The strengths of the IRS campaigns in Benin
The four strengths of the IRS campaigns in Benin which caught our attention and have already been mentioned  were: i) the great support from the population; ii) the skill and expertise of the international NGOs responsible for implementation; iii) the high coverage of protected households exceeding 80 % of the sleeping quarters; and, iv) a significant reduction in malaria transmission. The drastic decrease of the EIR mentioned from 2008 to 2016 in Ouémé and Atacora was also observed from 2017 to 2018 in Alibori and Donga. This article does not aim at emphasizing these results. It rather presents the shortcomings related to the IRS monitoring which remained silent about concerns raised by Benin NMCP and the partner funding IRS in Benin, the challenges and attempts to suggest solutions. Beside the suggestions already mentioned  in relation with appropriate locations for IRS implementation and management of Anopheles resistance to insecticides, we would like to draw attention to challenges affecting IRS impact.
Challenges affecting IRS impact
Challenge related to the residual malaria transmission in communes under IRS: lack of evident correlation between the decrease of EIR and malaria incidence
Despite the coverage of more than 80% of the structures by IRS and a drastic reduction of EIR, there is still significant residual malaria transmission in the four regions. In Ouémé region, despite two rounds of IRS in 2010 and the reduction of 90% of EIR, the 10% remaining representing the residual malaria transmission should be considered relatively high. Therefore, despite IRS in Adjohoun, Dangbo and Sèmè in Ouémé region, this residual transmission was estimated between 12.91 and 55.14 An. gambiae infected bites per year . Theoretically, in this region, despite IRS, everybody under protection of IRS was at risk of being in contact with 1 to 4 An. gambiae carrying P. falciparum in the salivary glands per month. The same situation was noticed in Alibori and Donga where a residual transmission of around 2 infecting bites per month after 80% reduction of EIR occurred; this was also the same in Atacora only in 2011 . In view of the complaints from the stakeholders about the lack of evidence of IRS epidemiological impact and the absence of significant reduction in the number of malaria cases suggested by the routine health facility data although these areas have been under annual IRS for many years, the reduction of 80-90% of EIR is encouraging, but should be observed with caution: it may be insufficient to decrease epidemiological indicators given that the residual EIR in IRS communes was still higher than it was in some regions of stable malaria. By other hand, EIR calculated for the communes under IRS might be overestimated. As a matter of fact, bendiocarb and pirimiphos methyl make mosquitoes fall from treated walls among which most of them blood fed. Since the houses of the study area were not surrounded by ant traps, such mosquitoes could be removed before PSC mosquito collections carried out by the technicians of CREC. Thus, the number of mosquitoes entering the houses and dead after blood meal might be underestimated. However, as EIR was calculated using mosquitoes collected using HLC, the number of mosquitoes which could escape to the ability of mosquito collectors would be low. Regarding the 2 graphs representing the variation of the malaria case incidence in the Atacora and Alibori-Donga regions being under IRS from 2011 to 2018 as mentioned by the health facilities, there is no reduction (Fig. 4). As a matter of fact, a retrospective investigation was carried out based on the indicators recorded by Atacora and Alibori-Donga health’s centers on the one hand, and malaria incidence calculated by the statisticians of the NMCP using the formula of Kalala et al. (2014)  and Hellenbrand (1999)  during the period of IRS on the other hand. The goal of this investigation was to have an idea on the evolution of malaria incidence in the communes where communities were protected by IRS. In addition, a report of Gates and Melinda project implemented from 2011 to 2015 in Benin, Cameroon, Kenya, India and Sudan on the impact of vector resistance on the efficacy of vector control tools (LLINs and IRS) showed mitigated results. In this study, the authors found very little correlation between the decrease of entomological parameters and both parasitological and clinical parameters [14, 17]. Another study on the efficacy of four vector control interventions supported by the French Ministry of Cooperation in an area of high vector insecticide resistance in Benin did not show significant differences from one intervention to another [18, 19]. This study concerned a cluster randomized controlled trial carried out in 28 villages southern Benin villages to investigate on the efficacy of various vector control interventions: LLIN targeted coverage to pregnant women and children under 6 years (TLLIN, reference group), LLIN universal coverage of all sleeping units (ULLIN), TLLIN plus full coverage of carbamate-IRS applied every 8 months (TLLIN + IRS), and ULLIN plus full coverage of carbamate-treated plastic sheeting (CTPS) (ULLIN + CTPS). The clinical incidence density of malaria was reduced neither in the children from the ULLIN group nor in those from TLLIN+IRS group nor the ULLIN+CTPS group compared with reference group (TLLIN) despite reduction of HBR and EIR. Relations between P. falciparum incidence and Entomological Inoculation Rates (EIR) was already observed by Beier  through a study in western Kenya showing that measurements of either the EIRs or HBRs can be used to predict corresponding attack rates in children. Some years later (1999), according to a report from Beier  on entomological Inoculation Rates and P. falciparum malaria prevalence in Africa, the basic relationship between EIR and P. falciparum prevalence accross different ecological zones showed convincingly that substantial reductions in malaria prevalence are likely to be achieved only when EIRs are reduced to levels of less 1 infective bite per person per year. This report showed that some sites with EIRs < 5 infected bites per year had levels of P. falciparum prevalence exceeding 40% and when transmission exceeded 15 infected bites per year, there were no sites with prevalence rates <50%.
To curtail residual malaria transmission, additional interventions able to target vectors escaping IRS should be prioritized. For instance, people who sleep indoors should use LLINs. Those who remain outdoors for various activities (cooking, praying, listening to the radio, eating, doing household chores) should use anti-mosquito ointment or smoke coils). Another alternative would be wearing garments impregnated with repellent insecticides which are available in some countries.
The evolution of malaria incidence in the communes where communities were protected by IRS presented from data collected in the health facilities should be more precise. That is why, the NMCP has decided to implement, from next year, a research study on epidemiological data on P. falciparum incidence and prevalence in collaboration with the parasitological research unit of CREC and the Faculty of Health Sciences of the University of Abomey-Calavi, Benin.
The choice of the control district affects the estimation of the EIR reduction in the districts under intervention
The high rates of reduction of the entomological parameters mentioned for the districts under IRS should be taken with caution. They were estimated based on the data collected in the control districts. Yet, it is not easy to have control areas with the same environmental characteristics as the intervention areas. The M&E of IRS implemented in Alibori and Donga regions had an advantage. During this monitoring, we compared two periods of the same area: before (2016) IRS for source base and after (2017, 2018) IRS. However, between 2016 and after (2017/2018), other factors might have changed (season, socioeconomic conditions). For the other regions (Ouémé, Atacora), in absence of source data before IRS intervention, not choosing the same areas for both intervention and control was a limit to the study. As a matter of fact, for these types of control areas, it is difficult to find control sites with characteristics similar to those of intervention sites. The 2 methods have inconveniences which might have influenced the estimation of EIR reduction rate.
Choosing the best control is an important component. We wish we had collected one year source data before IRS in all regions. But sometimes, decisions are taken in haste to avoid, either losing the financing, or because the partners demand to start the implementation as quickly as possible. It was the case for the campaigns of Ouémé and Atacora. Given that the assessment of IRS impact is in connection with the control zone data, proper arrangements must be made to identify the best control.
Some key elements of Benin IRS strategic plan cas
Strategy of IRS extension
The National Strategic Plan of Vector Control developed by the National Malaria Control Program states that IRS should start in some eligible areas and be progressively extended to other areas. But what is observed is quite different. In fact, instead of a strategy of extension, IRS withdrawal and move from one area to another is what occurs. These moves are all due to the same reason: lack of evident correlation between the decrease of EIR and malaria incidence. To benefit from the epidemiological impact of IRS, it is desirable to redefine the criteria of the eligible areas by prioritizing areas where malaria transmission doesn't go beyond the period of residual effect of the insecticide.
How can IRS Benin benefit from Insecticide Resistance Management strategies and protocols?
The emergence of An. gambiae resistance to bendiocarb in Atacora region [4, 11, 13, 18] might have been linked not only to the IRS campaign, but also to the massive use of agricultural insecticides against cotton pests . The management of resistance cannot be perceived by replacing one insecticide with another. Two methods could be used as far as Atacora is concerned:
(i) the NMCPs should avoid exposing Anopheles to the same product over several years in order to weaken the mosquitoes carrying resistance genes (fitness cost). To achieve this, as proposed by Akogbeto et al. , alternation of IRS campaigns with LLIN distribution campaigns should be a way. This means that the implementation of IRS will take place two years after the distribution of LLINs. By that time, most LLINs will have lost their insecticidal effectiveness and some will have been torn. The loss of effectiveness of pyrethroids will then be covered by IRS.
(ii) a combination of two insecticides with different modes of action could be used in rotation with the first method. The method based on combination or mosaic use of 2 or 3 insecticides is well knowed to NMCPs but its implementation is quite hard due to the limited variety of available products recommeded for IRS. However, the release of new IRS insecticides for use (SumiShild of Sumitomo and Fludora Fusion of Bayer) should facilitate the implementation of insecticide resistance management.
The withdrawal of IRS for another region is often accompanied by displeasure from the beneficiary communities. This was the case in Ouémé and Atacora. And, this is the reason why, before starting IRS implementation in a region, the beneficiary community should be informed that IRS is not a permanent control measure and might be replaced by another intervention at a specific time. In conclusion, clear communication and expectations are important.
Financing of IRS intervention
Financing is a key factor. It is an important point for IRS sustainability in Africa. Most of IRS campaigns in this region are supported by PMI and other partners. Unfortunately, IRS in Africa cannot remain at the expense of international partners forever. If for any reasons, the financial partner withdraws, this could rebound in malaria transmission due to IRS termination. This is the case for IRS results in Benin: when IRS was moved from Ouémé to Atacora, and from Atacora to Alibori and Donga, a resurgence of all malaria entomological indicators was noted a year later . As mentioned by Akogbéto et al. , IRS must be made cheaper and more affordable. There are irreducible sections included in IRS cost planning. This is the case when environmental compliance, logistics and insecticide provisioning are involved. However, costs related to staff, management and technical assistance due to the presence of international NGOs supporting the implementation must be reduced. African countries should get ready to take over from the financing in areas where IRS gives satisfying results.
Community practices affecting IRS impact
Dealing with reductions in entomological parameters that have not necessarily been associated with similar reductions in epidemiological indicators in communes under IRS, it is theorized that community behaviors may account for continued high numbers of malaria cases despite the use of both IRS and reported high rates of use of insecticide treated nets (ITNs). To identify household (HH) practices that may impact vector control strategies, we performed through a parallel study, a nighttime socio-behavioral surveillance during the cold, dry and rainy seasons in three regions where IRS was as performed (Alibori, Donga, Atacora) . This investigation showed that during the cold season, 25.1% of household members (HHMs) slept outdoors without net, 23.8% during the rainy season and 41.1% during the hot period. More than 90% of the HHMs in the study were not protected against mosquito bites between 7 PM–10 PM during all the three seasons because most were performing activities outdoors without the protection of IRS, likely due to poor lighting and limited space within their HH. During the hot season, the majority of HHMs were not protected at all from mosquito bites after 10 PM because they either slept outside without protection from IRS or slept inside without an ITN. HHs were also noted to have large amounts of indoor belongings and clothing hanging on the walls that served as places for mosquitoes to rest other than IRS-treated walls.
Sleeping under LLINs in the districts under IRS as a supplement malaria control effort.
During the 10-year campaigns of IRS, the majority of mosquitoes entering the treated houses were found fed. As a matter of fact, when a mosquito succeeds in entering a treated house, it goes directly to its host and gets its blood meal even when the house is treated. After the meal, the mosquito either remains on the walls or on indoor belongings and clothing hanging on the walls or seeks to escape when the house is treated. This situation was explained since fed mosquitoes were collected by pyrethrum spray catch and in exit window traps in treated houses and experimental huts in Malanville  as well as at community level (Djènontin [34, 35].
The question related to the simultaneous use of ITN + IRS was several times asked. There is no doubt about this question. In Benin, despite IRS, nobody wants to abandon their net. Given the relative high blood feeding rate of mosquitoes in the communes under IRS, we suggest that the NMCPs educate communities that are protected by IRS, particularly children and pregnant women, to consider sleeping under LLINs to supplement malaria control efforts.
Indoor Residual Spraying program evaluation
M&E is an important component of IRS implementation. The propose of this study is to enlighten the stakeholders involved in vector control, especially PMI countries in Africa, about many challenges related to this intervention. M&E should go along with IRS implementation, but for a limited time. Two 3 year cycles of M&E after a year of source data collection is sufficient for any in use-insecticide.