What the impact of the economic crisis on the Brazilian’s lifestyle between 2013 and 2016? A long-term trend cross-sectional study.

Background: Brazil has been going through a period of economic crisis since 2015 and the impact on health indicators and the quality of life of the most recent crisis is unclear. This study sought to establish a relationship between the impact of the economic crisis experienced by Brazil during the period 2013-2016 and the lifestyle of adult populations in the 27 capitals of the Brazilian state. Results: According to the binary logistic regression model, at least completing secondary education increases the probability of individuals presenting higher LS by 4.9% (OR=1,049, 95% CI 1,029– 1,070); being involved in some type of relationship decreases the probability of presenting a higher LS by 5.3% (OR=0.947, 95% CI 0.029-0.965); having an adequate body mass index increases LS by 27.3% (OR=1.273, 95% CI 1,245.10.10). Living in a city with a lower unemployment rate increases the probability of having a higher LS score by 3.8%(OR=1,038, 95% CI 1,012-1.064); living in a city with a more basic food basket decreases the probability of having a higher LS score by 5.4% (OR=0.946, 95% CI 0.910-0.982). Presenting higher yields decreases the probability of having a higher LS by 12% (OR=0.880, 95% CI 0.860-0.900). In addition, there was an increase in the percentage of leisure-time PA in 2015 (20.9%), followed by a decrease to 18.1% in 2016, after a stabilization period in 2013 (19.5%) and 2014 (19.6%). The physical exercise showed a steady annual increase, with a score of 49.6% in 2013, followed by 50.9% and 51.9% in 2014 and 2015, respectively, and ending with 54.5% in 2016. The "Watch TV 5x per week" indicator recorded a sharp drop between 2013 (27.9%) and 2015 (22.7%), as well as the consumption of soft drinks, which went from stable values in 2013 (19.5%) and 2014 (19%)

such as public safety, education, and health, which have made progress in recent decades. The following are some milestones in the field of health: the implementation of the Unified Health System (1990), the Family Health Strategy (1994), and the National Health Promotion Policy (2006).
In addition to a decline in chronic malnutrition, Brazilian public health policies have resulted in a significant decrease in smoking habits, gaining international attention for the combat against tobacco use 1 . According to Malta et  Thus, this study aimed to establish a relation between the impact of the economic crisis experienced Lifestyle Score (ls) For the analysis of lifestyle, we opted for the creation of a "Lifestyle Score" (LS) based on the sum of the following indicators: physical activity during leisure time (yes = 1; no = 0);practicing physical exercises (yes = 1; no = 0); active form of transport to work (yes = 1; no = 0); active in terms of general transit (yes = 1; no = 0); regular consumption of fruits (yes = 1; no = 0); regular consumption of vegetables and legumes (yes = 1; no = 0); regular consumption of beans (yes = 1; no = 0); smoking (no = 1; yes = 0); excessive consumption of alcohol (no = 1; yes = 0); and excessive television viewing time (no = 1; yes = 0). Thus, the minimum and maximum score were0 and 11, respectively. In the case of micro and macroeconomic indicators, we used the strategy of dividing the data according to each variable's median, defining them as "higher value" and "lower value."

Statistical analysis
Descriptive statistics (i.e., means and standard deviations) of macro and microeconomic variables and relative values (i.e., absolute frequency, relative and 95% CI) of socioeconomic and lifestyle variables were performed. For the statistical treatment of the collected data, we opted for the use of binary logistic regression analysis, considering the LS as a dependent variable and the following as independent variables: vacancy rate, staple foods basket value, average worker income, GINI index of inequality and years (2013-2016).The following were used as covariates: years of schooling, marital status (accompanied), BMI (body mass index), gender, and capital region (north, northeast, midwest, southeast, south).

Behavior of socioeconomic indicators
The descriptive analysis of the socioeconomic indicators (Table 1)   Analyzing the data related to lifestyle ( Table 2 and Table 2   As for the regions of the country, with reference to the northern region, with the lowest HDI, an upward trend was shown in terms of the chance of achieving a higher LS score as the regional HDI increased (p < 0.001); all regions showed statistically significant associations with the southeastern region, with the highest HDI, evincing 59.8% more chances of obtaining higher LS. Residing in a city with a lower unemployment rate increases the chance of having a higher LS score On evaluating the score, 2015 is the year that presents a differentiated curve, reaching the mark of 53.4%, almost 20% higher in comparison to the other years.

Discussion
Maintaining a healthy lifestyle and being physically active is a multifactorial result and is also influenced by socioeconomic conditions. Therefore, investigating changes in a society's lifestyle from an epidemiological perspective requires the observation of a set of variables that may or may not be clearly perceptible.
According to Triaca 14 , studies covering themes such as health and economy use unemployment and mortality as main outcomes. Further, most studies showed a pro-cyclical trend in their findings (i.e., a decrease in certain behavior following worsening of economic conditions). This is the trend found by Hone et al 15 , who analyzed the effects of economic recession on mortality rate in 5565 Brazilian municipalities between 2012 and 2017, in a longitudinal study that found a direct correlation between unemployment and a percentage increase in national mortality rates, as well as between higher mortality rates among black individuals (black or brown) and greater protection against mortality in locations with better access to social protection programs.
However, in some cases,an anti-cyclic pattern is observed (i.e., inverse association of behavior with economic indicators). So, if the economy is shrinking and smoking increases, this is an anti-cyclical pattern. However, if the economy faces a downward turn and smoking also decreases, it is a procyclical pattern 4, 5 .
In the study developed by Xu 6 ,using  What these studies have in common is the use of a comprehensive approach, avoiding the option to use mortality and unemployment as main variables and using other factors in the analysis. We also opted for this standard in our study. The present study also points to oscillations between anti-cyclic and pro-cyclic patterns depending on the historical cut adopted. In this analysis, 2013 and 2014 presented anti-cyclical tendencies; although these years preceded the crisis period, they still presented favorable economic data. Thus, with a favorable economic situation, there was a slight improvement in the score. However, after 2014, even with the onset of recession in the Brazilian economy, the increase in the unemployment rate, and the already sharp rate of informality, the trends in quality of life standards of the general population were not sharply affected, especially in regions with higher MHDI, a fact corroborated by the findings demonstrated in Table 2 and Fig. 1. Thus, we found a prominent pro-cyclical trend in 2015, with a significant improvement in health indicators that make up the LS even during the acute period of the economic crisis. Another important factor to be highlighted is that trends in 2015 were directly influenced by two major sporting events year that escaped the anti-cyclical pattern presented in the other years. In the midst of a political and economic crisis, 2015 presented an improvement in most health-related indicators with which we formulated the LS. Consequently, the reason for the curve shift was descriptively identified without, however, identifying a causal relation that could explain these findings. In the case of Brazil, the economic crisis that initially began in 2015 had a negative impact on the population over time, but these effects were only observed in the second year of crisis, being preceded by a temporary improvement in indicators. Therefore, years of socioeconomic development can lessen the negative effects of a crisis on the lifestyle of the population for a short period.

Ethics approval and consent to participate.
This study was done with secondary data and obtained from a public domain database obtained from the website of federal, state, and municipal government agencies in Brazil. Therefore, according to Brazilian national guidelines, approval is not applied to the ethics committee.

Consent for publication.
Not applicable.

Availability of data and materials
All data generated or analyzed during this study are included in this published article. The Debt data of the States and Municipalities were visualized by the public database https://www4.bcb.gov.br/fis/dividas/lmunicipios.asp.
DataSUS/Vigitel data were obtained from the Public Database of the Ministry of Health at the address http://tabnet.datasus.gov.br/cgi/vigitel/vigteldescr.html.
Finally, all data related to federal socioeconomic indicators were obtained directly from the database of the Brazilian Institute of Geography and Statistics (IBGE) available at: https://downloads.ibge.gov.br/downloads_estatisticas.html .

Competing interest
The authors declare that they have no competing interests.

Funding
None.

Author's contributions
Sonny Bezerra, Denilson Menezes and Michele Oliveira compiled data. Sonny Bezerra, Denilson Menezes and Emanuel Salvador conceptualized the study, analyzed data, produced the figures and tables, and wrote the original draft of the manuscript. Emanuel Salvador provided statistical guidance for interpreted data and input on the analysis and edited early drafts of the manuscript. All authors contributed to subsequent drafts of the manuscript and provided critical interpretation of the results and approved the manuscript for publication and Levy Rezende also translated the manuscript after review.