiliary costs related to malaria treatment. In 2004, artemisinin-based combination therapy (ACT) using Coartem® (artemether + lumefantrine) was first introduced as a pilot intervention for the first-line treatment of uncomplicated falciparum malaria in three southern provinces of Laos. Since 2008, the use of this ACT has gradually scaled-up to cover the whole public health sector including village health volunteers in the country. Quinine and artesunate injectables are available for the treatment of severe malaria in district and provincial hospitals [
8], while only chloroquine and quinine are available in the private sector. Since 2009, ACT has also been expanded to the private sector by the pilot Public Private Mix project in Laos. In Vietnam, artemisinin was used for the first time in 1989 in Binh Phuoc province after which various kinds of artemisinin derivatives have been used for malaria treatment. The National Malaria Control Programme first recommended the use of locally produced “CV8” which was a combination of dihydroartemisinin with piperaquine and primaquine in 2003. In 2007, the recommendation was changed to only dihydroartemisinin plus piperaquine (as ‘Artekin’ and some other local trade names) after which this ACT became the standard first- line treatment of
Plasmodium falciparum infections [
12]. Chloroquine plus primaquine remained stipulated for
Plasmodium vivax infections.
As border malaria is one of the biggest obstacles for malaria control and elimination in the Greater Mekong sub-region (GMS) countries, coordination and collaboration among neighboring countries is very important [
13]. Although cross border studies are in place in the GMS countries to monitor drug quality and drug resistance, few attempts have been made to assess socio-economic and demographic characteristics, preventive and treatment-seeking behaviour, and cross-border movements of people living in a border area against the background of different health systems and malaria control policies on both sides of that border.
Therefore, a cross border survey on malaria was jointly designed, organized and analysed among the responsible preventive health services of Savannakhet province, Laos, and Quang Tri province, Vietnam. The objectives of this study were to obtain a better joint understanding of the main factors affecting the malaria situation along the border and, on the basis of that, improve and harmonize the local malaria control methods through more effective cooperation between the preventive health staff of the two countries.
Discussion
This study demonstrated some quite large differences between the Laos and Vietnamese respondents. At first sight this seemed surprising in view of the facts that 1) they belong to the same ethnic group (often with family relations across the border), 2) their villages are only a few kilometres from each other and 3) they easily and often cross the border. This applied to some socio-economic factors (e.g., the possession of cattle was much larger in Vietnam) and health related behavioural factors (e.g., health-seeking behaviour). On further reflection however, it is clear that an international border between two very different countries itself can create contrasts within the communities divided by it. The more easily accessible and more attractive health facilities on the Vietnamese side naturally caused the asymmetry in health-seeking behaviour. This was already known long before and one of the reasons for the need to improve cross-border collaboration in this area. The records of the Commune Health Centres in Vietnam have, for many years, shown a high proportion of Lao out-patients. It was clear from the Lao respondents that they preferred Vietnamese health service due to a combination of easier to reach and more affordable treatment. Other factors, like the perception about the quality of treatment and the language, are also likely to play a role. In comparison, more than half (68%) of the border crossing population on the Vietnamese side reported crossing the border to Laos for business purposes and none for seeking health care. Ironically, while borders are per definition designed to restrict people’s movements, they also create certain special livelihood opportunities (trading/smuggling) that stimulate people to cross it.
The environmental/biological differences over a short distance were in fact the more unexpected ones: higher malaria prevalence, different age distribution of infections, higher mosquito densities, different vector composition and higher forest cover on the Lao side. The aim of this study was to measure and, if possible, explain these differences in relation to the environmental and human factors, with a view to improve malaria control in the border area.
The blood survey showed that malaria prevalence was lower than expected on both sides, but it was significantly higher in Laos than in Vietnam. Both
P. falciparum and
P. vivax were found but the proportion of
P. vivax to
P. falciparum (equal) was greater than is usually found in the nationwide records in both countries. In Laos especially,
P. falciparum is reported to account for 95% of all recorded malaria cases [
21,
22]. However, only provincial and district hospitals, and a few health centres, have microscopists who can perform species-specific diagnosis in Laos. In addition, as complications and severe disease resulting from
P. vivax infection are rare and patients probably seek treatment less frequently, the passive case detection system is likely to underestimate the prevalence of
P. vivax infection [
23]. To improve case management at the community level the NMCP of Laos introduced between 2005 and 2008 a
P. falciparum -specific malaria rapid diagnostic test that does not detect
P. vivax[
22], so that the number of
P. vivax cases was underestimated even more since that time. However, in 2010, NMCP of Laos has revised the national strategy for malaria control and pre-elimination from 2011 and 2015. One of the objectives in this strategy is to improve access to early and accurate diagnosis for malaria by strengthening the public sector microscopy network and provide RDTs that can distinguish
P. falciparum and
P. vivax for the diagnosis at all “stratum 3 villages” (with high malaria incidence). Through the implementation of this strategy the determination of the
P. vivax infection rate as well as species-specific treatment are likely to be improved in Laos [
8].
Malaria infection rates in Laos were found to be higher in children under 15 years of age than in adults and the prevalence was highest in the age group from 5 to 14 years, whereas in the sample from Vietnam children and adults were equally likely to be infected. This suggests that there is a difference in acquired immunity status between the two populations as a result of a more intensive exposure to malaria infections in Laos. If this is true we would also expect a higher rate of asymptomatic malaria on the Laos side, which was not systematically checked to allow a comparison in this study.
Overnighting in the forest was only slightly more reported by respondents from Laos (17%) than from Vietnam (14%) and was on both sides mostly done by the males. However, in both countries the infection rates were not significantly different by sex. In combination with the high prevalence among children this suggests that at least on the Lao side most malaria is transmitted in the villages and not mainly incurred during work in the forest. A limitation of the study is however that we cannot know the proportion and infection status of people that were missed due to being away in the forest at the time of the survey. In Vietnam, it has been reported that regular sleeping in the forest increased the risk of malaria infection up till eight times and was more incurred by men than women [
24]. However, this increased risk factor often only appeared after the transmission risk in the villages had been reduced, for instance by effective control of
An. minimus[
25].
With respect to malaria prevention measures, this study showed that the possession of bed nets by households was high on both sides, but that the coverage of insecticide-treated nets was much higher in Laos than in Vietnam. ITN also form the mainstay of prevention by the Vietnamese Malaria Control Program but when it is observed that too few people (less than 80–75%) actually sleep under a net in highly malarious area, or when this is considered a remote or otherwise problematic area, indoor residual spraying is applied. This border area is one of those areas where Vietnam continued to use IRS and where, while most people reported to have slept under a net, these were not insecticide treated. Not using IRS may be one of the reasons for the higher malaria prevalence on the Laos side, even though it has been reported that untreated bed nets also can have a significant protective effect on the risk of clinical malaria and malaria infection [
24,
26].
Another factor is the utilization rate of the available nets. During this survey no direct observations of bed net usage were made but the number of interviewed householders who reported that people had slept under the nets during the previous night was lower in Laos (83%) than in Vietnam (98%). Since 1999, the NMCP of Laos has distributed only the x-family size bed nets to the target villages with a target average coverage of 2.5 persons per net. This distribution ratio may not have provided enough bed nets for everyone in the family, especially in the poor ethnic minority villages where they could not buy an extra bed nets from the local market, and also where the sleeping behaviour is different from other groups due to local custom and culture. For instance, it has been noted from this area that the husband and wife do not regularly want to share the same bed net. This may explain the lower rate of people who reported to have slept under a bed net during the previous night at the Laos side.
However, more important is that the time of usually going to bed was reported to be between 8:00 and 11:00 pm on both sides. This still leaves some hours during which they are at risk of biting by Anopheles mosquitoes, especially by
An.dirus which is often reported to start to bite early in the evening [
25,
27]. Therefore, even full utilization of bed nets may still be ineffective in reducing malaria transmitted by
An. dirus.
The short entomological survey showed the presence of malaria vector mosquitoes on both sides of the border, but that their density was considerably higher in Laos than in Vietnam. Especially, the occurrence of a high density of
An. dirus in the villages in an area at the Laos side where most malaria infections were found, while none were found during this survey in Vietnam, leaves no doubt about the role of this vector in explaining the difference in malaria prevalence. Less easy is it to explain the difference in the occurrence of
An. dirus between villages that are near to each other just across the border, such as the Denvilay and Oi villages in Laos next to the A Ho and Xung villages in Vietnam. The indoor residual spraying that was applied at the Vietnamese side could be a factor, but this is not certain because the types of pyrethroid insecticides used in both IRS and LLITN are not known to strongly deter mosquitoes from entering houses. This study suggests that the presence of LLITN/ITN on the Lao side does not prevent
An. dirus from entering houses and be captured by indoor light traps. This points to the urgent need to study the actual utilization of bed nets by people and the effectiveness of the used insecticides on reducing the exposure to mosquito bites, none of which were studied here. In general too little is known about the behaviour of
An. dirus, as the most effective malaria vector in Southeast Asia [
4,
25,
28‐
30] in response to different applications of insecticide. Its well-documented exophily [
31‐
33] would make it less affected by any indoor insecticide use, certainly in comparison with
An. minimus. The high density of
An. dirus in indoor light traps found in this study is especially disturbing because of the sole reliance on LLITN for vector control in the Laos malaria control programme.
The geographic distribution of malaria prevalence obtained through the blood survey showed that the highest intensity of malaria transmission (on both sides of the border) occurred in the southern part of the study area. This was also the area that had the highest forest cover at the Laos side and is continuous with a relatively large, still uninhabited forest in Laos. This indicated a priority area for intensified control measures, especially at the Laos side.
The difference between the villages at the Laos and Vietnamese side of the border in respect of forest cover (by analysis of satellite pictures and the average distance of houses to the forest edge) found in this study fully matches with the strong association of
An. dirus with the forest [
25,
34,
35] and this may yet be the overriding factor for its presence in the Lao villages.
A limitation of this study was that the time at which the survey was conducted in the two countries differed by almost one month due to internal logistical factors of each country. However, according to the local meteorological data of each country, there was little seasonal variation in this border area during the period of study. Another limitation was that, although it was tried to recruit equal numbers of male and female respondents to participate in the interview, in Laos often the wife preferred her husband to answer when he was present (more men than women can speak the Lao or Vietnamese language). It is not known to what extent this has influenced the results of the interviews, nor can we assess whether the fact that all interviewers were male has influenced this.
Although it was not an objective of this study to assess the quality of the RDT diagnosis, a considerably proportion of false negatives was noted (13 out of the 35 persons microscopically positive for
P. falciparum on the Laos side; the sensitivity was 63% and specificity was 99%). This implies that these microscopically positive people were not promptly treated during the survey if they had no symptoms either. It is, therefore, strongly recommended to study this separately in the future, preferably with the new RDTs that will be introduced. It is recommended to continue to strengthen the role and function of the VHVs, especially those at the Lao side, where they are more needed due to the scarcity of Health Zone Centres. The interview data showed that VHVs are under-utilized on both sides (Table
3). A prime necessity is that they are trusted by the local population and perceived as useful by the health sector, which remains difficult if they can not perform reliable diagnosis nor always have free anti-malarial medicines of the correct type.
This survey was the first cross-country survey jointly designed and conducted by staff from two provincial branches of the Malaria control services in Laos and Vietnam. Despite the limitations and obstacles encountered during the survey, this proved to be an effective initiative in cross border collaboration. It helped to focus the malaria staff of the two neighboring countries on a common evidence base in stead of each side talking from their own, often difficult to compare, data, ideas and prejudices. It led to more mutual appreciation of the particular constraints the other side had to face and set their minds to overcome or accept their differences and find a common solution to battle the malaria problem in their border area. It has led to a concrete intensification of the coordination (sharing malaria data) and even in an actual adjustment of the control methodology used on the Laos side (an IRS campaign in a hot spot along the border). Therefore we think this approach can be a good model to mediate effective cross-border collaboration also in other areas. In this case the intermediary role of an NGO (Medical Committee Netherlands-Vietnam) with experience of working in both countries was instrumental to facilitate the process while guarding the equivalence of the two partners in the joint project, each of which had their own strengths and weaknesses.
To improve the situation of malaria along this border area, an effective package of cross-border malaria control interventions needs to be developed and piloted on both sides. Based on the results of study it is recommended to conduct IRS in Laos not only in response to outbreaks but also allow more routine use of IRS in some villages in remote areas with continuous intense malaria transmission. Secondly, in addition to free distribution of LLITN, the use of insect repellents should be tested and popularized to reduce the biting risk in the hours that people are not yet under the nets. Thirdly, to improve bed net utilization both x-family and single sizes should be provided. Self-evidently this should all go hand in hand with a continued strengthening of routine anti-malaria activities such as the full provision of free malaria diagnosis and effective treatment in the public sector and the development of materials for information, education and communication (IEC) for malaria prevention and control that can be understood in Lao, Vietnamese and (for local radio and TV) also in the ethnic minority languages.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
TP, HH, LT and RPM were responsible for the entire process. DN supported data analysis and helped to review the manuscript. PP contributed to the development of the study design. KM and JK and TB contributed to the development of the study design, analyse and write the results of forest cover by using the satellite images and review the manuscript. All authors read and approved the final manuscript.