A Telephone Surveillance System for Noncommunicable Diseases in Brazil

Noncommunicable diseases (NCDs) are major causes of morbidity and mortality worldwide and are a major challenge for public health. According to the World Health Organization (WHO), NCDs cause 70% of all deaths worldwide ¹ and more than 50% of the global years of potential life lost from fatal and nonfatal conditions combined. ² Of the 10 major factors associated with years of potential life lost to disease or premature death in the Americas, 5 are also risk factors for NCDs: alcohol abuse, obesity, smoking, low fruit and vegetable intake, and a sedentary lifestyle. ³ In Brazil, NCDs are among the greatest health problems and cause 75% of all deaths. ⁴

Population-based surveys, as part of a surveillance plan for NCDs, are essential to identify and monitor health profiles; the distribution, incidence, and prevalence of risk factors; and health inequalities. Such ongoing surveys can also evaluate programs that serve specific groups and help plan and manage health interventions. ^5,6

Inspired by the Behavioral Risk Factor Surveillance System (BRFSS), begun by the Centers for Disease Control and Prevention in 1984, the Brazilian Ministry of Health started a surveillance program of the risks and protective factors for NCDs in 2003. ⁷ Beginning in 2006, the surveillance program included an annual telephone survey, VIGITEL (Vigilância de Fatores de Risco e Proteção para Doenças Crônicas por Inquérito Telefônico [Surveillance of Risk and Protective Factors for Chronic Diseases by Telephone Survey]). ⁸ The VIGITEL program provides recent national estimates of self-reported health behaviors and risk factors for NCDs. ⁹

Here, we describe the implementation of the VIGITEL surveillance program and the challenges of conducting such a survey in Brazil.

Implementing the VIGITEL Telephone Surveillance System

In 1998, faced with the growing burden of NCDs and the need to provide an urgent and effective public health response, the WHO’s 51st World Health Assembly recognized the need for a global strategy to prevent and control NCDs. ¹⁰ This strategy was subsequently developed and endorsed by the WHO and formally adopted in 2000. This global strategy was the beginning of almost 2 decades of increased focus on NCDs, during which NCDs became a priority at the highest political levels. ^11,12

Given changes in demographic, nutritional, and epidemiologic characteristics (eg, aging population, increase in prevalence of obesity, increased prevalence of chronic diseases) in Brazil at the beginning of the 21st century, the Ministry of Health implemented a surveillance system for NCDs. The first incentives in implementing the surveillance system for NCDs were to (1) fund a study to better estimate the burden of the most prevalent diseases in 2002 and (2) establish a health surveillance secretary position to coordinate NCD-related efforts. ^7,9

Another incentive was the III Global Forum on Integrated Noncommunicable Disease Prevention & Control, held in Rio de Janeiro in November 2003. The Forum offered an important overview of worldwide progress in developing integrated strategies for preventing and controlling NCDs and provided technical training for staff members of the NCD health surveillance secretary. In 2004, scientific meetings were held in various cities in Brazil to discuss the surveillance of NCDs; in the same year, forums reached consensus on methodology and defined monitoring indicators. ^7,11,13

The NCD surveillance program incorporated information from health information systems and implemented population surveys, such as household surveys, school-based surveys, and telephone surveys. VIGITEL started in 2006 with an annual sample of 54 000 telephone lines in the 26 Brazilian state capitals and the federal district (Box). By 2018, VIGITEL had completed 12 years of continuous data collection. Among all public health research initiatives in Brazil, we believe that VIGITEL is the best maintained. It allows us to monitor temporal trends and to support policies on priorities, such as controlling tobacco, reducing alcohol abuse, and promoting physical activity.

Box

Characteristics of the Brazilian VIGITEL surveillance system^a

Managed by the Ministry of Health

Began in 2006

Annual landline telephone survey

Covers 26 Brazilian state capitals and the federal district

Survey participants are adults aged ≥18

Includes 54 000 interviews per year

Standard questionnaire used for the entire sample each year

Sample weights are calculated by using the raking method,^b adopted after 2012

• Abbreviation: VIGITEL, Vigilância de Fatores de Risco e Proteção para Doenças Crônicas por Inquérito Telefônico (Surveillance of Risk and Protective Factors for Chronic Diseases by Telephone Survey).

• ^aVIGITEL allows for the monitoring of temporal trends and accompanies policies that are priorities, such as tobacco control, reduction of alcohol use, and promotion of physical activity.

• ^bThe raking method uses the simple frequency distribution of each variable, such as age group, sex, and education level, for each state capital, and allows the use of various external sources in the intercensus period to calculate the weights.

Methodological Characteristics of the VIGITEL System

VIGITEL is a population-based, landline telephone survey designed to monitor the prevalence of NCDs among adults in Brazil. Data are collected on sociodemographic characteristics, food consumption, physical activity, smoking, alcohol abuse, self-reported weight and height, and other variables.

A 2-stage probabilistic sample is selected in each of the 27 municipalities. In the first stage, landline telephone numbers stratified by postal code are systematically sampled from the electronic register of households served by the telephone companies. In the second stage, a simple random sample of prospective interviewees is selected from each household. ⁸

The program has had only one major methodological change since 2006. The poststratification weighting method was replaced in 2012 with a more advanced method called “raking.” ¹⁴ The raking method uses the simple frequency distribution of each variable, such as age group, sex, and education level, for each state capital (n = 26) and federal district, but it uses data collected from various external sources between surveys to calculate the weights. This method has the advantage of using univariate distributions from various external sources during various time periods. This revision was necessary to increase the accuracy of the estimates when compared with those obtained by the cell-by-cell weighting method, which uses the strata of sociodemographic variables to match the sample composition of persons with fixed telephones to the total population. The revised weighting method adopted was important to improve data accuracy. ¹⁴

A question about the type of health system (public or private) mainly used by the interviewee was added in 2008. This question is important because Brazil has a mixed health system in which services are financed and provided by the state and by a private company, with health plans and insurance policies funded mostly by employers. ¹⁵ Annually, approximately 54 000 persons are interviewed in VIGITEL, and the survey response rate is 70%.

Limitations and Challenges of the Survey

Surveillance systems for NCDs are the main sources of information for monitoring and triggering public health actions to prevent and control these diseases. Telephone surveys, as part of these surveillance systems, have several advantages over other types of surveys, including practicality, lower cost, the ability to detect changes and trends, and speed in data collection. ¹⁶ In particular, VIGITEL data support the preparation and evaluation of public policies in Brazil by contributing to the Pluri-Annual Development Plan, which establishes the guidelines, objectives, and goals to be followed by the federal, state, or municipal government during a 4-year period. Data on alcohol and food consumption have informed interventions for current public health issues, such as accidents and diseases caused by poor diets. ^17,18

Survey limitations include the small number of questions that can be asked compared with the BRFSS (about 90 questions), that data are self-reported, the lack of clinical data, and the survey’s cross-sectional design. Also, because VIGITEL is restricted to landline telephones, the sample may be biased. The proportion of landlines available ranges from 38% in the northern region to 74% in the southeastern region, according to data from the 2010 Demographic Census (the most recent data available). Thus, at least 62% of the study population in the northern region is not included, which can bias estimates. The low landline coverage is especially relevant in the capitals of the northern, northeastern, and midwestern regions. In contrast, in the southern and southeastern regions, which have high rates of landline use, the biases may be negligible. ¹⁹ Data originating from the population residing only in state capitals are also limited because lifestyles vary widely from capitals to smaller municipalities, especially the behavioral risk factors for NCDs.

Despite these limitations, VIGITEL is valuable for monitoring the risk and protective factors for NCDs and for characterizing the burden and distribution of adverse health events, setting priorities for public health actions, and assessing the effect of disease control measures.

The BRFSS is a model used not only in Brazil but also in several other countries, such as Australia, Canada, China, and Italy, ²⁰ although the methodology and scope of the surveys differ. For example, despite including cell phones in their samples, the systems implemented in Italy (the Italian Behavioral Risk Factor Surveillance System) and in Australia (the South Australian Monitoring and Surveillance System) use various methodologies to sample the telephone lines to be interviewed. ^21,22 In Italy, a nationwide survey uses local health units as a unit of data collection, and the eligible participants are residents who have a telephone number (landline or cell phone) and who are capable of being interviewed. ²¹ In Australia, all households in South Australia with a landline listed in the electronic white pages are eligible for selection. ²²

VIGITEL was a pioneer in incorporating the raking method for poststratification weighting that was implemented in the BRFSS. The challenge of using telephone-based surveillance systems to collect data on chronic diseases and behavioral risk factors is to ensure that the methodology is effective and efficient in obtaining and providing representative and reliable population data. The raking method produces more accurate estimates of population risk factors than does the cell-by-cell weighting method. ¹⁴

In part because of its relatively recent implementation and the socioeconomic differences between Brazil and the United States, VIGITEL still has challenges. For example, the sample must be expanded to include the interior regions of the states to ensure representative results. Including cell phone numbers in the sample is another challenge because of the increased number of Brazilian households with only cell phones. In the United States, the proportion of adults living in cell phone–only households increased by more than 700% from 2003 to 2009 ²³ ; in 2016, an estimated 61% of households in Brazil had only cell phones. ²⁴ The exclusive use of cell phones is especially high in younger age groups and in certain racial/ethnic minority groups. ²³ Including cell phone numbers is even more necessary where landline coverage is limited, as in the northern and northeastern regions, ¹⁹ because of differences in the characteristics of persons living in households with or without landline telephones.

In Brazil, a 2008 pilot study ²⁵ and a 2017 simulation study assessed the feasibility of including cell phone interviews in surveys. ¹⁹ Although these studies indicated the benefits of including a subsample of adults with only mobile phones in the VIGITEL sample, especially in the capitals of the northern and northeastern regions, it is not known whether cell phones will be included in the sample.

Another challenge of VIGITEL is how to increase the number of variables studied so that other chronic diseases (eg, chronic respiratory diseases, depression) and other populations of greater vulnerability (eg, children, adolescents) can be included. A final challenge is maintaining funding for the next few years, considering that Brazil has few resources for nationally representative research, despite the value of the information it provides. The recovery of Brazil’s economic crisis may help guarantee investment in surveys such as VIGITEL.