Correlates of influenza vaccine uptake among community-dwelling older adults in Brazil

ABSTRACT

This study aimed at assessing the factors associated with vaccine uptake in a representative sample of community-dwelling Brazilian older adults, specifically focusing on differences in socioeconomic factors among the country regions. We conducted a cross-sectional, population-based study, using a probabilistic household sample in 2013. Individuals aged 60 years or more answered a structured questionnaire informing on vaccination status and sociodemographic and behavioral covariates. Associations between variables were evaluated using prevalence ratios estimated by Poisson regression models. The overall vaccination coverage (72.6%) in older adults ranked lower than the goal of 80% stipulated by the Brazilian health authority; vaccine uptake differed significantly among the country regions. The prevalence of vaccination was lower in black individuals in Brazil than that in their white counterparts. The prevalence of vaccine uptake was significantly associated with covariates on current life style, use of health care, and socioeconomic determinants. Compared with individuals with 0–3 years of education, more schooled individuals had higher prevalence of vaccine uptake in the North, Northeast, and South regions of the country. This study showed that there is room for increasing vaccination coverage among the elderly in Brazil. The knowledge previously obtained on factors significantly associated with vaccine uptake has not prevented them to continue influencing this outcome. The socioeconomic inequality in vaccination in some Brazilian regions reinforces the need of targeting the intervention toward the most vulnerable groups.

KEYWORDS:

• aging
• coverage
• epidemiology
• inequalities
• Influenza vaccination
• uptake

Introduction

Seasonal influenza is a public health problem with significant socioeconomic implications.¹ It is a respiratory infectious disease that affects 5–10% of the world's adult population,^1,2 causing 250,000–500,000 deaths in annual epidemics.² Severe illness, complications, and hospitalizations are more frequent in groups including children aged less than 5 years, people with chronic diseases, and older individuals.^1,3,4 In particular, older adults are considered to be at a higher risk for influenza-associated excess of mortality.^5,6

Influenza vaccination is still the most effective way to prevent the disease and reduce the mortality rate among elderly individuals, albeit its efficacy in this age group is modest.¹ Antunes et al.⁷ observed a reduction of 26% in the overall mortality due to influenza and pneumonia after the introduction of annual mass vaccination. Daufenbach et al. (2014)⁸ reported a decrease in the influenza-related hospitalization rate (0.75/1,000 elderly/year) among Brazilian elderly population between 1992 and 2006. In addition to reducing the risk of hospitalization and death, influenza vaccination may benefit older adults by preventing secondary complications (e.g., severe respiratory infections and cardiopulmonary disease).^9,10

Although factors related to vaccine uptake have been widely investigated,^4,11-13 continuous research is needed to evaluate whether the knowledge previously acquired has impacted on subsequent vaccination campaigns or the factors that have been identified as constraints to vaccination still continue influencing this outcome. The Advisory Group of the World Health Organization for Influenza Vaccines stated that the policy-making process demands studies assessing how attitudes are shaped and identifying the factors that are related to the decision to vaccinate.¹⁴ These studies became even more necessary with the emergence of the H1N1 (2009) pandemic, a more lethal virus, which reinforces vaccination as an essential public health intervention.

Influenza vaccination has been implemented in Brazil since 1999 through nationwide campaigns that are usually conducted in April/May, immediately before the cold season. The Brazilian National Immunization Program provides free-of-charge influenza vaccination to the following groups: children aged 6 months to 5 years, pregnant women, old adults (aged 60 years or more), health professionals, indigenous population, and people with chronic diseases and other diseases that compromise immunity. Vaccine delivery is primarily conducted at primary health care units. The health system produces the vaccine and distributes it to each state, municipality, and public health service. In 2013, nearly 43 million doses of influenza vaccine were delivered at an estimated cost of almost 150 million US dollars; an additional financial transfer of more than 11 million dollars was demanded for the organization of the annual vaccination campaign.¹⁵

Vaccine uptake varies widely between and within regions, independent of the country's level of economic development.¹¹ In Brazil, previous studies have reported evidence on vaccination coverage in specific sites^16-18 or in the biggest city of the country.^12,19 However, no previous study has reported information regarding either the national level or its macroregions. This study aimed at assessing the factors associated with vaccine uptake in a representative sample of community-dwelling Brazilian older adults and evaluating the differences in the uptake related to socioeconomic factors among the country regions.

Results

In this study of a representative sample of Brazilian older individuals, the overall prevalence of influenza vaccine uptake was 72.6% (95% CI: 71.1–74.1). Vaccination coverage was significantly different (p < 0.001) among the country regions, the Northeast region being the one with the lowest uptake (66.6%). The other regions showed vaccination uptake of more than 70% [North (71.2%); Southeast (73.6%), South (78.7%), and Midwest (75.4%)] (Fig. 1).

The most frequent reason for not taking the vaccine in Brazil was the fear of vaccine side effects (25.8%), followed by the report that they rarely get the flu (24.9%) and that they do not believe that the vaccine protects against influenza (12.7%; Table 1).

Table 1. Reasons for not taking influenza vaccine.

		Country region
	Brazil	North	Northeast	Southeast	South	Midwest
Reasons	%	%	%	%	%	%
Afraid of vaccine side effect	25.8	29.2	25.5	26.2	27.2	18.4
Rarely gets the flu	24.9	15.8	19.2	31.5	17.6	26.6
Do not believe the vaccine protects against flu	12.7	10.1	12.7	13.1	13.5	9.8
Afraid of shot	7.1	8.9	9.1	4.6	12.1	4.4
Didn't know it was necessary to take the vaccine	5.2	7.4	4.4	6.5	2.1	3.0
Medical contraindication	3.4	2.6	2.9	3.0	6.3	4.3
Vaccine was not available in the health care center	2.8	1.6	3.5	2.4	2.3	3.9
Didn't know where to take the vaccine	1.4	3.9	1.7	0.9	0.3	3.3
Transport issues	1.4	0.9	1.7	1.1	1.6	1.2
Health care center was far away	1.1	4.6	1.6	0.2	1.3	1.6
No one to accompany to the health service	0.6	2.2	1.0	0.0	0.0	2.7
Financial difficulties	0.1	0.0	0.1	0.1	0.2	0.0
Other	13.6	12.7	16.6	10.3	15.3	20.8

Table 2 depicts the distribution of population and shows the unadjusted assessment of covariates for vaccination in the whole sample. According to the results, vaccine uptake was significantly associated with 2 demographic variables (race and current marital relationship). A lower proportion of vaccination was observed among black individuals (66.9%) and among single, widowed, or divorced people (70.6%). Vaccine uptake was significantly associated with use of health services and life style factors (Table 2).

Table 2. Descriptive statistics and Unadjusted PR according to total sample and country regions. Brazil, 2013 (N=11,175).

		Influenza vaccine uptake
Variables	%	Yes (%)	No (%)	p-value^*
Gender				0.456
Male	43.6	71.9	28.1
Female	56.4	73.1	26.9
Age (years)				0.149
60 – 69	56.4	71.6	28.4
70 – 79	30.0	74.8	25.2
80 or more	13.6	72.1	27.9
Race / skin color				0.032
White	53.6	74.1	25.9
Black	9.2	66.9	33.1
Brown	35.6	71.4	28.6
Yellow	1.3	83.9	16.1
Indigenous	0.3	71.4	28.6
Marital relationship				0.021
No	42.6	70.6	29.4
Yes	57.4	74.1	25.9
Education (years)				0.197
0–3	44.9	70.9	29.1
4–8	25.1	73.5	26.5
8–11	19.8	74.6	25.4
12 or more	10.2	74.1	25.9
Binge drinking				<0.001
No	95.8	73.2	26.8
Yes	4.2	59.9	40.1
Smoking				<0.001
No	87.3	74.3	25.7
Yes	12.7	61.0	39.0
Physical activity				<0.001
No	86.1	71.4	28.6
Yes	13.9	79.9	20.1
Number of chronic diseases				<0.001
0 diseases	33.6	65.1	34.9
1 diseases	34.8	74.7	25.3
2+ diseases	31.6	78.3	21.7
Depression				0.173
No	90.5	72.2	27.8
Yes	9.5	76.0	24.0
Disability				0.270
No	83.8	72.9	27.1
Yes	16.2	70.8	29.2
Recent health care				<0.001
No	15.5	56.0	44.0
Yes	84.5	75.6	24.4
Health insurance				<0.001
No	68.0	70.5	29.5
Yes	32.0	77.0	23.0
Register in the Family Health Program				0.011
No	34.8	69.9	30.1
Yes	56.1	74.5	25.5
Don't know	9.1	71.5	28.5

* Rao-Scott test.

According to the multiple regression model, the prevalence of influenza vaccine uptake was higher among older individuals and those currently married [PR = 1.05 (1.01–1.10)]. Black elders had 9% lower prevalence of vaccine uptake when compared to whites. As regards current life style, the prevalence of vaccination was significantly higher among those who reported practicing physical activities [PR = 1.11 (1.06–1.16)] and about 17% lower among elders with unhealthy life style (i.e., binge drinking and smoking). The number of chronic diseases was the only general health variable related to vaccination. Older individuals with more than one chronic disease had more chance of being vaccinated. Being registered in the family health program, having had a recent health care visit, and having a health insurance plan significantly increased the prevalence of vaccine uptake, respectively, in 27%, 6%, and 8% of the individuals (Table 3).

Table 3. Multiple variables analysis: Poisson regression model for influenza vaccination among the elderly (n= 11,175). Brazil, 2013.

Variables	Unadjusted PR (95% CI)	Adjusted PR (95% CI)
Gender	Demographic (1)
Male	1.00
Female	1.02 (0.97–1.01)
Age (years)
60 – 69	1.00	1.00
70 – 79	1.05 (1.00–1.09)	1.05 (1.00–1.10)^*
80 or more	1.01 (0.95–1.07)	1.02 (0.95–1.08)
Race / skin color
White	1.00	1.00
Black	0.90 (0.83–0.98)^*	0.91 (0.83–0.99)^*
Brown	0.96 (0.92–1.01)	0.96 (0.92–1.01)
Yellow	1.13 (0.97–1.32)	1.12 (0.97–1.31)
Indigenous	0.96 (0.73–1.27)	0.97 (0.74–1.29)
Marital relationship
No	1.00	1.00
Yes	1.05 (1.01–1.09)^*	1.05 (1.01–1.10)^*
	Socioeconomic (2)
Education (years)
0–3	1.00
4–7	1.04 (0.98–1.09)
8–11	1.05 (1.00–1.11)^*
12 or more	1.04 (0.98–1.12)
	Life style (3)
Binge drinking
No	1.00	1.00
Yes	0.82 (0.71–0.93)^**	0.82(0.72–0.94)^**
Smoking
No	1.00	1.00
Yes	0.82 (0.76–0.89)^***	0.83 (0.76–0.90)^***
Physical activity
No	1.00	1.00
Yes	1.12 (1.07–1.17)^***	1.11 (1.06–1.16)^***
	General health (4)
Number of chronic diseases
0 diseases	1.00	1.00
1 diseases	1.14 (1.09–1.20)^***	1.14 (1.08–1.19)^***
2+ diseases	1.20 (1.14–1.26)^***	1.19 (1.13–1.25)^***
Depression
No	1.00
Yes	1.05 (0.98–1.13)
Disability
No	1.00
Yes	0.97 (0.92–1.02)
	Use of health services (5)
Recent health care
No	1.00	1.00
Yes	1.35 (1.26–1.45)^***	1.27 (1.18–1.37)^***
Health insurance
No	1.00	1.00
Yes	1.09 (1.04–1.14)^***	1.06 (1.01–1.10)^*
Register in the Family Health Program
No	1.00	1.00
Yes	1.06 (1.02–1.11)^**	1.08 (1.03–1.12)^***
Don't know	1.02 (0.95–1.10)	1.04 (0.97–1.12)

* p < 0.050,

** p < 0.010,

*** p < 0.001.

(1) Demographic variables adjusted for themselves.

(2) Socioeconomic variable adjusted for demographic variables.

(3) Covariates assessing life style adjusted for socio-demographic characteristics.

(4) Covariates assessing general health adjusted for socio-demographic and life style characteristics.

(5) Covariates on use of health services adjusted for socio-demographic and life style and general health characteristics.

Table 4 shows the adjusted association of vaccine uptake and socioeconomic status according to the 5 Brazilian regions. Except in the Southeast and Midwest regions, the prevalence of vaccination was higher among more schooled individuals. In the North, the prevalence of vaccine uptake among individuals with 4–11 years of education was nearly 20% higher than that in individuals with 0–3 years of education. In the Northeast and South, this association was observed in individuals with 12 or more years of education, with an increase of 15% and 12%, respectively.

Table 4. Adjusted prevalence ratio (PR) for factors related to influenza vaccination uptake among elderly individuals according to country regions. Brazil, 2013.

	Influenza Vaccine Uptake
	Country Region
Education	North	Northeast	Southeast	South	Midwest
(years)	PR (95%CI)^†	PR (95%CI)^†	PR (95%CI)^†	PR (95%CI)^†	PR (95%CI)^†
0–3	1.00	1.00	1.00	1.00	1.00
4–7	1.22 (1.06–1.40)^**	1.09 (0.99–1.19)	0.98 (0.89–1.08)	0.98 (0.89–1.08)	0.94 (0.85–1.04)
8–11	1.17 (1.02–1.35)^*	1.03 (0.92–1.15)	1.02 (0.93–1.11)	1.07 (0.96–1.18)	1.01 (0.91–1.12)
12 or more	1.17 (0.95–1.43)	1.15 (1.03–1.28)^*	0.95 (0.85–1.08)	1.12 (1.03–1.22)^**	0.92 (0.79–1.07)

^† PR adjusted for demographic variables (age, race and marital relationship).

* p < 0.050,

** p < 0.010.

Discussion

This study showed that vaccine uptake in Brazil ranked lower than the goal of 80% stipulated by the Brazilian Ministry of Health,²⁰ although it was close to the 75% coverage suggested by the World Health Organization.²¹ Prevalence significantly differed between the country's regions, and uptake was associated with race, lifestyle, and health services in Brazil. Schooling significantly increased the prevalence of vaccine uptake in the North, Northeast, and South regions. To the best of our knowledge, this was the first study to produce information regarding the factors related to vaccine uptake at the national level and its macroregions. The continuous monitoring of information on vaccine coverage and associated factors is relevant for the evaluation and planning of the vaccination program, primarily in Brazil, which encompasses a very large territory, with intense socioeconomic differences across its macroregions, and is committed in achieving a high population coverage.

The results reported here indicate that the prevalence of vaccine uptake also ranked lower than the administrative estimates made by the Brazilian Ministry of Health, which used information on doses delivered and effectively administered in each region and are not sensitive to individuals who have effectively received the shot.²² Furthermore, the administrative record of the Brazilian health authority includes elders living in long-stay institutions, whereas this study assessed a probabilistic sample of older people living in the community. However, the administrative record of the Ministry of Health also depicted differences in vaccine coverage between the country regions, and the Northeast was the region with the poorest performance.

Difficulties to achieve the goals of vaccination coverage stipulated by the health authorities have been widely reported in the literature.^3,11,23 In the European Union, vaccination coverage ranged from 1.0% to 77.4% (median 44.7%) of older people, in 2012–2013. The European Union target of 75% was reached in only 2 Member States.³ In 2013, the coverage rate among older adults in the Americas ranged from 5% to 100%.²³ Furthermore, vaccine uptake varied widely between and within regions.¹¹ Our results showed a reduced, though significant, difference in vaccination coverage among the country regions, with lower figures in the North and Northeast regions, those with poorer indices of socioeconomic development. These five macroregions in Brazil have significant differences regarding socioeconomic conditions. The South and Southeast are the most developed, with a higher population density. A lower population density is observed in the North region, followed by the Midwest and Northeast. These regions are the ones with lower socioeconomic conditions.²⁴

In spite of having not achieved the goal established by the Ministry of Health, the prevalence of vaccine uptake reported here is high, which reflects the success of the Brazilian program of vaccination. Indeed, financial difficulties were the least prevalent motive alleged for having not been vaccinated, which is in line with the universal access to vaccines in the country. The most prevalent motives for having not taken the vaccine reported here are consistent with previous studies.^18,25 Moreover, our results concur with the concern that accessibility for older adults is not limited to free-of-cost access and includes several dimensions, such as the distance to the health care center, means of transport, and health literacy.²⁶ The success of vaccination campaigns in the country is supported by the credibility of the Brazilian National Immunization Program. This is recognized by the population due to its results in infant vaccination that led to control and eradication of several diseases. The expansion of health services and the decentralization process are also considered as factors that directly or indirectly contribute to an increased access to vaccination, thereby increasing the coverage.²⁷

Regarding the factors associated with vaccine uptake, several studies have shown that it is related to individual characteristics,^{4,12,25,28,29} such as gender,^13,29 socioeconomic conditions,^4,25 and general health status.^4,12,13 The adjusted analysis confirmed the importance of current life style, use of health care, and socioeconomic determinants.

The association between socioeconomic factors and influenza vaccine uptake is controversial. Some studies reported a positive relationship,^4,11 others a negative²⁸ or absent association.¹³ Findings reported here suggest that universality and equity in vaccination was observed in the Southeast and Midwest of the country, because vaccination was not more prevalent in more schooled individuals. However, vaccination was more prevalent in whites than in blacks in the whole country and in older adults with more years of education in the North, Northeast, and Southeast. Previous studies have reported similar associations with education^4,11 and race.^4,30,31 Since vaccine delivery is provided free of charge by the Brazilian National Immunization Program, socioeconomic inequalities in access to the vaccine cannot be attributed to the lack of money to afford it. They rather reflect the lower level of health literacy that affects lower socioeconomic and minority strata.³² Accordingly, Bennet et al.³⁰ found that health literacy significantly mediates racial/ethnic and education-related disparities in receiving influenza vaccination.

Another explanation for the abovementioned association is the so-called Inverse Care Law, which states that good medical care tends to vary inversely with the need of the population served,³³ as richer people often benefit more from public spending on health.³⁴ Further specifying this explanation, Victora et al.³⁵ raised the inverse equity hypothesis, stating that new public health interventions and programs initially reach individuals in the higher socioeconomic groups. According to their results, the reduction of health inequalities is solely observed in a later stage, when the rich strata already has achieved a new minimum level for morbidity and mortality and the poor gains greater access to the interventions.

The role of current lifestyle in predicting vaccine uptake has been reported earlier^4,12,18 in studies highlighting a higher prevalence of vaccination among older adults who practice physical activities and who did not smoke and drink. These associations may be explained by the fact that these individuals are more committed to healthy behaviors and have more access to health care facilities.¹² Health literacy may have made them more likely to receive advice on vaccination.

Having received recent health care, having health insurance, and being registered in the Family Health Program were directly associated with vaccination. These associations are also coherent with previous findings in different cities.^12,18,28 The contact with health professionals and services increases the likelihood of receiving influenza vaccine.^12,36,37

Finally, as previously observed,^{4,12,18,36-38} vaccine uptake was more prevalent among individuals with chronic diseases. According to the evidence,^26,36 having comorbidities may increase the severity of influenza, possibly making this group more motivated to seek vaccination. In addition, those affected by chronic diseases may have a more intense interaction with health services, which increases the likelihood of being appropriately informed of the beneficial effects of vaccination.

Having assessed a large, probabilistic sample, representative of older people living in the community in the whole country and each of its geographic regions, is the primary strength of this study. Having exclusively assessed self-reported information on vaccination status is its primary limitation. The assessment of self-reported information is subject to recall bias; this study did not assess the vaccination card or other written records on the subject. In spite of this observation, we argue that previous studies considered the self-report of vaccine uptake as highly sensitive and moderately specific.^39,40 However, it is important to mention that misclassification may be related to demographic characteristics. The specificity of self-report of influenza vaccine uptake was lower among individuals aged 65 years or more when compared with individuals of younger ages. Moreover, under and over-reporting of vaccine uptake related to a combination of 7 vaccines were found to be associated with socioeconomic factors.⁴⁰

Conclusions

This study concluded that the prevalence of vaccine uptake in Brazil is related to sociodemographic and behavioral factors that had already been acknowledged in the literature. Having a previous knowledge on those factors that influence vaccine uptake has not prevented them to continue impacting on the decision to take the shot. Providing universal access to free-of-cost vaccination is not sufficient to achieve the goal of vaccination coverage, and socioeconomic inequalities in vaccination persist in some Brazilian macroregions. The results reported here are supportive of modifications in the vaccination program targeting the intervention toward the most vulnerable groups.

Methods

We conducted a cross-sectional study with data from the Brazilian National Health Survey. This survey gathered a probabilistic household sample and was performed by governmental agencies (Brazilian Institute of Geography and Statistics and the Ministry of Health) in 2013. Data collection started in August and was completed in a 6-month period. The sample comprised 64,348 households and was stratified into 3 cluster stages. The primary sampling units were census tracts, households were the second-stage units, and the adults in these households (18 years of age and over) were the third-stage units. The sample size took into consideration the level of precision desired for the estimation of some indicators at different levels of group subdivision and population groups.^41,42 The final sample size was adjusted considering the design effect. The number of households estimated in the sampling was 81,167, with data being collected from 64,348 of them, thus accounting for a 20.8% of sample loss (8.1% of non-response). Sample weights were defined for the primary survey units, the households, and all of their residents. Four age groups were used: 18–24 years, 25–39 years, 40–59 years, and 60 years or older.⁴³ The final weighting was computed as the product of the inverse probability of being selected at each stage of the sampling plan, subsequently corrected for non-response. Detailed information regarding the study sample design has been published elsewhere.⁴⁴

This study included all participants aged 60 years and more who were selected to answer the section dedicated to older individuals (n = 11,175) and had complete information on the outcome variable. All participants were interviewed in their own homes by health professionals specifically trained to apply a comprehensive questionnaire on sociodemographic and behavioral characteristics, lifestyle, quality of life, use of health services, and general health information.

Outcome of interest

Influenza vaccine uptake was evaluated according to the direct response to the question on whether the subject had received the influenza vaccine in the preceding 12 months.

Non-vaccinated individuals were asked to inform the motive why they did not take the shot. Answers were classified into the following categories: 1) rarely get the flu; 2) did not know it was necessary to take the vaccine; 3) did not know where to take the vaccine; 4) afraid of vaccine side effects; 5) afraid of the shot; 6) did not have anyone to accompany them to the health service; 7) financial difficulties; 8) transport difficulties; 9) health care center was too far; 10) the vaccine was not available at the health care center; 11) medical contraindication; 12) do not believe that the vaccine protects against flu; and 13) other motives.

Covariates

Independent variables assessed demographic characteristics [age (60–69 years, 70–79 years, 80+ years), sex (male, female), race/skin color (white, black, brown, yellow, indigenous), and current marital relationship (yes, no)]; socioeconomic status [education (0–3, 4–7, 8–11, and 12 years or more of formal education)]; current life style [tobacco smoking (yes, no), abusive consumption of alcohol (yes, no), and physical activity (yes, no)]; general health [number of self-reported chronic diseases (0 disease, 1 disease, and 2+ diseases including hypertension, diabetes, cardiovascular disease, asthma, pulmonary disease, and kidney disease), self-reported depression (yes, no), and disabilities (yes, no)]; and health care [health insurance (yes, no), recent use of health care (yes, no), and self-report about registering in the Family Health Strategy (no, yes, do not know)].

Individuals considered as practicing physical activities were those reporting at least either 150 min or more of mild or moderate physical activity per week or 75 min of vigorous activity per week.⁴⁵ The activities included sports or planned exercise. Abusive consumption of alcohol was defined as 5 standard drinks or more for men and 4 standard drinks or more for women in one single occasion in the last 30 days.⁴⁶ Recent use of health care was defined as the visit to a physician in the last year. Disability was defined as difficulties in one or more Basic⁴⁷ (bathing, dressing, walking, using the toilet, feeding, and moving in and out of bed and chairs) or Instrumental Activities of Daily Living⁴⁸ (managing finances, using transportation, shopping, and taking medication).

Statistical analysis

Statistical analysis involved descriptive and inferential analyses, with a 5% significance level and 95% confidence interval (CI). Associations between categorical variables were assessed using the Rao-Scott chi-square test. All independent variables with p < 0.20 in the unadjusted assessment of associations were included into a multiple Poisson regression model in the following order: sociodemographic factors, current life style, general health, and use of health services.

Data analysis was performed using the Stata 13.0 software (Stata Corporation, College Station, TX, USA). All analyses considered weighting and correction for the design effect, using the “survey” command to analyze data originating from a complex sample.

Ethical considerations

The survey was approved by the Brazilian National Commission on Ethics in Research (CONEP) of the National Health Council (CNS).

Figure 1. Vaccination coverage and confidence interval of 95% by country region, Brazil, 2013.

Disclosure of potential conflicts of interest

No potential conflicts of interest were disclosed.

Funding

Dr Antunes was supported by the Brazilian Council for the Scientific and Technological Development (CNPq), Grant N. 301968/2014–4.