The 2008 Financial Crisis and Changes in Lifestyle-Related Behaviors in Italy, Greece, Spain, and Portugal: A Systematic Review

Italy, Greece, Spain, and Portugal have all been strongly affected by the 2008 financial crisis, which has had a negative impact on health. We systematically evaluated the effects of the crisis on lifestyle and socioeconomic inequalities. We conducted a literature search using MEDLINE, Embase, the Cochrane Library, and health economics databases for studies reporting quantitative comparisons before and after (or during) the crisis on the following risk behaviors: alcohol consumption, smoking habit, healthy diet, physical activity, and psychotropic drugs and substance abuse, without setting any age restrictions. We selected 34 original articles published between 2011 and 2020. During/after the crisis, alcohol consumption and substance abuse decreased, while psychotropic drug use increased. We also observed a deterioration in healthy eating behavior, with a reduction in fruit and vegetable consumption. Smoking habit and physical activity showed a more complex, controversial trend. Socioeconomic inequalities were affected by the recession, and the negative effects on unhealthy lifestyle tended to be more pronounced among the disadvantaged. These results suggest the need to implement health policies and interventions aimed at monitoring risk behaviors, with special regard to disadvantaged people, and considering the potential additional impact of the COVID-19 pandemic.


Introduction
Before the COVID-19 pandemic, the 2008 Great Recession was the most severe crisis experienced by Europe since the Second World War, particularly in the southern European countries of Italy, Greece, Spain, and Portugal [1]. The considerable amount of literature on the topic shows that the financial crisis has had a strong impact on the health of most European populations [2]. Despite the trend towards reduced mortality, a deterioration in mental health, an increase in the number of suicides and, to a varying extent, in some non-communicable and communicable diseases, and a worsening in perceived health has been observed in most European populations [3,4]. However, an increase in mortality due to alcohol-related causes and to the consumption of drugs [5] has been observed in some countries.
Italy, Greece, Spain, and Portugal have been affected both by the direct effects of the financial crisis on the health of their populations and by the barriers to healthcare access imposed by the austerity policies introduced by governments to pay off the public debt [2]. In Greece, the austerity measures implemented to contain public spending contributed to an increase in forgoing health care due to economic reasons, especially among the poor, people with lower incomes, and the unemployed [6]; similar results were observed in Italy [7]. A deeper analysis by Karanikolos [2] suggests that, although recession poses risks to health, the interaction between fiscal austerity with economic shocks and weak social protection is what ultimately seems to escalate health and social crises in Europe.
The impact of the financial crisis on the health of the populations has been disproportionate. In fact, the more limited decrease in the number of deaths among the more disadvantaged social groups compared to the general population has determined a widening of mortality inequalities [8]. A recent systematic review has found an increase in socioeconomic inequalities [9]. Historical evidence supports the hypothesis that recession periods are associated with worse lifestyle, including increased alcohol consumption [10] or drug use [11], but the association between the Great Recession and worse lifestyle has not yet been systematically evaluated.
In this context, the COVID-19 pandemic began while the effects of the 2008 financial crisis were still manifesting; in addition to the dramatic impact on mortality and the direct long-term effects on the health of those who have recovered, there have also been indirect effects due to the cancellation or postponement of non-urgent assistance or interventions to decongest overwhelmed care facilities, technologies, and personnel.
The purpose of our review was to systematically evaluate the effects of the 2008 financial crisis on lifestyle and socioeconomic inequalities in Italy, Greece, Spain, and Portugal.

Materials and Methods
The review protocol was registered (CRD42019129105) in the PROSPERO open access database of systematic reviews (Available online: http://www.crd.york.ac.uk/PROSPERO) (accessed on 12 August 2021). Conducting and reporting are in accordance with PRISMA guidelines. The Covidence systematic review software (Veritas Health Innovation, Melbourne, VIC, Australia. Available online: http://www.covidence.org) (accessed on 12 August 2021) was used as the reference manager tool for the phases of importation, deduplication, and selection.

Information Sources and Search Strategy
The literature search was performed using MEDLINE (via Ovid), Embase, The Cochrane Library and health economics databases (EconLit). Search terms for financial crisis were combined with the terms "eating behavior", "smoking habit", "alcohol consumption", "psychotropic drug use", "drug abuse", or "gambling". Supplementary  Table S1 shows the full search strategy for MEDLINE. To identify additional relevant documents, the grey literature was searched for using OpenGrey and through the screening of the websites of the following referral organizations on population health and healthcare: The World Health Organization, the Organization for Economic Co-operation and Development, the European Observatory on Health Systems and Policies, the European Commission, and the European Centre for Disease Prevention and Control.
The references of included articles were also screened to identify potentially eligible articles for inclusion.
We included observational studies reporting quantitative comparisons before and after, before and during, or during and after the crisis of the following key health behaviors: diet, smoking, physical activity, alcohol consumption, and psychotropic drug use or substance abuse affecting individuals of any age. We considered studies published between January 2008 and November 2020 in English, Italian, Spanish, Portuguese, and Catalan. Multicountry studies were included when individual country data were available.
The exclusion criteria were: 1. Type of study: publications lacking primary data and/or explicit descriptions of the methods. Abstracts, editorials, correspondence, and commentaries were deemed acceptable for inclusion if they reported sufficient data; 2.
Methodology: pre-post comparison missing;

3.
Data not suitable for extraction (e.g., reporting data on a group of countries); 4.
Study population overlap.
Studies were selected through a three-step selection procedure based on: (1) screening of title and abstract, (2) screening of full-text article, and (3) final screening during the data extraction phase. Two independent researchers with experience in reviews and in the topic filtered and selected the references. In cases of discordance, a third researcher was consulted to determine inclusion or exclusion of the reference.

Data Extraction, Quality Evaluation, and Synthesis of Results
The following data were extracted from each included article: study population (number, age range, sex), population characteristics (e.g., students, household members), data source (e.g., national registry, questionnaires), study design, outcome definition, results (if available, by sex and socioeconomic status), main conclusions.
Two independent researchers judged the quality of each eligible study using a modified Newcastle-Ottawa Scale (NOS) for cross-sectional studies [12]: a study is assigned a maximum total score (stars) of nine for the following domains: selection, comparability, and outcome. We considered scores of 0-3, 4-6, and 7-9 as indications of low, medium, and high quality, respectively. Any doubt was resolved by consulting with a third reviewer. The score of the included articles are shown in the Supplementary Table S2. The results, organized by country, report study population (number and age range) and data sources, outcome definition, results, and effects on inequalities, if reported. Given the highly heterogeneous nature of the studies, we did not attempt to conduct a metaanalysis, and we report the results narratively.

Results
Our search identified 2325 unduplicated records; after the selection process, 31 crosssectional studies  were identified, as detailed in the PRISMA flow diagram (Figure 1). The main characteristics of the eligible studies are reported in Table 1. The articles, published between 2011 and 2020, were conducted primarily in Spain (n = 19); the remainder were conducted in Italy (n = 6), Greece (n = 6), and Portugal (n = 4). Two studies were multi-country [19,34]. Of the studies included, six also considered a juvenile population (age ≤ 15 years) [21,29,33,34,41,43], while the remaining were on adolescents aged > 15 years and on adults. The studies used validated questionnaires or administrative registries as data sources.
Regarding the outcomes considered, most of the studies concerned diet (n = 18) and smoking habit (n = 17), followed by alcohol consumption (n = 15) and physical activity (n = 13). Antidepressant/anxiolytic/antipsychotic drug use was examined by nine studies and substance abuse by six. In order to report briefly the main findings of the studies from a public health perspective, we created Table 2, which shows the variation in health behavior during or after the 2008 crisis by means a symbol for each risk behavior.
Regarding study quality, the available evidence was affected by a high risk of bias for exposure and for outcome assessment due to the study design, use of self-reported measures, and lack of adjusting for potential confounding factors. Supplementary Table S2 summarizes the NOS assessment of the included studies, which obtained scores between 2 and 7, with an average score of 4.7; four studies appeared to be of very low quality (total score 2 or 3). Confidence in ascertainment of exposure and of outcome assessment was very low in most studies. Even though most studies used a large sample of individual-level data, most data were collected by surveys that collected information using questionnaires on self-reported changes in several indicators on health-related behaviors as well as on consumption of medications. Some studies used aggregate data, which could mask individual-level effects, or self-reported questionnaires that had not been previously validated. Only six studies received a total score of >7, which was considered high quality.

The Financial Crisis and Alcohol Consumption
Compared to the pre-crisis period, alcohol consumption decreased during or after among adults in Spain, Greece, and Italy [16,19,20,[25][26][27][28][29]35], and among young Spanish people [13,43]. Two studies showed a significant upward overall trend in binge drinking [20]. Conversely, two studies provided evidence of increased prevalence of moderate [18] and heavy alcohol intake after the recession [16], while a Spanish study showed no differences [17].
When stratifying by socioeconomic level, controversial results were observed. Two studies suggested a widening of socioeconomic inequalities, although this effect is due to contrasting results: one study showed that heavy alcohol consumption increased during the crisis among the least educated men [16], while another showed an increase among the most educated people [17]. Finally, a more relevant reduction in alcohol intake among the lower class was also observed elsewhere, showing a narrowing of inequalities [25].

Time-trend analysis
Greek  +4.6%) *: for stratified analysis, we report results of association when at least one strata is statistically significant.

The Financial Crisis and Smoking
Similarly to alcohol consumption, smoking habit also appeared to decrease markedly in most studies conducted in different settings or subgroups: young Spanish men aged 16-24 [13], adolescents in Portugal [21] and Spain [43], and adults in Greece [23,24] and Spain (only men) [25]. A reduction in smoking prevalence was observed in Spain also when stratifying by the size of the municipality [30], in a sample of patients hospitalized for cardiac catheterization [36], and in a sample of Spanish mothers [40]. A few studies reported a slight decrease [27,32], no difference [37], or a slight increase in the percentage of smokers after or during the crisis, for only women [16] in Spain and for both sexes in Italy [29,34], Portugal, and Spain [34].
Regarding socioeconomic status, controversial associations were reported during or after the crisis: a reduction in inequalities was identified by some studies, where percentages of daily or occasional smokers decreased more for those with a lower educational level [16], lower socioeconomic status [23], or having a manual occupation [40]. Conversely, other studies found an increase in socioeconomic inequalities due to an increase in smoking consumption among the unemployed [13,28,37,43] and among people belonging to a lower social class [25].
Several studies reported an increase in socioeconomic inequalities in healthy diet, especially regarding fruits and vegetables. In Spain, the probability of declaring eating fruit daily decreased more among unemployed men and least educated men, and the probability of declaring eating vegetables daily among unemployed men and women and the least educated women [16]. Two other Spanish studies reported an increase in inequalities in the consumption of fruits and vegetables [22,25].
Similar results were observed regarding fish consumption [22]. During the crisis in Italy, the socioeconomic differences in adherence to the Mediterranean diet widened, becoming less probable among people with a low wealth index score, those with a lower education level, and those performing manual labor [18]. One study conducted in Spain on young people below the age of 15 reported that the prevalence of junk food consumption increased in families with low maternal education level [33].

The Financial Crisis and Physical Activity
Most of the studies highlighted an increase in physical activity in the adult population when comparing the periods after vs before the financial crisis, which was more robust and statistically significant in Spain [25,35] and Greece [23,24], and slighter in Portugal [21], while a limited reduction was observed in the few studies in adults [37] and in children [33,41].
Two Spanish studies reported an increase in socioeconomic inequalities in physical activity during and after the crisis. In one, an increase in the prevalence of physical activity was observed in all social classes, but this was slighter in the lower class, resulting in an increase in socioeconomic differences [25]. The other study observed that physical activity increased during or after the crisis among more educated women and decreased among the less educated, causing a widening of socioeconomic inequalities [16].
Instead, a Greek study showed a reduction in inequalities due to a significant increase in percentages of adults only, with those with a middle or lower socioeconomic status reporting high or moderate level of physical activity [24].

The Financial Crisis and Use of Antidepressant, Anxiolytic, and/or Antipsychotic Drugs
Most studies reported an increase in any type of psychotropic drug use among all individuals considered [31,42] or only among women [16]. In a Spanish study performed on the economically active population, heavy use of hypnotics/sedatives among men and women increased in the period examined [20]. In a population study conducted in Portugal, the odds of consuming any psychotropic drug was estimated to be 1.5 times higher than before the crisis, and when evaluating the interaction effect of the year with sex and age, men and younger individuals reported higher odds of consuming any psychotropic drug [38]. Bartoll et al. [16] observed a stable trend in tranquilizer tablet use among men and a decrease among women.
This general increasing trend was not homogeneously reported for all psychotropic drugs: specifically, Marquez Calderon et al. [27] found an increase in sedative, tranquilizer, and hypnotic drugs, while antidepressant use decreased. According to Arroyo [15], only sedative use increased, and Madianos et al. evaluated only antidepressant use, finding an increase in its use in Greece [26].
When considering socioeconomic status, according to Arroyo et al. [15], the probability of consuming antidepressants or sedatives depended on employment status: in the case of individuals in short-term unemployment, both men and women showed that between 2006-2007 and 2011-2012, there was an increase in the risk of using sedatives. However, this increase was greater for women than for men. For the long-term unemployed, however, the differences between both sexes widened between 2006-2007 and 2011-2012: the risk of using sedatives in women increased, whereas it decreased in men, while antidepressant consumption decreased overall, and more markedly among short-and long-term unemployed subjects. Regarding education level, there was an increase in the intake of tranquilizers among those without any qualification and a slight decrease for men with high secondary education; among women, there was a drop-in intake, which was greater among those employed and those without any qualification [16].
Finally, an increase in psychotropic drug consumption was observed among households whose socioeconomic status was most affected by the crisis [42].

The Financial Crisis and Substance Abuse
One of the included studies described a stable or a slight downward trend in drug use [28]. A study performed in Portugal on adolescents reported a slight decrease in monthly drug use [21]. A decrease in cocaine, marijuana, ecstasy, and hard drug use caused by the economic downturn, which could have affected the prices of these drugs, was observed by a Spanish study [28]. According to another Spanish study, overall cannabis use remained stable during the crisis, but unemployed men and women were more likely to have increased sporadic use compared to their employed counterparts [20]. Given a 10% increase in the provincial unemployment rate in Spain, an increase in the probability of using marijuana and cocaine in the last 30 days and also over the previous 12 months was observed [28].

Discussion
Since the financial crisis of 2008, southern European countries, in particular Italy, Greece, Spain, and Portugal, have been affected by the economic recession [44][45][46].
According to the findings of our extensive review, we observed controversial effects on healthy behaviors in the period characterized by the 2008 financial crisis.
A deterioration in proper eating habits was also observed [14,16,[21][22][23][24][25], but some comments regarding diet are necessary. In fact, although most of the studies report an overall worsening of eating behavior during the crisis, the evaluation of its impact is more complex than it is for the other lifestyle-related behaviors. A reduction in the consumption of fish and meat, probably due to the reduction in available income, was observed, as was a reduction in fruit and vegetable consumption; both of these phenomena could cause a shift toward a worse diet. The crisis effect could have also contributed to determine the decrease in consumption of fruits and vegetables and the increase in junk food, sweets, and dessert consumption. Social inequality in proper eating habits generally increased because of the economic crisis [16,18,22,25,33]: several studies observed a decrease in the consumption of fruits and vegetables and an increase in junk food consumption among people with a low socioeconomic status.
However, even before the crisis, a decline in adherence to the Mediterranean diet had been observed in the younger population, so that a reduction in the mean consumption of fruit and vegetables was expected. Therefore, this decline may not be solely attributed to the economic crisis [23].
Less income available to purchase alcohol might have been behind the decrease in heavy drinking, while binge drinking could have increased as a means to deal with anxiety and emotional distress related to job loss, whether real or threatened, and to financial hardship [20]. However, it has been argued that a process of change in the pattern of alcohol use is taking place in Mediterranean countries such as Spain, where alcohol has traditionally been embedded in daily life, with wine drunk regularly with meals [20].
The decrease in tobacco use during times of economic downturn is related to price increases; it is likely that tobacco control measures may have interacted synergistically with the decline in disposable income. Therefore, austerity may have been a driving force in the decline among low-income individuals, along with the other public health measures [23].
Most of the studies found a slight increase in physical activity during the crisis [16,21,[23][24][25]35], while other studies found a slight increase in sedentary habits [27,33,37,39,41]. There was an increase in socioeconomic differences in the prevalence of physical activity [16,25], with a heterogeneous gradient according to socioeconomic status.
The reduction in substance abuse observed during the crisis was slight, and generally referred to soft drugs [20,21,27,28], without any significant differences in terms of socioeconomic status.
Most studies reported an increase in the use of any psychotropic drug [20,26,27,31,38,42]. The decreased availability of income during the recession may have determined a reduction in the purchase of tobacco, alcohol, and drugs, but also of more expensive and healthier food. On the contrary, the increase in the use of psychotropic drugs can be considered a kind of coping mechanism against the insecurity and stress related to the economic crisis.
Our review appears to confirm previous evidence [47,48] that the financial crisis had an impact on socioeconomic inequalities and that negative effects on health tended to be more pronounced among the culturally, economically, and socially disadvantaged.
The low socioeconomic strata experienced inequalities in access to cultural and material resources (e.g., education, working conditions, income), which determined worse health and limited access to appropriate health care. These inequalities grew over the course of the global crisis, and the recession period could have accelerated the accumulation of such disadvantages [10,49,50].
The crisis itself may have played an independent, additional role, acting as a chronic stressor. Furthermore, it is possible that unemployed people and those at risk of unemployment or who experienced work instability may have had less time to dedicate to themselves and to their lifestyle, including food choice and physical activity [29].
In this scenario, the COVID-19 pandemic has exacerbated inequalities with a cumulative effect of the risks [51]. In England, as already shown in the Marmot Review, COVID-19 infection and mortality risks have been much higher for those living in more deprived areas, in overcrowded housing, in key workers in close proximity to others, in those from minority groups, in those with underlying health conditions, and in those who are older and/or male [52]. Furthermore, due to the effects of the mobility restrictions and the periods of lockdown as well as to the impoverishment of many sections of the population, it is reasonable to expect a further deterioration in lifestyles, especially among the people most affected by the economic crisis generated by the pandemic [53].
In fact, most countries were forced to introduce confinement measures to minimize the propagation of the SARS-CoV-2 virus, and for many people, it was difficult to maintain a healthy lifestyle, in particular a proper diet, regular physical exercise, quality of sleep, and limited smoking and alcohol consumption [53][54][55][56][57][58][59][60][61][62][63][64][65][66][67][68][69]. However, similar to the analysis reporting on the 2008 recession, the effects on diet are not unidirectional. In fact, many people used the period of home isolation to improve their eating habits and to limit dietary excesses and bad eating behaviors. Having the opportunity to devote time daily to having breakfast and to cooking meals resulted in an increase in the consumption of fruit, vegetables, and pulses [56,60,64,67,68], and a general decrease in alcohol consumption [56,66,67].

Strengths and Limitations
Our review of the impact of the 2008 financial crisis on many health behaviors aimed to shed light on the links between changes in habits and health outcomes, an issue that has not yet been systematically investigated [3]. The focus on four southern European countries represents an added value, as they were hit harder by the crisis than were other European countries.
By focusing exclusively on health outcomes, our study did not look at the impact of the crisis on health systems, such as shortages in the health workforce or in medical supplies, for which several studies have shown a negative trend during the financial crisis [3,70,71].
The exact moment the financial crisis began is difficult to establish, and some studies, reporting data on different countries, have defined the duration of the crisis differently; this may have had an impact on the homogeneity of the reported results. The included studies had a high risk of bias in exposure and outcome assessment due to the study design, use of self-reported measures, and the lack of adjusting for potential confounding factors. Further, some observational studies did not apply any statistical tests [21,22,26,27,32,34,37,39,43]. Most importantly, although the studies included in our review investigated changes in population health status and health behaviors associated with the Great Recession, it cannot be established whether this was a causal relationship.
Due to the nature of the data of most of the included studies, which were not designed to measure exposure at the individual level, no causal relationship between the economic recession and changes in lifestyles can be established. We can, however, state that these changes occurred after the crisis. In fact, although a causal association between the financial crisis and trends in risk factors seems reasonable, we cannot exclude unmeasured confounding, which would provide alternative explanations for the observed trends. On the other hand, the financial crisis is a natural experiment at the population level and the possibility of its effects on the findings may be supposed [23].
Finally, self-reported information collected by questionnaire, as was the case for most of the included studies, may have been affected by information bias.

Conclusions
Our results seem to show that the crisis has had a negative effect on eating habits and a positive effect on alcohol consumption and on smoking, the consequence of mechanisms probably determined by decreased available income. Psychophysical stress linked to unemployment and job loss, as well as the worsening mental health observed in nu-merous studies, could explain the increase in the consumption of antidepressant and anxiolytic drugs.
These results suggest the need to implement health policies aimed at monitoring risk behaviors and for interventions aimed at contrasting the effects of the financial crisis in the countries studied. Inequalities in health behaviors should also be a priority area for action. Financial recommitment to public health system should accompany a substantial commitment to tackling the social determinants of poor health and wellbeing.
The identification of lifestyles and socioeconomic inequalities produced by the 2008 crisis may facilitate the understanding and the response to the possible effects of the current COVID-19 crisis.
The current pandemic provides harsh lessons on the societal vulnerabilities that arise from inequality. Investing in young people and supporting long-deprived regions and sectors of society are arguably the most powerful ways to break the chain of inequality transmitted from generation to generation. Adopting a broadened, equity-focused approach to population health should be an essential part of building a more resilient society that is better prepared to weather future pandemics.