Economic crises, behavioral changes and hospitalization due to affective disorders in Brazil between 2003 and 2017: a nationwide cross-sectional study

ABSTRACT Our aim was to analyze hospitalization due to affective disorders in Brazil from 2003 to 2017 and the possible association with economic indicators during crises. We used data on hospitalizations due to affective disorders within the Brazilian National Health System, obtained from DATASUS; data on health-related behavior (television-viewing and physical activity) from the VIGITEL database; and economic data from the World Bank database. We found that the numbers of hospitalizations increased one year after the 2009 crisis and one year after the 2016 crisis. Negative changes in health-related behavior also followed changes in the numbers of hospitalizations due to affective disorders.


INTRODUCTION
After the years of economic prosperity in Brazil in the early 2000s, with the success of the "real plan" and the reductions in inflation and unemployment and increases in healthcare investments, 1,2 Brazil experienced two economic crises. The first was the world crisis of [2008][2009] which, although not large in Brazil, was responsible for a 0.1% deflation in the gross domestic product (GDP) in 2009. This was the first deflation in Brazil since 1993. 1 After this, Brazil returned to a period of growth in GDP, with social, educational and public health achievements up to 2015, when the greatest recession in Brazilian history began. The 2016 crisis led to successive years with GDP deflation of more than 3%, 1 along with an important political crisis that included a presidential impeachment.
During these economic and political crises, the percentage of GDP expended on the healthcare system started to reduce, through the impact of several austerity policies. 2 With the economic crises, the levels of unemployment increased. Higher unemployment has the consequences of affecting economic power, social security and job stability (through labor-law deregulation).
This may have substantial associations with mental health. 3,4 OBJECTIVE Our aim was to analyze occurrences of hospitalizations due to affective disorders in Brazil between 2003 and 2017, along with the economic oscillations and the prevalences of different types of health-related behavior.

METHODS
This ecological study used data on hospitalizations due to affective disorders (International Classification of Diseases 10 th edition, ICD 10; codes F30-F39). These data were collected from the morbidity and mortality surveillance system (DATASUS), which records all hospitalizations within the Brazilian National Health System (Sistema Único de Saúde, SUS). DATASUS has national coverage, with registration of approximately 11 million hospital admissions per year. 5 Brazilian data on GDP and unemployment were collected from the World Bank database covering the years 2003-2017. 1 Data on health-related behavior (television-viewing and physical activity) were obtained from the telephone survey-based surveillance system for risk and protective factors for chronic diseases (VIGITEL). VIGITEL is a survey that has been conducted annu-  6 Data on hospitalizations, economic oscillations and behaviors were cross-referenced to analyze possible interrelationships.
Our statistical approach was to provide crude values for the numbers of hospitalizations due to affective disorders, prevalence of physical activity, prevalence of elevated time spent sitting down, GDP and unemployment rate, in order to obtain an ecological perspective. Moreover, we created an indicator for hospitalization ratio, which comprised the number of hospitalizations due to affective disorders per 1000 all-cause hospitalizations.

RESULTS
Descriptive statistics regarding the prevalence of hospitalizations due to affective disorders, economic oscillations and healthrisk behaviors are presented in Figure 1 and Table 1    Worldwide, evidence of several negative changes to health outcomes during crisis periods has been found, especially when these have been followed by austerity policies within healthcare systems. Studies on the most recent crisis in European countries found that increases in communicable and non-communicable diseases occurred. 3 One of the main examples comes from Greece, which suffered severe austerity policies in 2010.
This was the country that suffered the greatest number of public health consequences. 4 Although the greatest Brazilian crisis is still recent, the negative impact on public health can already be observed. In times with increased unemployment, labor-law deregulation and social insecurity, mental health seems be affected in different ways. 3,4 One possible explanation for our findings could be the migration of people from private healthcare to the public healthcare system. 2 However, we found similar results when taking into consideration the hospitalization ratio.
Specifically, regarding the impact of economic crises on mental health, previous studies found that economic crises were associated with increases in the prevalences of several negative mental health outcomes, especially stress, anxiety and depressive symptoms. These were correlated with recent economic crises in Italy and Spain. 8,9 Moreover, employment status seems to be an important mediator of the negative impact of economic crises on mental health.
Going beyond unemployment itself, a previous study investigating the impact of an economic crisis in the Netherlands found that working conditions became worse during the crisis and that work insecurity increased. There was also an association with negative health outcomes. 10 Economic crises can also affect lifestyle behaviors such as physical activity levels, television-viewing and diet. 11,12 Consequently, adoption of unhealthy lifestyles can also be correlated with negative mental health outcomes. 7,13,14 Given the ecological approach adopted, we do not intend to establish causality between economic crises, health-related behaviors and hospitalizations due to affective disorders. Naturally, many factors could explain changes in the parameters analyzed.
Nevertheless, these data show that a tendency towards changes in the same direction occurs, which may indicate co-occurrence.
These findings at least provide empirical evidence regarding the changes at population level that follow periods of economic recess.
They also highlight the need for further studies to clarify these interrelationships.

CONCLUSION
These findings confirm the potentially harmful effects of an economic crisis on public health. Similar to the experience of other countries, Brazil will probably feel the negative impacts of the last recession period for some time yet and will take time to recover.
Most importantly, this is a period for reflection and decision-