Study Design and Participants
This was a community-based, cross-sectional, descriptive study conducted in three countries: Brazil, Mexico and Colombia. The main survey was conducted in 2012 between March and May for the three countries. Additional interviews were conducted in Mexico and Colombia in July and August 2012 to get a representative sample. The most common interview technique used in each country was selected to attain a general population sample. In Mexico and Colombia, interviews were conducted in the participants’ homes, whereas in Brazil, interviews were conducted in the street. Due to different methodology used to collect the data in Brazil, the Brazilian data are provided in the supplementary appendix.
Adults aged 18 years or older were eligible for inclusion in the survey. A quota sampling approach was used to obtain a nationally representative sample. Census data from each country was used to set quota, with non-interlocking quotas set by age (age groups 18–24, 25–34, 35–44, 45–54, 55–64 years and ≥65 years in Mexico, and 18–25, 26–35, 36–40, 41–45, 46–50 and ≥51 years in Colombia), gender, socioeconomic classification and region (including endemic vs non-endemic regions for countries with mixed dengue endemicity, i.e., Mexico and Colombia). An endemic region was defined as a region where dengue cases are reported, the vector exists and environmental conditions allow transmission. Quotas were also set to ensure a representative split in terms of urban/rural areas in Mexico and Colombia. The definition of rural was ‘those communities with less than 2500 inhabitants’ (defined by National Institute of Statistics) [14].
Twenty-nine cities were included in the survey conducted in Mexico, nine in non-endemic regions (Aguascalientes, Chihuahua, Coahuila, Federal District, Durango, Estado de México, Guanajuato, Hidalgo and Querétaro) and 20 in endemic regions (Baja California Sur, Campeche, Chiapas, Colima, Guerrero, Jalisco, Michoacán, Morelos, Nayarit, Nuevo León, Oaxaca, Puebla, Quintana Roo, San Luis Potosí, Sinaloa, Sonora, Tabasco, Tamaulipas, Veracruz and Yucatán). All regions in Mexico included rural areas. In Colombia, the Bogotá region was included as a non-endemic area whereas the Caribe, Oriental, Ori/Amazonia, Sur Oriental, Central and Pacifica regions were included as endemic areas. Twenty cities were included in the survey conducted in Colombia with 19 in endemic areas (Arauca, Barranquilla, Boyacá, Cali, Cartagena, Cartago, Casanare, Caucasia, Cundinamarca, Guainía, Guaviare, Huila, Quindío, Magangue, Malambo, Medellin, Putumayo, Risaralda, Santander) and one in a non-endemic area (Bogota D.C).
Survey Questionnaire and Interviews
The survey questionnaire was drafted in English and then translated in Spanish. A pretest (or survey validation) of the questionnaire was undertaken with four participants (two in each country) within the same target population to assess the suitability and test for any possible vagueness in the questions formulated. The questionnaire comprised 58 questions, taking an average of 25 min to complete. It was divided into eight sections (see supplementary appendix): (1) demographics; (2) health profile, perceptions and behavior in seeking health advice; (3) general knowledge of infectious diseases; (4) perceptions of dengue disease; (5) personal experience with dengue; (6) family experience with dengue; (7) attitudes toward vaccinations; and (8) perceptions of dengue vaccination. The current manuscript focuses on the first seven sections of the survey only because the questions in Section 8 are purely hypothetical at present as there is currently no licensed dengue vaccine available.
The interviews (one per household) were conducted with residents of households in their homes. The first house was randomly selected according to the geographical localization followed by a systematic selection (i.e., skipping a set number of houses to identify the next household) of remaining houses within the localization. Participation in the study was voluntary and no incentives were provided.
Country-specific profiling questions were included to enable the interviewer to determine the socioeconomic level of the participants to ensure that the correct quota of participants was recruited to attain a general population sample. Social–economic class was categorized according to official national data [15, 16]. In Mexico, this was based on the National Association of Market Research criteria, which used characteristics of the household (number of rooms, bathrooms, lighting and education level). Utility usage was the basis for socioeconomic classification in Colombia.
Data Analysis
No formal statistical power calculation was used in this descriptive study; the sample size in each country was chosen to include at least 800 participants. The survey results were analyzed using Latent Gold version 4.5 (Boston, USA). Participant responses to each question were expressed as relative percentages.
All data were weighted according to age, gender and region to correct for differences between the demographic profiles of the samples surveyed in each country relative to their general adult population. This was done to ensure that the results would be broadly representative of the adult population in each country.
This article does not contain any new studies with human or animal subjects performed by any of the authors.