Tweet
The Lancet Infectious Diseases 2005; 5:695-708
DOI:10.1016/S1473-3099(05)70268-1
Reducing the burden of malaria in different eco-epidemiological settings with environmental management: a systematic review
Jennifer Keiser (Research fellow)
a 
, Burton H Singer (Professor) b and J?rg Utzinger (Assistant professor) a
Summary
Introduction
Background
Methods
Results
Discussion
Search strategy and selection criteria
References
Summary
The public health and economic significance of malaria is enormous, and its control remains a great challenge. Many established malaria control methods are hampered by drug resistance and insecticide-resistant vectors. Malaria control measures built around environmental management are non-toxic, cost-effective, and sustainable. However, there has been no comprehensive review of the literature or meta-analysis examining the effect of these interventions. We therefore did a systematic literature review and identified 40 studies that emphasised environmental management interventions and reported clinical malaria variables as outcome measures. Of these 40 studies, environmental modification (measures aiming to create a permanent or long-lasting effect on land, water, or vegetation to reduce vector habitats?eg, the installation and maintenance of drains) was the central feature in 27 studies, environmental manipulation (methods creating temporary unfavourable conditions for the vector?eg, water or vegetation management) in four, and nine quantified the effect of modifications of human habitation. Most of the studies (n=34, 85%) were implemented before the Global Malaria Eradication Campaign (1955?69), which mainly relied on indoor residual spraying with dichlorodiphenyltrichloroethane (DDT). In 16 studies that applied environmental modification and in eight studies on modification of human habitation, the risk ratio of malaria was reduced by 88?0% (95% CI 81?7?92?1) and 79?5% (95% CI 67?4?87?2), respectively. We conclude that malaria control programmes that emphasise environmental management are highly effective in reducing morbidity and mortality. Lessons learned from these past successful programmes can inspire sound and sustainable malaria control approaches and strategies.
Back to top
Introduction
Malaria continues to be a major public-health challenge. Globally, an estimated 300?660 million clinical Plasmodium falciparum cases occurred in 2002.1 In sub-Saharan Africa, malaria is responsible for 1?5?2?7 million deaths each year, of which at least 0?7 million are among children aged under 5 years.2 Acute febrile illness, anaemia, chronic debilitation, complications in the course and outcome of pregnancy, and delays in cognitive and physical development contribute to a heavy public-health burden, which has a negative impact on the social and economic development of affected countries.3?6
Key strategies in the control of malaria include early diagnosis and treatment, indoor residual spraying, and the use of insecticide-treated nets.7 Unfortunately, these control strategies are becoming less effective with the rapid development and spread of resistance to widely used drugs and insecticides. Exploiting the genomic sequences of the malaria parasites and vectors8 might ultimately lead to new generations of drugs and insecticides, the development of an effective vaccine,9 or genetically modified mosquitoes.10 However, these tools are not likely to be available for at least 10 years.10
Malaria control methods that involve environmental management, which is a modification or manipulation of the environment to reduce malaria transmission (eg, through the installation, cleaning, and maintenance of drains, or the systematic elimination of standing pools of water), currently receive far less attention. There is historical evidence that environmental management programmes effectively reduced malaria.11?13 In the early decades of the 20th century, engineering personnel worked alongside malaria control officers to ensure the implementation of environmental management that would deal simultaneously with waste removal and malaria suppression. However, environmental management interventions almost disappeared with the advent of dichlorodiethyltrichloroethane (DDT), which offered a standardised single attack, during the Global Malaria Eradication Campaign, spearheaded by the WHO. Environmental management is still readily available for implementation. Important features of environmental management strategies are their non-toxicity, relative ease of application, cost-effectiveness, and sustainability.14?16 Environmental management can have a major role within integrated vector management,17 adding resilience to the results of individual control strategies and reducing costs, and the likelihood of development and spread of drug and insecticide resistance.18
To our knowledge, there has been no systematic literature review and meta-analysis to examine the effectiveness of malaria control programmes that emphasise environmental management. Our aim is to fill this gap, and hence to establish an evidence base of the impact that environmental management has in preventing malaria-attributable ill health and death.
Back to top
Background
Environmental management
Drainage interventions to reduce fevers in settings near swamps or stagnant waters date back more than two millennia.19 However, an internationally recognised definition of environmental management for mosquito control was published less than 3 decades ago.20 In 1979, WHO put forth the statement that, ?Environmental management is the planning, organization, carrying out and monitoring of activities for the modification and/or manipulation of environmental factors or their interaction with man with a view to preventing or minimizing vector propagation and reducing man?vector?pathogen contact?.20 The major approaches of environmental management comprise: (1) environmental modification, which aims to create a permanent or long-lasting effect on land, water, or vegetation to reduce vector habitats; (2) environmental manipulation, which produces temporary unfavourable conditions for the vector; and (3) modification or manipulation of human habitation or behaviour, which reduces man?vector contact.20 These somewhat cumbersome articulations become more transparent if one thinks of environmental management interventions as the installation, cleaning, and maintenance of drains, vegetation management and river boundary modifications to promote flowing water, intermittent irrigation, and systematic elimination of standing pools of water (figures 1, 2, and 3). The specificities must, of necessity, vary with local ecosystem structure; therefore there is no single uniform environmental management recipe that is appropriate in all settings.
Click to enlarge image
Figure 1. Well designed and maintained drainage canals WHO
Click to enlarge image
Figure 2. Effective canal maintenance with community participation to ensure that canals are in a good shape and generally free from vegetation and silting (Thailand) WHO
Click to enlarge image
Figure 3. Concrete lined rice irrigation canal in central Laos Swiss Tropical Institute, 2004
Regarding malaria, environmental management can be coarsely grouped into four distinct eco-epidemiological settings: (1) malaria of deep forests, forest fringe, and hills; (2) rural malaria attributable to water resources development and management (eg, irrigation and large dams); (3) rural malaria attributable to wetlands, rivers, streams, coasts, and non-agricultural man-made habitats; and (4) urban and peri-urban malaria. Typical features of these eco-epidemiological settings and possible environmental manipulation and modification approaches, including changes in the human habitations, are summarised in table 1.
Table 1. Stratification of malaria into different eco-epidemiological settings and possible environmental management strategies
Malaria burden in different eco-epidemiological settings
Table 2 shows the total population, the population at risk of malaria, and the estimated number of malaria cases in sub-Saharan Africa, and the most severely affected countries outside of Africa. These variables are stratified into the four eco-epidemiological settings outlined above. The list of countries most affected by malaria was compiled by consultation of the regional WHO office websites.
Table 2. Estimated populations at risk of malaria and annual number of clinical malaria cases stratified by malaria eco-epidemiological settings in sub-regions and countries that carried the main burden of malaria in 2002
For the years 2002?03, we estimate that more than 1?9 billion people were at risk of the disease in sub-Saharan Africa and in 11 malaria-endemic countries outside Africa. Among these, at least 500 million live in urban and peri-urban settings, and 145?673 million people live in close proximity to irrigation and large dam sites. No precise information is available on the number of people living in rural or forested malaria-endemic areas. We estimate that 245?406 million clinical malaria cases occur in the countries considered here each year, with most occurring in rural areas. In urban and peri-urban settings at least 26?4 million clinical malaria events occurred. According to our estimates, 15?1?17?1 million cases are attributable to forest malaria in just five countries for which we could retrieve data (Brazil, India, Indonesia, Burma, and Bangladesh).
Back to top
Methods
Search strategy and selection criteria
We searched the following electronic databases, with no restriction on year or language of publication, to identify studies that assessed the effect of environmental management on pertinent malaria outcome measures: Medline, Web of Science, WHO Library Database, WorldCat and FirstSearch, and ScienceDirect. ScienceDirect includes full text articles of The Lancet and the Transactions of the Royal Society of Tropical Medicine and Hygiene dating back to 1823 and 1907, respectively. In addition, we hand-searched 15 volumes of the Tropical Diseases Bulletin (1930?45), which is a leading abstract database.
We used a broad search strategy to optimise sensitivity. We applied the following keywords: ?malaria?, ?alternate wet dry irrigation?, ?drainage?, ?engineering?, ?environmental management?, ?environmental manipulation?, ?environmental modification?, ?filling?, ?fluctuation?, ?flushing?, ?grading?, ?housing?, ?integrated vector control?, ?integrated malaria control?, ?intermittent irrigation?, ?levelling?, ?prevention?, ?prophylaxis?, ?sanitation?, ?sprinkler irrigation?, ?vector control?, ?vegetation management?, and ?water management?. We included all studies that used environmental management interventions exclusively or as the main feature of control, and that assessed malaria prevalence, incidence, spleen rates (prevalence of splenomegaly), or mortality rates. Bibliographies of the obtained articles were screened for additional relevant publications. Studies analysing entomological variables only and studies examining the effect of insecticide-treated nets on malaria were excluded. The key findings of the latter have been systematically reviewed in a recently updated Cochrane review.31
Statistical analysis
Studies obtained were stratified into three main groups of interventions: environmental modification, environmental manipulation, and modifications of human habitation. Descriptive studies and studies without defined sample size groups were excluded from the meta-analysis, but their key features are summarised in the tables. Data on intervention effectiveness (ie, malaria incidence, spleen rates, and mortality) were extracted and analysed by use of StatsDirect software (version 2.4.5; StatsDirect Ltd, Cheshire, UK). We calculated the risk ratio and 95% CI. We examined heterogeneity between trials with Cochrane's Q statistics and calculated I2 figures, which measure the proportion of variation in study outcomes due to between-study heterogeneity. We examined whether there was a publication bias by use of Egger's test (a small-study bias is evident if p
0?1).32 A random-effects model was used for calculation of the pooled risk ratio for environmental manipulation and modification, and modifications of human habitation, because the test of heterogeneity was highly significant for both groups of interventions (p0?001).33 Studies with a risk ratio below 1?0 indicate a reduction in clinical malaria variables in the intervention groups compared with controls.
Back to top
Results
Our systematic review yielded 40 studies with an emphasis on environmental management interventions that reported clinical malaria variables as outcome measures. Environmental modification was the central feature in 27 studies (29 datasets; table 3) and environmental manipulation in four studies (table 4). Nine studies (ten datasets) measured the effect of modifications of human habitation (table 5).
Table 3. Studies that used environmental modification and appraisal of outcome in terms of clinical malaria manifestations
Table 4. Studies using environmental manipulation and appraisal of outcome in terms of clinical malaria manifestations
Table 5. Studies that used modification or manipulation of human habitation and appraisal of outcome in terms of clinical malaria manifestations
The identified malaria control programmes took place in 18 different countries or territories, in different eco-epidemiological settings, involving different anopheline vectors and different levels of endemicities. Figure 4 shows that most of the environmental management studies (n=34; 85%) were launched before the Global Malaria Eradication Campaign, which was implemented between 1955 and 1969, and mainly relied on indoor residual spraying with DDT. We could not identify a single study that applied environmental management as the central feature for malaria control during this period. Only six studies were initiated after 1969. The studies identified various outcome measures, such as malaria incidence, spleen rates, hospital admissions, or mortality rates. Overall, 24 datasets could be derived from 21 malaria control programmes to calculate the risk ratio. Among those, 16 applied a before-and-after evaluation design (comparing the pre-programme situation with the malaria situation during or after implementation of the interventions), and seven studies (eight datasets) compared an environmental management intervention group (all in the category modification or manipulation of human habitation) with a control group in which no intervention was implemented.
Click to enlarge image
Figure 4. Temporal examination of studies with environmental management (n=40) as the main feature of malaria control The year when the studies were launched is used for the analysis. Light blue bars: environmental modification (n=27); red bars: environmental manipulation (n=4); dark blue bars: modification or manipulation of human habitation (n=9).
Environmental modification
Prominent environmental modifications were as follows: drainage; filling of swamps, borrow pits, pools, and ponds; modification of river boundaries; or other engineering approaches. Multiple interventions were implemented in several studies, such as environmental modification in combination with oiling and the application of Paris Green and, for part of the population, distribution of bed nets, quinine prophylaxis, and treatment, as well as health education. The risk ratio estimates of environmental modification were found to range between 0?017 and 0?45 (table 3). Heterogeneity of the studies was very pronounced (p0?001; I2=99%). The random summary risk ratio measure of 16 environmental modification studies plus one environmental manipulation study was calculated as 0?120 (95% CI 0?079?0?183), giving a summary protective efficacy of 88% (figure 5). There is evidence of a publication bias (Egger's test: −15?07; 95% CI −29?76 to −0?38; p=0?045).
Click to enlarge image
Figure 5. Random effects meta-analysis of environmental modification and manipulation interventions against malaria Squares=relative risk. The size of each square denotes the sample size. Diamond=combined relative risk. Horizontal lines=95% confidence interval.
William Gorgas, Malcolm Watson, and Nicolaas H Swellengrebel were three of the pioneers who implemented environmental modification measures in rural and urban areas of the Panama Canal zone, peninsular Malaysia, and the Indonesian archipelago, respectively. We briefly summarise these successful malaria programmes below.
Panama
In 1878, Ferdinand de Lesseps (the French engineer who led the construction of the Suez Canal) and his team began to dig a canal across the Isthmus of Panama, which was then part of Colombia. However, yellow fever, malaria, and engineering challenges halted construction of the canal. Between 1881 and 1889, an estimated 22
189 labourers of the French Canal Company died. The area was covered with dense jungle, streets were unpaved, and there were many puddles. Furthermore, the towns were surrounded by swamps, and there was no water supply or sanitation. After the Americans took over the project and Gorgas was placed in charge of the health programme, environmental management was adopted as the mainstay of malaria control. This mainly consisted of temporary and permanent drainage infrastructure (ie, installation of concrete-bottomed drains and subterranean pipes) and vegetation management. The malaria incidence among the employees decreased from 821 per 1000 in 1906, to 14 per 1000 in 1917.11,40
Malaysia
In 1901, the burden of malaria among the 4000 inhabitants of Klang Town and Port Swettenham, in present-day Malaysia, was enormous: half of the patients in the hospitals had malaria, which led many employees to leave Port Swettenham.12,19,56?59 Although the two main malaria vectors were not known at that time (Anopheles umbrosus in Klang and Anopheles sundaicus in Port Swettenham), detailed entomological surveys and examination of the spatial distribution of malaria cases were carried out. The principal anopheles breeding habitats in Klang, for example, were permanent swamps, open wells, and pools along the roads. Due to a lack of compliance and difficulties with supervision, the use of bed nets, house screening, and quinine prophylaxis were thought to be impractical for the Chinese and Indian rubber planters.19 Consequently, engineering approaches were implemented. The drainage of swamps complemented with vegetation management (ie, clearing of undergrowth, forest, and mangroves) successfully reduced mortality and malaria incidence (risk ratio estimate of 0?12) and malaria-attributable mortality. Following these effective environmental management control approaches, environmental modification works were widely and successfully applied in other parts of Malaysia, notably in the rubber estates,12 Singapore,12,61 and Hong Kong.11
Indonesia
Between 1920 and 1935, species sanitation was widely practised on the Indonesian archipelago. This concept emphasises the selection of environmental management strategies after studying the ecological niche and breeding habits of the principal malaria vectors.13,61 Hence, the antimalarial campaigns in Indonesia consisted of several stages. First, the spatial distribution of clinical malaria variables was examined by collecting adult anopheles to draw correlations between malaria distribution and its principal vector. Second, the natural infection rate of the local vector(s), including important breeding habitats, was examined. Environmental management was then targeted to settings where a principal malaria vector with a distinct breeding habitat had been identified. Six Indonesian studies showed the effect of environmental modification, namely drainage or filling of swamps, tidal zones, irrigation schemes, or other water bodies (table 3). In all six settings spleen rates were substantially reduced.13
Importantly, environmental modification in combination with the use of bed nets and quinine among European workers was also found to be largely successful in two programmes carried out in high transmission areas of sub-Saharan Africa, among the most efficient malaria vectors (ie, Anopheles funestus and Anopheles gambiae).16,44,45 Both programmes covered relatively small populations: mining communities in Northern Rhodesia (present-day Zambia), and military personnel at the Royal Air Force station in Apapa, Nigeria. The risk ratio estimate of environmental management programmes in four Zambian copper mining communities was calculated as 0?35. These studies were also remarkably cost-effective and spurred unprecedented economic development.16,44
Back to top
Environmental manipulation
Several studies have assessed the effect of environmental manipulation on the anopheles population, with most focusing on water management, such as intermittent irrigation (alternate wet-dry irrigation).62?64 With regard to clinical variables of malaria, we found only four studies that assessed the effect of environmental manipulation, three studies that adopted water management, and one study that reported the effect of vegetation management (table 4). These programmes were implemented in China, India, North Borneo, and USA. Only one study presented data that could be used to estimate the risk ratio. Intermittently irrigated rice fields in six villages in India, in combination with environmental modification, such as the filling or draining of borrow pits in the villages, reduced the spleen rate in children from 48% in 1937, to 4% in 1941, giving a risk ratio estimate of 0?08. These interventions also reduced the parasite rate from 42% in 1937, to 0% in 1941.47
Interestingly, more than 50 years later, intermittent irrigation is being promoted extensively in Sichuan province, China. For example, in Xindu district and Quionglai county, 99?7% and 93?4% of the rice fields, respectively, are now irrigated in this way. Vector breeding has been reduced accordingly, and no malaria cases have been observed in these two settings.50
Modification or manipulation of human habitation
The relation between the quality of housing and good health has a long history. The first trials showing that people could be protected from malaria by mosquito proofing their homes were done between 1899 and 1904 along two railway lines in Italy (Prenestina-Cervara and Pontegalera). Houses in the intervention group had their windows covered with frames of tulle or muslin, and doors were screened with wire gauze, whereas the houses of the control groups were left unprotected. In 1899, 17% (four of 24) of the individuals in the intervention group compared with 96% (exact numbers not known) in the control group contracted malaria. These findings were translated into broader public health action, and by 1904, 12378 people already lived in mosquito-proofed houses. Protection from malaria was observed, even if only parts of the houses were screened. Whereas 25?96% of the inhabitants of non-mosquito-proofed houses had malaria, only 1?9% of individuals in completely screened houses and 10?9% in partly mosquito-proofed houses became infected.65 The practice of mosquito proofing houses began to make inroads into the tropics and subtropics with European settlers and military personnel. For example, malaria cases were greatly reduced among British troops stationed in India who lived in mosquito-proofed barracks, and among South African children living in screened houses (table 5). Heterogeneity of these studies was significant (p0?001; I2=87%). The risk ratio estimates of these studies were found to range between 0?06 and 0?54. A pooled random risk ratio estimate was calculated as 0?205 (95% CI 0?128?0?326; figure 6). There is evidence of a publication bias (Egger's test: −5?5; 95% CI −10?7 to −0?44; p=0?037).
Click to enlarge image
Figure 6. Random effects meta-analysis of modification of the human habitation against malaria Squares=relative risk. The size of each square denotes the sample size. Diamond=combined relative risk. Horizontal lines=95% confidence interval.
By contrast with the aforementioned studies, Indonesia showed an increase in malaria-related mortality after the implementation of house improvement programmes. Malaria outbreaks were observed in 44 of 71 districts with improved housing. Possible explanations included the creation of borrow pits accompanying house renovations, the neglect of cleaning up operations, and the immigration of workers carrying parasites in their blood. The improved houses had better ventilations and were often smoke-free, which in turn might represent more suitable resting places for adult anopheles mosquitoes.66
A negative aspect of the Global Malaria Eradication Campaign was that during and after the campaign little attention had been paid to the relatively straightforward task of modifying human habitation. We are not aware of any recent malaria control programmes that have used human habitation modification in a systematic manner to reduce malaria morbidity and mortality. However, several studies have analysed the relation between house design and mosquito densities. Recently, Lindsay and colleagues51 published an excellent summary of these entomological studies. In Sao Tome, twice as many A gambiae were found in houses built at ground level compared with houses built on stilts.67 Another set of studies identified housing risk factors for malaria. These studies compared clinical and entomological variables of malaria in either poor or well-constructed housing. Houses with closed eaves, ceilings, and in good condition had fewer mosquitoes, and inhabitants had lower malaria incidence than those living in houses with open eaves, no ceilings, or that were poorly constructed.51,68,69
Back to top
Discussion
There is growing recognition that systematic reviews and meta-analyses have important roles in evidence-based medicine and therefore contribute to decision making with the aim of using scarce public-health resources in a cost-effective manner.70 However, by contrast with malaria chemotherapy, the use of insecticide-treated nets, or indoor residual spraying, the impact of environmental management control strategies to reduce the malaria burden has not been systematically reviewed to date. We made an initial attempt to fill this gap through a systematic review and a meta-analysis of malaria control programmes, emphasising environmental management as their main feature. We found very high summary protective efficacies on clinical malaria parameters (79?5?88?0%) regardless of whether environmental modification, environmental manipulation, or modification or manipulation of human habitation or behaviour was used. Furthermore, high protective efficacies were also achieved in high transmission areas, with the most efficient malaria vectors. It is important to note that most of the studies reviewed here were site specific, often quite large programmes, implemented over a period of several years, with adaptive tuning of interventions over time, but covering a relatively small subset of the national level at-risk populations. This is by contrast with most clinical trials, where interventions are standardised, fixed in advance, but only last for relatively short periods of time.
We found only one study, the malaria control programme in Mian Mir (now Lahore, Pakistan) implemented from 1901?09, for which no conclusion could be drawn on the effect of environmental management on clinical malaria parameters.34,71,72 The experiences made in Mian Mir continue to be debated in the contemporary literature.19,44,72 It was argued, for example, that the malaria control programme in Mian Mir failed because inappropriate control measures were adopted, inexperienced workers employed, insufficient larvicides used, and no protection zone for dwellings established.34
All the other studies reported very high protective efficacies when environmental management was used for malaria control. The results and the potential for contemporary malaria control programmes require scrutiny. The following issues are offered for discussion. First, we cannot rule out that we have overestimated the effectiveness of environmental management interventions. One reason is that the number of environmental management control studies identified is relatively small. Second, our statistical results suggest that there might be publication bias, and thus successful interventions were more likely to be reported in the literature. In addition, it was impossible to scrutinise the methodological quality of the studies as most of them were implemented 50?100 years ago. Furthermore, most of the malaria control programmes based on environmental management were to some degree accompanied by other control measures, which probably contributed to their overall success. For example, oil-based antilarval products were applied in combination with environmental management in a number of studies. Since several control measures were put in place concurrently, it is intrinsically impossible to tease out the independent effect of environmental management. However, in several places, as for example in Apapa, Nigeria, no change in malaria rates could be observed by applying quinine, house screening, and bed nets alone. Once environmental management was put in place, together with the former control programme, malaria incidence sharply decreased.45
Even if the true protective efficacy of environmental management is somewhat lower than reported here, it is crucial that we are aware of the lessons that can be learned from past campaigns. Their integrated nature, inter-relations with water and vegetation management, and organisational frameworks are of central importance for malaria control today.
The success of the malaria control programmes reviewed can be summarised by the following features. First, most programmes emphasised the permanent elimination of breeding sites. Second, environmental management tools used were idiosyncratic to the individual settings and vector. For example, if the local vector favoured a wide variety of remote and temporary breeding sites, then modification of human habitation would be the environmental management strategy of choice. Third, most studies lasted several years, allowing continuing performance evaluations of the control tools and adaptive adjustment to enhance performance.73 Fourth, community participation and usually also health education were integral parts of these programmes. Finally, in general a package of multiple interventions guided by a specifically trained multisectorial staff was used with expertise in malaria epidemiology and entomology, vector ecology, and land and water engineering. Recently developed malaria transmission models strengthen the evidence that substantial reductions of the entomological inoculation rate are possible when an integrated malaria control programme with multiple interventions (eg, environmental management tools and larvicide application) implemented simultaneously is used.74,75
Lessons learned from these successful malaria control programmes are readily available to guide effective and sustainable malaria control today. Besides modification of the human habitat, which can be applied with ease in all eco-epidemiological settings or environmental manipulation programmes in irrigation schemes, environmental management interventions are particularly appealing in urban areas, due to the highly focal nature of urban malaria transmission23 and in light of rapid urbanisation in malaria-endemic countries.76
Interdisciplinary collaborations between the health, agricultural, water, and infrastructure development sectors in the design of integrated intervention packages will be a feature of programmes that reduce the burden of malaria. Collaboration, in this context, means that the higher expenses of environmental management are not borne by the health sector and that malaria control per se is not the driving force. For example, the vector control impact of engineering works such as drains comes nearly free of charge to the health sector and is simply a byproduct of the design of the drainage system to also serve the needs of malaria control.
Finally, health systems have to be strengthened or rehabilitated and local personnel recruited and trained. The relevant personnel must include a good hydrologist, as in the successful control programmes during the early decades of the 20th century. This can be accomplished by training some engineering personnel in malariology, thereby building a natural inter-sectoral linkage. Effective control programmes should encourage community participation, and contain careful monitoring and adaptive modification to the intervention programme over time. With regard to evaluating the contribution of environmental management, as a component of a multi-intervention integrated programme, it is important to emphasise that here is an instance where historical controls at a given locality will be the basis for comparison with the malaria indicators following the introduction of a control programme at a designated locality.44 Cross-community comparisons are essentially ruled out due to the highly site-specific implementation of environmental management and, correlatively, the virtual non-existence of matched natural ecosystems. Systematic evaluation of environmental management is needed on a broad scale, and this activity will also provide opportunity for considerable methodological innovation. We look forward to seeing the results.
Back to top
Search strategy and selection criteria
These are described in detail in the Methods section.
Back to top
Conflicts of interest
We declare that we have no conflicts of interest.
Acknowledgments
J Keiser (project number PMPDB-10622) and J Utzinger (project number PPOOB-102883) are grateful to the Swiss National Science Foundation for financial support.
DOI:10.1016/S1473-3099(05)70268-1
Reducing the burden of malaria in different eco-epidemiological settings with environmental management: a systematic review
Jennifer Keiser (Research fellow)
a , Burton H Singer (Professor) b and J?rg Utzinger (Assistant professor) aSummary
Introduction
Background
Methods
Results
Discussion
Search strategy and selection criteria
References
Summary
The public health and economic significance of malaria is enormous, and its control remains a great challenge. Many established malaria control methods are hampered by drug resistance and insecticide-resistant vectors. Malaria control measures built around environmental management are non-toxic, cost-effective, and sustainable. However, there has been no comprehensive review of the literature or meta-analysis examining the effect of these interventions. We therefore did a systematic literature review and identified 40 studies that emphasised environmental management interventions and reported clinical malaria variables as outcome measures. Of these 40 studies, environmental modification (measures aiming to create a permanent or long-lasting effect on land, water, or vegetation to reduce vector habitats?eg, the installation and maintenance of drains) was the central feature in 27 studies, environmental manipulation (methods creating temporary unfavourable conditions for the vector?eg, water or vegetation management) in four, and nine quantified the effect of modifications of human habitation. Most of the studies (n=34, 85%) were implemented before the Global Malaria Eradication Campaign (1955?69), which mainly relied on indoor residual spraying with dichlorodiphenyltrichloroethane (DDT). In 16 studies that applied environmental modification and in eight studies on modification of human habitation, the risk ratio of malaria was reduced by 88?0% (95% CI 81?7?92?1) and 79?5% (95% CI 67?4?87?2), respectively. We conclude that malaria control programmes that emphasise environmental management are highly effective in reducing morbidity and mortality. Lessons learned from these past successful programmes can inspire sound and sustainable malaria control approaches and strategies.
Back to top
Introduction
Malaria continues to be a major public-health challenge. Globally, an estimated 300?660 million clinical Plasmodium falciparum cases occurred in 2002.1 In sub-Saharan Africa, malaria is responsible for 1?5?2?7 million deaths each year, of which at least 0?7 million are among children aged under 5 years.2 Acute febrile illness, anaemia, chronic debilitation, complications in the course and outcome of pregnancy, and delays in cognitive and physical development contribute to a heavy public-health burden, which has a negative impact on the social and economic development of affected countries.3?6
Key strategies in the control of malaria include early diagnosis and treatment, indoor residual spraying, and the use of insecticide-treated nets.7 Unfortunately, these control strategies are becoming less effective with the rapid development and spread of resistance to widely used drugs and insecticides. Exploiting the genomic sequences of the malaria parasites and vectors8 might ultimately lead to new generations of drugs and insecticides, the development of an effective vaccine,9 or genetically modified mosquitoes.10 However, these tools are not likely to be available for at least 10 years.10
Malaria control methods that involve environmental management, which is a modification or manipulation of the environment to reduce malaria transmission (eg, through the installation, cleaning, and maintenance of drains, or the systematic elimination of standing pools of water), currently receive far less attention. There is historical evidence that environmental management programmes effectively reduced malaria.11?13 In the early decades of the 20th century, engineering personnel worked alongside malaria control officers to ensure the implementation of environmental management that would deal simultaneously with waste removal and malaria suppression. However, environmental management interventions almost disappeared with the advent of dichlorodiethyltrichloroethane (DDT), which offered a standardised single attack, during the Global Malaria Eradication Campaign, spearheaded by the WHO. Environmental management is still readily available for implementation. Important features of environmental management strategies are their non-toxicity, relative ease of application, cost-effectiveness, and sustainability.14?16 Environmental management can have a major role within integrated vector management,17 adding resilience to the results of individual control strategies and reducing costs, and the likelihood of development and spread of drug and insecticide resistance.18
To our knowledge, there has been no systematic literature review and meta-analysis to examine the effectiveness of malaria control programmes that emphasise environmental management. Our aim is to fill this gap, and hence to establish an evidence base of the impact that environmental management has in preventing malaria-attributable ill health and death.
Back to top
Background
Environmental management
Drainage interventions to reduce fevers in settings near swamps or stagnant waters date back more than two millennia.19 However, an internationally recognised definition of environmental management for mosquito control was published less than 3 decades ago.20 In 1979, WHO put forth the statement that, ?Environmental management is the planning, organization, carrying out and monitoring of activities for the modification and/or manipulation of environmental factors or their interaction with man with a view to preventing or minimizing vector propagation and reducing man?vector?pathogen contact?.20 The major approaches of environmental management comprise: (1) environmental modification, which aims to create a permanent or long-lasting effect on land, water, or vegetation to reduce vector habitats; (2) environmental manipulation, which produces temporary unfavourable conditions for the vector; and (3) modification or manipulation of human habitation or behaviour, which reduces man?vector contact.20 These somewhat cumbersome articulations become more transparent if one thinks of environmental management interventions as the installation, cleaning, and maintenance of drains, vegetation management and river boundary modifications to promote flowing water, intermittent irrigation, and systematic elimination of standing pools of water (figures 1, 2, and 3). The specificities must, of necessity, vary with local ecosystem structure; therefore there is no single uniform environmental management recipe that is appropriate in all settings.
Click to enlarge image
Figure 1. Well designed and maintained drainage canals WHO
Click to enlarge image
Figure 2. Effective canal maintenance with community participation to ensure that canals are in a good shape and generally free from vegetation and silting (Thailand) WHO
Click to enlarge image
Figure 3. Concrete lined rice irrigation canal in central Laos Swiss Tropical Institute, 2004
Regarding malaria, environmental management can be coarsely grouped into four distinct eco-epidemiological settings: (1) malaria of deep forests, forest fringe, and hills; (2) rural malaria attributable to water resources development and management (eg, irrigation and large dams); (3) rural malaria attributable to wetlands, rivers, streams, coasts, and non-agricultural man-made habitats; and (4) urban and peri-urban malaria. Typical features of these eco-epidemiological settings and possible environmental manipulation and modification approaches, including changes in the human habitations, are summarised in table 1.
Click to view table
Table 1. Stratification of malaria into different eco-epidemiological settings and possible environmental management strategies
Malaria burden in different eco-epidemiological settings
Table 2 shows the total population, the population at risk of malaria, and the estimated number of malaria cases in sub-Saharan Africa, and the most severely affected countries outside of Africa. These variables are stratified into the four eco-epidemiological settings outlined above. The list of countries most affected by malaria was compiled by consultation of the regional WHO office websites.
Click to view table
Table 2. Estimated populations at risk of malaria and annual number of clinical malaria cases stratified by malaria eco-epidemiological settings in sub-regions and countries that carried the main burden of malaria in 2002
For the years 2002?03, we estimate that more than 1?9 billion people were at risk of the disease in sub-Saharan Africa and in 11 malaria-endemic countries outside Africa. Among these, at least 500 million live in urban and peri-urban settings, and 145?673 million people live in close proximity to irrigation and large dam sites. No precise information is available on the number of people living in rural or forested malaria-endemic areas. We estimate that 245?406 million clinical malaria cases occur in the countries considered here each year, with most occurring in rural areas. In urban and peri-urban settings at least 26?4 million clinical malaria events occurred. According to our estimates, 15?1?17?1 million cases are attributable to forest malaria in just five countries for which we could retrieve data (Brazil, India, Indonesia, Burma, and Bangladesh).
Back to top
Methods
Search strategy and selection criteria
We searched the following electronic databases, with no restriction on year or language of publication, to identify studies that assessed the effect of environmental management on pertinent malaria outcome measures: Medline, Web of Science, WHO Library Database, WorldCat and FirstSearch, and ScienceDirect. ScienceDirect includes full text articles of The Lancet and the Transactions of the Royal Society of Tropical Medicine and Hygiene dating back to 1823 and 1907, respectively. In addition, we hand-searched 15 volumes of the Tropical Diseases Bulletin (1930?45), which is a leading abstract database.
We used a broad search strategy to optimise sensitivity. We applied the following keywords: ?malaria?, ?alternate wet dry irrigation?, ?drainage?, ?engineering?, ?environmental management?, ?environmental manipulation?, ?environmental modification?, ?filling?, ?fluctuation?, ?flushing?, ?grading?, ?housing?, ?integrated vector control?, ?integrated malaria control?, ?intermittent irrigation?, ?levelling?, ?prevention?, ?prophylaxis?, ?sanitation?, ?sprinkler irrigation?, ?vector control?, ?vegetation management?, and ?water management?. We included all studies that used environmental management interventions exclusively or as the main feature of control, and that assessed malaria prevalence, incidence, spleen rates (prevalence of splenomegaly), or mortality rates. Bibliographies of the obtained articles were screened for additional relevant publications. Studies analysing entomological variables only and studies examining the effect of insecticide-treated nets on malaria were excluded. The key findings of the latter have been systematically reviewed in a recently updated Cochrane review.31
Statistical analysis
Studies obtained were stratified into three main groups of interventions: environmental modification, environmental manipulation, and modifications of human habitation. Descriptive studies and studies without defined sample size groups were excluded from the meta-analysis, but their key features are summarised in the tables. Data on intervention effectiveness (ie, malaria incidence, spleen rates, and mortality) were extracted and analysed by use of StatsDirect software (version 2.4.5; StatsDirect Ltd, Cheshire, UK). We calculated the risk ratio and 95% CI. We examined heterogeneity between trials with Cochrane's Q statistics and calculated I2 figures, which measure the proportion of variation in study outcomes due to between-study heterogeneity. We examined whether there was a publication bias by use of Egger's test (a small-study bias is evident if p
0?1).32 A random-effects model was used for calculation of the pooled risk ratio for environmental manipulation and modification, and modifications of human habitation, because the test of heterogeneity was highly significant for both groups of interventions (p0?001).33 Studies with a risk ratio below 1?0 indicate a reduction in clinical malaria variables in the intervention groups compared with controls.Back to top
Results
Our systematic review yielded 40 studies with an emphasis on environmental management interventions that reported clinical malaria variables as outcome measures. Environmental modification was the central feature in 27 studies (29 datasets; table 3) and environmental manipulation in four studies (table 4). Nine studies (ten datasets) measured the effect of modifications of human habitation (table 5).
Click to view table
Table 3. Studies that used environmental modification and appraisal of outcome in terms of clinical malaria manifestations
Click to view table
Table 4. Studies using environmental manipulation and appraisal of outcome in terms of clinical malaria manifestations
Click to view table
Table 5. Studies that used modification or manipulation of human habitation and appraisal of outcome in terms of clinical malaria manifestations
The identified malaria control programmes took place in 18 different countries or territories, in different eco-epidemiological settings, involving different anopheline vectors and different levels of endemicities. Figure 4 shows that most of the environmental management studies (n=34; 85%) were launched before the Global Malaria Eradication Campaign, which was implemented between 1955 and 1969, and mainly relied on indoor residual spraying with DDT. We could not identify a single study that applied environmental management as the central feature for malaria control during this period. Only six studies were initiated after 1969. The studies identified various outcome measures, such as malaria incidence, spleen rates, hospital admissions, or mortality rates. Overall, 24 datasets could be derived from 21 malaria control programmes to calculate the risk ratio. Among those, 16 applied a before-and-after evaluation design (comparing the pre-programme situation with the malaria situation during or after implementation of the interventions), and seven studies (eight datasets) compared an environmental management intervention group (all in the category modification or manipulation of human habitation) with a control group in which no intervention was implemented.
Click to enlarge image
Figure 4. Temporal examination of studies with environmental management (n=40) as the main feature of malaria control The year when the studies were launched is used for the analysis. Light blue bars: environmental modification (n=27); red bars: environmental manipulation (n=4); dark blue bars: modification or manipulation of human habitation (n=9).
Environmental modification
Prominent environmental modifications were as follows: drainage; filling of swamps, borrow pits, pools, and ponds; modification of river boundaries; or other engineering approaches. Multiple interventions were implemented in several studies, such as environmental modification in combination with oiling and the application of Paris Green and, for part of the population, distribution of bed nets, quinine prophylaxis, and treatment, as well as health education. The risk ratio estimates of environmental modification were found to range between 0?017 and 0?45 (table 3). Heterogeneity of the studies was very pronounced (p
0?001; I2=99%). The random summary risk ratio measure of 16 environmental modification studies plus one environmental manipulation study was calculated as 0?120 (95% CI 0?079?0?183), giving a summary protective efficacy of 88% (figure 5). There is evidence of a publication bias (Egger's test: −15?07; 95% CI −29?76 to −0?38; p=0?045).Click to enlarge image
Figure 5. Random effects meta-analysis of environmental modification and manipulation interventions against malaria Squares=relative risk. The size of each square denotes the sample size. Diamond=combined relative risk. Horizontal lines=95% confidence interval.
William Gorgas, Malcolm Watson, and Nicolaas H Swellengrebel were three of the pioneers who implemented environmental modification measures in rural and urban areas of the Panama Canal zone, peninsular Malaysia, and the Indonesian archipelago, respectively. We briefly summarise these successful malaria programmes below.
Panama
In 1878, Ferdinand de Lesseps (the French engineer who led the construction of the Suez Canal) and his team began to dig a canal across the Isthmus of Panama, which was then part of Colombia. However, yellow fever, malaria, and engineering challenges halted construction of the canal. Between 1881 and 1889, an estimated 22
189 labourers of the French Canal Company died. The area was covered with dense jungle, streets were unpaved, and there were many puddles. Furthermore, the towns were surrounded by swamps, and there was no water supply or sanitation. After the Americans took over the project and Gorgas was placed in charge of the health programme, environmental management was adopted as the mainstay of malaria control. This mainly consisted of temporary and permanent drainage infrastructure (ie, installation of concrete-bottomed drains and subterranean pipes) and vegetation management. The malaria incidence among the employees decreased from 821 per 1000 in 1906, to 14 per 1000 in 1917.11,40Malaysia
In 1901, the burden of malaria among the 4000 inhabitants of Klang Town and Port Swettenham, in present-day Malaysia, was enormous: half of the patients in the hospitals had malaria, which led many employees to leave Port Swettenham.12,19,56?59 Although the two main malaria vectors were not known at that time (Anopheles umbrosus in Klang and Anopheles sundaicus in Port Swettenham), detailed entomological surveys and examination of the spatial distribution of malaria cases were carried out. The principal anopheles breeding habitats in Klang, for example, were permanent swamps, open wells, and pools along the roads. Due to a lack of compliance and difficulties with supervision, the use of bed nets, house screening, and quinine prophylaxis were thought to be impractical for the Chinese and Indian rubber planters.19 Consequently, engineering approaches were implemented. The drainage of swamps complemented with vegetation management (ie, clearing of undergrowth, forest, and mangroves) successfully reduced mortality and malaria incidence (risk ratio estimate of 0?12) and malaria-attributable mortality. Following these effective environmental management control approaches, environmental modification works were widely and successfully applied in other parts of Malaysia, notably in the rubber estates,12 Singapore,12,61 and Hong Kong.11
Indonesia
Between 1920 and 1935, species sanitation was widely practised on the Indonesian archipelago. This concept emphasises the selection of environmental management strategies after studying the ecological niche and breeding habits of the principal malaria vectors.13,61 Hence, the antimalarial campaigns in Indonesia consisted of several stages. First, the spatial distribution of clinical malaria variables was examined by collecting adult anopheles to draw correlations between malaria distribution and its principal vector. Second, the natural infection rate of the local vector(s), including important breeding habitats, was examined. Environmental management was then targeted to settings where a principal malaria vector with a distinct breeding habitat had been identified. Six Indonesian studies showed the effect of environmental modification, namely drainage or filling of swamps, tidal zones, irrigation schemes, or other water bodies (table 3). In all six settings spleen rates were substantially reduced.13
Importantly, environmental modification in combination with the use of bed nets and quinine among European workers was also found to be largely successful in two programmes carried out in high transmission areas of sub-Saharan Africa, among the most efficient malaria vectors (ie, Anopheles funestus and Anopheles gambiae).16,44,45 Both programmes covered relatively small populations: mining communities in Northern Rhodesia (present-day Zambia), and military personnel at the Royal Air Force station in Apapa, Nigeria. The risk ratio estimate of environmental management programmes in four Zambian copper mining communities was calculated as 0?35. These studies were also remarkably cost-effective and spurred unprecedented economic development.16,44
Back to top
Environmental manipulation
Several studies have assessed the effect of environmental manipulation on the anopheles population, with most focusing on water management, such as intermittent irrigation (alternate wet-dry irrigation).62?64 With regard to clinical variables of malaria, we found only four studies that assessed the effect of environmental manipulation, three studies that adopted water management, and one study that reported the effect of vegetation management (table 4). These programmes were implemented in China, India, North Borneo, and USA. Only one study presented data that could be used to estimate the risk ratio. Intermittently irrigated rice fields in six villages in India, in combination with environmental modification, such as the filling or draining of borrow pits in the villages, reduced the spleen rate in children from 48% in 1937, to 4% in 1941, giving a risk ratio estimate of 0?08. These interventions also reduced the parasite rate from 42% in 1937, to 0% in 1941.47
Interestingly, more than 50 years later, intermittent irrigation is being promoted extensively in Sichuan province, China. For example, in Xindu district and Quionglai county, 99?7% and 93?4% of the rice fields, respectively, are now irrigated in this way. Vector breeding has been reduced accordingly, and no malaria cases have been observed in these two settings.50
Modification or manipulation of human habitation
The relation between the quality of housing and good health has a long history. The first trials showing that people could be protected from malaria by mosquito proofing their homes were done between 1899 and 1904 along two railway lines in Italy (Prenestina-Cervara and Pontegalera). Houses in the intervention group had their windows covered with frames of tulle or muslin, and doors were screened with wire gauze, whereas the houses of the control groups were left unprotected. In 1899, 17% (four of 24) of the individuals in the intervention group compared with 96% (exact numbers not known) in the control group contracted malaria. These findings were translated into broader public health action, and by 1904, 12
378 people already lived in mosquito-proofed houses. Protection from malaria was observed, even if only parts of the houses were screened. Whereas 25?96% of the inhabitants of non-mosquito-proofed houses had malaria, only 1?9% of individuals in completely screened houses and 10?9% in partly mosquito-proofed houses became infected.65 The practice of mosquito proofing houses began to make inroads into the tropics and subtropics with European settlers and military personnel. For example, malaria cases were greatly reduced among British troops stationed in India who lived in mosquito-proofed barracks, and among South African children living in screened houses (table 5). Heterogeneity of these studies was significant (p0?001; I2=87%). The risk ratio estimates of these studies were found to range between 0?06 and 0?54. A pooled random risk ratio estimate was calculated as 0?205 (95% CI 0?128?0?326; figure 6). There is evidence of a publication bias (Egger's test: −5?5; 95% CI −10?7 to −0?44; p=0?037).Click to enlarge image
Figure 6. Random effects meta-analysis of modification of the human habitation against malaria Squares=relative risk. The size of each square denotes the sample size. Diamond=combined relative risk. Horizontal lines=95% confidence interval.
By contrast with the aforementioned studies, Indonesia showed an increase in malaria-related mortality after the implementation of house improvement programmes. Malaria outbreaks were observed in 44 of 71 districts with improved housing. Possible explanations included the creation of borrow pits accompanying house renovations, the neglect of cleaning up operations, and the immigration of workers carrying parasites in their blood. The improved houses had better ventilations and were often smoke-free, which in turn might represent more suitable resting places for adult anopheles mosquitoes.66
A negative aspect of the Global Malaria Eradication Campaign was that during and after the campaign little attention had been paid to the relatively straightforward task of modifying human habitation. We are not aware of any recent malaria control programmes that have used human habitation modification in a systematic manner to reduce malaria morbidity and mortality. However, several studies have analysed the relation between house design and mosquito densities. Recently, Lindsay and colleagues51 published an excellent summary of these entomological studies. In Sao Tome, twice as many A gambiae were found in houses built at ground level compared with houses built on stilts.67 Another set of studies identified housing risk factors for malaria. These studies compared clinical and entomological variables of malaria in either poor or well-constructed housing. Houses with closed eaves, ceilings, and in good condition had fewer mosquitoes, and inhabitants had lower malaria incidence than those living in houses with open eaves, no ceilings, or that were poorly constructed.51,68,69
Back to top
Discussion
There is growing recognition that systematic reviews and meta-analyses have important roles in evidence-based medicine and therefore contribute to decision making with the aim of using scarce public-health resources in a cost-effective manner.70 However, by contrast with malaria chemotherapy, the use of insecticide-treated nets, or indoor residual spraying, the impact of environmental management control strategies to reduce the malaria burden has not been systematically reviewed to date. We made an initial attempt to fill this gap through a systematic review and a meta-analysis of malaria control programmes, emphasising environmental management as their main feature. We found very high summary protective efficacies on clinical malaria parameters (79?5?88?0%) regardless of whether environmental modification, environmental manipulation, or modification or manipulation of human habitation or behaviour was used. Furthermore, high protective efficacies were also achieved in high transmission areas, with the most efficient malaria vectors. It is important to note that most of the studies reviewed here were site specific, often quite large programmes, implemented over a period of several years, with adaptive tuning of interventions over time, but covering a relatively small subset of the national level at-risk populations. This is by contrast with most clinical trials, where interventions are standardised, fixed in advance, but only last for relatively short periods of time.
We found only one study, the malaria control programme in Mian Mir (now Lahore, Pakistan) implemented from 1901?09, for which no conclusion could be drawn on the effect of environmental management on clinical malaria parameters.34,71,72 The experiences made in Mian Mir continue to be debated in the contemporary literature.19,44,72 It was argued, for example, that the malaria control programme in Mian Mir failed because inappropriate control measures were adopted, inexperienced workers employed, insufficient larvicides used, and no protection zone for dwellings established.34
All the other studies reported very high protective efficacies when environmental management was used for malaria control. The results and the potential for contemporary malaria control programmes require scrutiny. The following issues are offered for discussion. First, we cannot rule out that we have overestimated the effectiveness of environmental management interventions. One reason is that the number of environmental management control studies identified is relatively small. Second, our statistical results suggest that there might be publication bias, and thus successful interventions were more likely to be reported in the literature. In addition, it was impossible to scrutinise the methodological quality of the studies as most of them were implemented 50?100 years ago. Furthermore, most of the malaria control programmes based on environmental management were to some degree accompanied by other control measures, which probably contributed to their overall success. For example, oil-based antilarval products were applied in combination with environmental management in a number of studies. Since several control measures were put in place concurrently, it is intrinsically impossible to tease out the independent effect of environmental management. However, in several places, as for example in Apapa, Nigeria, no change in malaria rates could be observed by applying quinine, house screening, and bed nets alone. Once environmental management was put in place, together with the former control programme, malaria incidence sharply decreased.45
Even if the true protective efficacy of environmental management is somewhat lower than reported here, it is crucial that we are aware of the lessons that can be learned from past campaigns. Their integrated nature, inter-relations with water and vegetation management, and organisational frameworks are of central importance for malaria control today.
The success of the malaria control programmes reviewed can be summarised by the following features. First, most programmes emphasised the permanent elimination of breeding sites. Second, environmental management tools used were idiosyncratic to the individual settings and vector. For example, if the local vector favoured a wide variety of remote and temporary breeding sites, then modification of human habitation would be the environmental management strategy of choice. Third, most studies lasted several years, allowing continuing performance evaluations of the control tools and adaptive adjustment to enhance performance.73 Fourth, community participation and usually also health education were integral parts of these programmes. Finally, in general a package of multiple interventions guided by a specifically trained multisectorial staff was used with expertise in malaria epidemiology and entomology, vector ecology, and land and water engineering. Recently developed malaria transmission models strengthen the evidence that substantial reductions of the entomological inoculation rate are possible when an integrated malaria control programme with multiple interventions (eg, environmental management tools and larvicide application) implemented simultaneously is used.74,75
Lessons learned from these successful malaria control programmes are readily available to guide effective and sustainable malaria control today. Besides modification of the human habitat, which can be applied with ease in all eco-epidemiological settings or environmental manipulation programmes in irrigation schemes, environmental management interventions are particularly appealing in urban areas, due to the highly focal nature of urban malaria transmission23 and in light of rapid urbanisation in malaria-endemic countries.76
Interdisciplinary collaborations between the health, agricultural, water, and infrastructure development sectors in the design of integrated intervention packages will be a feature of programmes that reduce the burden of malaria. Collaboration, in this context, means that the higher expenses of environmental management are not borne by the health sector and that malaria control per se is not the driving force. For example, the vector control impact of engineering works such as drains comes nearly free of charge to the health sector and is simply a byproduct of the design of the drainage system to also serve the needs of malaria control.
Finally, health systems have to be strengthened or rehabilitated and local personnel recruited and trained. The relevant personnel must include a good hydrologist, as in the successful control programmes during the early decades of the 20th century. This can be accomplished by training some engineering personnel in malariology, thereby building a natural inter-sectoral linkage. Effective control programmes should encourage community participation, and contain careful monitoring and adaptive modification to the intervention programme over time. With regard to evaluating the contribution of environmental management, as a component of a multi-intervention integrated programme, it is important to emphasise that here is an instance where historical controls at a given locality will be the basis for comparison with the malaria indicators following the introduction of a control programme at a designated locality.44 Cross-community comparisons are essentially ruled out due to the highly site-specific implementation of environmental management and, correlatively, the virtual non-existence of matched natural ecosystems. Systematic evaluation of environmental management is needed on a broad scale, and this activity will also provide opportunity for considerable methodological innovation. We look forward to seeing the results.
Back to top
Search strategy and selection criteria
These are described in detail in the Methods section.
Back to top
Conflicts of interest
We declare that we have no conflicts of interest.
Acknowledgments
J Keiser (project number PMPDB-10622) and J Utzinger (project number PPOOB-102883) are grateful to the Swiss National Science Foundation for financial support.
Comment