Income, Educational Level, and Depressive Symptoms in a Time of Multiple Crises

Trends revealed by high-frequency mental health surveillance in Germany, 2019–2024

Background: An increase in depressive symptoms among adults was observed in Germany from late 2020 onward, during a time of multiple collective stressors. In view of the uneven distribution across society of vulnerability to depressive disorders, we studied the varying impact of this trend on different socioeconomic groups.

Methods: Using population-based data from the German Health Update (Gesundheit in Deutschland aktuell) study of the Robert Koch Institute (n = 95 267, data collected from April 2019 to February 2024), we calculated time series analyses for the proportion of the population screening positive for possible depressive disorder (Patient Health Questionnaire 2 summed score of 3 or above) with the aid of moving 3-month estimators and smoothing curves, stratified by socioeconomic status (educational level and income). Absolute and relative inequalities were quantified with regression-based methods.

Results: The proportion of the population screening positive for possible depressive disorder was higher in groups with lower educational levels and lower incomes. Particularly from 2022 onward, it rose to a much larger extent in these groups than in those with higher educational levels and higher incomes. From 2019 to 2024, the absolute inequalities, i.e., the differences between the groups with lowest and highest educational level and between the groups with lowest and highest income, rose from 10 [7;14] to 22 [8;36] and from 12 [7;17] to 30 [17;44] percentage points, respectively (means and 95% confidence intervals).

Conclusion: Socioeconomic inequalities in depressive symptoms in adults rose over the period 2019–2024. Such inequalities in mental health should be systematically monitored, and measures should be developed to improve equity with respect to health.

Cite this as: Kersjes C, Junker S, Mauz E, Beese F, Walther L, Müters S, Schnitzer S, Hoebel J: Income, educational level, and depressive symptoms in a time of multiple crises: Trends revealed by high-frequency mental health surveillance in Germany, 2019–2024. Dtsch Arztebl Int 2025; 122: 573–8. DOI: 10.3238/arztebl.m2025.0130

Collective stressors, such as natural disasters, economic crises, epidemics/pandemics, and armed conflicts, can have an impact on the health of entire populations (1, 2). Over the past years, the population in Germany and many other countries experienced the consequences of collective stressors, including the COVID-19 pandemic, the Russia–Ukraine war and the Middle East conflict, as well as their economic fallout in the form of short-time work, unemployment, recessions, and record inflation (3, 4, 5).

In the context of such events, a number of European countries saw a deterioration in mental health at the population level, e.g., during and as the result of the COVID-19 pandemic (1) or associated with economic crises (1, 6, 7). Even before the emergence of the above mentioned stressors, mental health was poorer in individuals belonging to socially disadvantaged groups, such as persons with lower educational level and/or lower income (8, 9). In Germany, for example, the 2013 prevalence rates of depressive symptoms were 21%, 10% and 5% among adults with low, medium and high socioeconomic status, respectively (10). In-depth analyses reveal that socially disadvantaged persons are more likely to experience a deterioration in mental health in times of crisis (6, 7, 11). During the global financial crisis of 2007–2008, for example, persons in the general population who were hit by unemployment (1, 6, 7), faced loss of income/financial insecurity (6) or precarious working conditions and debts (7) suffered more frequently from depression, anxiety disorders and sleep disorders, up to and including suicide. Accordingly, inequalities of mental health widened between the various socioeconomic groups during this period (12). In addition, financial crises have a negative effect on the income distribution of a country (13), which, in turn, is associated with an increased risk of depression, in particular among persons with low income (14). During the COVID-19 pandemic, too, increases in mental health inequalities, disadvantaging persons with low socioeconomic status, were observed in a number of countries (11).

In Germany, high-frequency mental health surveillance at the Robert Koch Institute (RKI) also revealed a significant increase in depressive symptoms among adults during the period 2019–2024 (15, 16, 17). However, it is not yet known how the extent of inequality with regard to depressive symptoms between different socioeconomic groups developed in Germany over this period.

Thus, our aim was to carry out in-depth analyses of these nation-wide RKI surveillance data and (1) to analyze monthly time series for the proportion of the population screening positive for possible depressive disorder, stratified by educational level and income, as well as (2) to quantify the annual extent of socioeconomic inequalities in the distribution of depressive symptoms.

Methods

Underlying data

The analyses are based on data from the RKI’s German Health Update (Gesundheit in Deutschland aktuell, GEDA) study which, designed to be representative, was conducted from April 2019 to February 2024. Telephone surveys were carried out using the dual-frame method (random sampling by means of a combination of landline and mobile phone numbers), capturing approximately 1000 and 2000–4000 respondents per month during the periods 2019–2021 and 2022–2024, respectively (for details, see eMethods).

Variables

The outcome variable was the proportion of persons screening positive for possible depressive disorder as determined using the Patient Health Questionnaire-2 (PHQ-2) (18), an internationally used short screening instrument for depressive disorders which has been validated for population studies (19). The PHQ-2 captures the frequency of being affected by depressive symptoms (“not at all” (0) to “nearly every day” [3]) over the last two weeks based on two of the nine criteria for major depression listed in the DSM-5, i.e., the core symptoms of depressive mood and loss of interest. In line with the established screening cutoff, a summed score of 3 or above was considered to indicate a significant burden due to depressive symptoms (no clinical diagnosis) (18).

Information about educational level and income was used to establish the socioeconomic status. The formal level of education was determined based on questions about school and vocational qualifications and classified into the categories low (primary to lower secondary education), medium (intermediate/higher secondary education) and high (tertiary education) level of education according to the Comparative Analysis of Social Mobility in Industrial Nations (CASMIN) classification scheme (20).

Income was determined by computing the net equivalent income, which was calculated by dividing the monthly net household income by the needs-adjusted household size according to the modified Organization for Economic Cooperation and Development (OECD) scale (21). By taking into account the number and ages of household members as well as cost savings through shared household economy in multi-person households, this method allows for comparability of income across different households. Missing income values (25.3%) were imputed using a multiple regression model (22). Distribution-based income quintiles were formed to contrast low and high incomes:

• Low (quintile 1)
• Medium (quintile 2–4)
• High (quintile 5).

The comparison of the top quintile and the bottom quintile is commonly used to describe income inequalities (23).

Sex (male/female) and age (18–29, 30–44, 45–64, ≥ 65 years) were also included in the analysis. Table 1 and eTable 1 show the distribution of characteristics in the study population.

Table 1

The descriptive statistics of the study population

eTable 1

The descriptive statistics of the study population by year

Statistical analysis

The R Statistical Software (version 4.3.0; R Foundation for Statistical Computing, Vienna, Austria) was used for the analyses. The samples were weighted by region, age, education, and sex to adjust for design effects and different participation probabilities (24).

Estimation of moving 3-month averages

Using logistic regression models, moving three-month estimators (monthly proportion of persons with PHQ-2 ≥ 3 of the respective month and its preceding and following months) were computed for all educational level and income groups, standardized by age and sex, to illustrate the development of the proportion of the population screening positive for possible depressive disorder (24).

Estimation of inequality measures

Absolute and relative inequalities were analyzed to estimate educational and income inequalities in the distribution of depressive symptoms for each year. Absolute inequality is quantified using the difference in prevalence between the bottom and top position in the socioeconomic distribution, relative inequality using the prevalence ratio of these positions. Higher values indicate greater inequality (25).

In this study, we calculated regression-based inequality measures that took the distribution of health characteristics across the entire socioeconomic spectrum into account: the Slope Index of Inequality (SII) for absolute inequality and the Relative Index of Inequality (RII) for relative inequality. We used generalized linear models to model the presence of positive screening results for possible depressive disorder by ridit scores for educational level and income, adjusting for the covariables survey year, age, and sex, as well as their interactions (see eMethods for details).

Results

Across the entire observation period, the proportion of the population screening positive for possible depressive disorder was higher, the lower the level of education and income (eTable 2). In 2019, 13.3%, 10.7% and 5.6% of persons with low, medium and high levels of education, respectively, screened positive for possible depressive disorder as well as 16.0%, 10.3% and 6.0% of persons with low, medium and high income, respectively.

eTable 2

Annual proportions of the population screening positive for possible depressive disorder (PHQ-2 ≥ 3)—total, by sex as well as by level of education and income

Following an initial drop in the proportion of persons screening positive for possible depressive disorder in all education and income groups during the first months of the pandemic (Figure), a continuous increase was noted in all socioeconomic groups from autumn 2020 onward (eTable 2; Figure). In 2024, 29.3%, 21.9% and 11.2% of persons with low, medium and high levels of education, respectively, screened positive for possible depressive disorder as well as 32.9%, 22.1% and 8.4% of persons with low, medium and high income, respectively. Starting in 2022, a significant increase in prevalence rates was observed, particularly in the lower education and income groups. When applying the criterion of non-overlapping confidence intervals, there were no significant differences between men and women with regard to the educational level- and income-stratified proportions of the population screening positive for possible depressive disorder (eTable 2).

Figure

High-frequency time series of the proportion of the population screening positive for possible depressive disorder by characteristics of socioeconomic status, with 95% confidence intervals and smoothing curves

The absolute difference in the proportions of the population screening positive for possible depressive disorder between the highest and lowest levels of education (SII, Slope Index of Inequality) increased from 10 percentage points in 2019 to 22 percentage points in 2024 and between the highest and lowest income positions from 12 percentage points to 30 percentage points (Table 2). The absolute education and income inequalities initially remained stable in 2020 and 2021, but rose sharply from 2022 onward.

Table 2

Absolute & relative inequalities in the proportion of the population screening positive for possible depressive disorder by education & income for the years 2019–2024

With regard to relative inequalities (RII, Relative Index of Inequality), the risk of screening positive for possible depressive disorder was consistently 2 to 3 times higher among persons with a low level of education and 2 to 4 times higher for those with a low income compared to those with a high level of education or high income. The relative inequality initially decreased in 2021, but then started to increase in 2022. Although this observation was made with respect to both level of education and income, it only reached statistical significance for education.

In men, absolute educational and income inequalities initially decreased until 2021 and then increased from 2022 onward, while a steady increase was observed in women. A sex-stratified analysis did not find a significant trend in relative educational and income inequalities (eTable 3).

eTable 3

Annual absolute and relative inequalities in the proportion of the population screening positive for possible depressive disorder—total, by sex as well as by level of education and income

The adjusted logistic regression models with level of education and income showed a minor improvement compared to the null model (McFaddens R² of 0.04 in both cases).

Discussion

The internationally described social gradient in mental health disadvantaging adults with low socioeconomic status (8, 11) was also evident in our study across the entire observation period. In all educational and income groups, a drop in the proportion of the population screening positive for possible depressive disorder was noted during the first months of the pandemic, followed by an increase from autumn 2020 onward. Starting in 2022, the proportions rose more sharply overall compared to the previous years, with a particularly marked increase in the low education and income groups. Correspondingly, from 2022 onwards, both the absolute and relative inequality measures in education and income groups climbed, although the rise in relative inequality reached only statistical significance for educational inequality.

The decrease in the proportion of persons screening positive for possible depressive disorder noted in all education and income groups at the start of the pandemic and the subsequent increase until 2024 are in line with the general development, based on GEDA data of the PHQ-2 and the more comprehensive screening instrument PHQ-8 (15, 16, 17). A continuous rapid review identified six surveys on the first lockdown that were suitable to assess trends in the population reliably, of which only the GEDA study found a decrease, while two studies found stable trends and three an increase in depressive symptoms—presumably attributable to differences in survey design (26). In the absence of stratified results from other studies, no conclusions about differences in level of education and income could be drawn from the literature. Analyses of individual symptoms of the PHQ-8 found a symptom reduction in the groups with low and medium levels of education, in particular in stress-associated symptoms of depression, in the first months of the pandemic, possibly due to changes in work and everyday life (27). The fact that the trends were overall stable or even opposite in the corresponding months of 2019 suggests that seasonal effects did not play a role.

Contrary to expectations, no significant increase in absolute and relative inequalities was found in the proportion of the population screening positive for possible depressive disorder in the first two years of the pandemic. The relative inequalities even decreased slightly in 2021. While the low and high education and income groups experienced a comparably large increase in the proportion of persons screening positive for possible depressive disorder in 2021 compared to 2019 and 2020 (i.e., no change in absolute inequality), the relative increase was significantly higher for the high education and income groups as a result of the lower baseline prevalence. Given the socially unequal distribution of some risk factors for mental health problems, the finding that the inequality did not increase was unexpected. In the first year of the pandemic, job loss and job insecurity, reduced working hours without short-time allowance, extra work, working from home, and deterioration of financial situation were identified as risk factors for poorer mental health (28). People with a low socioeconomic status may have been more severely affected by some of these factors. For example, the low income group experienced loss of income more frequently than the high income group (29). However, other risk factors, such as the switch to working from home, may have more affected higher education and income groups (30). Correspondingly, a study revealed that persons with a low level of education felt a great burden with regard to their financial situation, while persons with a high level of education felt a greater burden in the domains of social life, leisure, and family (31).

Only in women, the absolute inequality in the proportion of the population screening positive for possible depressive disorder started to increase already during the COVID-19 pandemic. This finding is in line with the results of a scoping review, highlighting that women with a low level of education and low income represent a particularly affected group with higher scores for depression and anxiety (32). During the pandemic, the burden on women was greater than that on men (32, 33). Due to the gender pay gap, they frequently received lower short-term allowances or unemployment benefits (34) and performed more care work (33). This applied in particular to women with low income who were more likely to take on more housework and to reduce their working hours to care for children than women with higher income (33).

The increase in the proportions of persons screening positive for possible depressive disorder in all socioeconomic groups observed in spring 2022 coincided with the emergence of additional collective stressors. In the weeks after Russia’s attack on Ukraine, overall anxiety levels in Germany were reported to have risen even further compared to a period when strict COVID-19 containment measures were in place (35). Low-income households were less able to absorb the intensified household energy and food price increases that started in 2022 (5), because these expenses account for a larger share of disposable income in this group (36). This may have contributed to an increase in depressive symptoms in low socioeconomic groups in 2022 and thus to a rise in absolute and relative inequalities.

In both relative and absolute terms, income inequalities in the prevalence of depressive symptoms were more pronounced than educational inequalities. Even though different indicators of socioeconomic status are interrelated, they can capture different health risks and resources: The level of education has an impact on later employment due to unequal opportunities in the labor market, economic circumstances, health behavior, and health literacy (9); in addition, it is associated with higher levels of psychosocial resources, such as self-efficacy (37), control belief, resilience, and cultural participation which, in turn, can lower the risk of depression (38). Income secures the livelihood and material standard of living and is a key pillar of old-age provision, social security and social participation (9). Low income can have a negative impact on mental health by causing chronic psychosocial stress, e.g. related to financial difficulties (39). In comparison to level of education, income is often more strongly associated with (mental) health (40), as was also evident during the COVID-19 pandemic (11). The underlying mechanisms are diverse. For example, stress (e.g., due to pandemic-related job loss or short-time work, financial difficulties or inflation) is likely to be related to income in particular (39, 40) and associated with mental health problems (39).

One limitation of our study is that the study design did not allow to identify causal relationships between socioeconomic status, collective stressors und depressive symptoms. In addition, our study is limited by the exclusive use of the PHQ-2 screening instrument which does not allow for a comprehensive diagnostic assessment and clinical diagnosis of a depressive disorder and only captures one aspect of mental health. In order to fully understand the development of social inequality with regard to mental health, future research should study additional indicators, such as suicidal tendency, anxiety symptoms and chronic stress. Furthermore, health inequalities due to other forms of structural disadvantage, e.g. due to migration background, should also be analyzed.

The strengths of our analyses include the representativeness of the underlying data, the long observation period and the high survey frequency for trend analyses. In addition, the use of regression-based measures of inequality allowed to make more robust statements on the development of health inequalities compared to basic comparisons of individual socioeconomic groups.

Conclusion

In the years 2020–2024, health inequality increased among adults in Germany. The presented findings indicate that persons with socioeconomically disadvantaged backgrounds are particularly vulnerable to the mental health consequences of multiple collective stressors. This highlights the need for continuous monitoring of trends in key core indicators of mental health, their influencing factors and the socially unequal distribution in order to identify problem situations promptly and develop measures to improve health equity before and during periods of collective stressors.

Funding

This study was funded by the German Research Foundation (Deutsche Forschungsgemeinschaft, DFG) (INHECOV project; funding code: 458531028), the German Federal Ministry of Health (Bundesministerium für Gesundheit, BMG) (project for the “Establishment of Mental Health Surveillance”; funding code: ZMI5–2519FSB402), as well as the Robert Koch Institute and the German Federal Ministry of Health (GEDA study).

Conflict of interest statement

JH is co-spokesperson for the joint working group “Social Epidemiology” of the DGMS, DGEpi, and DGSMP.

The remaining authors declare that they have no conflicts of interest.

Manuscript received on 27 February 2025, revised version accepted on 10 July 2025

Translated from the original German by Ralf Thoene, M.D.

Corresponding authors
Christina Kersjes, M.Sc. Psych.

Dr. PH Jens Hoebel