**4. Brazilian case studies**

Two types of traps associated with georeferencing systems were developed and evaluated continuously in Brazil: (1) ovitraps associated with the surveillance platform MSCP-*Aedes* (Monitoring System and Population Control for urban *Aedes*) and (2) sticky traps for gravid *Aedes* mosquitoes associated with a real-time, large-scale surveillance system known as MI-*Aedes* platform (from Portuguese "Monitoramento Integrado do *Aedes*"). Both systems will be described below, with emphasis on the adult trapping technology—MI-*Aedes*—since it has been used in the last 13 years in hundreds of Brazilian cities.

### **4.1 Monitoring system and population control for urban** *Aedes* **(MSCP-***Aedes***)**

The MSCP-*Aedes* platform was developed by the National Institute for Space Research (INPE) and Research Center Aggeu Magalhães (CPqAM), Oswaldo Cruz Foundation (Fiocruz), located in Recife city, Pernambuco State, Brazil. The potential of ovitraps in reducing the population of *Aedes* spp. was evaluated for 1 year (April 2014–April 2015), during all seasons of the year (summer, autumn, winter, and spring), by the deployment of 464 georeferenced traps in five areas of Recife, the capital city of the state of Pernambuco, located in northeastern Brazil. Thirteen egg collection cycles were performed with 98.5% of the ovitraps being positive for *Aedes* eggs. At the end of the study, more than 4 million eggs were collected from the environment, and the *Ae. aegypti* population in one of the five localities evaluated was significantly reduced. The platform provided information on the spatialtemporal distribution of *Aedes* spp. eggs. Using this data, maps generated within a GIS environment helped the health authorities to prioritize the city areas in most need of vector control actions [40] (**Figure 4**).

Another pilot trial of the MSCP-*Aedes* system was carried out from March 2008 to October 2011 in two other cities of Pernambuco State, Brazil: Ipojuca and Santa Cruz [37]. After the first 2 years of evaluation, a significant decrease in the density of eggs was observed in both cities showing the potential of the MSCP-*Aedes* platform associated with the vector control actions conducted by the health authorities to reduce mosquito abundance (**Figure 5**). However, the MSCP-*Aedes* platform required a great number of people to accomplish the field and laboratory activities, which is not realistic to use in a large-scale scenario.

**129**

38, 41, 47, 48] (**Figure 6**).

(4) production of entomological indices.

*New Cost-Benefit of Brazilian Technology for Vector Surveillance Using Trapping System*

*DOI: http://dx.doi.org/10.5772/intechopen.78781*

*April 2005). Adapted from Regis et al. [40].*

**Figure 4.**

**Figure 5.**

*(Adapted from Regis et al. [37]).*

**4.2 Integrated** *Aedes* **surveillance system (MI-***Aedes***)**

The innovative MI-*Aedes* platform was developed in Brazil by a universitycompany partnership between the Federal University of Minas Gerais and the university's "spin-off" Ecovec, in Belo Horizonte, Minas Gerais State. University-Company partnerships have been stimulated by the Brazilian Innovation Law, which aims to foster the generation of innovations and dissemination of new technologies aiming to solve national (and international) problems [44, 45]. The World Health Organization has praised this new technology for the surveillance and generation of

*Spatial distribution of Aedes eggs at the high-density period (May–June) in Ipojuca (Recife, Brazil), Pernambuco State, 2008–2011. The mass suppression of eggs using 2700 control ovitraps started in October 2009* 

*Spatial distribution of Aedes spp. eggs in Brasília Teimosa (Recife, Pernambuco State, Brazil), in June–July 2004 (A), December 2004 (B), and March 2005 (C). Green, low infestation; yellow, intermediate infestation; red, high infestation. The map at the right shows the mean number of eggs for the whole period (April 2004 to* 

entomological indices [46]. More details about the platform are given below.

The MI-*Aedes* platform consists of (a) the sticky trap MosquiTRAP (baited with a Atr*Aedes* to generate mosquito abundance indices), which is placed within blocks of urban areas 250 m equidistant from each other and inspected weekly, (b) the recording of entomological data on electronic spreadsheets or by cell phone during trap inspection, and (c) an Internet site that integrates real-time adult mosquito surveillance data and GIS technology to provide entomological indices [12, 31, 36,

The information used for vector control relies on (1) weekly surveillance of gravid *Ae. aegypti* infestation of the municipality street blocks, (2) re-infestation surveillance of the monitored blocks, (3) identification of hotspot areas, and

The MI-*Aedes* Web-data system consists of three integrated software developed to simplify information gathering and processing: (a) the "'geo-mosquito *New Cost-Benefit of Brazilian Technology for Vector Surveillance Using Trapping System DOI: http://dx.doi.org/10.5772/intechopen.78781*

**Figure 4.**

*Malaria*

ruses [31, 37, 38].

infestation, as shown in Brazil [12].

the last 13 years in hundreds of Brazilian cities.

need of vector control actions [40] (**Figure 4**).

which is not realistic to use in a large-scale scenario.

**4. Brazilian case studies**

**3. Use of geographic information system (GIS) for vector surveillance**

sis, and display of spatial data that offer expanding opportunities for epidemiology because it allows a spatial perspective on a disease. The integration of vector surveillance with the mosquito traps and georeferencing technologies has emerged as an important tool for fighting *Ae. aegypti* and transmission of arbovi-

GIS is a powerful automated system for the capture, storage, retrieval, analy-

By georeferencing the ovitrap and sticky traps, the egg collection and adult catching data obtained during *Ae. aegypti* surveillance was used to generate maps that show the areas of high and low infestation [31, 39–42]. This information provides real-time data and allows spatial analyses to determine vector control actions and to evaluate their impact on mosquito populations and infection with arboviruses [31, 39, 43]. The continuous surveillance of *Aedes* population allied with mathematical modeling strategies (described below) allows reliable predictions of

Two types of traps associated with georeferencing systems were developed and evaluated continuously in Brazil: (1) ovitraps associated with the surveillance platform MSCP-*Aedes* (Monitoring System and Population Control for urban *Aedes*) and (2) sticky traps for gravid *Aedes* mosquitoes associated with a real-time, large-scale surveillance system known as MI-*Aedes* platform (from Portuguese "Monitoramento Integrado do *Aedes*"). Both systems will be described below, with emphasis on the adult trapping technology—MI-*Aedes*—since it has been used in

**4.1 Monitoring system and population control for urban** *Aedes* **(MSCP-***Aedes***)**

The MSCP-*Aedes* platform was developed by the National Institute for Space Research (INPE) and Research Center Aggeu Magalhães (CPqAM), Oswaldo Cruz Foundation (Fiocruz), located in Recife city, Pernambuco State, Brazil. The potential of ovitraps in reducing the population of *Aedes* spp. was evaluated for 1 year (April 2014–April 2015), during all seasons of the year (summer, autumn, winter, and spring), by the deployment of 464 georeferenced traps in five areas of Recife, the capital city of the state of Pernambuco, located in northeastern Brazil. Thirteen egg collection cycles were performed with 98.5% of the ovitraps being positive for *Aedes* eggs. At the end of the study, more than 4 million eggs were collected from the environment, and the *Ae. aegypti* population in one of the five localities evaluated was significantly reduced. The platform provided information on the spatialtemporal distribution of *Aedes* spp. eggs. Using this data, maps generated within a GIS environment helped the health authorities to prioritize the city areas in most

Another pilot trial of the MSCP-*Aedes* system was carried out from March 2008 to October 2011 in two other cities of Pernambuco State, Brazil: Ipojuca and Santa Cruz [37]. After the first 2 years of evaluation, a significant decrease in the density of eggs was observed in both cities showing the potential of the MSCP-*Aedes* platform associated with the vector control actions conducted by the health authorities to reduce mosquito abundance (**Figure 5**). However, the MSCP-*Aedes* platform required a great number of people to accomplish the field and laboratory activities,

**128**

*Spatial distribution of Aedes spp. eggs in Brasília Teimosa (Recife, Pernambuco State, Brazil), in June–July 2004 (A), December 2004 (B), and March 2005 (C). Green, low infestation; yellow, intermediate infestation; red, high infestation. The map at the right shows the mean number of eggs for the whole period (April 2004 to April 2005). Adapted from Regis et al. [40].*

**Figure 5.** *Spatial distribution of Aedes eggs at the high-density period (May–June) in Ipojuca (Recife, Brazil), Pernambuco State, 2008–2011. The mass suppression of eggs using 2700 control ovitraps started in October 2009 (Adapted from Regis et al. [37]).*

#### **4.2 Integrated** *Aedes* **surveillance system (MI-***Aedes***)**

The innovative MI-*Aedes* platform was developed in Brazil by a universitycompany partnership between the Federal University of Minas Gerais and the university's "spin-off" Ecovec, in Belo Horizonte, Minas Gerais State. University-Company partnerships have been stimulated by the Brazilian Innovation Law, which aims to foster the generation of innovations and dissemination of new technologies aiming to solve national (and international) problems [44, 45]. The World Health Organization has praised this new technology for the surveillance and generation of entomological indices [46]. More details about the platform are given below.

The MI-*Aedes* platform consists of (a) the sticky trap MosquiTRAP (baited with a Atr*Aedes* to generate mosquito abundance indices), which is placed within blocks of urban areas 250 m equidistant from each other and inspected weekly, (b) the recording of entomological data on electronic spreadsheets or by cell phone during trap inspection, and (c) an Internet site that integrates real-time adult mosquito surveillance data and GIS technology to provide entomological indices [12, 31, 36, 38, 41, 47, 48] (**Figure 6**).

The information used for vector control relies on (1) weekly surveillance of gravid *Ae. aegypti* infestation of the municipality street blocks, (2) re-infestation surveillance of the monitored blocks, (3) identification of hotspot areas, and (4) production of entomological indices.

The MI-*Aedes* Web-data system consists of three integrated software developed to simplify information gathering and processing: (a) the "'geo-mosquito

**Figure 6.**

*The MI-Aedes platform consists of (A) sticky trap (MosquiTRAP), (B) cell phone for sampling GIS mosquito data, (C) MI-Virus (optional—see text for detail), and (D) Internet georeferenced maps at real time that produces automatic data for entomological indices.*

collection," which is installed in portable devices (e.g., cell phones) to record household information, placement of the trap within the residence, and *Ae. aegypti* field capture data; (b) the "monitoring," which processes the field data to produce tables with entomological indices and graphs showing trends; and (c) the "geo-*Aedes*," which produces georeferenced maps of mosquitoes captured with the sticky MosquiTRAP and makes them available to users on the Internet on a weekly basis.

There are several advantages of using electronic spreadsheets or mobile phone over conventional data acquisition systems. The field data can be accessed immediately (premises visited and scheduled for visits, trap locations, residents' names, and so on), and the entomological indices can be produced automatically. Also, there is no delay between the data that is reported to the database and the database that is available for web mapping and public health access.

The MI-*Aedes* platform was evaluated in hundreds of Brazilian cities for more than 10 years and showed to greatly reduce arbovirus transmission [41]. The georeferenced maps presented weekly on the Internet by the MI-*Aedes* platform allowed health managers to identify the infestation status of city blocks by the colors green, yellow, orange, and red, according to the number of adult *Ae. aegypti* females captured (**Figure 7**). The weekly data evaluating vector infestation levels became an important information for dengue control programs because it helped public health managers to optimize *Ae. aegypti* control activities with improved precision of the target activities to the infested blocks. Indeed, a study analyzing three Brazilian municipalities revealed that following implementation of the MI-*Aedes* platform, the weekly vector control indicator established by the entomological "mean female *Aedes* index "(MFAI) was reduced (**Figure 7**) and so was the number of dengue cases [31]. Further research showed how the health authorities used the platform to evaluate the performance of the control measures employed by them within the area covered by the MI-*Aedes* [12, 41].

### *4.2.1 Virus detection of trapped gravid Aedes spp. collected by the sticky trap MosquiTRAP*

The detection of gravid mosquitoes infected with arboviruses such as DENV, CHIKV, and ZIKV is an important information for public health managers looking to control *Ae. aegypti* infestation and the spread of arboviral diseases in hotspot areas. The inclusion of a strategy to identify the arbovirus present in infected mosquitoes trapped in the MosquiTRAP into the MI-*Aedes* platform was intended to provide additional information regarding the spread of arboviruses and serve as an early warning system for epidemics since viral detection in mosquitoes can precede detection in humans. Accordingly, a rapid and well-established method for arbovirus identification [49] was associated with the MI-*Aedes* platform to create an

**131**

**Figure 7.**

**Figure 8.**

*Adapted from Eiras and Resende [31].*

*New Cost-Benefit of Brazilian Technology for Vector Surveillance Using Trapping System*

Integrated Monitoring Virus (MI-Virus) platform. The trapped *Ae. aegypti* and *Ae. albopictus* are placed in Eppendorf tubes with guanidine and sent by mail for virus detection and identification by reverse transcriptase RT-PCR (**Figure 8**). In Brazil, the MI-Virus platform was tested in hundreds of municipalities to detect and map not only *Ae. aegypti* abundance but also the presence of mosquito populations infected with different arboviruses such as DENV, CHIKV, and ZIKV. The use of MosquiTRAP to detect DENV-infected gravid *Ae. aegypti* trapped by was performed in Brazil [41, 50] and Colombia [51]. In In 2017, the MI-*Aedes* and MI-Virus platforms were used during an outbreak of chikungunya in Governador Valadares city, located in the southeastern state of Minas Gerais, Brazil (data not published). The real-time data obtained by the MI-*Aedes* platform and the confirmation of CHIKV by the MI-Virus (**Figure 9**) led the health authorities to act quickly and employ additional vector control activities to target areas with the

*GIS trap position. Source: Ecovec Company, Belo Horizonte, Minas Gerais State, Brazil.*

*Logistic of the Integrated Monitoring Virus (MI-Virus). Weekly, caught female Aedes aegypti by MosquiTRAP are place in barcoded Eppendorf tube and send by post to laboratory for RT-PCR analysis. Bar code provides* 

*Temporal and spatial analysis of neighborhoods in the municipality of Presidente Epitácio (São Paulo State, Brazil) (2008). Colored maps are classified according to the entomological indices from epidemiological week 7–16 (February–April, summer). In epidemiological weeks 7–10, there were 44 and 55% of neighborhoods classified as "dengue alert" and "critical," respectively. Weeks 14–17 were 88.9 and 11.1% of the municipality's territory as "risk-free" and "dengue alert," respectively, indicating a strong seasonal variation in the Aedes aegypti population density that was probably influenced by the climate conditions or targeted control measures.* 

*DOI: http://dx.doi.org/10.5772/intechopen.78781*

*New Cost-Benefit of Brazilian Technology for Vector Surveillance Using Trapping System DOI: http://dx.doi.org/10.5772/intechopen.78781*

#### **Figure 7.**

*Malaria*

**Figure 6.**

collection," which is installed in portable devices (e.g., cell phones) to record household information, placement of the trap within the residence, and *Ae. aegypti* field capture data; (b) the "monitoring," which processes the field data to produce tables with entomological indices and graphs showing trends; and (c) the "geo-*Aedes*," which produces georeferenced maps of mosquitoes captured with the sticky MosquiTRAP and makes them available to users on the Internet on a weekly basis. There are several advantages of using electronic spreadsheets or mobile phone over conventional data acquisition systems. The field data can be accessed immediately (premises visited and scheduled for visits, trap locations, residents' names, and so on), and the entomological indices can be produced automatically. Also, there is no delay between the data that is reported to the database and the database

*The MI-Aedes platform consists of (A) sticky trap (MosquiTRAP), (B) cell phone for sampling GIS mosquito data, (C) MI-Virus (optional—see text for detail), and (D) Internet georeferenced maps at real time that* 

The MI-*Aedes* platform was evaluated in hundreds of Brazilian cities for more than 10 years and showed to greatly reduce arbovirus transmission [41]. The georeferenced maps presented weekly on the Internet by the MI-*Aedes* platform allowed health managers to identify the infestation status of city blocks by the colors green, yellow, orange, and red, according to the number of adult *Ae. aegypti* females captured (**Figure 7**). The weekly data evaluating vector infestation levels became an important information for dengue control programs because it helped public health managers to optimize *Ae. aegypti* control activities with improved precision of the target activities to the infested blocks. Indeed, a study analyzing three Brazilian municipalities revealed that following implementation of the MI-*Aedes* platform, the weekly vector control indicator established by the entomological "mean female *Aedes* index "(MFAI) was reduced (**Figure 7**) and so was the number of dengue cases [31]. Further research showed how the health authorities used the platform to evaluate the performance of the control measures employed by them within the

*4.2.1 Virus detection of trapped gravid Aedes spp. collected by the sticky trap* 

The detection of gravid mosquitoes infected with arboviruses such as DENV, CHIKV, and ZIKV is an important information for public health managers looking to control *Ae. aegypti* infestation and the spread of arboviral diseases in hotspot areas. The inclusion of a strategy to identify the arbovirus present in infected mosquitoes trapped in the MosquiTRAP into the MI-*Aedes* platform was intended to provide additional information regarding the spread of arboviruses and serve as an early warning system for epidemics since viral detection in mosquitoes can precede detection in humans. Accordingly, a rapid and well-established method for arbovirus identification [49] was associated with the MI-*Aedes* platform to create an

that is available for web mapping and public health access.

area covered by the MI-*Aedes* [12, 41].

*produces automatic data for entomological indices.*

*MosquiTRAP*

**130**

*Temporal and spatial analysis of neighborhoods in the municipality of Presidente Epitácio (São Paulo State, Brazil) (2008). Colored maps are classified according to the entomological indices from epidemiological week 7–16 (February–April, summer). In epidemiological weeks 7–10, there were 44 and 55% of neighborhoods classified as "dengue alert" and "critical," respectively. Weeks 14–17 were 88.9 and 11.1% of the municipality's territory as "risk-free" and "dengue alert," respectively, indicating a strong seasonal variation in the Aedes aegypti population density that was probably influenced by the climate conditions or targeted control measures. Adapted from Eiras and Resende [31].*

#### **Figure 8.**

*Logistic of the Integrated Monitoring Virus (MI-Virus). Weekly, caught female Aedes aegypti by MosquiTRAP are place in barcoded Eppendorf tube and send by post to laboratory for RT-PCR analysis. Bar code provides GIS trap position. Source: Ecovec Company, Belo Horizonte, Minas Gerais State, Brazil.*

Integrated Monitoring Virus (MI-Virus) platform. The trapped *Ae. aegypti* and *Ae. albopictus* are placed in Eppendorf tubes with guanidine and sent by mail for virus detection and identification by reverse transcriptase RT-PCR (**Figure 8**).

In Brazil, the MI-Virus platform was tested in hundreds of municipalities to detect and map not only *Ae. aegypti* abundance but also the presence of mosquito populations infected with different arboviruses such as DENV, CHIKV, and ZIKV. The use of MosquiTRAP to detect DENV-infected gravid *Ae. aegypti* trapped by was performed in Brazil [41, 50] and Colombia [51]. In In 2017, the MI-*Aedes* and MI-Virus platforms were used during an outbreak of chikungunya in Governador Valadares city, located in the southeastern state of Minas Gerais, Brazil (data not published). The real-time data obtained by the MI-*Aedes* platform and the confirmation of CHIKV by the MI-Virus (**Figure 9**) led the health authorities to act quickly and employ additional vector control activities to target areas with the

#### *Malaria*

#### **Figure 9.**

*Data obtained by the MI-Aedes platform and MI-Virus deployed in Governador Valadares city, Minas Gerais State, Brazil (epidemiological week 11, March 05–11, 2017). Dark dots at colored circles mean MosquiTRAP (GIS) position. Colored circles mean infestation status of mosquito abundance (see text for further detail). Black circles mean infected Aedes aegypti by chikungunya virus captured by MosquiTRAP (data not published). Source: Ecovec company, Belo Horizonte, Minas Gerais State, Brazil.*

highest mosquito densities. As a result, abundance of *Ae. aegypti* and chikungunya cases reduced significantly (data not published). Futures studies should be conducted of arbovirus detections in other areas.

#### *4.2.2 Modeling the population dynamics of Aedes aegypti using MI-Aedes*

Once vector surveillance and control are established as the recommended approach to manage vector-borne diseases, the ecological problem of the population dynamics of mosquitoes arises as a fundamental question [52]. In such context, mathematical modeling has a twofold role: to assist the validation of these novel technologies by providing methods to predict the population dynamics of adult mosquitoes and to offer ways to improve the surveillance indices. As an ecological problem, the infestation by mosquitoes is influenced by many anthropic (everything that results from human action such as sanitation and mosquito breeding container) and non-anthropic variables (temperature and rainfall) [53]. It has been well established that the population dynamics of different stages of *Ae. aegypti* and viral transmission are influenced by environmental variables, especially those of climate: temperature, humidity, and precipitation [54–58]. The vector-virus-human system can generate multiple sources of complexity for modeling. The vector management approach of considering the female population as a risk of infection indicator helps to simplify the modeling efforts, which can decouple the complex

**133**

*New Cost-Benefit of Brazilian Technology for Vector Surveillance Using Trapping System*

MI-*Aedes* at the neighborhood level for mosquito surveillance [62].

ecological vector-virus-human system and focus on the mosquito population. In addition, the surveillance platform MI-*Aedes* generates a huge amount of sampling data from many localities [12, 59, 60]. These huge data banks provided basis for many modeling studies such as a novel stochastic point process pattern algorithm that identify the spatial and temporal association between DENV-infected mosquitoes and human cases. This process showed a strong and significant association between high DENV incidence in mosquitoes and the onset of symptoms in humans at specific spatial and temporal windows [61]. Also the model goodness-of-fit studies based on the number of sticky traps and suggests a minimum of 16 traps for the

Decades of studies regarding the effects of temperature, precipitation, and humidity on vector population and the occurrence of infectious disease cases generated controversial conclusions, suggesting that the phenomenon depends on local specificities, as extensively demonstrated for dengue [54, 63]. Nevertheless, it is well established that temperature affects the physiology of the mosquito and virus and, consequently, is associated with the vector population size and dengue cases [13, 58, 64–66]. Humidity greatly affects the development of vector stages and the number of dengue infections [67, 68]. Although precipitation is strongly correlated with humidity, due to its complex pattern and unpredictable influence in the environment, it figures as the most complex meteorological variable [69]. Notwithstanding, precipitation is a good explanatory variable for dengue cases and mosquito population size [70]. Hence, the construction of models to explain the effects of climatological variables cannot disregard the complete set of these

The problem of describing or even predicting a response time series such as disease cases or infestation can be approached with descriptive models, which provide a model time series solution by fitting coefficients and/or functions in accordance with past lagged time series of a set of explanatory variables. Belonging to that class are the regressive models, which have been used for predicting or describing the number of dengue cases or the degree of adult *Ae. aegypti* infestation [12, 56, 59, 71]. Through a descriptive model, time series of temperature, precipitation, wind velocity, and humidity were analyzed as explanatory variables for the adult mosquito abundance index MFAI in the subhumid tropical climate of the city of Governador Valadares, Brazil [59]. In the study, generalized linear models (GLM) with time lags and interaction terms between explanatory variables were used to identify the following significant associations: interaction between lagged temperature and humidity with the mosquito abundance data obtained on the previous week. Transient associations were mapped in a periodogram using wavelets and revealed significant effects for precipitation and wind velocity. Interestingly, the wavelet technique identified non-stationary effects on the relationship between meteoro-

Another study using descriptive models was conducted in the city of Porto Alegre, located in a humid subtropical region of southern Brazil. It used data derived from monitoring the *Ae. aegypti* adult female population in the course of MI-Dengue (nowadays, MI-*Aedes*) surveillance platform [12]. As described above, the platform employs sticky traps to capture adult *Ae. aegypti* females to provide a weekly infestation index. To predict mosquito abundance in subsequent weeks, time series data from previous weeks regarding the maximum, minimum, and mean temperature, precipitation, humidity, and mosquito abundance were fitted in a set of proposed models using generalized additive models (GAM). The best power of prediction was achieved when previous values of minimum temperatures and adult females were included in the set of explanatory variables (**Figure 10**). Precipitation was not a significant explanatory variable for the humid temperate climate of Porto

*DOI: http://dx.doi.org/10.5772/intechopen.78781*

influential variables.

logical variables and infestation.

#### *New Cost-Benefit of Brazilian Technology for Vector Surveillance Using Trapping System DOI: http://dx.doi.org/10.5772/intechopen.78781*

ecological vector-virus-human system and focus on the mosquito population. In addition, the surveillance platform MI-*Aedes* generates a huge amount of sampling data from many localities [12, 59, 60]. These huge data banks provided basis for many modeling studies such as a novel stochastic point process pattern algorithm that identify the spatial and temporal association between DENV-infected mosquitoes and human cases. This process showed a strong and significant association between high DENV incidence in mosquitoes and the onset of symptoms in humans at specific spatial and temporal windows [61]. Also the model goodness-of-fit studies based on the number of sticky traps and suggests a minimum of 16 traps for the MI-*Aedes* at the neighborhood level for mosquito surveillance [62].

Decades of studies regarding the effects of temperature, precipitation, and humidity on vector population and the occurrence of infectious disease cases generated controversial conclusions, suggesting that the phenomenon depends on local specificities, as extensively demonstrated for dengue [54, 63]. Nevertheless, it is well established that temperature affects the physiology of the mosquito and virus and, consequently, is associated with the vector population size and dengue cases [13, 58, 64–66]. Humidity greatly affects the development of vector stages and the number of dengue infections [67, 68]. Although precipitation is strongly correlated with humidity, due to its complex pattern and unpredictable influence in the environment, it figures as the most complex meteorological variable [69]. Notwithstanding, precipitation is a good explanatory variable for dengue cases and mosquito population size [70]. Hence, the construction of models to explain the effects of climatological variables cannot disregard the complete set of these influential variables.

The problem of describing or even predicting a response time series such as disease cases or infestation can be approached with descriptive models, which provide a model time series solution by fitting coefficients and/or functions in accordance with past lagged time series of a set of explanatory variables. Belonging to that class are the regressive models, which have been used for predicting or describing the number of dengue cases or the degree of adult *Ae. aegypti* infestation [12, 56, 59, 71]. Through a descriptive model, time series of temperature, precipitation, wind velocity, and humidity were analyzed as explanatory variables for the adult mosquito abundance index MFAI in the subhumid tropical climate of the city of Governador Valadares, Brazil [59]. In the study, generalized linear models (GLM) with time lags and interaction terms between explanatory variables were used to identify the following significant associations: interaction between lagged temperature and humidity with the mosquito abundance data obtained on the previous week. Transient associations were mapped in a periodogram using wavelets and revealed significant effects for precipitation and wind velocity. Interestingly, the wavelet technique identified non-stationary effects on the relationship between meteorological variables and infestation.

Another study using descriptive models was conducted in the city of Porto Alegre, located in a humid subtropical region of southern Brazil. It used data derived from monitoring the *Ae. aegypti* adult female population in the course of MI-Dengue (nowadays, MI-*Aedes*) surveillance platform [12]. As described above, the platform employs sticky traps to capture adult *Ae. aegypti* females to provide a weekly infestation index. To predict mosquito abundance in subsequent weeks, time series data from previous weeks regarding the maximum, minimum, and mean temperature, precipitation, humidity, and mosquito abundance were fitted in a set of proposed models using generalized additive models (GAM). The best power of prediction was achieved when previous values of minimum temperatures and adult females were included in the set of explanatory variables (**Figure 10**). Precipitation was not a significant explanatory variable for the humid temperate climate of Porto

*Malaria*

**Figure 9.**

**132**

highest mosquito densities. As a result, abundance of *Ae. aegypti* and chikungunya cases reduced significantly (data not published). Futures studies should be con-

*Data obtained by the MI-Aedes platform and MI-Virus deployed in Governador Valadares city, Minas Gerais State, Brazil (epidemiological week 11, March 05–11, 2017). Dark dots at colored circles mean MosquiTRAP (GIS) position. Colored circles mean infestation status of mosquito abundance (see text for further detail). Black circles mean infected Aedes aegypti by chikungunya virus captured by MosquiTRAP (data not* 

Once vector surveillance and control are established as the recommended approach to manage vector-borne diseases, the ecological problem of the population dynamics of mosquitoes arises as a fundamental question [52]. In such context, mathematical modeling has a twofold role: to assist the validation of these novel technologies by providing methods to predict the population dynamics of adult mosquitoes and to offer ways to improve the surveillance indices. As an ecological problem, the infestation by mosquitoes is influenced by many anthropic (everything that results from human action such as sanitation and mosquito breeding container) and non-anthropic variables (temperature and rainfall) [53]. It has been well established that the population dynamics of different stages of *Ae. aegypti* and viral transmission are influenced by environmental variables, especially those of climate: temperature, humidity, and precipitation [54–58]. The vector-virus-human system can generate multiple sources of complexity for modeling. The vector management approach of considering the female population as a risk of infection indicator helps to simplify the modeling efforts, which can decouple the complex

*4.2.2 Modeling the population dynamics of Aedes aegypti using MI-Aedes*

*published). Source: Ecovec company, Belo Horizonte, Minas Gerais State, Brazil.*

ducted of arbovirus detections in other areas.

#### **Figure 10.**

*Observed adult infestation index MFAI and predicted generalized additive models (GAM) with meteorological and infestation data as explanatory variables for the city of Porto Alegre, Rio Grande do Sul, Brazil. From September 2012 to January 2016 (Figure extracted from da Cruz Ferreira et al. [12]).*

**Figure 11.**

*Mechanistic model comprising the populations of the stages of development of Aedes aegypti: E(t), A(t), F1(t), and F2(t), which are the populations of eggs, aquatic forms, and females pre and post blood meal, respectively. The rates of development of the model are set to depend on precipitation p. The curves are generated after solving a system of nonlinear differential equations, the preferred framework to represent that class of models (data not published).*

Alegre, presumably because precipitation is less seasonal in this region. The association between mosquito infestation and the number of dengue cases was positive, indicating that the infestation index MFAI is a good indicator for the risk of arboviral transmission [12].

Mechanistic models have the same goal of describing or even predicting a time response series as the descriptive models but differ from the latter because they are structured with realism based on the natural phenomena, for example, the population model comprises the biological cycle of *Ae. aegypti*. Hence, through these models, from deviations and corrections for adjustment to the data, it is possible to reveal the cause-effect relations of the underlying phenomena.

Mechanistic models have been used to study and predict vector infestation and the number of dengue cases [60, 72–78]. One such model was developed to account for the effect of precipitation on the stages of development of *Ae. aegypti* by setting the model parameters as dependent on the precipitation index (in millimeters) (**Figure 11**) [60]. **Figure 12** illustrates the model result considering the infestation index MFAI and the precipitation from June 2009 to December 2010 for the city of Sete Lagoas, Minas Gerais, Brazil.

#### *4.2.3 Cost-benefit of the MI-Aedes platform*

Dengue epidemics pose a heavy burden on health services and the economy of any country. Recently, studies in eight countries in the Americas and Asia have shown that the cost of epidemics in these countries reached approximately US\$ 1.8 billion per year [79, 80]. This number only refers to the money spent on outpatient and hospital expenses and did not consider costs such as those related to

**135**

practices.

on pooled mosquitoes (source: Ecovec. Ltd).

months to years [82].

**Figure 12.**

*New Cost-Benefit of Brazilian Technology for Vector Surveillance Using Trapping System*

surveillance and vector control activities. The economic losses imposed by arboviral diseases involve the patient's withdrawal from productive activities, drug expenses, hospitalization, medical consultations, treatment of sequelae, and death [81]. The time needed to treat and recover from arboviruses varies. On average, dengue removes the affected patient for 10–12 days from their work activities. The ZIKV may lead to birth defects including microcephaly and other severe brain malformations, which impose lifetime incapacity. Recovery from chikungunya varies from

*Comparison of 83 epidemiological weeks covering the time series data of Aedes aegypti females captured divided by the number of traps (MFAI), the model result, and the precipitation index in the municipality of Sete Lagoas, Minas Gerais, Brazil. From June 2009 to December 2010 (J.L. Acebal—data not published).*

Following the guidelines of the National Program for Dengue Control (PNCD), the MI-*Aedes* and MI-Virus technologies were adopted by the health authorities of the state of Minas Gerais (Brazil) and implemented in 21 cities with a high incidence of dengue in the period 2009–2011. The total cost of the program for all 21 cities for 2 years of work was less than US\$ 1.5 million, making an average of US\$ 71,428 per city. It included 4700 sticky traps, 115,000 sticky cards, synthetic oviposition attractants, RT-PCR on all mosquitoes caught in the traps, Web software licensing, cell phones, and technical support, among other items. The number of people benefited by the program was approximately 2 million, making the cost per habitant per year around US\$ 0.70. The cost-effectiveness was calculated as the cost of running the MI-*Aedes* and MI-Virus platforms divided by the number of cases of prevented arboviral diseases compared with cities that did not use the MI-*Aedes* platform and relied only in the PNCD guidelines. The MI-*Aedes* and MI-Virus platforms prevented a total of 27,191 cases at a total cost of US\$ 7.5 million, thus saving approximately US\$ 0.4 million in direct costs (health care and vector control) and US\$ 7.1 million in lost wages (societal impact) annually [41]. The cost-effectiveness of the platforms MI-*Aedes* and MI-Virus in cities with high mosquito infestation levels emphasizes the power of using these new technologies in vector control

Currently, the MI-*Aedes* platform is running simultaneously in 154 Brazilian cities targeting approximately 7.5 million people, using about 12,000 sticky traps, and performing 625,000 trap inspections and around 8200 RT-PCR analysis per month

Investing more effort into integrating MI-*Aedes* strategies and costs with vector control operations, and standardizing the MI-*Aedes*-based control system across

*DOI: http://dx.doi.org/10.5772/intechopen.78781*

*New Cost-Benefit of Brazilian Technology for Vector Surveillance Using Trapping System DOI: http://dx.doi.org/10.5772/intechopen.78781*

#### **Figure 12.**

*Malaria*

**Figure 10.**

**Figure 11.**

**134**

ral transmission [12].

*(data not published).*

Sete Lagoas, Minas Gerais, Brazil.

*4.2.3 Cost-benefit of the MI-Aedes platform*

Alegre, presumably because precipitation is less seasonal in this region. The association between mosquito infestation and the number of dengue cases was positive, indicating that the infestation index MFAI is a good indicator for the risk of arbovi-

*Mechanistic model comprising the populations of the stages of development of Aedes aegypti: E(t), A(t), F1(t), and F2(t), which are the populations of eggs, aquatic forms, and females pre and post blood meal, respectively. The rates of development of the model are set to depend on precipitation p. The curves are generated after solving a system of nonlinear differential equations, the preferred framework to represent that class of models* 

*Observed adult infestation index MFAI and predicted generalized additive models (GAM) with meteorological and infestation data as explanatory variables for the city of Porto Alegre, Rio Grande do Sul, Brazil. From* 

*September 2012 to January 2016 (Figure extracted from da Cruz Ferreira et al. [12]).*

Mechanistic models have the same goal of describing or even predicting a time response series as the descriptive models but differ from the latter because they are structured with realism based on the natural phenomena, for example, the population model comprises the biological cycle of *Ae. aegypti*. Hence, through these models, from deviations and corrections for adjustment to the data, it is possible to

Mechanistic models have been used to study and predict vector infestation and the number of dengue cases [60, 72–78]. One such model was developed to account for the effect of precipitation on the stages of development of *Ae. aegypti* by setting the model parameters as dependent on the precipitation index (in millimeters) (**Figure 11**) [60]. **Figure 12** illustrates the model result considering the infestation index MFAI and the precipitation from June 2009 to December 2010 for the city of

Dengue epidemics pose a heavy burden on health services and the economy of any country. Recently, studies in eight countries in the Americas and Asia have shown that the cost of epidemics in these countries reached approximately US\$ 1.8 billion per year [79, 80]. This number only refers to the money spent on outpatient and hospital expenses and did not consider costs such as those related to

reveal the cause-effect relations of the underlying phenomena.

*Comparison of 83 epidemiological weeks covering the time series data of Aedes aegypti females captured divided by the number of traps (MFAI), the model result, and the precipitation index in the municipality of Sete Lagoas, Minas Gerais, Brazil. From June 2009 to December 2010 (J.L. Acebal—data not published).*

surveillance and vector control activities. The economic losses imposed by arboviral diseases involve the patient's withdrawal from productive activities, drug expenses, hospitalization, medical consultations, treatment of sequelae, and death [81]. The time needed to treat and recover from arboviruses varies. On average, dengue removes the affected patient for 10–12 days from their work activities. The ZIKV may lead to birth defects including microcephaly and other severe brain malformations, which impose lifetime incapacity. Recovery from chikungunya varies from months to years [82].

Following the guidelines of the National Program for Dengue Control (PNCD), the MI-*Aedes* and MI-Virus technologies were adopted by the health authorities of the state of Minas Gerais (Brazil) and implemented in 21 cities with a high incidence of dengue in the period 2009–2011. The total cost of the program for all 21 cities for 2 years of work was less than US\$ 1.5 million, making an average of US\$ 71,428 per city. It included 4700 sticky traps, 115,000 sticky cards, synthetic oviposition attractants, RT-PCR on all mosquitoes caught in the traps, Web software licensing, cell phones, and technical support, among other items. The number of people benefited by the program was approximately 2 million, making the cost per habitant per year around US\$ 0.70. The cost-effectiveness was calculated as the cost of running the MI-*Aedes* and MI-Virus platforms divided by the number of cases of prevented arboviral diseases compared with cities that did not use the MI-*Aedes* platform and relied only in the PNCD guidelines. The MI-*Aedes* and MI-Virus platforms prevented a total of 27,191 cases at a total cost of US\$ 7.5 million, thus saving approximately US\$ 0.4 million in direct costs (health care and vector control) and US\$ 7.1 million in lost wages (societal impact) annually [41]. The cost-effectiveness of the platforms MI-*Aedes* and MI-Virus in cities with high mosquito infestation levels emphasizes the power of using these new technologies in vector control practices.

Currently, the MI-*Aedes* platform is running simultaneously in 154 Brazilian cities targeting approximately 7.5 million people, using about 12,000 sticky traps, and performing 625,000 trap inspections and around 8200 RT-PCR analysis per month on pooled mosquitoes (source: Ecovec. Ltd).

Investing more effort into integrating MI-*Aedes* strategies and costs with vector control operations, and standardizing the MI-*Aedes*-based control system across

cities, should help to increase the platform cost-effectiveness. Future studies should be conducted for developing new predictive model of serotype dynamics across cities for accurate arboviruses transmission.
