The Risk of Aedes aegypti Breeding and Premises Condition in South Mexico Author(s): Pablo Manrique-Saide, Clive R. Davies, Paul G. Coleman, Azael Che-Mendoza, Felipe Dzul-Manzanilla, Mario Barrera-Pérez, Silvia HernándezBetancourt, Guadalupe Ayora-Talavera, Miguel Pinkus-Rendón, Pierre BurciagaZúñiga, Gustavo Sánchez Tejeda, and Juan I. Arredondo-Jiménez Source: Journal of the American Mosquito Control Association, 29(4):337-345. 2013. Published By: The American Mosquito Control Association DOI: http://dx.doi.org/10.2987/13-6350.1 URL: http://www.bioone.org/doi/full/10.2987/13-6350.1
BioOne (www.bioone.org) is a nonprofit, online aggregation of core research in the biological, ecological, and environmental sciences. BioOne provides a sustainable online platform for over 170 journals and books published by nonprofit societies, associations, museums, institutions, and presses. Your use of this PDF, the BioOne Web site, and all posted and associated content indicates your acceptance of BioOne’s Terms of Use, available at www.bioone.org/page/ terms_of_use. Usage of BioOne content is strictly limited to personal, educational, and non-commercial use. Commercial inquiries or rights and permissions requests should be directed to the individual publisher as copyright holder.
BioOne sees sustainable scholarly publishing as an inherently collaborative enterprise connecting authors, nonprofit publishers, academic institutions, research libraries, and research funders in the common goal of maximizing access to critical research.
Journal of the American Mosquito Control Association, 29(4):337–345, 2013 Copyright E 2013 by The American Mosquito Control Association, Inc.
THE RISK OF AEDES AEGYPTI BREEDING AND PREMISES CONDITION IN SOUTH MEXICO PABLO MANRIQUE-SAIDE,1 CLIVE R. DAVIES,2 PAUL G. COLEMAN,2 AZAEL CHE-MENDOZA,3 ´ REZ,4 SILVIA HERNA ´ NDEZ-BETANCOURT,1 FELIPE DZUL-MANZANILLA,4 MARIO BARRERA-PE ´ N,6 PIERRE BURCIAGA-ZU ´N ˜ IGA,7 GUADALUPE AYORA-TALAVERA,5 MIGUEL PINKUS-RENDO ´ NCHEZ TEJEDA7 AND JUAN I. ARREDONDO-JIME ´ NEZ7 GUSTAVO SA ABSTRACT. A recent innovation instrumented for the Dengue Prevention and Control program in Mexico is the use of the premises condition index (PCI) as an indicator of risk for the vector Aedes aegypti infestation in dengue-endemic localities of Mexico. This paper addresses whether further improvements for the dengue control program could be made if the prevalence and productivity of Ae. aegypti populations could be reliably predicted using PCI at the household level, as well as medium-sized neighborhoods. We evaluated the use of PCI to predict the infestation with Aedes aegypti (breeding sites and immature productivity) in Merida, Mexico. The study consisted of a cross-sectional survey based on a clusterrandomized sampling design. We analyzed the statistical association between Aedes infestation and PCI, the extent to which the 3 components of PCI (house maintenance, and tidiness and shading of the patio) contributed to the association between PCI and infestation and whether infestation in a given premises was also affected by the PCI of the surrounding ones. Premises with the lowest PCI had significantly lower Aedes infestation and productivity; and as PCI scores increased infestation levels also tended to increase. Household PCI was significantly associated with Ae. aegypti breeding, largely due to the effect of patio untidiness and patio shade. The mean PCI within the surroundings premises also had a significant and independent explanatory power to predict the risk for infestation, in addition to individual PCI. This is the 1st study in Me´xico showing evidence that premises condition as measured by the PCI is related to Ae. aegypti breeding sites and immature productivity. Results suggest that PCI could be used to streamline surveys to inform control efforts at least where Ae. aegypti breeds outdoors, as in Merida. The effect of individual premises, neighborhood condition, and the risk of Aedes infestation imply that the risk for dengue vector infestation can only be minimized by the mass effect at the community level. KEY WORDS
Aedes aegypti, premises condition, breeding sites, productivity, Mexico
INTRODUCTION Although temperature, rainfall, and altitude are environmental macrodeterminants of Ae. aegypti (L.) abundance and distribution in Mexico (Iba´n˜ez-Bernal and Go´mez-Dante´s 1995, Peterson et al. 2005), monitoring of these variables is weakly informative for mosquito surveillance and targeted control strategies at lower scale resolutions. Meso- and microdeterminants of Ae. aegypti risk of infestation (i.e., the distribution and productivity of breeding sites within localities,
1 Departamento de Zoologı´a, Campus de Ciencias Biolo´gicas y Agropecuarias, Universidad Auto´noma de Yucata´n, Merida, Me´xico. 2 London School of Hygiene and Tropical Medicine, London, United Kingdom. 3 Servicios de Salud de Yucata´n, Merida, Me´xico. 4 Servicios de Salud del Estado de Guerrero, Chilpancingo, Me´xico. 5 Centro de Investigaciones Regionales ‘‘Dr. Hideyo Noguchi,’’ Universidad Auto ´ noma de Yucata´n, Merida, Me´xico. 6 Centro Peninsular en Humanidades y Ciencias Sociales, Universidad Nacional Auto ´ noma de Me´xico, Merida, Me´xico. 7 Programa de Enfermedades Transmitidas por Vector, Centro Nacional de Programas Preventivos y Control de Enfermedades, Secretarı´a de Salud, Mexico DF, Me´xico.
neighborhoods, and households) are seldom studied and hence, rarely employed for streamlining Aedes surveys and control. Consequently, targeting control against the dengue vector immature stages still depends on ground surveys as the most reliable tool for identifying larval habitats. Identification of these habitats sites is a costly and labor-intensive field, making sustainability a serious concern where resources are limited. The implication of an influential study (Tun-Lin et al. 1995a) in North Queensland, Australia, was that Aedes control can be more effective if certain breeding sites and houses are given priority. That study noted the existence of productive/key containers (those that produce a high proportion of pupae) as well as key premises (those with persistent or productive containers over time). The premises condition index (PCI) was developed based on these findings and proposed as a rapid, cost-effective indicator of risk of infestation of Ae. aegypti at the household level. The PCI combined conditions of the property, such as tidiness of houses and yards, with the degree of shade (Tun-Lin et al. 1995b). The use of pupal productivity surveys to identify and subsequently apply targeted control on key breeding places has been demonstrated by multicenter studies comparing survey techniques and the cost-effectiveness of targeted interventions versus holistic or blanket interventions
337
338
JOURNAL OF THE AMERICAN MOSQUITO CONTROL ASSOCIATION
(Focks 2003, Focks and Alexander 2006, Tun-Lin et al. 2009). However, few research projects have documented the use of PCI as a tool for the identification of key premises (Tun-Lin et al. 1995a, Moloney et al. 1998, Espinoza-Go´mez et al. 2001, Nogueira et al. 2005, Maciel-de-Freitas et al. 2008, Macoris et al. 2009, Basker and Ezhil 2012). A recent innovation instrumented for the Dengue Prevention and Control program in Mexico is the use of PCI as an indicator of risk for the vector Ae. aegypti infestation in dengueendemic localities of Mexico (DOF 2011). However, there is only 1 published report documenting the association of PCI and the presence of breeding sites in Mexico (Espinoza-Go´mez et al. 2001), and the characterization of house conditions that can promote the breeding of Ae. aegypti have not been fully explored or yet to be determined for particular dengue-prone urban settings in this country. This paper addresses whether further improvements for the dengue control program could be made if the prevalence and productivity of Ae. aegypti populations could be reliably predicted using PCI at the household level as well as medium-sized neighborhoods. Using data collected after extensive entomological surveys during the rainy season of 2003 (Manrique-Saide et al. 2008), we analyzed the statistical association between Aedes infestation and PCI, the extent to which the 3 components of PCI (house maintenance, and tidiness and shading of the patio/ backyard) contributed to the association between PCI and infestation, and finally, whether infestation in a given premises was also affected by the PCI of surrounding ones. If PCI significantly correlates with Ae. aegypti infestation, then PCI may represent a practical and cost-effective survey tool to 1) stratify areas by entomological risk, 2) target vector control interventions on those areas that are most at risk, and 3) focus health education messages on recommending changes at the household level that will have the greatest effect on vector abundance, and therefore disease incidence. MATERIALS AND METHODS Sampling design and study site The general characteristics of the sampling design and study area were comprehensively described by Manrique-Saide et al. (2008). Briefly, the study consisted of a cross-sectional survey based on a cluster randomized sampling design (Bennett et al. 1991). Twenty-nine clusters, corresponding to Basic Geo-Statistical Areas defined by the Mexican Institute of Geography and Statistics (AGEBs; 302 AGEBs in Merida) (INEGI 2000) from different sectors of the city
VOL. 29, NO. 4
(North, South, East, and West), were randomly selected as primary sampling units (Fig. 1). The AGEB is similar to a neighborhood and is a medium-sized area that would typically be used for regular Aedes surveillance. The northern sector has habitually been where upper and upper-middle classes live and where, correspondingly, most roads are paved. There is the largest coverage of street lighting, large houses and better schools, and commercial centers in the city. The southern city sector is the opposite, with low-income residents, more unpaved roads, incomplete coverage of street lighting, small one-roomed concrete blockhouses, and overcrowded schools. The eastern and western parts are lower-middle class with the city government undertaking highly subsidized housing development projects in these sectors. There were anecdotal reports relating levels of mosquito infestation risk to the differing socioeconomic characteristics; previous studies in Merida had documented Ae. aegypti house indices (percentage of houses positive for Aedes larvae) to be around 40 in the southern sector, and around 25 in some eastern and western neighborhoods (Lloyd 1992, Winch et al. 1992, Lloyd et al. 1994). No published studies are available documenting infestation levels from the North, but it was expected to be less than those reported in the other areas. Therefore, since different sectors of Merida were expected to differ in infestation rates, a separate randomization was made for each, which resulted in 4, 10, 7, and 8 AGEBs being chosen from the North, South, East, and West sectors, respectively. Within each selected AGEB, 40 premises were chosen, by starting at the center of the AGEB and sampling every 3rd one along transects on each cardinal direction (N–S and E–W); thus, sampling a total of 1,160 premises in all 29 AGEBs. The climate in Merida is mainly warm, with an annual average temperature of 20–27uC (minimum 18uC to maximum 36uC). Two seasons can be clearly distinguished: a rainy season in May to October (with most of the rainfall from June to October), and a dry season from November to April. This study was carried out during the rainy season in 2003. The average figures for the whole period were 203.2 mm of rainfall, 27.8uC (maximum 38.4uC), and 80.5% RH (data from the National Water Commission). The rainy season is considered the ‘‘dengue risk’’ season and marks the starting point for vector control activities. The city of Merida has historically concentrated .60% of all cases in the Mexican state of Yucatan, with continuous dengue transmission throughout the year with increased transmission (80% approximately) during the rainy season. Current vector control activities are: Temephos is used for larval control, but not extensively,
DECEMBER 2013
AEDES AEGYPTI
AND
PREMISES CONDITION IN SOUTH MEXICO
339
Fig. 1. Basic Geo-Statistic Areas (AGEB) and corresponding neighborhood name sampled from sectors in Merida (polygons represent each AGEB). North: 1) Chuburna, 2) Jardines de Merida, 3) Francisco de Montejo, 4) Maya; South: 5) 5 Colonias, 6) Castilla Camara, 7) Delio Moreno Canton, 8) Dolores Otero, 9) Mercedes Barrera, 10) Plan de Ayala, 11) San Antonio Xluch, 12) Santa Rosa, 13) Los Cocos, 14) Serapio Rendon; East: 15) Cortes Sarmiento, 16) Brisas, 17) Pacabtun, 18) Poligono 108, 19) San Antonio Kahua, 20) Vergel I, 21) Unidad Habitacional Morelos; and West: 22) Bojorquez, 23) Ibe´rica, 24) Sambula, 25) Xoclan, 26) Juan Pablo II, 27) Mulsay, 28) Yucalpeten, 29) Residencial Pensiones.
depending on the type of larval habitats present. Previous studies on productive container types for Ae. aegypti immatures in Merida have incriminated disposable containers and buckets/ pots, mostly rain-filled left in the backyards (Winch et al. 1992, Manrique-Saide et al. 2008, Garcı´a-Rejo ´ n et al. 2011). Regarding adulticides, from 2000 to 2010, permethrin was used as ultralow volume (ULV) space application and later replaced with sumithrin in 2011. Since January 2012, the Ministry of Health (MOH) of Yucatan replaced pyrethroid-based insecticides with chlorpyrifos for ULV space-spraying. For indoor spraying, pyrethroids were replaced with the carbamate propoxur. Premises condition index Premises quality was classified according to the categorical variables in the PCI (Tun-Lin et al. 1995a) that considers the following aspects: 1) house maintenance, 2) tidiness, and 3) shade, where each one itself it is given a value of 1, 2, or 3 (1 5 good , 2 5 indifferent, and 3 5 bad). In this study we classified house maintenance as 1) well-maintained, i.e., newly painted; 2) moder-
ately well-maintained house; and 3) not wellmaintained house, i.e., paint peeling, windows or doors broken or nonexistent, dilapidated. Because all the larval sites were found outdoors, tidiness and shade values focused on the patio (backyard) condition: 1) tidy, i.e., no rubbish evident, well-maintained vegetation; 2) moderately tidy; 3) untidy/or a mess; and shade: 1) very little or no shade (,25%), no major trees or bushes; 2) some shade (25–50%); 3) shady (.50%), e.g., large trees. Categories 1 and 3 were straightforward to score because they represented the opposites; category 2 was selected by default. Hence, the premises with the best characteristics scored a minimum PCI of 3 points, while the worst with the highest risk scored a maximum 9 points. The classification of each of the components of the PCI index was made by 4 data collectors and was standardized prior to the surveys, carrying out a pilot test in some premises of the city. Entomological variables Data on Ae. aegypti larvae and pupae were generated by direct inspection inside and around
(1.0) (2.0) (2.2) (4.4) (2.2) (4.9) (3.9) 0.7 1.4 1.5 3.1 1.5 3.4 2.7 (1.0) (2.9) (1.6) (1.5) (1.5) (5.0) (5.5) 0.1 0.3 0.2 0.2 0.2 0.5 0.6 (1.0) (2.3) (2.5) (3.6) (3.1) (4.5) (4.0) 6.8 15.3 16.7 24.6 20.9 30.4 27.3 62 230 381 903 308 391 60 2,335 9 48 39 45 30 57 12 240 6 25 42 72 42 35 6 228 (1.0) (2.3) (3.9) (6.2) (6.0) (7.9) (21.3) 5.3 12.4 20.4 33.1 32.0 41.7 113.1 (1.0) (0.5) (1.0) (1.0) (1.1) (3.2) (4.6) 0.5 0.3 0.6 0.6 0.6 1.7 2.6 (1.0) (1.5) (1.7) (2.4) (2.3) (2.9) (2.6) 15.9 24.5 27.1 38.2 35.8 46.1 40.9 467 2,015 5,128 9,701 6,440 4,791 2,488 31,030 48 48 144 168 120 200 56 784 14 40 68 112 72 53 9 368 88 163 251 293 201 115 22 1,133
Pupae-positive containers per Pupae per house house % pupaepositive houses
VOL. 29, NO. 4
3 4 5 6 7 8 9 Total
Statistical analysis of the relationships between PCI and Aedes presence/abundance (dictated by the resolution of the entomological data) were performed in StataH statistical software (version 8.1; Stata Corporation, College Station, TX). Three main sets of analyses were conducted to determine the utility of PCI as an indicator of Aedes infestations in premises. A 1st set of analyses tested whether each of the entomological measures increased significantly with PCI. The 2nd set of analyses addressed the extent to which each of the 3 components of PCI contributed to any observed association with overall household PCI. Finally, we investigated whether Aedes infestations in a given premises were not only affected by the PCI of the house in question, but also by the average PCI of the surrounding locality (i.e., AGEB). Using data from each premises, negative binomial regression analyses were undertaken to test for associations between PCI (fitted as categorical variables) and continuous Aedes outcomes (total number of positive containers for immature and only for pupae, and Ae. aegypti total larvae and pupae). Logistic regression analyses (and chi-square tests for trend) were performed to test for associations between PCI and categorical outcomes of infestation (i.e., presence/absence of larvae and/or pupae). Both sets of analyses adjusted for survey (beginning and end of the rainy season) and were clustered by AGEB to reduce the impact of spatial autocorrelation. The model with the best fit was considered the one with the lowest value for the Akaike Information Criterion (AIC), attempting to find the minimal model that correctly explained the data (Burnham and Anderson 2004). All analyses were then repeated replacing PCI with the 3 constituent indices (tidiness, shade, and
Immatures Pupae-positive PupaeTotal per house houses containers pupae
Data management and analysis
Positive No. houses Positive Positive Total % positive containers PCI surveyed houses containers immatures houses per house
all the selected premises following the recommendations of the Pan American Health Organization (PAHO 1994) and Focks (2003). Six entomological measures recorded were: 1) presence or absence of immature Aedes stages, i.e., larvae or pupae, 2) presence or absence of Aedes pupae, 3) number of containers with Aedes immatures, 4) number of containers with Aedes pupae, 5) number of Aedes immatures per house, and 6) number of Aedes pupae per house. Two entomological surveys were carried out in the city of Merida, at the start and end of the wet season in 2003 (June 22 to July 5, and July 7 to September 15). Both surveys were combined for analysis. Both premises condition and entomological surveys required a verbal informed consent of the residents, to whom the procedures were explained.
Relationship between premises condition index (PCI) and different immature entomological indicators in Merida (rainy season 2003). Numbers in parentheses are positivity expressed as a rate ratio relative to lowest PCI score.
JOURNAL OF THE AMERICAN MOSQUITO CONTROL ASSOCIATION
Table 1.
340
DECEMBER 2013 Table 2.
AEDES AEGYPTI
AND
PREMISES CONDITION IN SOUTH MEXICO
341
Association between premises condition index (PCI) and different entomological indicators for Aedes immature breeding. P-values and incidence rate ratios (IRR) are with respect to PCI 5 3. R2
PCI score
IRR
P
95% confidence intervals
Abundance of positive containers per house
0.018
Abundance of pupaepositive containers per house
0.018
No. of immatures (larvae/pupae) per house
0.005
No. of pupae per house
0.009
Presence of immaturepositive containers (positive)
0.049
Presence of pupaepositive containers
0.029
4 5 6 7 8 9 4 5 6 7 8 9 4 5 6 7 8 9 4 5 6 7 8 9 4 5 6 7 8 9 4 5 6 7 8 9
1.55 1.97 2.72 2.50 3.46 3.08 1.65 2.55 3.40 2.94 4.50 3.60 2.25 4.00 6.21 6.08 7.68 19.28 2.38 2.99 6.46 3.12 6.36 5.29 1.75 1.91 3.23 3.14 4.60 4.51 2.49 2.70 4.42 3.67 5.95 5.54
0.161 0.069 0.005*1 0.015* ,0.001* 0.021* 0.269 0.051 0.007* 0.034* ,0.001* 0.042* 0.084 0.001* 0.001* ,0.001* ,0.001* ,0.001* 0.153 0.091 0.009* 0.125 0.004* 0.090 0.066 0.095 0.002* 0.003* ,0.001* 0.023* 0.035* 0.043* 0.001* 0.011* ,0.001* 0.015*
0.84–2.88 0.95–4.08 1.35–5.45 1.20–5.24 1.82–6.58 1.19–8.02 0.68–4.00 1.00–6.60 1.40–8.26 1.08–7.96 2.02–10.04 1.05–2.33 0.88–5.74 1.77–9.06 2.19–17.64 2.22–16.68 2.87–20.57 5.10–73.45 0.73–7.80 0.84–10.68 1.58–26.44 0.73–13.36 1.83–22.09 0.77–36.27 0.97–3.16 0.89–4.08 1.51–6.93 1.46–6.76 2.07–10.23 1.23–16.56 1.07–5.79 1.03–7.09 1.88–10.68 1.35–10.01 2.34–15.02 1.40–21.99
1
* Significant P-values (P , 0.05).
Fig. 2. (a) Premises condition index (PCI) and the odds of premises being positive with immatures (OR with 95% confidence interval [CI] versus PCI 5 3); and (b) association of mean PCI (at Basic Geo-Statistic Area [AGEB] level) and the house index (proportion of houses positive for immatures in the same AGEB).
342
JOURNAL OF THE AMERICAN MOSQUITO CONTROL ASSOCIATION
VOL. 29, NO. 4
Table 3. Association between premises condition index (PCI) components and the 6 Aedes measurements evaluated in Merida houses. R2 values are given for the 6 full models, i.e., multivariate and for each of the univariate models, i.e., incorporating only 1 of the 3 PCI components, and adjusting for survey. P-values shown are for the comparisons shown within the appropriate multivariate model. P-values in parentheses are not significant (P $ 0.05).1 PCT
PUPCT
Total immatures
Total pupae
Positive houses
Pupaepositive houses
0.022
0.024
0.005
0.010
0.054
0.030
House maintenance (R ) 1 vs. 3 1 vs. 2 2 vs. 32
0.011 (0.842) 0.043 0.021
0.011 (0.988) (0.053) 0.021
0.002 (0.088) 0.001 0.021
0.006 (0.852) 0.044 0.043
0.033 (0.650) (0.060) (0.060)
0.013 (0.561) (0.077) (0.191)
Patio tidiness (R2) 1 vs. 3 1 vs. 2 2 vs. 3
0.013 0.003 (0.217) 0.008
0.011 0.007 (0.325) 0.023
0.004 ,0.001 ,0.001 (0.067)
0.006 0.011 (0.279) 0.011
0.037 0.003 (0.191) 0.015
0.016 0.009 (0.340) (0.053)
0.011 0.001 0.025 0.041
0.009 ,0.001 (0.062) 0.050
0.001 (0.055) (0.089) (0.334)
0.003 (0.078) (0.658) (0.111)
0.034 0.004 0.002 (0.159)
0.014 0.004 0.038 (0.081)
Full model (R2) 2
Patio shade (R2) 1 vs. 3 1 vs. 2 2 vs. 3 1 2
PCT, positive containers; PUPCT, pupae positive containers. Indicates that in all 6 comparisons, the Aedes infestation levels were higher in premises with house maintenance 2 in comparison to 3.
maintenance) also as categorical variables, to investigate their relative explanatory power. All the PCI regressions were repeated with the addition of the mean PCI at AGEB level as a further continuous explanatory variable. As before, all analyses adjusted for survey and were clustered by AGEB. RESULTS Data on individual premises grouped by their PCI scores (Table 1), showed significant positive correlations between PCI and 5 out of 6 of the Aedes infestation measurements: the presence of immatures (x2 for trend 5 29.3, P , 0.001), the presence of pupae (x2 for trend 5 19.7, P , 0.001), mean number of positive containers with immatures per house (r 5 0.825, P 5 0.02), mean number of immatures per house (r 5 0.846, P 5 0.017), and mean number of pupae per house (r 5 780, P 5 0.043). The mean number of containers with pupae per house was marginally nonsignificant (r 5 0.741, P 5 0.056). In all cases, the association with PCI was highly significant (Table 2), although the association between PCI score and number of pupae/house was the least significant. Premises with the lowest PCI (3) had significantly lower Aedes infestations measurements than those with PCI score .5, and as PCI score increased from 5, Aedes infestation levels tended to increase (Table 1). The best association between PCI values and infestation levels was observed with the odds of a premises being positive with immatures (R2 5 0.049; Table 2 and Fig. 2). Patio tidiness and patio shade provided independent significant explanatory power for Aedes distributions when the ‘‘worst’’ condition classes
were compared against the ‘‘best’’ (i.e., 1 vs. 3). The worst house maintenance had no discernible association with any of the infestation measurements. As observed in Table 3, the untidiest patios were positively associated with all 6 Aedes measurements, while the most shaded patios were positively associated with all but two, the number of Aedes immatures or pupae/house, with which it had no detectable correlation. Nevertheless, house maintenance cannot be ruled out for having no discernible association with any of the infestation measurements; its lack of explanatory power could be related with its association with patio tidiness (r 5 0.546, P , 0.001). When patio tidiness was extracted from the multivariate model, the worst maintained houses were positively associated with all 6 Aedes measurements except one, the number of Aedes pupae/house. The strongest entomological associations with the values of the PCI constituents (either combined in the full multivariate model, or singly in the 3 univariate models) were consistent with the odds of a premises being positive for immatures (Table 2). For all 6 Aedes infestation measures, the strongest trend was always with patio tidiness. The weakest associations were with house maintenance, which provided relatively little additional explanatory power for any of the Aedes measures after adjusting for premises shade or tidiness. The univariate effect of each constituent variable on the odds of a premises infestation is presented in Fig. 3. These confirm that the odds at household level for infestation increased 1.5and 2.2-fold as patio tidiness values rose from 1 to 2 or 3, respectively. Similarly, the odds for
DECEMBER 2013
AEDES AEGYPTI
AND
PREMISES CONDITION IN SOUTH MEXICO
343
explanatory power. Components integrated in as an index, PCI predicted better, the number of positive containers (AIC 5 2.12 vs. 2.50) and the odds for being positive with immatures (AIC 5 1.46 vs.1.65). Actually, the AIC values in the 4 remaining models did not differ (.0.03). Finally, after the regressions of PCI and the mean PCI (at AGEB level) on all the selected outcomes (Aedes measurements), the only significant effect of the average PCI of the surrounding locality (i.e., the AGEB) was observed on the number of houses positive for Aedes immatures, i.e., house index (OR 5 1.37; 95% confidence interval [CI] 1.04 to 1.72; P 5 0.007). As observed in Fig. 3, it was evident that AGEBs with poorest house conditions had the highest house indices in Merida. DISCUSSION
Fig. 3. (a) House maintenance, (b) patio tidiness, and (c) patio shade and the odds of premises being positive with immatures versus each of the premises condition index components 5 1.
infestation increased 1.5- and 2.2-fold as the patio shade values increased from 1 to 2 or 3, respectively. In contrast, the odds for infestation were higher for premises with a house maintenance value of 2, as compared with either 1 or 3. However, the PCI 3 constituent indices, individually or combined, did not provide significant
In Merida, nearly all Aedes infestation data were significantly associated with PCI, including immature and pupal productivity. This is the 1st study in Me´xico showing evidence that premises condition as measured by PCI is related to Ae. aegypti habitats and immature productivity. Maciel-de-Freitas et al. (2008) and Macoris et al. (2009) 1st reported, from Rio de Janeiro and Sao Paulo, Brazil, that premises with higher PCI scores were more productive for pupae, to the extent that .70% of pupae found on premises with PCI values $6. The same pattern was observed in Merida where premises with PCI values $6 contributed 60% of total pupaepositive containers and 71% of total pupae collected. Other previous studies reported an association between PCI and either the presence and the number of breeding sites, or the proportion of houses with immature habitats (Tun-Lin et al. 1995a, Espinoza-Go´mez et al. 2001, Basker and Ezhil 2012) and the occurrence of Ae. aegypti oviposition (using ovitraps) (Nogueira et al. 2005). Two main messages can arise from this study. First, the importance of untidy and shaded patios, and 2nd, that community efforts are required to reduce dengue vector infestation. Importantly, this study clearly shows that as PCI ‘‘worsens’’ (i.e., PCI .5), there is a significant increase at the household level in the risk of having positive containers for Ae. aegypti and the size of immatures harbored. These results suggest, as proposed by the other authors that PCI could be used to streamline surveys or control effort, at least where Ae. aegypti–productive containers are outdoors, as in Merida (Manrique-Saide et al. 2008, Garcı´a-Rejo ´ n et al. 2011). The association of Ae. aegypti population levels with PCI was mostly due to the impact of patio untidiness. House maintenance provided
344
JOURNAL OF THE AMERICAN MOSQUITO CONTROL ASSOCIATION
little additional predictive power, not least because in Merida it was strongly confounded by patio tidiness. The important finding was that there was no significant advantage in separating the PCI components as predictors of Aedes infestations—it was simpler and better to use single integrated PCI value. Nevertheless, and it deserves further exploration, a simpler PCI excluding ‘‘house maintenance’’ and focusing on untidy patios, particularly if they are shaded, could be operationally much easier to apply for vector control staff in Merida and other places, as mosquito productivity is associated with the abundance of unused and unprotected outdoor containers in shaded areas (Arunachalam et al. 2010). In addition to the strong association between PCI and the risk of Aedes infestation within a particular premises, this study demonstrated that PCI at the neighborhood or community level has an additional explanatory power in predicting the risk of infestation of an individual premises. According to results from this study, the probability of a premises for being positive was positively associated with the mean PCI value in the surrounding neighborhood (AGEB). This is a challenge for Aedes/dengue control programs, as it implies that controlling the mosquitoes in your patio would fail to completely prevent your household from being infested, unless the surrounding ones also took the same approach. Hence, to ensure personal protection from Aedes infestations the whole community needs to act, but ‘‘community participation’’ has to be understood as ‘‘social participation’’ involving the general public, but also authorities and different sectors on the planning and effective implementation of preventive and counteractive practices for dengue and Ae. aegypti control at individual and collective levels. A current key strategy for Ae. aegypti and dengue control adopted by MOH is ‘‘Patio limpio y cuidado del agua almacenada’’ (Tidy backyard and care for stored water). The effective use of resources for this program could be enhanced by 1) focusing vector control interventions on those areas that are most at risk, and 2) provide health education messages on recommending social and environmental changes at the household level that will have the greatest effect on vector abundance, and therefore disease incidence. The PCI, as far as we are aware, is currently not used systematically in any dengue surveillance program. The MOH is currently developing pilot projects on the utility and practicability of a ‘‘modified’’ PCI for entomological risk including both immature but also adult infestation, particularly with respect to Ae. aegypti adult females inside premises. Both studies previously cited from Brazil evaluating PCI as a tool to prioritize houses for vector control measures have reported
VOL. 29, NO. 4
a potential of PCI as an adult proxy for adult infestation. Macoris et al. (2009) found that houses in Sao Paulo with PCI values $6 accounted for 69% of adult-positive premises and 76% of total adult collections. Maciel-deFreitas et al. (2008) also observed that that keypremises distribution seemed to be correlated with areas where more adult Ae. aegypti females were collected. This reinforces the importance of further studies on the utility of PCI in Mexico as a proxy for adult Aedes distribution as well as being an easy-to-use and effective tool for directing surveillance and control activities in urban settings. The PCI information is being gathered in 65% (21/32) of the Mexican states and uploaded into a national platform (Plataforma Nacional de Vigilancia Entomolo´gica y Control Integral de Dengue) since 2012. In the near future, it is expected to use information derived from PCI in 2 ways: 1) to focalize integrated dengue vector control on houses/city blocks/neighborhoods/areas with high levels of PCI (6–9); and 2) to promote public health policies, at municipal, state, and national levels, to improve house quality, and ultimately to reduce vector infestation and the risk of disease transmission.
ACKNOWLEDGMENTS We express our sincere gratitude to Consejo Nacional de Ciencia y Tecnologı´a (CONACYT Mexico, Fondo Sectorial de Investigacio´n en Salud y Seguridad Social SSA/IMSS/ISSSTECONACYT SALUD-2011-1-161551 and the Programa de Impulso y Orientacion a la Investigacion (PRIORI-UADY) for their financial support of this project. REFERENCES CITED Arunachalam N, Tana S, Espino F, Kittayapong P, Abeyewickreme W, Wai KT, Tyagi BK, Kroeger A, Sommerfeld J, Petzold M. 2010. Eco-bio-social determinants of dengue vector breeding: a multicountry study in urban and periurban Asia. Bull WHO 88:173–184. Basker P, Ezhil R. 2012. Study in the correlation of premises condition index and the presence of larvae of Aedes species mosquitoes in human dwellings of the Cuddalore District of Tamil Nadu, India. Osong Public Health Res Perspect 3:3–7. Bennett S, Woods T, Liyanage WM, Smith DL. 1991. A simplified general method for cluster-sample surveys of health in developing countries. World Health Stat Q 44:98–106. Burnham KP, Anderson DR. 2004. Multimodel inference: understanding AIC and BIC in model selection. Sociol Method Res 33:261–304. Diario Oficial de la Federacion [DOF]. 2011. Para la vigilancia epidemiolo´gica, prevencio´n y control de enfermedades transmitidas por vector. Norma Oficial
DECEMBER 2013
AEDES AEGYPTI
AND
PREMISES CONDITION IN SOUTH MEXICO
Mexicana de Emergencia NOM-EM-003-SSA2-2010. Mexico DF, Mexico: Diario Oficial de la Federacion. Espinoza-Go ´ mez F, Hernandez CM, Coll R. 2001. Factors that modify the larval indices of Aedes aegypti in Colima, Mexico. Rev Panam Salud Publica 10:6–12. Focks DA. 2003. A review of entomological sampling methods and indicators for dengue vectors [Internet]. Geneva, Switzerland: World Health Organization on behalf of the Special Programme for Research and Training in Tropical Diseases [accessed September 2013]. Available from: http://whqlibdoc.who.int/hq/ 2003/TDR_IDE_DEN_03.1.pdf. Focks DA, Alexander N. 2006. Multicountry study of Aedes aegypti pupal productivity survey methodology: findings and recommendations [Internet]. Geneva, Switzerland: World Health Organization on behalf of the Special Programme for Research and Training in Tropical Diseases [accessed September 2013]. Available from: http://www.who.int/tdr/publications/ documents/aedes_aegypti.pdf. Garcı´a-Rejo ´ n JE, Lo ´ pez-Uribe MP, Loron˜o-Pino MA, Farfa´n-Ale JA, Del Najera-Vazquez MR, LozanoFuentes S, Beaty BJ, Eisen L. 2011. Productive container types for Aedes aegypti immatures in Me´rida, Me´xico. J Med Entomol 48:644–650. Iba´n˜ez-Bernal S, Go ´ mez-Dante´s H. 1995. Los vectores del dengue en Me´xico: una revisio´n crı´tica. Salud Publica Mex 37:53–63. Instituto Nacional de Estadı´stica, Geografı´a e Informa´tica [INEGI]. 2000. SCINCE, Sistema para la consulta de informacio´n censal 2000. Me´xico, DF: Instituto Nacional de Estadı´stica, Geografı´a e Informa´tica. Lloyd LS. 1992. Results of a community-based Aedes aegypti control intervention. In: Halstead SB, Go ´ mez–Dante´s H, eds. Dengue. A worldwide problem, a common strategy. Proceedings of the International Conference on Dengue and Aedes aegypti Community-Based Control Mexico, DF: Mexican Ministry of Health–Rockefeller Foundation. p 287–291. Lloyd LS, Winch P, Ortega-Canto J, Kendall C. 1994. The design of a community-based health education intervention for the control of Aedes aegypti. Am J Trop Med Hyg 50:401–411. Maciel-de-Freitas R, Peres RC, Souza-Santos R, Lourencode Oliveira R. 2008. Occurrence, productivity and spatial distribution of key premises in two dengue-endemic areas of Rio de Janeiro and their role in adult Aedes aegypti spatial infestation pattern. Trop Med Int Health 13:1488–1494.
345
Macoris M, Galvani K, da Grac¸a Macoris M. 2009. Evaluation of premise condition index in the context of Aedes aegypti control in Marı´lia, Sa˜o Paulo, Brazil. Dengue Bull 33:167–175. Manrique-Saide P, Davies CR, Coleman PG, RebollarTellez E, Che-Medoza A, Dzul-Manzanilla F, Zapata-Peniche A. 2008. Pupal surveys for Aedes aegypti surveillance and potential targeted control in residential areas of Me´rida, Me´xico. J Am Mosq Control Assoc 24:289–298. Moloney JM, Skelly C, Weinstein P, Maguire M, Ritchie S. 1998. Domestic Aedes aegypti breeding site surveillance: limitations of remote sensing as a predictive surveillance tool. Am J Trop Med Hyg 59:261–264. Nogueira LA, Gushi LT, Miranda JE, Madeira NG, Ribolla PEM. 2005. Application of an alternative Aedes species (Diptera: Culicidae) surveillance method in Botucatu city, Sa˜o Paulo, Brazil. Am J Trop Med Hyg 73:309–311. PAHO [Pan American Health Organization]. 1994. Dengue and dengue hemorrhagic fever in the Americas: guidelines for prevention and control. Washington, DC: Pan American Health Organization, Regional Office of the World Health Organization. Peterson AT, Martı´nez-Campos C, Nakazawa Y, Martı´nez-Meyer E. 2005. Time-specific ecological niche modelling predicts spatial dynamics of vector insects and human dengue cases. Trans R Soc Trop Med Hyg 99:647–655. Tun-Lin W, Kay BH, Barnes A. 1995a. Understanding productivity, a key to Aedes aegypti surveillance. Am J Trop Med Hyg 53:595–601. Tun-Lin W, Kay BH, Barnes A. 1995b. The Premise Condition Index: a tool for streamlining surveys of Aedes aegypti. Am J Trop Med Hyg 53:591–594. Tun-Lin W, Lenhart A, Nam VS, Rebollar-Te´llez E, Morrison AC, Barbazan P, Cote M, Midega J, Sanchez F, Manrique-Saide P, Kroeger A, Nathan M, Meheus F, Petzold M. 2009. Reducing costs and operational constraints of dengue vector control by targeting productive breeding places: a multi-country non-inferiority cluster randomized trial. Trop Med Int Health 14:1143–1153. Winch PJ, Barrientos-Sanchez G, Puigserver-Castro E, Manzano-Cabrera L, Lloyd LS, Mendez-Galvan J. 1992. Variation in Aedes aegypti larval indices over one year period in a neighborhood of Merida, Yucatan, Mexico. J Am Mosq Control Assoc 8: 193–195.
BioOne (www.bioone.org) is a nonprofit, online aggregation of core research in the biological, ecological, and environmental sciences. BioOne provides a sustainable online platform for over 170 journals and books published by nonprofit societies, associations, museums, institutions, and presses. Your use of this PDF, the BioOne Web site, and all posted and associated content indicates your acceptance of BioOne’s Terms of Use, available at www.bioone.org/page/ terms_of_use. Usage of BioOne content is strictly limited to personal, educational, and non-commercial use. Commercial inquiries or rights and permissions requests should be directed to the individual publisher as copyright holder.
BioOne sees sustainable scholarly publishing as an inherently collaborative enterprise connecting authors, nonprofit publishers, academic institutions, research libraries, and research funders in the common goal of maximizing access to critical research.
Journal of the American Mosquito Control Association, 29(4):337–345, 2013 Copyright E 2013 by The American Mosquito Control Association, Inc.
THE RISK OF AEDES AEGYPTI BREEDING AND PREMISES CONDITION IN SOUTH MEXICO PABLO MANRIQUE-SAIDE,1 CLIVE R. DAVIES,2 PAUL G. COLEMAN,2 AZAEL CHE-MENDOZA,3 ´ REZ,4 SILVIA HERNA ´ NDEZ-BETANCOURT,1 FELIPE DZUL-MANZANILLA,4 MARIO BARRERA-PE ´ N,6 PIERRE BURCIAGA-ZU ´N ˜ IGA,7 GUADALUPE AYORA-TALAVERA,5 MIGUEL PINKUS-RENDO ´ NCHEZ TEJEDA7 AND JUAN I. ARREDONDO-JIME ´ NEZ7 GUSTAVO SA ABSTRACT. A recent innovation instrumented for the Dengue Prevention and Control program in Mexico is the use of the premises condition index (PCI) as an indicator of risk for the vector Aedes aegypti infestation in dengue-endemic localities of Mexico. This paper addresses whether further improvements for the dengue control program could be made if the prevalence and productivity of Ae. aegypti populations could be reliably predicted using PCI at the household level, as well as medium-sized neighborhoods. We evaluated the use of PCI to predict the infestation with Aedes aegypti (breeding sites and immature productivity) in Merida, Mexico. The study consisted of a cross-sectional survey based on a clusterrandomized sampling design. We analyzed the statistical association between Aedes infestation and PCI, the extent to which the 3 components of PCI (house maintenance, and tidiness and shading of the patio) contributed to the association between PCI and infestation and whether infestation in a given premises was also affected by the PCI of the surrounding ones. Premises with the lowest PCI had significantly lower Aedes infestation and productivity; and as PCI scores increased infestation levels also tended to increase. Household PCI was significantly associated with Ae. aegypti breeding, largely due to the effect of patio untidiness and patio shade. The mean PCI within the surroundings premises also had a significant and independent explanatory power to predict the risk for infestation, in addition to individual PCI. This is the 1st study in Me´xico showing evidence that premises condition as measured by the PCI is related to Ae. aegypti breeding sites and immature productivity. Results suggest that PCI could be used to streamline surveys to inform control efforts at least where Ae. aegypti breeds outdoors, as in Merida. The effect of individual premises, neighborhood condition, and the risk of Aedes infestation imply that the risk for dengue vector infestation can only be minimized by the mass effect at the community level. KEY WORDS
Aedes aegypti, premises condition, breeding sites, productivity, Mexico
INTRODUCTION Although temperature, rainfall, and altitude are environmental macrodeterminants of Ae. aegypti (L.) abundance and distribution in Mexico (Iba´n˜ez-Bernal and Go´mez-Dante´s 1995, Peterson et al. 2005), monitoring of these variables is weakly informative for mosquito surveillance and targeted control strategies at lower scale resolutions. Meso- and microdeterminants of Ae. aegypti risk of infestation (i.e., the distribution and productivity of breeding sites within localities,
1 Departamento de Zoologı´a, Campus de Ciencias Biolo´gicas y Agropecuarias, Universidad Auto´noma de Yucata´n, Merida, Me´xico. 2 London School of Hygiene and Tropical Medicine, London, United Kingdom. 3 Servicios de Salud de Yucata´n, Merida, Me´xico. 4 Servicios de Salud del Estado de Guerrero, Chilpancingo, Me´xico. 5 Centro de Investigaciones Regionales ‘‘Dr. Hideyo Noguchi,’’ Universidad Auto ´ noma de Yucata´n, Merida, Me´xico. 6 Centro Peninsular en Humanidades y Ciencias Sociales, Universidad Nacional Auto ´ noma de Me´xico, Merida, Me´xico. 7 Programa de Enfermedades Transmitidas por Vector, Centro Nacional de Programas Preventivos y Control de Enfermedades, Secretarı´a de Salud, Mexico DF, Me´xico.
neighborhoods, and households) are seldom studied and hence, rarely employed for streamlining Aedes surveys and control. Consequently, targeting control against the dengue vector immature stages still depends on ground surveys as the most reliable tool for identifying larval habitats. Identification of these habitats sites is a costly and labor-intensive field, making sustainability a serious concern where resources are limited. The implication of an influential study (Tun-Lin et al. 1995a) in North Queensland, Australia, was that Aedes control can be more effective if certain breeding sites and houses are given priority. That study noted the existence of productive/key containers (those that produce a high proportion of pupae) as well as key premises (those with persistent or productive containers over time). The premises condition index (PCI) was developed based on these findings and proposed as a rapid, cost-effective indicator of risk of infestation of Ae. aegypti at the household level. The PCI combined conditions of the property, such as tidiness of houses and yards, with the degree of shade (Tun-Lin et al. 1995b). The use of pupal productivity surveys to identify and subsequently apply targeted control on key breeding places has been demonstrated by multicenter studies comparing survey techniques and the cost-effectiveness of targeted interventions versus holistic or blanket interventions
337
338
JOURNAL OF THE AMERICAN MOSQUITO CONTROL ASSOCIATION
(Focks 2003, Focks and Alexander 2006, Tun-Lin et al. 2009). However, few research projects have documented the use of PCI as a tool for the identification of key premises (Tun-Lin et al. 1995a, Moloney et al. 1998, Espinoza-Go´mez et al. 2001, Nogueira et al. 2005, Maciel-de-Freitas et al. 2008, Macoris et al. 2009, Basker and Ezhil 2012). A recent innovation instrumented for the Dengue Prevention and Control program in Mexico is the use of PCI as an indicator of risk for the vector Ae. aegypti infestation in dengueendemic localities of Mexico (DOF 2011). However, there is only 1 published report documenting the association of PCI and the presence of breeding sites in Mexico (Espinoza-Go´mez et al. 2001), and the characterization of house conditions that can promote the breeding of Ae. aegypti have not been fully explored or yet to be determined for particular dengue-prone urban settings in this country. This paper addresses whether further improvements for the dengue control program could be made if the prevalence and productivity of Ae. aegypti populations could be reliably predicted using PCI at the household level as well as medium-sized neighborhoods. Using data collected after extensive entomological surveys during the rainy season of 2003 (Manrique-Saide et al. 2008), we analyzed the statistical association between Aedes infestation and PCI, the extent to which the 3 components of PCI (house maintenance, and tidiness and shading of the patio/ backyard) contributed to the association between PCI and infestation, and finally, whether infestation in a given premises was also affected by the PCI of surrounding ones. If PCI significantly correlates with Ae. aegypti infestation, then PCI may represent a practical and cost-effective survey tool to 1) stratify areas by entomological risk, 2) target vector control interventions on those areas that are most at risk, and 3) focus health education messages on recommending changes at the household level that will have the greatest effect on vector abundance, and therefore disease incidence. MATERIALS AND METHODS Sampling design and study site The general characteristics of the sampling design and study area were comprehensively described by Manrique-Saide et al. (2008). Briefly, the study consisted of a cross-sectional survey based on a cluster randomized sampling design (Bennett et al. 1991). Twenty-nine clusters, corresponding to Basic Geo-Statistical Areas defined by the Mexican Institute of Geography and Statistics (AGEBs; 302 AGEBs in Merida) (INEGI 2000) from different sectors of the city
VOL. 29, NO. 4
(North, South, East, and West), were randomly selected as primary sampling units (Fig. 1). The AGEB is similar to a neighborhood and is a medium-sized area that would typically be used for regular Aedes surveillance. The northern sector has habitually been where upper and upper-middle classes live and where, correspondingly, most roads are paved. There is the largest coverage of street lighting, large houses and better schools, and commercial centers in the city. The southern city sector is the opposite, with low-income residents, more unpaved roads, incomplete coverage of street lighting, small one-roomed concrete blockhouses, and overcrowded schools. The eastern and western parts are lower-middle class with the city government undertaking highly subsidized housing development projects in these sectors. There were anecdotal reports relating levels of mosquito infestation risk to the differing socioeconomic characteristics; previous studies in Merida had documented Ae. aegypti house indices (percentage of houses positive for Aedes larvae) to be around 40 in the southern sector, and around 25 in some eastern and western neighborhoods (Lloyd 1992, Winch et al. 1992, Lloyd et al. 1994). No published studies are available documenting infestation levels from the North, but it was expected to be less than those reported in the other areas. Therefore, since different sectors of Merida were expected to differ in infestation rates, a separate randomization was made for each, which resulted in 4, 10, 7, and 8 AGEBs being chosen from the North, South, East, and West sectors, respectively. Within each selected AGEB, 40 premises were chosen, by starting at the center of the AGEB and sampling every 3rd one along transects on each cardinal direction (N–S and E–W); thus, sampling a total of 1,160 premises in all 29 AGEBs. The climate in Merida is mainly warm, with an annual average temperature of 20–27uC (minimum 18uC to maximum 36uC). Two seasons can be clearly distinguished: a rainy season in May to October (with most of the rainfall from June to October), and a dry season from November to April. This study was carried out during the rainy season in 2003. The average figures for the whole period were 203.2 mm of rainfall, 27.8uC (maximum 38.4uC), and 80.5% RH (data from the National Water Commission). The rainy season is considered the ‘‘dengue risk’’ season and marks the starting point for vector control activities. The city of Merida has historically concentrated .60% of all cases in the Mexican state of Yucatan, with continuous dengue transmission throughout the year with increased transmission (80% approximately) during the rainy season. Current vector control activities are: Temephos is used for larval control, but not extensively,
DECEMBER 2013
AEDES AEGYPTI
AND
PREMISES CONDITION IN SOUTH MEXICO
339
Fig. 1. Basic Geo-Statistic Areas (AGEB) and corresponding neighborhood name sampled from sectors in Merida (polygons represent each AGEB). North: 1) Chuburna, 2) Jardines de Merida, 3) Francisco de Montejo, 4) Maya; South: 5) 5 Colonias, 6) Castilla Camara, 7) Delio Moreno Canton, 8) Dolores Otero, 9) Mercedes Barrera, 10) Plan de Ayala, 11) San Antonio Xluch, 12) Santa Rosa, 13) Los Cocos, 14) Serapio Rendon; East: 15) Cortes Sarmiento, 16) Brisas, 17) Pacabtun, 18) Poligono 108, 19) San Antonio Kahua, 20) Vergel I, 21) Unidad Habitacional Morelos; and West: 22) Bojorquez, 23) Ibe´rica, 24) Sambula, 25) Xoclan, 26) Juan Pablo II, 27) Mulsay, 28) Yucalpeten, 29) Residencial Pensiones.
depending on the type of larval habitats present. Previous studies on productive container types for Ae. aegypti immatures in Merida have incriminated disposable containers and buckets/ pots, mostly rain-filled left in the backyards (Winch et al. 1992, Manrique-Saide et al. 2008, Garcı´a-Rejo ´ n et al. 2011). Regarding adulticides, from 2000 to 2010, permethrin was used as ultralow volume (ULV) space application and later replaced with sumithrin in 2011. Since January 2012, the Ministry of Health (MOH) of Yucatan replaced pyrethroid-based insecticides with chlorpyrifos for ULV space-spraying. For indoor spraying, pyrethroids were replaced with the carbamate propoxur. Premises condition index Premises quality was classified according to the categorical variables in the PCI (Tun-Lin et al. 1995a) that considers the following aspects: 1) house maintenance, 2) tidiness, and 3) shade, where each one itself it is given a value of 1, 2, or 3 (1 5 good , 2 5 indifferent, and 3 5 bad). In this study we classified house maintenance as 1) well-maintained, i.e., newly painted; 2) moder-
ately well-maintained house; and 3) not wellmaintained house, i.e., paint peeling, windows or doors broken or nonexistent, dilapidated. Because all the larval sites were found outdoors, tidiness and shade values focused on the patio (backyard) condition: 1) tidy, i.e., no rubbish evident, well-maintained vegetation; 2) moderately tidy; 3) untidy/or a mess; and shade: 1) very little or no shade (,25%), no major trees or bushes; 2) some shade (25–50%); 3) shady (.50%), e.g., large trees. Categories 1 and 3 were straightforward to score because they represented the opposites; category 2 was selected by default. Hence, the premises with the best characteristics scored a minimum PCI of 3 points, while the worst with the highest risk scored a maximum 9 points. The classification of each of the components of the PCI index was made by 4 data collectors and was standardized prior to the surveys, carrying out a pilot test in some premises of the city. Entomological variables Data on Ae. aegypti larvae and pupae were generated by direct inspection inside and around
(1.0) (2.0) (2.2) (4.4) (2.2) (4.9) (3.9) 0.7 1.4 1.5 3.1 1.5 3.4 2.7 (1.0) (2.9) (1.6) (1.5) (1.5) (5.0) (5.5) 0.1 0.3 0.2 0.2 0.2 0.5 0.6 (1.0) (2.3) (2.5) (3.6) (3.1) (4.5) (4.0) 6.8 15.3 16.7 24.6 20.9 30.4 27.3 62 230 381 903 308 391 60 2,335 9 48 39 45 30 57 12 240 6 25 42 72 42 35 6 228 (1.0) (2.3) (3.9) (6.2) (6.0) (7.9) (21.3) 5.3 12.4 20.4 33.1 32.0 41.7 113.1 (1.0) (0.5) (1.0) (1.0) (1.1) (3.2) (4.6) 0.5 0.3 0.6 0.6 0.6 1.7 2.6 (1.0) (1.5) (1.7) (2.4) (2.3) (2.9) (2.6) 15.9 24.5 27.1 38.2 35.8 46.1 40.9 467 2,015 5,128 9,701 6,440 4,791 2,488 31,030 48 48 144 168 120 200 56 784 14 40 68 112 72 53 9 368 88 163 251 293 201 115 22 1,133
Pupae-positive containers per Pupae per house house % pupaepositive houses
VOL. 29, NO. 4
3 4 5 6 7 8 9 Total
Statistical analysis of the relationships between PCI and Aedes presence/abundance (dictated by the resolution of the entomological data) were performed in StataH statistical software (version 8.1; Stata Corporation, College Station, TX). Three main sets of analyses were conducted to determine the utility of PCI as an indicator of Aedes infestations in premises. A 1st set of analyses tested whether each of the entomological measures increased significantly with PCI. The 2nd set of analyses addressed the extent to which each of the 3 components of PCI contributed to any observed association with overall household PCI. Finally, we investigated whether Aedes infestations in a given premises were not only affected by the PCI of the house in question, but also by the average PCI of the surrounding locality (i.e., AGEB). Using data from each premises, negative binomial regression analyses were undertaken to test for associations between PCI (fitted as categorical variables) and continuous Aedes outcomes (total number of positive containers for immature and only for pupae, and Ae. aegypti total larvae and pupae). Logistic regression analyses (and chi-square tests for trend) were performed to test for associations between PCI and categorical outcomes of infestation (i.e., presence/absence of larvae and/or pupae). Both sets of analyses adjusted for survey (beginning and end of the rainy season) and were clustered by AGEB to reduce the impact of spatial autocorrelation. The model with the best fit was considered the one with the lowest value for the Akaike Information Criterion (AIC), attempting to find the minimal model that correctly explained the data (Burnham and Anderson 2004). All analyses were then repeated replacing PCI with the 3 constituent indices (tidiness, shade, and
Immatures Pupae-positive PupaeTotal per house houses containers pupae
Data management and analysis
Positive No. houses Positive Positive Total % positive containers PCI surveyed houses containers immatures houses per house
all the selected premises following the recommendations of the Pan American Health Organization (PAHO 1994) and Focks (2003). Six entomological measures recorded were: 1) presence or absence of immature Aedes stages, i.e., larvae or pupae, 2) presence or absence of Aedes pupae, 3) number of containers with Aedes immatures, 4) number of containers with Aedes pupae, 5) number of Aedes immatures per house, and 6) number of Aedes pupae per house. Two entomological surveys were carried out in the city of Merida, at the start and end of the wet season in 2003 (June 22 to July 5, and July 7 to September 15). Both surveys were combined for analysis. Both premises condition and entomological surveys required a verbal informed consent of the residents, to whom the procedures were explained.
Relationship between premises condition index (PCI) and different immature entomological indicators in Merida (rainy season 2003). Numbers in parentheses are positivity expressed as a rate ratio relative to lowest PCI score.
JOURNAL OF THE AMERICAN MOSQUITO CONTROL ASSOCIATION
Table 1.
340
DECEMBER 2013 Table 2.
AEDES AEGYPTI
AND
PREMISES CONDITION IN SOUTH MEXICO
341
Association between premises condition index (PCI) and different entomological indicators for Aedes immature breeding. P-values and incidence rate ratios (IRR) are with respect to PCI 5 3. R2
PCI score
IRR
P
95% confidence intervals
Abundance of positive containers per house
0.018
Abundance of pupaepositive containers per house
0.018
No. of immatures (larvae/pupae) per house
0.005
No. of pupae per house
0.009
Presence of immaturepositive containers (positive)
0.049
Presence of pupaepositive containers
0.029
4 5 6 7 8 9 4 5 6 7 8 9 4 5 6 7 8 9 4 5 6 7 8 9 4 5 6 7 8 9 4 5 6 7 8 9
1.55 1.97 2.72 2.50 3.46 3.08 1.65 2.55 3.40 2.94 4.50 3.60 2.25 4.00 6.21 6.08 7.68 19.28 2.38 2.99 6.46 3.12 6.36 5.29 1.75 1.91 3.23 3.14 4.60 4.51 2.49 2.70 4.42 3.67 5.95 5.54
0.161 0.069 0.005*1 0.015* ,0.001* 0.021* 0.269 0.051 0.007* 0.034* ,0.001* 0.042* 0.084 0.001* 0.001* ,0.001* ,0.001* ,0.001* 0.153 0.091 0.009* 0.125 0.004* 0.090 0.066 0.095 0.002* 0.003* ,0.001* 0.023* 0.035* 0.043* 0.001* 0.011* ,0.001* 0.015*
0.84–2.88 0.95–4.08 1.35–5.45 1.20–5.24 1.82–6.58 1.19–8.02 0.68–4.00 1.00–6.60 1.40–8.26 1.08–7.96 2.02–10.04 1.05–2.33 0.88–5.74 1.77–9.06 2.19–17.64 2.22–16.68 2.87–20.57 5.10–73.45 0.73–7.80 0.84–10.68 1.58–26.44 0.73–13.36 1.83–22.09 0.77–36.27 0.97–3.16 0.89–4.08 1.51–6.93 1.46–6.76 2.07–10.23 1.23–16.56 1.07–5.79 1.03–7.09 1.88–10.68 1.35–10.01 2.34–15.02 1.40–21.99
1
* Significant P-values (P , 0.05).
Fig. 2. (a) Premises condition index (PCI) and the odds of premises being positive with immatures (OR with 95% confidence interval [CI] versus PCI 5 3); and (b) association of mean PCI (at Basic Geo-Statistic Area [AGEB] level) and the house index (proportion of houses positive for immatures in the same AGEB).
342
JOURNAL OF THE AMERICAN MOSQUITO CONTROL ASSOCIATION
VOL. 29, NO. 4
Table 3. Association between premises condition index (PCI) components and the 6 Aedes measurements evaluated in Merida houses. R2 values are given for the 6 full models, i.e., multivariate and for each of the univariate models, i.e., incorporating only 1 of the 3 PCI components, and adjusting for survey. P-values shown are for the comparisons shown within the appropriate multivariate model. P-values in parentheses are not significant (P $ 0.05).1 PCT
PUPCT
Total immatures
Total pupae
Positive houses
Pupaepositive houses
0.022
0.024
0.005
0.010
0.054
0.030
House maintenance (R ) 1 vs. 3 1 vs. 2 2 vs. 32
0.011 (0.842) 0.043 0.021
0.011 (0.988) (0.053) 0.021
0.002 (0.088) 0.001 0.021
0.006 (0.852) 0.044 0.043
0.033 (0.650) (0.060) (0.060)
0.013 (0.561) (0.077) (0.191)
Patio tidiness (R2) 1 vs. 3 1 vs. 2 2 vs. 3
0.013 0.003 (0.217) 0.008
0.011 0.007 (0.325) 0.023
0.004 ,0.001 ,0.001 (0.067)
0.006 0.011 (0.279) 0.011
0.037 0.003 (0.191) 0.015
0.016 0.009 (0.340) (0.053)
0.011 0.001 0.025 0.041
0.009 ,0.001 (0.062) 0.050
0.001 (0.055) (0.089) (0.334)
0.003 (0.078) (0.658) (0.111)
0.034 0.004 0.002 (0.159)
0.014 0.004 0.038 (0.081)
Full model (R2) 2
Patio shade (R2) 1 vs. 3 1 vs. 2 2 vs. 3 1 2
PCT, positive containers; PUPCT, pupae positive containers. Indicates that in all 6 comparisons, the Aedes infestation levels were higher in premises with house maintenance 2 in comparison to 3.
maintenance) also as categorical variables, to investigate their relative explanatory power. All the PCI regressions were repeated with the addition of the mean PCI at AGEB level as a further continuous explanatory variable. As before, all analyses adjusted for survey and were clustered by AGEB. RESULTS Data on individual premises grouped by their PCI scores (Table 1), showed significant positive correlations between PCI and 5 out of 6 of the Aedes infestation measurements: the presence of immatures (x2 for trend 5 29.3, P , 0.001), the presence of pupae (x2 for trend 5 19.7, P , 0.001), mean number of positive containers with immatures per house (r 5 0.825, P 5 0.02), mean number of immatures per house (r 5 0.846, P 5 0.017), and mean number of pupae per house (r 5 780, P 5 0.043). The mean number of containers with pupae per house was marginally nonsignificant (r 5 0.741, P 5 0.056). In all cases, the association with PCI was highly significant (Table 2), although the association between PCI score and number of pupae/house was the least significant. Premises with the lowest PCI (3) had significantly lower Aedes infestations measurements than those with PCI score .5, and as PCI score increased from 5, Aedes infestation levels tended to increase (Table 1). The best association between PCI values and infestation levels was observed with the odds of a premises being positive with immatures (R2 5 0.049; Table 2 and Fig. 2). Patio tidiness and patio shade provided independent significant explanatory power for Aedes distributions when the ‘‘worst’’ condition classes
were compared against the ‘‘best’’ (i.e., 1 vs. 3). The worst house maintenance had no discernible association with any of the infestation measurements. As observed in Table 3, the untidiest patios were positively associated with all 6 Aedes measurements, while the most shaded patios were positively associated with all but two, the number of Aedes immatures or pupae/house, with which it had no detectable correlation. Nevertheless, house maintenance cannot be ruled out for having no discernible association with any of the infestation measurements; its lack of explanatory power could be related with its association with patio tidiness (r 5 0.546, P , 0.001). When patio tidiness was extracted from the multivariate model, the worst maintained houses were positively associated with all 6 Aedes measurements except one, the number of Aedes pupae/house. The strongest entomological associations with the values of the PCI constituents (either combined in the full multivariate model, or singly in the 3 univariate models) were consistent with the odds of a premises being positive for immatures (Table 2). For all 6 Aedes infestation measures, the strongest trend was always with patio tidiness. The weakest associations were with house maintenance, which provided relatively little additional explanatory power for any of the Aedes measures after adjusting for premises shade or tidiness. The univariate effect of each constituent variable on the odds of a premises infestation is presented in Fig. 3. These confirm that the odds at household level for infestation increased 1.5and 2.2-fold as patio tidiness values rose from 1 to 2 or 3, respectively. Similarly, the odds for
DECEMBER 2013
AEDES AEGYPTI
AND
PREMISES CONDITION IN SOUTH MEXICO
343
explanatory power. Components integrated in as an index, PCI predicted better, the number of positive containers (AIC 5 2.12 vs. 2.50) and the odds for being positive with immatures (AIC 5 1.46 vs.1.65). Actually, the AIC values in the 4 remaining models did not differ (.0.03). Finally, after the regressions of PCI and the mean PCI (at AGEB level) on all the selected outcomes (Aedes measurements), the only significant effect of the average PCI of the surrounding locality (i.e., the AGEB) was observed on the number of houses positive for Aedes immatures, i.e., house index (OR 5 1.37; 95% confidence interval [CI] 1.04 to 1.72; P 5 0.007). As observed in Fig. 3, it was evident that AGEBs with poorest house conditions had the highest house indices in Merida. DISCUSSION
Fig. 3. (a) House maintenance, (b) patio tidiness, and (c) patio shade and the odds of premises being positive with immatures versus each of the premises condition index components 5 1.
infestation increased 1.5- and 2.2-fold as the patio shade values increased from 1 to 2 or 3, respectively. In contrast, the odds for infestation were higher for premises with a house maintenance value of 2, as compared with either 1 or 3. However, the PCI 3 constituent indices, individually or combined, did not provide significant
In Merida, nearly all Aedes infestation data were significantly associated with PCI, including immature and pupal productivity. This is the 1st study in Me´xico showing evidence that premises condition as measured by PCI is related to Ae. aegypti habitats and immature productivity. Maciel-de-Freitas et al. (2008) and Macoris et al. (2009) 1st reported, from Rio de Janeiro and Sao Paulo, Brazil, that premises with higher PCI scores were more productive for pupae, to the extent that .70% of pupae found on premises with PCI values $6. The same pattern was observed in Merida where premises with PCI values $6 contributed 60% of total pupaepositive containers and 71% of total pupae collected. Other previous studies reported an association between PCI and either the presence and the number of breeding sites, or the proportion of houses with immature habitats (Tun-Lin et al. 1995a, Espinoza-Go´mez et al. 2001, Basker and Ezhil 2012) and the occurrence of Ae. aegypti oviposition (using ovitraps) (Nogueira et al. 2005). Two main messages can arise from this study. First, the importance of untidy and shaded patios, and 2nd, that community efforts are required to reduce dengue vector infestation. Importantly, this study clearly shows that as PCI ‘‘worsens’’ (i.e., PCI .5), there is a significant increase at the household level in the risk of having positive containers for Ae. aegypti and the size of immatures harbored. These results suggest, as proposed by the other authors that PCI could be used to streamline surveys or control effort, at least where Ae. aegypti–productive containers are outdoors, as in Merida (Manrique-Saide et al. 2008, Garcı´a-Rejo ´ n et al. 2011). The association of Ae. aegypti population levels with PCI was mostly due to the impact of patio untidiness. House maintenance provided
344
JOURNAL OF THE AMERICAN MOSQUITO CONTROL ASSOCIATION
little additional predictive power, not least because in Merida it was strongly confounded by patio tidiness. The important finding was that there was no significant advantage in separating the PCI components as predictors of Aedes infestations—it was simpler and better to use single integrated PCI value. Nevertheless, and it deserves further exploration, a simpler PCI excluding ‘‘house maintenance’’ and focusing on untidy patios, particularly if they are shaded, could be operationally much easier to apply for vector control staff in Merida and other places, as mosquito productivity is associated with the abundance of unused and unprotected outdoor containers in shaded areas (Arunachalam et al. 2010). In addition to the strong association between PCI and the risk of Aedes infestation within a particular premises, this study demonstrated that PCI at the neighborhood or community level has an additional explanatory power in predicting the risk of infestation of an individual premises. According to results from this study, the probability of a premises for being positive was positively associated with the mean PCI value in the surrounding neighborhood (AGEB). This is a challenge for Aedes/dengue control programs, as it implies that controlling the mosquitoes in your patio would fail to completely prevent your household from being infested, unless the surrounding ones also took the same approach. Hence, to ensure personal protection from Aedes infestations the whole community needs to act, but ‘‘community participation’’ has to be understood as ‘‘social participation’’ involving the general public, but also authorities and different sectors on the planning and effective implementation of preventive and counteractive practices for dengue and Ae. aegypti control at individual and collective levels. A current key strategy for Ae. aegypti and dengue control adopted by MOH is ‘‘Patio limpio y cuidado del agua almacenada’’ (Tidy backyard and care for stored water). The effective use of resources for this program could be enhanced by 1) focusing vector control interventions on those areas that are most at risk, and 2) provide health education messages on recommending social and environmental changes at the household level that will have the greatest effect on vector abundance, and therefore disease incidence. The PCI, as far as we are aware, is currently not used systematically in any dengue surveillance program. The MOH is currently developing pilot projects on the utility and practicability of a ‘‘modified’’ PCI for entomological risk including both immature but also adult infestation, particularly with respect to Ae. aegypti adult females inside premises. Both studies previously cited from Brazil evaluating PCI as a tool to prioritize houses for vector control measures have reported
VOL. 29, NO. 4
a potential of PCI as an adult proxy for adult infestation. Macoris et al. (2009) found that houses in Sao Paulo with PCI values $6 accounted for 69% of adult-positive premises and 76% of total adult collections. Maciel-deFreitas et al. (2008) also observed that that keypremises distribution seemed to be correlated with areas where more adult Ae. aegypti females were collected. This reinforces the importance of further studies on the utility of PCI in Mexico as a proxy for adult Aedes distribution as well as being an easy-to-use and effective tool for directing surveillance and control activities in urban settings. The PCI information is being gathered in 65% (21/32) of the Mexican states and uploaded into a national platform (Plataforma Nacional de Vigilancia Entomolo´gica y Control Integral de Dengue) since 2012. In the near future, it is expected to use information derived from PCI in 2 ways: 1) to focalize integrated dengue vector control on houses/city blocks/neighborhoods/areas with high levels of PCI (6–9); and 2) to promote public health policies, at municipal, state, and national levels, to improve house quality, and ultimately to reduce vector infestation and the risk of disease transmission.
ACKNOWLEDGMENTS We express our sincere gratitude to Consejo Nacional de Ciencia y Tecnologı´a (CONACYT Mexico, Fondo Sectorial de Investigacio´n en Salud y Seguridad Social SSA/IMSS/ISSSTECONACYT SALUD-2011-1-161551 and the Programa de Impulso y Orientacion a la Investigacion (PRIORI-UADY) for their financial support of this project. REFERENCES CITED Arunachalam N, Tana S, Espino F, Kittayapong P, Abeyewickreme W, Wai KT, Tyagi BK, Kroeger A, Sommerfeld J, Petzold M. 2010. Eco-bio-social determinants of dengue vector breeding: a multicountry study in urban and periurban Asia. Bull WHO 88:173–184. Basker P, Ezhil R. 2012. Study in the correlation of premises condition index and the presence of larvae of Aedes species mosquitoes in human dwellings of the Cuddalore District of Tamil Nadu, India. Osong Public Health Res Perspect 3:3–7. Bennett S, Woods T, Liyanage WM, Smith DL. 1991. A simplified general method for cluster-sample surveys of health in developing countries. World Health Stat Q 44:98–106. Burnham KP, Anderson DR. 2004. Multimodel inference: understanding AIC and BIC in model selection. Sociol Method Res 33:261–304. Diario Oficial de la Federacion [DOF]. 2011. Para la vigilancia epidemiolo´gica, prevencio´n y control de enfermedades transmitidas por vector. Norma Oficial
DECEMBER 2013
AEDES AEGYPTI
AND
PREMISES CONDITION IN SOUTH MEXICO
Mexicana de Emergencia NOM-EM-003-SSA2-2010. Mexico DF, Mexico: Diario Oficial de la Federacion. Espinoza-Go ´ mez F, Hernandez CM, Coll R. 2001. Factors that modify the larval indices of Aedes aegypti in Colima, Mexico. Rev Panam Salud Publica 10:6–12. Focks DA. 2003. A review of entomological sampling methods and indicators for dengue vectors [Internet]. Geneva, Switzerland: World Health Organization on behalf of the Special Programme for Research and Training in Tropical Diseases [accessed September 2013]. Available from: http://whqlibdoc.who.int/hq/ 2003/TDR_IDE_DEN_03.1.pdf. Focks DA, Alexander N. 2006. Multicountry study of Aedes aegypti pupal productivity survey methodology: findings and recommendations [Internet]. Geneva, Switzerland: World Health Organization on behalf of the Special Programme for Research and Training in Tropical Diseases [accessed September 2013]. Available from: http://www.who.int/tdr/publications/ documents/aedes_aegypti.pdf. Garcı´a-Rejo ´ n JE, Lo ´ pez-Uribe MP, Loron˜o-Pino MA, Farfa´n-Ale JA, Del Najera-Vazquez MR, LozanoFuentes S, Beaty BJ, Eisen L. 2011. Productive container types for Aedes aegypti immatures in Me´rida, Me´xico. J Med Entomol 48:644–650. Iba´n˜ez-Bernal S, Go ´ mez-Dante´s H. 1995. Los vectores del dengue en Me´xico: una revisio´n crı´tica. Salud Publica Mex 37:53–63. Instituto Nacional de Estadı´stica, Geografı´a e Informa´tica [INEGI]. 2000. SCINCE, Sistema para la consulta de informacio´n censal 2000. Me´xico, DF: Instituto Nacional de Estadı´stica, Geografı´a e Informa´tica. Lloyd LS. 1992. Results of a community-based Aedes aegypti control intervention. In: Halstead SB, Go ´ mez–Dante´s H, eds. Dengue. A worldwide problem, a common strategy. Proceedings of the International Conference on Dengue and Aedes aegypti Community-Based Control Mexico, DF: Mexican Ministry of Health–Rockefeller Foundation. p 287–291. Lloyd LS, Winch P, Ortega-Canto J, Kendall C. 1994. The design of a community-based health education intervention for the control of Aedes aegypti. Am J Trop Med Hyg 50:401–411. Maciel-de-Freitas R, Peres RC, Souza-Santos R, Lourencode Oliveira R. 2008. Occurrence, productivity and spatial distribution of key premises in two dengue-endemic areas of Rio de Janeiro and their role in adult Aedes aegypti spatial infestation pattern. Trop Med Int Health 13:1488–1494.
345
Macoris M, Galvani K, da Grac¸a Macoris M. 2009. Evaluation of premise condition index in the context of Aedes aegypti control in Marı´lia, Sa˜o Paulo, Brazil. Dengue Bull 33:167–175. Manrique-Saide P, Davies CR, Coleman PG, RebollarTellez E, Che-Medoza A, Dzul-Manzanilla F, Zapata-Peniche A. 2008. Pupal surveys for Aedes aegypti surveillance and potential targeted control in residential areas of Me´rida, Me´xico. J Am Mosq Control Assoc 24:289–298. Moloney JM, Skelly C, Weinstein P, Maguire M, Ritchie S. 1998. Domestic Aedes aegypti breeding site surveillance: limitations of remote sensing as a predictive surveillance tool. Am J Trop Med Hyg 59:261–264. Nogueira LA, Gushi LT, Miranda JE, Madeira NG, Ribolla PEM. 2005. Application of an alternative Aedes species (Diptera: Culicidae) surveillance method in Botucatu city, Sa˜o Paulo, Brazil. Am J Trop Med Hyg 73:309–311. PAHO [Pan American Health Organization]. 1994. Dengue and dengue hemorrhagic fever in the Americas: guidelines for prevention and control. Washington, DC: Pan American Health Organization, Regional Office of the World Health Organization. Peterson AT, Martı´nez-Campos C, Nakazawa Y, Martı´nez-Meyer E. 2005. Time-specific ecological niche modelling predicts spatial dynamics of vector insects and human dengue cases. Trans R Soc Trop Med Hyg 99:647–655. Tun-Lin W, Kay BH, Barnes A. 1995a. Understanding productivity, a key to Aedes aegypti surveillance. Am J Trop Med Hyg 53:595–601. Tun-Lin W, Kay BH, Barnes A. 1995b. The Premise Condition Index: a tool for streamlining surveys of Aedes aegypti. Am J Trop Med Hyg 53:591–594. Tun-Lin W, Lenhart A, Nam VS, Rebollar-Te´llez E, Morrison AC, Barbazan P, Cote M, Midega J, Sanchez F, Manrique-Saide P, Kroeger A, Nathan M, Meheus F, Petzold M. 2009. Reducing costs and operational constraints of dengue vector control by targeting productive breeding places: a multi-country non-inferiority cluster randomized trial. Trop Med Int Health 14:1143–1153. Winch PJ, Barrientos-Sanchez G, Puigserver-Castro E, Manzano-Cabrera L, Lloyd LS, Mendez-Galvan J. 1992. Variation in Aedes aegypti larval indices over one year period in a neighborhood of Merida, Yucatan, Mexico. J Am Mosq Control Assoc 8: 193–195.
