1. Introduction
Depression is currently the most prevalent mental disorder [
1] and is characterized by profound sadness and loss of interest as well as a wide range of emotional, cognitive, physical and behavioral symptoms [
2].
Depression is a major contributor to the global burden of disease [
3], causing a substantial burden on health systems [
4]. At its most extreme, it can lead to suicide [
3]. Public health policies focus their efforts on preventing this disorder, which affects one in ten people in the world [
1], 6.4% of Europeans [
5] and 6.68% of Spanish [
6]. There is a need to scale up strategies on early interventions [
7], because treatments with antidepressants are not only costly but also unsuccessful [
8]. This is done through health promotion campaigns, advice on good practices, etc. [
1]. Health promotion through physical activity and exercise programs for adults is an intervention with a proven beneficial effect on reducing the risk of depression [
1].
It is universally accepted that regular physical activity is one of the priorities in Public Health as a means of preventing chronic diseases [
9] and the decrease in premature mortality [
10]. By physical activity we mean any bodily movement that is produced by skeletal muscles that produces an energy expenditure that is greater than that at rest [
11]. Regular physical activity improves both physical and mental health [
12]. Closely linked to physical activity and Public Health is physical exercise, which is defined as planned, structured, and repeated physical activity that is aimed at acquiring, maintaining, or improving physical fitness [
11]. In relation to PA and PE we can find strength work, which is physical ability to perform work or a movement, being one of the essential performance factors [
13]. Muscle strengthening is closely related to health [
14].
When it comes to depression, PA is an effective preventive tool against depressive disorders [
15] and has a positive effect on physical and mental health due to the release of endorphins, reducing anxiety, depression, and stress [
12].
There is a large body of evidence on the benefits of PA and PE for people with depression, in prevention [
16,
17] as a protective factor against depression [
18], as a preventive strategy [
19], effective in primary prevention [
20], and treatment [
21,
22,
23,
24,
25], and as an adjuvant treatment [
19,
26,
27] from severe to moderate depression with an exercise program [
28], where strengthening exercises [
28] could reduce the intake of antidepressants [
16,
17]. There were two meta-analyses of structured exercise programs that were recently published that conclude that heart rate and muscle strength improvement (based on resistance exercise) are moderators of depression improvement in middle-aged and older adults [
29,
30]. However, there is a lack of documentation on the recommended frequency of physical activity depending on the type of exercise that is performed (aerobic and/or strength) to prevent or reduce the prevalence of depression and to establish strategies for different age groups.
The aim of this study is, therefore, to establish relationships between physical activity and the prevalence of depression and antidepressant intake in the Spanish population, and to be able to establish a recommended dose of physical exercise to prevent and reduce depression.
2. Materials and Methods
2.1. Study Design
In this cross-sectional investigation, a descriptive correlational study was performed that was based on data that were obtained from the public files included from the European Health Interview Survey of Spain 2020 (EESE 2020). The EESE is a survey that is conducted every five years by the Ministry of Health and the Spanish National Institute of Statistics (INE), forming part of the European Health Interview Survey (EHIS). This survey is coordinated by Eurostat and is regulated by Regulation (EC) 1338/2008 and Commission Regulation 141/2013. The interviews were conducted between 15 July 2019 to 24 July 2020 by previously trained and accredited interviewers.
2.2. Participants
The initial sample constituted of 22,072 participants, aged between 15 and 104, residing in Spain, who were interviewed on the occasion of the EESE 2020, individual adult questionnaire. For this study, data from people aged between 18 and 85 years were taken into account, so that those over 85 years of age (1282 participants) and those under 18 (503 participants) were excluded. The final sample was 20,287 persons: 9731 men and 10,556 women.
The participants were selected using a three-stage sampling system with stratification: census sections (first stage units, municipalities were grouped into seven strata, according to number of inhabitants, selecting the sections in each stratum, taking into account the probability proportional to its size, according to the main family dwellings belonging to it), main family dwellings (second stage units, selecting living with equal probability, through systematic sampling with random start), and surveyable persons (third stage units, one person per dwelling, using the Kish random method. The selection is random, with equal probability for each adult in the dwelling).
2.3. Ethics
Not required, since the data were obtained from anonymous, non-confidential public files.
2.4. Variables and Procedures
The variables that were used from EESE 2020 and the procedures that were followed with these variables are shown in the following table (
Table 1).
2.5. Statistical Analysis
The data analysis was carried out with the IBM SPSS Statistics v.25 statistical analysis software.
Initially, the distribution that was followed by the data of the study variables was checked by performing a Kolgomorov–Smirnov test. When insufficient evidence was found to assume the normality of the variable distributions, it was accepted that the data did not follow a normal distribution for subsequent analyses. For this reason, in the descriptive analysis to characterize the sample, the median, the interquartile range (continuous variables), and the absolute and relative frequencies (ordinal variables) were used to present the data of the different groupings that were formed. Similarly, for this reason, nonparametric statistical tests were used to evaluate the possible intergroup differences using the Mann–Whitney U test (for the continuous variable: age). The chi-square statistic (to analyze possible dependency relationships), and a pairwise z-test for independent proportions, using the Bonferroni correction when it was necessary, (to analyze possible differences between proportions presented by sexes) were used for the categorical variables: age group, lifetime depression, 12-month depression, diagnosed depression, frequency of PA, days of PA, and days of muscle strengthening. The effect size was presented with Cramer’s V o phi as necessary.
The level of significance that was established in this investigation was less than 0.05.
3. Results
The median age of the population that was studied was 54 years, being lower in men (53) than in women (55), with significant differences between the two (
p < 0.001). Similarly, dependency relationships were found between the age group and sex (
x2 = 57.97,
p < 0.001,
V = 0.053), with women presenting a higher proportion of the population than men in the elderly group (
Table 2).
The prevalence of depression that was suffered at some time in life was 9.5% in the general population. Dependence relationships were found between the prevalence of depression and sex (
x2 = 234.80,
p < 0.001,
Φ = 0.108), with women (12.5%) having a higher prevalence than men (6.2%), with significant differences between them (
p < 0.05). Similar results were found in the prevalence of depression in the last 12 months (7.0%) and diagnosed by a physician (8.6%) in the general population, with dependence relationships between the prevalence and the sex of the participants (
x2 = 187.90,
p < 0.001,
Φ = 0.096 and
x2 = 234.28,
p < 0.001,
Φ = 0.108, respectively). In both, significant differences were found between the sexes (
p < 0.05), with 6 and 4.6 percentage points of difference between the men and the women, with higher prevalence in women, both in depression at 12 months, as well as in diagnosed depression. These dependency relationships were also found between the prevalence of antidepressant use and sex (
x2 = 206.74,
p < 0.001,
Φ = 0.101). The prevalence of taking antidepressants in the general population was 5.5%, being more than twice as high in women (7.7%) than in men (3.1%), with statistically significant differences in prevalence (
p < 0.05) (
Table 2).
A total of 74.6% of the general population reported not doing PA on a regular frequency: never (35.4%) and occasionally (39.2%). Dependency relationships were found between the frequency of PA and sex (
x2 = 144.17,
p < 0.001,
V = 0.084), the proportion of inactive women (38.3%) was 5.2 percentage points higher than that of men (33.1%), and the difference between the proportions of people with a frequency of several times a month or more between men (28.8%) and women (22.2%) was 6.6 points; all of these differences were significant (
p < 0.05). Dependency relationships were also found between the days of PA per week and sex (
x2 = 82.96,
p < 0.001,
V = 0.064). A total of 51.6% of the population reported performing PA zero days a week, this proportion being higher in women (54.7%) than in men (48.8%), with significant differences between the two (
p < 0.05). Something that was also found in the proportions of people with 5+ days of PA, being 18.9% in the general population, and finding a difference of 4 percentage points between men (21%) and women (17%). Much higher were the proportions of people who reported not performing muscle-strengthening exercises on any day of the week. In this sense, 81.8% of the general population performed these activities on zero days, and there was a relationship of dependence between these activities and sex (
x2 = 118.97,
p < 0.001,
V = 0.077). In women, this prevalence reached 83.9% compared to 79.6% in men, with
p < 0.05 (
Table 2).
Dependency relationships were found between the lifetime prevalence of depression and the frequency of PA, both in the general population (
x2 = 235.15,
p < 0.001,
V = 0.108) and in both sexes (Men:
x2 = 76.73,
p < 0.001,
V = 0.089; Women:
x2 = 126.85,
p < 0.001,
V = 0.110). The prevalence of lifetime depression was 13.4% in the inactive general population, being 4.8 percentage points lower in the population that performed PA occasionally (8.6%) and about 8 points lower in the more active groups (5.2% in the group that performed PA several times a month and 5.5% in the group that performed PA several times a week), finding significant differences between the groups (
p < 0.05). The same occurred in men and women, with significant differences between the prevalence of inactive and occasional people, and between these and the higher levels (
p < 0.05). In women, the highest prevalence of lifetime depression was found in inactive women (16.7%), with women who performed PA several times a month having the lowest prevalence (7.1%). This was similar for men, with the lowest prevalence found among men who performed PA several times a month (3.6%). These dependency relationships, both in the general population and in both sexes, were found in the prevalence of depression in the last 12 months and of diagnosed depression. Thus, significant differences were also found between the prevalence of depression among the different PA groups, being higher in the inactive than in the occasional, and in both groups and higher PA levels, with
p < 0.05. On the other hand, the prevalence of taking antidepressants showed dependence relationships with the frequency of PA, both in the general population (
x2 = 198.95,
p < 0.001,
V = 0.099) and in both sexes (Men:
x2 = 72.52,
p < 0.001,
V = 0.086; Women:
x2 = 102.33,
p < 0.001,
V = 0.099). The prevalence in the inactive population was more than three times higher than the prevalence in the group with the highest frequency of PA (8.2% vs. 2.5), with significant differences in proportions between these groups (
p < 0.05). This finding in the general population also occurred in both men and women. Men who performed PA several times a month had the lowest prevalence of taking antidepressants (1.3%), while the highest prevalence was found in inactive men (5.1%). On the other hand, inactive women had a prevalence that was 6.9 points higher than women with the highest frequency of PA (10.7% vs. 3.8) (
Table 3).
The prevalence of lifetime, past 12 months, and physician-diagnosed depression also presented dependence relationships with the days of PA that were performed per week, both in the general population and in both sexes. The highest prevalence in the three depression conditions were in the groups that performed zero days of PA per week compared to the groups that performed PA, at least one day, with
p < 0.05; this was found in the general population, in men and in women. The lowest prevalence was found in the groups that performed PA between three and four days a week, with significant differences, in most cases, with respect to the prevalence of the other groups (
p < 0.05). The highest prevalence was found in the inactive women: lifetime depression (15.7%), 12-month depression (12.1%), and diagnosed depression (14.5%); with differences in proportions between 6–7 points with respect to the women who performed PA three to four days a week (lifetime: 7.8%; 12-months: 5.3%; diagnosed: 7.1%) who presented the lowest prevalence. With smaller differences in proportions, the same was found in men, who presented differences of around 4–5 points between inactive and men with PA three to four days a week. These dependency relationships and these differences between the PA groups were also found in the prevalence of taking antidepressants, with prevalence that doubled or even tripled between inactive persons and those who took PA three to four days a week (
Table 4).
The existence of dependency relationships between the prevalence of the three conditions of depression that were analyzed in this study and the number of days per week that muscle-strengthening activities were also found, both in the general population and in both sexes. In men, significant differences were found between the prevalence of depression of those who did not perform, at least one day a week, muscle strengthening exercises, compared to the rest of the groups, although no differences were found between the rest of the groups among themselves (
p < 0.05). In women, the lowest prevalence of depression in the three conditions was found in those who performed these activities one to two days a week. Finally, dependency relationships were also found between the prevalence of taking antidepressants and the days per week that muscle strengthening exercises were performed, finding significant differences between those who did not perform, at least one day, and the rest (
p < 0.05) in the general population (
Table 5).
In each age group, the prevalence of depression that was suffered at any time in life was found to be related to the frequency of PA. In the young general population, the prevalence of depression reached 3.2%. In inactive young people, the prevalence was found to be 2.1 percentage points higher than the average (5.3%), while the prevalence decreased to 1.3% in those who performed PA several times a month. In the elderly, the prevalence in inactive persons was 19.8%, 10.6 percentage points higher than in the elderly with PA several times a month (9.2%). In any age group, significant differences were found in the prevalence of depression at any time of life between inactive persons and the rest of the PA groups, although not between them (
p < 0.05). The same relationships and differences were found in the prevalence of diagnosed depression. On the other hand, although the previous dependency relationships were again found in the prevalence of depression in the last 12 months in all age groups, in adults and the elderly, differences were found between performing PA occasionally and performing PA several times a month, or more, with lower prevalence in the latter (
p < 0.05). In the elderly, the prevalence of depression in the last 12 months was tripled in inactive persons (15.7%) compared to those who performed PA several times a month (4.7%), with
p < 0.05. Regarding the prevalence of taking antidepressants, dependence relationships were found with the frequency of PA in all age groups. In adults and the elderly, significant differences were found between the prevalence of inactive people and the rest of the groups (
p < 0.05) (
Table 6).
The results were not very different when analyzing the relationships between the prevalence of depression in each of the conditions and the PA groups, according to the days per week. Dependency relationships were found in all of them and in all age groups. A lower prevalence was found in all groups with one PA day per week, or more, compared to the inactive group, with
p < 0.05. In young people, between the inactive (5.3%) and the groups with the highest number of PA days per week reached 3.6 percentage points, these differences reached 8 percentage points in the elderly (0 days/week: 17.2% vs. 5+days/week: 9.2%) in the prevalence of depression that was suffered throughout life, with similar differences in the prevalence of depression in the last 12 months and in diagnosed depression. The prevalence of taking antidepressants was also related to the days of PA per week in all age groups. Although significant differences were found between then inactive and the rest of the groups, no differences were found in the latter among themselves (
p < 0.05). However, in young people and the elderly, no differences were found between the inactive people and those who performed PA 1–2 days a week, or between these and the rest of the groups (
p < 0.05) (
Table 7).
Finally, dependency relationships were found between the prevalence of depression in the three conditions that were analyzed and the days per week of muscle strengthening activities in all the age groups. In all the age groups, significant differences were found between the prevalence of the inactive groups and those who performed strengthening activities one to two times per week (
p < 0.05), being lower in the latter. However, the same did not occur with the prevalence of all the groups with a greater number of days per week, finding increased prevalence with respect to those of the one to two days per week group. Similarly, dependence relationships were found between the prevalence of taking antidepressants and the number of days of strengthening activities in all the age groups, although significant differences were only found between some groups, with the highest prevalence in the group of young people in the group of five days or more per week, although without significant differences with the inactive group. In the rest of the age groups, although with lower prevalence, the differences between the two groups were not significant either, although they were significant with other groups (
Table 8).
4. Discussion
The first finding of our study was in relation to sex. Significant differences were found (
p < 0.05), as noted in most studies, where being female is a risk factor that is associated with depression [
31,
32]. Likewise, dependency relationships were found between age group and sex, with women presenting a higher proportion of the population with depression than men in the elderly group; this is consistent with the majority of studies indicating that depression is higher in older women [
33,
34,
35].
Regarding age, the median age of the population that was studied was 54 years, being lower in men (53) than in women (55), with significant differences between the two (
p < 0.001). It may be related to the onset of menopause in women, as indicated by other studies [
36,
37].
The prevalence of depression that was suffered at some point in life was 9.5% in the general population, the same as the ENSE 2017 survey, whose figure stands at 9.2%. Dependent relationships were found between the prevalence of depression and sex, with women (12.5%) presenting a prevalence twice that of men (6.2%), with significant differences between the two (
p < 0.05). Similar results were found in the prevalence of depression in the last 12 months and diagnosed by a physician (8.6%) in the general population, as noted in most studies, where being female is a risk factor that is associated with depression [
31,
32].
Also, dependence relationships were found between the prevalence of taking antidepressants and sex, being more than twice as high in women (7.7%) than in men (3.1%), with statistically significant differences between both (
p < 0.05). This is in line with other studies [
38,
39] where the prevalence of antidepressant use is twice as high in women as in men.
Similarly, dependency relationships were found between PA frequency and sex. The rate of inactive women is higher than that of men, as reported in other studies [
40,
41]. In terms of strength work, women perform less strengthening exercises with
p < 0.05 [
42].
Another important finding is that dependence relationships were found between the prevalence of depression (lifetime, in the last 12 months and diagnosed) and the frequency of PA, both in the general population and in both sexes. The higher the frequency of physical activity, the lower the prevalence of depression. There are several studies and systematic reviews supporting this finding [
43,
44,
45]. The prevalence of taking antidepressants showed dependence relationships with the frequency of PA; in this line there is another study where the probability of taking antidepressants increases with inactivity [
46].
The existence of dependency relationships between the prevalence of the three depression conditions that were analyzed in this research and the number of days per week that muscle strengthening activities were performed were also found, both in the general population and in both sexes; coinciding with the systematic review on physical exercise and depression [
47,
48]. In all the age groups, significant differences were found between the prevalence of the inactive groups and those who performed strengthening activities one to two times per week (
p < 0.05), being lower in the latter. However, the same did not occur with a higher number of days per week, with increased prevalence.
If we know the recommended frequency of physical activity and the type of exercise to prevent or reduce the prevalence of depression in different age groups, our findings could be a reference for monitoring the prevalence of current depressive disorders, planning health resources and services, and developing screening and preventive strategies in different age groups at the national level.
4.1. Theoretical and Practical Implications
Health education and promotion campaigns among the Spanish population could help to reduce the prevalence of depression in all the age groups. Increasing the frequency of PA in inactive people or those with low levels of PA to three to four days/week of moderate physical activity, including one to two days of strength work, could reduce the prevalence of depression and antidepressant use in the Spanish population.
Possible initiatives that governmental policies could carry out could be: from including active breaks in high schools and universities for the youth group; physical activity programs in companies for adults, to active aging programs for the elderly.
4.2. Limitations
This article has some limitations to take into account: it is cross-sectional in nature. The data were obtained through the information that was submitted by the participants. Cause-effect relationships cannot be established due to the methodology that was used. Only male and female sex is considered; non-binary sex is not taken into account. Physical activity measures were not objectively assessed. Only whether or not participants were taking antidepressants was recorded; neither the active ingredient nor the amount taken by each participant who reported taking antidepressants was recorded. Other variables that could affect depression, such as sociodemographic, socioeconomic, and the sociocultural biases of the participants, were also not included, because although there are studies in which these variables seem to affect depression [
49,
50], this greatly reduced the sample by including new divisions in the groups, losing statistical power.