Introduction

Chronic non-communicable diseases (NCDs) are an important public health problem worldwide. In Brazil, in addition to compromising the quality of life of a significant portion of the population and contributing to health service overload, NCDs accounted for approximately 70% of deaths in 2016. Population surveys, such as the Chronic Disease Risk and Protective Factors Surveillance Telephone Survey (Vigitel) and the National Health Survey (PNS), help in the monitoring of NCDs in Brazilian municipalities, although they present relevant methodological differences.^1-3

The PNS was a population-based survey, using face-to-face data collection, while Vigitel was conducted by telephone and, therefore, implied lower cost and operational complexity.

In the municipality of Rio Branco, capital city of the state of Acre - one of the Brazilian capitals with the lowest landline coverage - there are concerns about the representativeness of Vigitel data and possible bias resulting from this limitation. However, the comparative assessment of the databases of both surveys can contribute to a better understanding of the methodological problems in these studies in the municipal context.⁴

The objective of this study was to compare the prevalence of NCDs and ultra-processed foods, alcohol and tobacco consumption, estimated by Vigitel and PNS in Rio Branco, state of Acre, Brazil.

Methods

This was a cross-sectional study with secondary data retrieved from Vigitel and PNS in 2019, made available by the Ministry of Health, with open access, in their respective websites.^5,6 Both databases were accessed on May 4, 2021.

Rio Branco is located in the Northern region of Brazil and it is the leading financial, corporate, political and cultural center in the state of Acre, with a high human development index (HDI= 0.727), according to the last Atlas of Human Development in Brazil, in 2010.^7,8 In 2020, Rio Branco had a population of 413,418 inhabitants,² 46.2% of the total population of the state in the period, and occupied a territory of 8,835,154 km². In 2019, 91.8% of the municipality's population lived in the urban area and, according to the 2010 Demographic Census, the schooling rate in the age group 6 to 14 was 95.1%.^7,9 Regarding health indicators, the infant mortality rate was 11.1 deaths per 1,000 live births in 2019. In 2020, Rio Branco had 837 health facilities and of these, 232 were within the Brazilian National Health System (SUS).^7,10

Vigitel is a telephone survey, conducted in two stages: first, (i) a systematic draw of local landline telephone, and then, (ii) the identification of the lines that were drawn, regarding their eligibility. Business and inactive phone lines or those that did not answer the phone call, were not eligible. Thus, a probabilistic sample of the adult population (≥18 years) residing in Brazilian capitals was obtained. With regard to the PNS, it is a household survey, conducted in three stages: (i) a draw of census tracts [primary sampling units (PSU)], (ii) households (secondary units) and (iii) a resident aged 15 years or older (tertiary units). Further information about the PNS method is available in another publication.²

Respondents aged 18 years or older were eligible for this study. Those under 18 years of age, pregnant women and women who did not know if they were pregnant were excluded. In addition, the individuals who were taking part in both surveys (Vigitel and PNS) but did not have complete records, were excluded.

The sociodemographic variables analyzed were sex (female; male), age group (in years: 18 to 24; 25 to 39; 40 to 59; 60 or older), race/skin color (White; Black; Asian; Brown; Indigenous) and schooling (in years of study: 0 to 8; 9 to 11; 12 and more). The variables related to health status were self-reported, based on a previous diagnosis of diabetes mellitus (no; yes), hypertension (no; yes), overweight (no; yes) and obesity (no; yes), with a physician confirmation. The lifestyle variables were smoking (no; yes), alcohol abuse (no; yes), ultra-processed foods (no; yes) and soft drink (no; yes) (Box 1) consumption.

Data were analyzed using a statistical software for analyzing complex samples, the Statistical Package for the Social Sciences (SPSS), version 26.0. The overall prevalence of sociodemographic, health and lifestyle characteristics, stratified by sex, and 95% confidence intervals (95%CI), were estimated. The prevalence of Vigitel was weighted (rake method) using estimates of sex, age and education of the population, for the municipality of Rio Branco in 2019, aiming to standardize the distribution of the interviewed population with that estimated by the Demographic Census.³ The PNS estimates were weighted by means of household weight, adjustment for non-response, sex and distribution of the total population by sex and age. The differences were calculated in percentage points, considering the proportion of each sample, Vigitel and PNS, and this difference was observed when there was no overlapping 95%CI.^1,11,12

The secondary data used in this article are in the public domain, made available by the Ministry of Health in an unrestricted manner and without personal identification, therefore it was not necessary to submit the research project to a Research Ethics Committee.

RESULTS

A total of 4,262 individuals (2,829 from Vigitel and 1,433 from PNS) were considered eligible. Of these, 3,037 individuals aged 18 years or older who had completed the interview (1,812 from Vigitel and 1,225 from PNS) were included in this study. 52.4% (95%CI 48.3;56.5) in Vigitel and 53.3% (95%CI 48.9;57.7) in PNS were female, and it could be seen a difference in the prevalence of -0.9%, with overlapping 95%CI (Table 1).

There was a difference between the surveys regarding the prevalence of race/skin color. The prevalence of people of mixed race/skin color was lower in Vigitel (60.3%; 95%CI 56.2;64.3), compared to that found in PNS (70.8%; 95%CI 67.4;73.9). There was no difference between the prevalence estimated by the surveys regarding the other sociodemographic variables such as health and lifestyle, and there was no overlapping of 95%CI values.

When the variables were stratified by sex, no differences in prevalence between Vigitel and PNS were observed in any of the variables investigated (Table 2). Although there was overlapping 95%CI, the difference in the prevalence of obesity between the two surveys was 3.2%, and it could be seen prevalence of 6.5% among males and 0.4% among females.

The differences in the prevalence of smoking and alcohol abuse, between Vigitel and PNS surveys, were 4.0% and 6.9%, in males, and -1.5% and -2.5% in females, respectively.

DISCUSSION

The analysis of 2019 Vigitel and PNS data showed that the prevalence related to NCDs, ultra-processed food consumption, alcohol abuse and smoking were similar, with overlapping 95%CI for the variables analyzed. Only the comparison of the prevalence of the race/skin color indicator showed a difference between the two surveys in relation to participants of Brown race/skin color, possibly related to difficulties regarding self-perception and self-declaration of this characteristic.¹³

Overlapping confidence interval for the prevalence of overweight and obesity suggests that there are no differences in the comparison between the two surveys, and corroborates studies that suggest self-reported weight and height as valid measures to determine nutritional status.^14,15

We did not observe any differences between the prevalence of smoking and alcohol abuse, in this study. However, other investigations suggest that, because it is a survey, in which there is no eye contact, respondents may feel more comfortable answering sensitive questions such as those related to the consumption of licit drugs.^16,17

The similarities identified in the prevalence of ultra-processed food and soft drink consumption, between both surveys, and in the prevalence of most of the variables evaluated in this study, are similar to those of another study that analyzed data from 2013 Vigitel and PNS surveys,¹² when the variables related to health and lifestyle in Brazilian capitals were compared.

This study has some limitations with regard to the method adopted by each of the surveys, this is the reason why the comparisons made are subject to possible biases of information and selection. There may be selection bias regarding Vigitel data, considering the need for landline coverage. Furthermore, it is worth highlighting the possibility of information bias due to the way Vigitel data was collected, by means of telephone interview, compared to PNS data collection. However, weighting methods were used and the differences in the information between the surveys showed overlapping 95%CI, and it can be inferred that telephone survey results were similar to those of the face-to-face survey. These findings corroborate evidence from other studies, which did not observe statistical differences in the comparison of parameter estimation, based on telephone and face-to-face surveys, of chronic diseases such as hypertension and diabetes mellitus, and sociodemographic variables such as sex, age (in years) and schooling.^18,19

Estimates of the prevalence of NCDs, ultra-processed food consumption, alcohol abuse and smoking in Rio Branco, state of Acre, in 2019, overlapped in both the Vigitel and PNS suggesting that even for the most distant regions from large centers and with low landline coverage such as the municipality of Rio Branco, telephone surveys may be a more cost-effective option when compared to face-to-face surveys.