Version classiqueVersion mobile

La lutte antivectorielle en France

Didier Fontenille
Christophe Lagneau
Sylvie Lecollinet
et al.

Assessment of vector control

Practices and methods1

Texte intégral

  • 1 In this chapter, references to the CD-ROM concern question 7: “What are the vector control assessm (...)

1Assessment consists of checking that the objectives of actions undertaken are being met, with the aid of the appropriate tools (see Box). The indicators used to assess the impact of vector control can therefore be direct (such as the morbidity index) or indirect ones (decrease in entomological indices, acquisition of protective behaviours), although the latter cannot necessarily be associated with a fall in morbidity;

2Therein lies the great difficulty of vector control. The current indicators of effectiveness are not sufficiently well defined for every vector system and often have not been formally validated. The indicators for measuring the process performance and the cover achieved must therefore be adapted to the type of intervention (biomedically based, environmental, educational, behavioural or social). The spectrum of different vectors covered by control operations is wide and hence indicator choice will vary according with the nature of the measures employed.

Objectives of assessment
■ Check that the actions undertaken have been carried out exactly as specified (process assessment).
■ Check that the target population (human or vector) has been reached (target cover).
■ Make sure that the desired health impact among the public (reduction of entomological indices, incidence of the disease and prevalence of parasitaemia and so on) has been produced (impact assessment).
■ Verification that no serious undesirable effect (side-effects from insecticide use, for instance) has arisen that could be attributed to the operation conducted.


3One of the issues concerning the assessment of the entomological and health impact of vector control is to know how much of a desired reduction in morbidity indicators (human or animal) has been achieved in the field, as a result of the actions performed. Another question is the quantification of such a decrease. Moreover, the proportions to which the observed impact is attributable to vector control must be determined.

4These issues lead to two major assessment objectives. The first entails following-up a whole series of standardized pre-defined indicators, when the vector control is implemented, to check that the goals of the actions undertaken have been reached (see examples on CD-ROM). The second consists in judging how far the results observed can be attributed to the control programme realized. Therefore it is necessary to draw up a specific, experimental or quasi-experimental, assessment plan in order to answer this question.

5These two major objectives relate to two different situations. The first is valid for cases where the action applied has already been drawn up under specific definition and standardization (involving indicators and thresholds) and where it is used regularly according to a definite plan. The second relates to a more experimental situation involving cases and controls, for which broad-scope innovatory actions are put in place (with new techniques, new insecticides or a new strategy) such as for large-scale treatment.

Theoretical model

6The ideal theoretical model in epidemiology is the experimental model that allows comparison between two groups of individuals (human or animal) selected following a randomisation principle. Only one of the groups is subjected to the recommended action, so that all the differences between the two groups are then attributed to the action conducted. This ideal setup is difficult to obtain, especially for vector control investigations. Either a case/control experimental system cannot be developed (an epidemic situation), or the study has to rely on indirect entomological indicators (with no close knowledge of relationships between entomological indicators, incidence and risk).

Table 8 – Basic set-up for experimental model

Table 8 – Basic set-up for experimental model

The values Oi designate observations made before and after programme X. Formulae of the type O1 – O2 must be interpreted as theoretical indications rather than deviations between the two observations. In the ideal assessment model, the result attributable to the programme is evidently O3 -O4.

7Nevertheless this model is useful for the purposes of comparison and explanation (see Table). It can serve as underlying proof for deciding whether or not results can be attributed to a programme for all the assessment models. If identification of the causal inferences is required, identical groups must be “compared”, except for exposure to a programme (before and after). In experimental (or random) models, an attempt is made to ensure the initial equivalence of the two groups by randomly distributing the individuals into two experimental groups, one of participants and the other a control group. And so the two groups to compare are equivalent (O1 = O2).

8Of course, the validity of the model, in other words its ability to provide elements for drawing conclusions on whether or not the result observed can be attributed to the programme, is strengthened by the fact that there is a control group and “before” and “after” measurements (see the details on the control groups on CD-ROM). If a non-random control group is used, the model is “quasi-experimental”.

Examples of application

9Examples of application of this “experimental” model to vector control are the subject of a number of publications (referenced in the database PubMed). This is particularly the case for trials of randomized community operations in the assessment of the health impact of a range of strategies, including assessment of insecticide-impregnated mosquito nets (against malaria, dengue), and an assessment project on a method of household protection by a screen against mosquito bites in Gambia. Some quasi-experimental assessments have also been conducted: assessment of an aerial adulticide treatment operation against Culex during a West Nile virus epidemic, assessment of impregnated and non-impregnated mosquito nets, of an insecticide treatment against malaria for housing, and of a combination of these measures.

10Other publications deal with the case where surveillance contributes to the pragmatic assessment (monitoring) in parallel with the implementation of a more-or-less complex control programme (for example, the treatment of bouts of malaria combined with the use of impregnated mosquito nets). Still others focus on the process or on the actions conducted for children (see the details on CD-ROM).

11There have also been some systematic reviews of scientific literature regarding vector control strategies (see CD-ROM). The conclusion of one of them challenges the evidence sometimes accepted for the efficacy of a community approach (on this approach, see also the Box on “Community participation” in the chapter “Perception of risk in relation to vector control communication”).


12The account that follows is the fruit of a detailed consultation of the scientific literature on the databases (Current Contents, PubMed, Medline, Science direct) and RESE (Réseau d’échanges santé environnement, ‘Health and Environment Exchange Network’) and also of an examination of mission reports of the IGAS (Inspection Générale des Affaires Sociales, ‘General Inspectorate for Social Affairs’ or the IRD group report on dengue in the French Departments of America (2003) and foreign literature. Finally, a report by a public health engineer published in October 2008 (source: Julien Fecherolle) seems to give the most exhaustive and all-embracing review of the French situation.

State of the art

13The State of the art of assessments on vector control in French territory shows up a relative paucity of literature. There are many studies on the efficiency of a measure (insecticide, impregnated mosquito net, repellent).The sector of research on the assessment of new biological means has been developing strongly over the past number of years. However, no article specifies clearly the methods and tools to use for assessing the effectiveness of vector control in the country. The existing documents come from WHO (protocols to follow, relevant indicators). The collaborative centre for WHO on French territory is IRD’s unit ‘Caractérisation de contrôle des populations de vecteurs’. Its major purpose is to determine, in the laboratory and through field work in Africa, the effectiveness and persistence of insecticides and materials that could be used in public health measures against mosquitoes.

What indicators?

14Most of the entomological indicators concern Aedes aegypti (and by extension Aedes albopictus). No indicator exists for anopheline vectors of malaria, ticks or culicoids and so on. Some RESE documents give reports on assessment practices, but these have often been applied under the pressure of crisis conditions, notably at the time of the 2005-2006 chikungunya epidemic in Reunion Island. However, the IRD group report cited previously gives the basic elements for assessments of control actions against Aedes aegypti. Although no detailed methodology is given, the report emphasizes the indicators to be used to measure vector density. In fact, if there are no indications on the epidemic risk, vector abundance estimation and follow-up are essential for subsequent assessment of vector control actions: variations in this parameter then represent a criterion of efficiency.

15In order to do this, there remains the question of sampling. Where human pathologies are concerned, sampling of females coming to bite an individual is a priori the sole method that enables investigators to count and examine them at the moment they might transmit the virus. Yet there are two major drawbacks to this way of sampling. For one thing, there must be a large number of surveillance points. The other disadvantage is the non-negligible risk of transmission of the dengue virus to the survey operators, hardly acceptable on ethical grounds. Experiments on traps substituting for the “human bait” have therefore been run (CO2 traps, egg-laying traps, mosquito net traps, and traps with animal baits). Although this is only an approximate form of follow-up, it gives some idea of the densities of populations and their variations in time and space. As for diseases transmitted by ticks to humans, and also especially to livestock, the sampling methods must still be improved and the indicators better defined (in terms of presence, density, threshold).

16Another means of assessing vector density is the sampling of pre-imaginal populations. That is possible essentially for mosquitoes. In this regard, exhaustive counting of larvae, even limited to samples form breeding sites, is extremely long and tiresome. The WHO criteria, used throughout the world, lie rather on the assessment of the number of breeding sites harbouring larvae (on the basis of sampling from districts or housing). The three criteria commonly used up to now are the house index, the Breteau index and the container index. Breeding site productivity is nevertheless highly disparate, depending on their location (flower pot, water container). This prompted the proposal of a new index, the productivity index or weighted Breteau index, whose generalized use has been recommended by the IRD expert report.

What regulations?

17The reform introduced by article 72 of law n° 2004-803 very precisely confers the responsibility for entomological follow-up to the State. It is moreover stipulated in article R. 3114-5 “the entomological surveillance of insect vectors and in particular the surveillance of resistance of these to insecticide products […] remains within the competence of the prefect”. It is for the State therefore to ensure the efficacy of these products. Nevertheless, the latter does not always have this competence at its disposal. It can therefore transfer this responsibility by order to the service in charge of control.

18An example is: article 11 of the prefectoral order of Haute-Corse bearing on the delimitation of mosquito control zones (mainly of Aedes albopictus) emphasizes that “the services in charge of mosquito control undertake a follow-up of control actions with monthly checking of the effectiveness of the treatment in terms of breeding site productivity, this information being recorded on the site report sheet. They send the prefect the information necessary for assessing the system and notably, the information […] regarding vector sensitivity” (source: DASS of Haute-Corse, 2007). Logically, this information must be sent to the Departmental Health and Social Affairs Service (DDASS).

19Finally, in the framework of the anti-proliferation plan against chikungunya and dengue in metropolitan France, the objective of entomological surveillanceis “in areas where the mosquito has been identified or indeed is installed, to estimate the density of vectors and follow up the effectiveness of proliferation control actions” (source: sheet 1 relating to entomological surveillance).

What deficiencies?

20The report of the interministerial mission, jointly involving the Inspection générale de l’administration (IGA), IGAS and the Inspection Générale de l’Environnement (IGE), which examined the reorganization of the vector control services (2006) observed “the flagrant lack of methods and tools for assessing the effectiveness of vector control”. It adds: “To date there is nothing to demonstrate that it is effective”.

21The same year, and during the chikungunya epidemic, another report from a support mission against the chikungunya epidemic in Reunion Island brought up the difficulty of assessing the effectiveness of the actions undertaken. The authors denounce “the lack of tools for following-up the vector densities” and “the absence of data on the sensitivity of the vector Aedes albopictus”.

Publication of a specialist review

22The first review giving an exhaustive account of the French situation was published recently (October 2008). The enquiry, conducted by Julien Fecherolle in the context of his health engineering report, had the main objectives of asking the services if and how they assessed the effectiveness of their actions, on the one hand, and to learn about their expectations regarding assessment, on the other. This enquiry was based on a questionnaire sent to the persons concerned (about 20 open questions, see CDROM) and on telephone interviews. The questions essentially dealt with assessment concerning the entomological aspects. Others related to assessment of the structure (in terms of equipment and facilities, also the level of knowledge of operators) and processes have been added.

23The services contacted were those most strongly involved in the operational side of vector control and mosquito destruction (according to the law of 13 August 2004), amounting to 16 in all: public mosquito-eradication organizations in mainland metropolitan France, département councils and bodies belonging to the Ministry of Health for Corse-du-Sud, Mayotte, Guadeloupe, Martinique and French Guiana. French Polynesia, New Caledonia, Wallis-and-Futuna were not called on to participate in the enquiry.

24This work was a means of asking services about the frequencies and expected results of effectiveness assessment of actions in favour of the operational objectives of vector control, on assessment methods of the efficacy of actions taken (entomological and sociological indicators, thresholds and measurement methods for Aedes aegypti, Aedes albopictus, the Anopheles and nuisance mosquitoes, sensitivity to insecticides and effectiveness of biocidal products) and on assessment of personnel and facilities and equipment (see the results of the enquiry by operator and by vector species on CD-ROM).


25The assessment of public health actions involves a culture of assessment and a specific methodological approach which have always been the prerogative of English-speaking countries. However, in France a structural deficiency exists in this domain, including for vector control.

26In the veterinary services field, everything remains to be done: there is no assessment structure for veterinary aspects of vector control. Local or individual vector control actions cannot be assessed, as no national or regional strategy or direction, or indicators, exist.

27There is no assessment culture in France, nor any experience sharing in public health, and even less so for vector control which was not a priority issue until recent years. Assessment involves multidisciplinary approaches with an external view: with actors in such fields as vector control, human and animal health, the environment, and also qualitative and quantitative disciplines.

28Different services have different conceptions of assessment, as well as widely differing knowledge and interpretations of efficacy assessment. This was one aspect revealed by Julien Fecherolle’s enquiry: services do not all use the same terms to explain the same things. Assessment of public health actions requires a culture and method that are not rooted in the vector control services in France.

29The actors confuse “follow-up” and “assessment”. The same enquiry shows that the actors involved do not distinguish between the continuous follow-up of indicators and the assessment of the effectiveness of actions taken. But the idea is not just to obtain entomological data and collect them, but also to analyse, compare and interpret them. In other words, it is to make the information “live” so that a judgment can be made. and a certain number of deficiencies are particularly detrimental: few tools exist, and regulatory texts are sometimes difficult to interpret (some talk of an “effectiveness follow-up”, others of “assessment of the system” and so on). In short, it is too rare that the operational objectives are clearly defined and measurable, which makes it difficult to fix a threshold for efficacy. Indicators are sometimes non-existent or unreliable and the degree to which they are really appropriate is rarely checked. They are also as varied as the local entomological and epidemiological situations that occur. Mosquito density and complaints about them are nevertheless relevant criteria, but need improving.

30The recommendations are not always applied. This is the case of the recommendations proposed by the group expertise report on dengue in the DFAs which were only partly put into practice. Yet these recommendations were relevant for the follow-up and assessment of the effectiveness of vector control actions against Aedes aegypti. The same applies to those put forward by the IGA-IGAS-IGE mission of 2006. For example, the use of the weighted Breteau index has still not been made general practice in the DFAs.

31Some encouraging experiences deserve to be mentioned, however. Assessments have been conducted sporadically: on “Kass moustique” in Reunion Island, the “Toussaint” events in Guadeloupe. These assessments have sometimes had an effect. One of them was to create the weighted Breteau index in Martinique. In metropolitan France the EID Méditerranée has promising prospects and is applying considerable means for assessing the efficacy and impact of mosquito destruction operations and as a result some operational research on trapping methods is under way. Another case is in Reunion where the vector control service is currently working on the choice of indicators.

32Several services have already practised assessment of the effectiveness of actions and have adjusted their protocols accordingly. Most of them are also aware that assessment is necessary for improving actions, and would like support with the implementation of assessment exercises.

Recall the basic rules of assessment
Even if no method of vector control or assessment has any universal value, a generic framework for operations assessment does exist. In particular, it is a three-phase process: information (collecting the information), evaluation (analyse the data, make a judgment), proposal (make recommendations for improvements).
Define the clearly measurable objectives and effectiveness indicators
Services cannot assess the efficacy of their actions unless they have clearly defined, measurable points of reference and operational objective, or without criteria or indicators of effectiveness that are recognized and validated by the scientific community. In the short term, the State, which defines the national strategy, should fix, according to the ARSs (and GIPs), priorities, measurable operational objectives that must be achieved, indicators and the means of assessing these actions, by calling on expert opinion. The operational actors must take the same rules and harmonized reference texts.
Reinforce research on indicators
The State must organize the setting-up of a multidisciplinary committee of experts in public health, veterinary health and the environment, in order to define the epidemiological and entomological indicators essential for assessing vector control in different vector systems.
Improve entomological surveillance tools and the follow-up of effectiveness
Improvement in data acquisition is a major point for optimizing the control of vector-borne diseases. Traceability of vector control actions, rigorous recording of data and their statistical analysis must all be assured.
Develop partnerships and networks
The processes involved in defining objectives and indicators, their interpretations, analyses and possible corrections are often complex. Moreover, some kinds of expertise are rare and should be shared. The same applies to training. More strongly formulated networks must therefore be put in place involving actors of research, vector control operators and decision-makers.
The Vector Centre currently being founded, on the initiative of the IRD, CIRAD, the University of Montepellier and EID Méditerranée, in partnership with the Department of Health (DGS), the ‘Agence Française de Sécurité Sanitaire Environmentale et du Travail, ‘French Agency for Environmental Health Security and Safety at Work’, (AFSSET) and the Institut de Veille Sanitaire, or ‘Health Surveillance Institute’ (InVS) should optimize the devel opment of entomological expertise and provide guidance for the services. The ADEGE could also play an important role (in training, information exchange, cross-disciplinary audits).


1 In this chapter, references to the CD-ROM concern question 7: “What are the vector control assessment processes and approaches used?”.

Table des illustrations

Titre Table 8 – Basic set-up for experimental model
Légende The values Oi designate observations made before and after programme X. Formulae of the type O1 – O2 must be interpreted as theoretical indications rather than deviations between the two observations. In the ideal assessment model, the result attributable to the programme is evidently O3 -O4.
Fichier image/jpeg, 50k

© IRD Éditions, 2009

Licence OpenEdition Books

Rechercher dans OpenEdition Search

Vous allez être redirigé vers OpenEdition Search