**6. Life history analyses to identify otherwise non-obvious vector control challenges and opportunities**

majority of human populations, most people sleep indoors at night for security reasons, so very little of the biting activity measured outdoors is relevant to normal human exposure patterns [60]. Taking the major African malaria vector *An. gambiae* as an example, the traditional narrative describing it as *endophilic* is inaccurate, because their biting rates indoors and outdoors are usually similar and they have no strong or consistent preference for attacking people in either location [50]. It is the timing of biting activity that caused most historical exposure to occur indoors. This human-specialized mosquito species usually exhibits biting activity peaks that occur in the middle of the night when most people are asleep indoors (**Figure 5A**), and therefore vulnerable to attack unless protected with an LLIN. It is therefore more accurate to say that these vectors are highly *nocturnal*, feeding mostly at times when humans exhibit strong *endophilic* tendencies [35], and that is why most human exposure occurred indoors at

Contrasting with vector populations like those of *Anopheles farauti* in the Solomon Islands, where humans are mostly exposed to outdoors, the most important feature of this behavioural interaction is again the timing of host-seeking activity. By feeding predominantly in the evening, this species can readily attack humans indoors or outdoors while they are still awake and active so bed net use is impractical (**Figure 5B** and **D**). Again, biting densities are similar indoors and outdoors (**Figure 5B**), so it is inaccurate to describe this vector as *exophagic* in the strict sense, and much more important to emphasize that so much exposure occurs outdoors because it is *crepuscular*, with feeding activity that peaks at dusk when people are awake,

The overall exposure distribution estimates represented by the areas under the curves in **Figure 5C** and **D**, can then be combined with direct field estimates for the proportion of bloodmeals obtained from humans to visualize the maximum limit of *biological coverage* [63] achievable with human-targeted measures like LLINs as simple box graphs (**Figure 6**).

**Figure 6.** Box diagrams illustrating how the two different vectors described in **Figure 5** differ in terms of their overall behavioural vulnerability to population suppression with long-lasting insecticidal nets, expressed as the maximum

achievable *biological coverage* of blood resources used by the mosquitoes [6].

night before the scale up of LLIN use (**Figure 5C**).

active and cannot use bed nets.

414 Towards Malaria Elimination - A Leap Forward

The use of models to look at entire life histories of mosquitoes was central to the very earliest work of Ross [64], and to the ethos of *epidemiological entomology* defined by Garrett-Jones half a century ago [65]. While direct interpretation of simple indicators should provide the essential core of evidence used to inform programmatic decisions, astute application of analytical models to examine the life histories of mosquito populations can also yield important insights that would not otherwise be obvious.

For example, the slow-acting toxicity of pyrethroids to mosquitoes that are clearly resistant to this insecticide class was only recently identified as being central to the sustained impacts of LLINs [59]. While most African *Anopheles* populations are now sufficiently resistant against pyrethroids to survive immediately after exposure, they do suffer increased mortality over the longer term, essentially all of which occurs within the 10 days required for the parasite to complete sporogonic development [59]. As a result, while pyrethroid resistance clearly does compromise the impacts of LLINs [66], it falls far short of abrogating them entirely, so they remain an invaluable tool for malaria vector control [59].

Also, fitting process-explicit models of mosquito population dynamics to vector density trends may yield some insights that cannot be obtained by direct interpretation. Such mechanistic modelling analyses have been successfully applied to field data to identify negative density-dependence of mosquito reproduction, which make vector populations more robust to control than would otherwise be expected [67]. Similar models have been fitted to the population trajectories of self-propagating populations in large cages, which were experimentally exposed to different vector control measures and combinations thereof. These biologically-informative analyses quantified impacts on specific target parameters like survival and fecundity, helped confirm that near-extinction of these small populations was achieved, and revealed a surprising mode of action for one of these emerging technologies (Ng'habi et al., Unpublished). Such approaches could be readily extended to data from routine population dynamics monitoring, allowing the complementarities, synergies and redundancies achieved by combinations of vector control measures to be understood at an unprecedented level of detail.

Relatively simple deterministic models have also been used to illustrate how insecticide resistance traits and intervention avoidance behaviours can interact synergistically, allowing resilient mosquito populations to persist despite widespread LLIN use without necessitating any major adaptations of their preferred feeding times [68]. By foraging cautiously and repeatedly inside houses, to maximize their feeding opportunities while minimizing their contact with LLINs, even nocturnal species like *An. arabiensis* can continually search around from one house to the next until an unprotected non-user is located [68]. By combining endophagy with exophily in this way, *An. arabiensis* can achieve feeding success rates despite high LLIN coverage that are only a quarter lower than in the absence of nets [68]. Furthermore, the resilience of such nocturnal but behaviourally plastic species may be further enhanced by physiological resistance to insecticides and opportunistically feeding upon animals, resulting in redistribution of feeding activity onto a combination of livestock and humans who either lack nets or are encountered outdoors at times when they are unprotected [68, 69].

However, life history analyses of how resilient mosquito species survive despite high LLIN coverage also identifies some exciting intervention opportunities that would not otherwise be obvious. For example, the most direct corollary of the observation that mosquitoes forage cautiously through several houses to find an unprotected human is that this creates enhanced opportunities to kill them if more effective indoor control methods can be deployed [7, 68, 69]. Emerging options for doing just that range from insecticidal eave tubes [70] and eave baffles [71] to untreated entry traps [72] and three-dimensional window screening [73].

More detailed consideration of life history distributions for the same vector population also reveals an even more counter-intuitive opportunity for such housing modifications to have an impact upon residual transmission. By the time a female *An. arabiensis* is old enough to have incubated malaria parasites through to infectious sporozoites, she will usually have completed at least 4 gonotrophic cycles, during which time she will most probably have been inside a house at least once [69]. So even though approximately half of all transmission events occur outdoors, they are all preceded by at least one house-entry event during which the guilty mosquito may be killed [69]. It is therefore possible to reduce levels of malaria transmission occurring outdoors using interventions that target mosquitoes when they enter or attempt to enter houses [69].

To a large extent, these geographic inequities of data analysis capacity are an understandable consequence of pre-existing global poverty, education and opportunity patterns. However, accepting the *status quo* illustrated in **Figure 7** is not consistent with the 'think global, act local' ethos of successful malaria elimination programmes, and these global capacity inequi-

**Figure 7.** The global geographic distribution of current members and collaborators in the Malaria Modelling Consortium (MMC), as well as analytical contributors to the World Health Organization 2015–2017 World Malaria Report (WHO-WMR),

Entomological Surveillance as a Cornerstone of Malaria Elimination: A Critical Appraisal

http://dx.doi.org/10.5772/intechopen.78007

417

If the strategic vision presented by the global modelling community in **Figure 8** continues to be implemented in the context of the world map in **Figure 7**, several consequences are

**1.** Malaria-related data will be collected in low-income countries but largely analyzed in

**2.** Collectors of malaria-related data will have insufficient opportunity and training support to analyze their own data, develop their analytical skills and influence policy and practice. The data interrogation processes essential to timely use and effective quality control of

**3.** Analysts of malaria data will continue to live far away from the point of data collection and the programme staff who collect it in the field, so their ability to critically analyze and interpret it will remain limited by lack of hands-on field experience and direct access to

**4.** These two communities will remain separated by thousands of kilometers, as well as their very different roles and perspectives (**Figure 9A**). The synergistic interface required between human beings to achieve optimal data collection processes, critical analyses and

surveillance data will remain underdeveloped where they are needed most.

appropriate programmatic responses (**Figure 9B**) will not be realized.

ties need to be addressed urgently if malaria is ever to be eradicated.

overlaid upon a map with contemporary malaria endemicity (white).

high-income countries with no malaria problem to speak of.

inevitable:

those who have it.

More strategically, this particular simulation analysis [69] also suggests a thematic perspective that may be useful to apply more broadly to life history analyses. It may often be more valuable to look for opportunities to intervene early in the life cycle of mosquitoes rather than targeting transmission events occurring when they are far older. The life histories of adult mosquitoes are cyclical so targeting mosquitoes when they engage in frequently repeated behaviours, in this case house entry, can have far greater impact than would be obvious from face-value interpretation of the fraction of single feeding events that occur indoors.
