EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Data Resource Profile: Harmonized health survey data for 240 cities across 11 countries in Latin America: the SALURBAL project

Kari Moore, Mariana Lazo, Ana Ortigoza iD, D Alex Quistberg, Brisa Sanchez, Binod Acharya, Tania Alfaro, Maria Fernanda Kroker-Lobos, Mariana Carvalho De Menezes, Olga Lucia Sarmiento, Amanda C de Souza Andrade, Carolina Perez Ferrer und 4 weitere

International Journal of Epidemiology · 16.12.2024 · 54 (1)

DOI: 10.1093/ije/dyae171 ↗

Kurz gefasst

Fachartikel in der Zeitschrift International Journal of Epidemiology, online erschienen am 06.01.2025. Verfasst von Kari Moore und 15 weiteren Autor:innen. Beteiligte Einrichtung: Urban Health Collaborative, Dornsife School of Public Health, Drexel University, Philadelphia, PA, USA. Die Arbeit wurde bisher 2-mal zitiert (Crossref) und nennt 22 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Wellcome Trust, FIC NIH HHS.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Harmonized health data sources in Latin America for research and action are limited. The SALURBAL (Salud Urbana en America Latina/Urban Health in Latin America) health survey data resource was created as part of an integrated, comprehensive resource to characterize and study the drivers of urban health in middle-income countries in Latin America. This resource is the largest in scope and size in cities in Latin America. It includes harmonized health survey data from 40 surveys from 11 countries. Survey years ranged from 2000 to 2021, the range of surveys per country is one to six, with seven out of the 11 countries having more than one survey. The total number of adults included in the resource is 721 099 (with sociodemographic data) and 542 336 with select health-related outcomes. For children (aged 0–17), the samples are 209 379 and 134 833, respectively. Data available for adults include individual-level demographic and socioeconomic characteristics, alcohol/tobacco use, anthropometry, diet, physical activity, diabetes, hypertension, mental health and self-reported health for adults. For children, data include household-level demographic and socioeconomic characteristics and individual-level anthropometry. City-level, model-based prevalence estimates for selected health risk factors are available as part of this resource. This data resource can be linked to built, natural and social environment data for select geographies [cities, subcity units (e.g. counties) and neighbourhoods). Researchers interested in the SALURBAL project data should visit: [https://data.lacurbanhealth.org]. Focused on one of the most highly urbanized regions of the world, Latin America, the SALURBAL (Salud Urban en America Latina/Urban Health in Latin America) Project was launched in 2017 to build a novel and unique multi-national, multi-level, multi-dimensional data infrastructure to allow the investigation of questions about how urban environments and policies affect health, health equity and environmental sustainability and to serve as a platform for research and action in the region.1,2 A detailed description of the SALURBAL project is available elsewhere.1,2 Briefly, the SALURBAL includes data from all urban agglomerations (‘cities’, total 371) with 100 000 residents or more in 2010 in 11 Latin American countries (Argentina, Brazil, Chile, Colombia, Costa Rica, El Salvador, Guatemala, Mexico, Nicaragua, Panama, and Peru). Of the 371 cities, 240 had health survey data available (Supplementary Figure S1, available as Supplementary data at IJE online). SALURBAL pooled and harmonized surveys and linked individual-level data to social, built and natural environment data for specific geographies, including countries, cities, sub-cities and ‘neighbourhoods’. SALURBAL defined cities as agglomerations of administrative units (i.e. municipios, comunas etc.) that encompass a city’s built-up area as defined using visual inspection of satellite images.2 These administrative units are referred to as sub-city units. ‘Neighbourhoods’ were defined as the smallest unit for which census data were available, similar to US census tracts (i.e. radio censal, sector censitario, zona censal, area geoestadistica basica, barrio etc.). The SALURBAL health survey component includes harmonized data from 40 national surveys from 11 countries, including 13 surveys (from seven countries) with data for children, with survey years ranging from 2000 to 2021. Another component included in this resource is model-based city-level prevalence estimates of selected health risk factors. The SALURBAL health survey component compiled, linked and harmonized data from 40 national surveys from 11 countries (Table 1; Supplementary Table S1 and Figure S2, available as Supplementary data at IJE online), including 13 surveys (from seven countries) with data for children. The majority of the data corresponds to surveys conducted between 2010 and 2021. The range of surveys per country is one to six, allowing for trend analyses (i.e. changes over time using repeat cross-sectional data) (see Data Resource Use section below). The total number of adults included in the resource is 721 099 (with sociodemographic data) and 542 336 with selected health-related outcomes. For children (aged 0–17), the samples are 209 379 and 134 833, respectively. A total of 240 unique SALURBAL-defined cities having at least one survey respondent is represented in the resource. The criteria for inclusion of surveys in the harmonized survey resource are shown in Supplementary Figure S3 (available as Supplementary data at IJE online) and described below. Summary of health surveys included in Salud Urbana en America Latina/Urban Health in Latin America (SALURBAL) data resource, by country name (alphabetical order) and survey year. For detailed characteristics, please see Supplementary Table S1 (available as Supplementary data at IJE online) CAMDI, Encuesta Multinacional de Diabetes mellitus y Factores de Riesgo (Multinational Survey of Diabetes Mellitus & Risk Factors, Central American Diabetes Initiative); DHS, Demographic and Health Survey, ELPI, Encuesta Longitudinal de Primera Infancia (Longitudinal Survey of Early Childhood); ENDES, Encuesta Nacional de Demografia y Salud (National Survey of Demographics and Health); ENECA, Encuesta Nacional de Enfermedades Cronicas no transmisibles en Poblacion Adulta de El Salvador (National Survey of Noncommunicable Chronic Diseases in the Adult Population of El Salvador); ENFR, Encuesta Nacional de Factores de Riesgo (National Risk Factors Survey); ENS, Encuesta Nacional de Salud (National Health Survey); ENSA, Encuesta National de Salud (National Health Survey); ENSANUT, Encuesta Nacional de Salud y Nutricion (National Survey for Health and Nutrition); ENSCAVI, Encuesta Nacional de Salud y Calidad de Vida (National Survey of Health and Quality of Life); ENSIN, Encuesta Nacional de la Situation Nutricional en Colombia (National Nutritional Situation in Colombia); FESAL, Encuesta Nacional de Salud Familiar (National Family Health Survey); L1AD, Level 1 administrative cities; L2, Level 2 administrative subcities; PNS, Pesquisa Nacional de Saúde (National Health Survey); SALURBAL, (Salud Urbana en America Latina/Urban Health in Latin America). Summary of health surveys included in Salud Urbana en America Latina/Urban Health in Latin America (SALURBAL) data resource, by country name (alphabetical order) and survey year. For detailed characteristics, please see Supplementary Table S1 (available as Supplementary data at IJE online) CAMDI, Encuesta Multinacional de Diabetes mellitus y Factores de Riesgo (Multinational Survey of Diabetes Mellitus & Risk Factors, Central American Diabetes Initiative); DHS, Demographic and Health Survey, ELPI, Encuesta Longitudinal de Primera Infancia (Longitudinal Survey of Early Childhood); ENDES, Encuesta Nacional de Demografia y Salud (National Survey of Demographics and Health); ENECA, Encuesta Nacional de Enfermedades Cronicas no transmisibles en Poblacion Adulta de El Salvador (National Survey of Noncommunicable Chronic Diseases in the Adult Population of El Salvador); ENFR, Encuesta Nacional de Factores de Riesgo (National Risk Factors Survey); ENS, Encuesta Nacional de Salud (National Health Survey); ENSA, Encuesta National de Salud (National Health Survey); ENSANUT, Encuesta Nacional de Salud y Nutricion (National Survey for Health and Nutrition); ENSCAVI, Encuesta Nacional de Salud y Calidad de Vida (National Survey of Health and Quality of Life); ENSIN, Encuesta Nacional de la Situation Nutricional en Colombia (National Nutritional Situation in Colombia); FESAL, Encuesta Nacional de Salud Familiar (National Family Health Survey); L1AD, Level 1 administrative cities; L2, Level 2 administrative subcities; PNS, Pesquisa Nacional de Saúde (National Health Survey); SALURBAL, (Salud Urbana en America Latina/Urban Health in Latin America). In this section we first describe the harmonization process and then the production of model-based prevalence estimates. The SALURBAL Health Survey data harmonization efforts were led by the SALURBAL Data and Methods Core with leadership at Drexel University and members from country teams. For the creation of this resource, SALURBAL followed four guiding principles: (i) use existing national health survey data administered by agencies within each country; (ii) restrict to surveys with geographical information that could be linked to SALURBAL sub-city level, publicly or through request to the agency implementing the survey; (iii) prioritize surveys with information on non-communicable health behaviours and risk factors; (iv) use of harmonization approaches that are rigorous but flexible to accommodate differences across surveys, described in more detail below.3,4 SALURBAL country teams identified national health surveys within each country and confirmed whether they met the inclusion criteria related to geographical linkage potential and focus (Supplementary Figure S3, available as Supplementary data at IJE online). Survey data were gathered from national bureaus of statistics or other relevant government ministries or agencies responsible for the survey use. Data were gathered either by directly downloading from the agencies’ public websites when available or by making formal requests to the agency. Once the data were gathered, we reviewed the data to ensure that geographical identifiers were available to link to the SALURBAL-defined sub-city or neighbourhood level. Only surveys that contained the necessary geographical linkages were further reviewed. Next, the survey forms and metadata were reviewed to determine if they contained information for at least one pre-selected non-communicable disease risk factors (anthropometry, diabetes, hypertension, tobacco use, alcohol use, diet, physical activity) and key sociodemographic information (age, sex or gender and educational attainment). First, we identified and collated the survey questions and response options by domain (Supplementary Figure S4 and Supplementary Table S2, available as Supplementary data at IJE online). We determined all skip patterns in the questions and respondent universe for each relevant question. For some questions, only a subset of respondents was included, such as a specified age range or sex/gender, or a randomly selected subset of the survey respondents. Because the surveys were administered in Spanish or Portuguese, the next step was translating into English as a common language. This was accomplished via project members who were fluent in both languages. Once the questions available were compiled by domain, we reviewed recommended definitions used by others such as the Centers for Disease Control and Prevention (CDC) or the World Health Organization (WHO). As part of harmonization, we addressed the following three major challenges: discrepancy in sets of questions, scales and questionnaires used for retrieving information about similar health behaviours or health outcomes, e.g. diabetes, hypertension, physical activity, diet, alcohol intake; lack of consistency in response categories or measurement units used. eg self-rated health; differences in respondent universe for selected health outcomes, e.g. gestational diabetes, use of medications for hypertension/diabetes, doctor/physician diagnosis of depression. To address these challenges, we used a flexible approach focused on creating different versions of variables based on scenarios. Surveys with questions that were asked in a similar manner in terms of response patterns and wording of the questions were grouped together as a scenario. For example, when harmonizing diabetes, some surveys asked, ‘Have you ever been diagnosed with diabetes or high blood sugar by a doctor?’ with response options of ‘yes’ and ‘no’, followed by an additional question (‘was this diagnosis during pregnancy?’) to women who responded ‘yes’ to the initial diabetes question. Other surveys asked the same leading question but included three response options of ‘yes’, ‘no’ and ‘yes, but only during pregnancy’, with no further questions. These were categorized into two different scenarios. Once the questions were grouped into scenarios, we developed three different approaches to harmonization, which we refer to as harmonization types. Type 1 occurred when surveys had questions that were asked in the same way with only slight differences in wording, such that there was only one scenario and it was possible to create a single harmonized version of the measure. Examples of this harmonization type are sex and depressive symptoms scales. Type 2 occurred when the survey questions were asked in different ways in terms of wording, response options and response patterns availabl,e such that there were multiple question scenarios but it was still possible to identify one harmonized version of the measure. Examples of this harmonization type are marital status and educational attainment. Type 3 occurred when multiple versions of the measure were created because we could not harmonize to a common definition, due to differences in the question wording. Examples of this harmonization type are hypertension and binge drinking. All proposed harmonization was reviewed by country teams and working groups with expertise in the domain, and all harmonization procedures are detailed in written protocols that accompany the datasets (see section below). After applying the harmonization protocol to the datasets, we reviewed descriptive statistics of initial harmonized variables. If any deviations from expected results (i.e. known patterns) were observed, we reviewed all programming and corrected errors. If deviations remained, we reviewed the harmonization protocol and modifications were made as needed. Whereas most of the surveys included in SALURBAL were based on complex, multi-stage probability designs to select nationally representative samples of the population of the country and provide nationally representative estimates, these data cannot be used directly to draw inferences about prevalence of diseases or exposures in other geographical units, for example SALURBAL cities. To address that issue and be able to provide city-specific prevalence estimates, SALURBAL employed a model-based prevalence estimates approach. Briefly, sex- and city-specific smoothed estimates were derived by information from the in the with information from other cities in the same using a with city-specific estimates provide more estimates for the cities. The approach the estimates to other (e.g. the SALURBAL age and sex This approach is similar to and are in the Figure 1 the prevalence estimates of for SALURBAL cities and countries. for prevalence of diabetes, hypertension, self-rated health status and are shown in Supplementary (available as Supplementary data at IJE online). In Supplementary Figure (available as Supplementary data at IJE online) we used of surveys as an example of the and potential use of prevalence estimates over For this example, we selected cities that were in each of the of Encuesta Nacional de Salud y Nutricion a total of and provide and prevalence estimates. prevalence of years and for SALURBAL cities and countries, by Costa El Salvador To accompany all the SALURBAL datasets, harmonization protocols were These protocols provide information about the domain, harmonized of the survey questions used in harmonization, across surveys, variables and of the include questions in both English and the response categories and by Table 2 the survey and across Supplementary and S3 (available as Supplementary data at IJE online) the of variables harmonized to related to model-based prevalence estimates includes information about the different model-based estimates, recommended use and and the Summary of harmonized survey and variables available in Salud Urbana en America Latina/Urban Health in Latin America For detailed characteristics of please see Supplementary Table (available as Supplementary data at IJE online) Only variables are included in the creation of this see Supplementary Table (available as Supplementary data at IJE online) for the Summary of harmonized survey and variables available in Salud Urbana en America Latina/Urban Health in Latin America For detailed characteristics of please see Supplementary Table (available as Supplementary data at IJE online) Only variables are included in the creation of this see Supplementary Table (available as Supplementary data at IJE online) for the to the of data from not all SALURBAL health survey data can be linked to all SALURBAL geographical sub-city and A of the linkage of survey respondents to the different SALURBAL geographical is shown in Table 3 and in the For example, in the of Brazil, to was only available for the of each of the leading to the of only cities for the The linkage of survey to the smallest SALURBAL geographical was possible in 11 of the surveys of Salud Urbana en America Latina/Urban Health in Latin America (SALURBAL) Health Survey respondents to SALURBAL Level Level and ‘Neighbourhoods’ Level 134 134 CAMDI, Encuesta Multinacional de Diabetes mellitus y Factores de Riesgo (Multinational Survey of Diabetes Mellitus & Risk Factors, Central American Diabetes Initiative); DHS, Demographic and Health Survey, ELPI, Encuesta Longitudinal de Primera Infancia (Longitudinal Survey of Early Childhood); ENDES, Encuesta Nacional de Demografia y Salud (National Survey of Demographics and Health); ENECA, Encuesta Nacional de Enfermedades Cronicas no transmisibles en Poblacion Adulta de El Salvador (National Survey of Noncommunicable Chronic Diseases in the Adult Population of El Salvador); ENFR, Encuesta Nacional de Factores de Riesgo (National Risk Factors Survey); ENS, Encuesta Nacional de Salud (National Health Survey); ENSA, Encuesta National de Salud (National Health Survey); ENSANUT, Encuesta Nacional de Salud y Nutricion (National Survey for Health and Nutrition); ENSCAVI, Encuesta Nacional de Salud y Calidad de Vida (National Survey of Health and Quality of Life); ENSIN, Encuesta Nacional de la Situation Nutricional en Colombia (National Nutritional Situation in Colombia); FESAL, Encuesta Nacional de Salud Familiar (National Family Health Survey); L1AD, Level 1 administrative cities; L2, Level 2 administrative subcities; Level 3 administrative PNS, Pesquisa Nacional de Saúde (National Health Survey); SALURBAL, (Salud Urbana en America Latina/Urban Health in Latin America). of Salud Urbana en America Latina/Urban Health in Latin America (SALURBAL) Health Survey respondents to SALURBAL Level Level and ‘Neighbourhoods’ Level 134 134 CAMDI, Encuesta Multinacional de Diabetes mellitus y Factores de Riesgo (Multinational Survey of Diabetes Mellitus & Risk Factors, Central American Diabetes Initiative); DHS, Demographic and Health Survey, ELPI, Encuesta Longitudinal de Primera Infancia (Longitudinal Survey of Early Childhood); ENDES, Encuesta Nacional de Demografia y Salud (National Survey of Demographics and Health); ENECA, Encuesta Nacional de Enfermedades Cronicas no transmisibles en Poblacion Adulta de El Salvador (National Survey of Noncommunicable Chronic Diseases in the Adult Population of El Salvador); ENFR, Encuesta Nacional de Factores de Riesgo (National Risk Factors Survey); ENS, Encuesta Nacional de Salud (National Health Survey); ENSA, Encuesta National de Salud (National Health Survey); ENSANUT, Encuesta Nacional de Salud y Nutricion (National Survey for Health and Nutrition); ENSCAVI, Encuesta Nacional de Salud y Calidad de Vida (National Survey of Health and Quality of Life); ENSIN, Encuesta Nacional de la Situation Nutricional en Colombia (National Nutritional Situation in Colombia); FESAL, Encuesta Nacional de Salud Familiar (National Family Health Survey); L1AD, Level 1 administrative cities; L2, Level 2 administrative subcities; Level 3 administrative PNS, Pesquisa N

Schlagwörter

Stichwörter der Autor:innen

Health surveyLatin Americahealth risk factorssocioeconomicurban health

Autor:innen und Institutionen

  • Urban Health Collaborative, Dornsife School of Public Health, Drexel University, Philadelphia, PA, USA.
  • Department of Community Health and Prevention, Dornsife School of Public Health, Drexel University, Philadelphia, PA, USA.
  • Department of Social and Environmental Determinants for Health Equity, Pan American Health Organization, Washington, DC, USA.
  • Department of Environmental and Occupational Health, Dornsife School of Public Health, Drexel University, Philadelphia, PA, USA.
  • Department of Epidemiology and Biostatistics, Dornsife School of Public Health, Drexel University, Philadelphia, PA, USA.
  • Programa de Doctorado en Salud Pública, Escuela de Salud Pública, Facultad de Medicina, Universidad de Chile, and Center for Cancer Prevention and Control (CECAN), Santiago, Chile.

ORCID-Profile: Ana Ortigoza, Ana V Diez Roux

Förderung

  • Wellcome Trust
  • FIC NIH HHS

Zitieren

APA

Moore, K., Lazo, M., Ortigoza, A., Quistberg, D. A., Sanchez, B., Acharya, B., Alfaro, T., Kroker-Lobos, M. F., Menezes, M. C., Sarmiento, O. L., de Souza Andrade, A. C., Perez Ferrer, C., Hernandez Vasquez, A., Caiaffa, W. T., Diez Roux, A. V., & SALURBAL Group (2024). Data Resource Profile: Harmonized health survey data for 240 cities across 11 countries in Latin America: the SALURBAL project. International Journal of Epidemiology, 54(1). https://doi.org/10.1093/ije/dyae171

Vancouver

Moore K, Lazo M, Ortigoza A, Quistberg DA, Sanchez B, Acharya B, et al. Data Resource Profile: Harmonized health survey data for 240 cities across 11 countries in Latin America: the SALURBAL project. International Journal of Epidemiology. 2024;54(1). doi:10.1093/ije/dyae171

BibTeX
@article{Moore2024,
  author = {Moore, Kari and Lazo, Mariana and Ortigoza, Ana and Quistberg, D Alex and Sanchez, Brisa and Acharya, Binod and Alfaro, Tania and Kroker-Lobos, Maria Fernanda and Menezes, Mariana Carvalho De and Sarmiento, Olga Lucia and de Souza Andrade, Amanda C and Perez Ferrer, Carolina and Hernandez Vasquez, Akram and Caiaffa, Waleska Teixeira and Diez Roux, Ana V and SALURBAL Group, SALURBAL Group},
  title = {Data Resource Profile: Harmonized health survey data for 240 cities across 11 countries in Latin America: the SALURBAL project},
  journal = {International Journal of Epidemiology},
  year = {2024},
  volume = {54},
  number = {1},
  doi = {10.1093/ije/dyae171}
}
Kostenlos starten →