Depressive Symptoms in Adolescents

Unipolar depressive disorders are among the most common mental health disorders in adolescents (1). According to the World Health Organization’s Global Burden of Disease study, the effects of this condition are already clearly apparent during adolescence (2). Given its widespread nature and potentially disastrous consequences (among others an increased risk of suicidality [3]; however, empirical data specific to adolescents are scarce), it has been suggested to treat adolescent depression as a Global Health Priority (4). Adolescent depression is thought to be caused by a combination of biological, psychological, and social factors. According to Naab et al. (5), the following key psychosocial risk factors have been identified:

• Parental neglect
• Problematic peer relations and
• Family problems.

A meta-analysis found a prevalence of depression among 13- to 18-year-olds of 5.6% (6). In the included primary studies, prevalences from acute prevalence to 12-month prevalence had been reported. According to the results of this meta-analysis (6), female adolescents were more frequently affected by depression than male adolescents (5.9% vs 4.6%). In numerous studies, this sex difference in the prevalence of depressive disorders has already been demonstrated empirically for children and adolescents. For Germany, the most recent prevalence estimate which is based on a representative sample is available from the BELLA study (7). Data were collected between May 2003 and May 2006. According to Bettge et al. (7), self-reports of 11- to 17-year-olds revealed that 9.7% of girls and 4.7% of boys were affected by depressive symptoms. However, such a positive screening finding does not mean that these respondents had a depressive disorder. New BELLA data from a longitudinal follow-up of the sample are expected to become available in 2018.

While Costello et al. (6) found no increase in the prevalence rates of depressive disorders among children and adolescents born between 1965 and 1996, several studies from various countries have been published in recent years, reporting increases in the occurrence of depressive symptoms among adolescents (8–11). These trend statements were typically based on data from representative cross-sectional studies, conducted at various survey time points. For the period between 1997 and 2006, an increase in self-reported depressive symptoms among Icelandic girls (but not boys) and for both sexes an increased utilization of healthcare services provided by psychiatrists, psychologists and social workers was reported (8). Among female Finnish adolescents, a significantly higher prevalence of severe depression was found when the depression scores for the years 2010/2011 (girls: 4.7%/boys: 2.2%) and 2000/2001 (girls: 4.0%/boys: 2.1%) were compared (9). Von Soest and Wichstrøm (10) found an increased proportion of Norwegian adolescents with severe depressive symptoms when comparing the years 2002 (girls: 11.4%/boys: 6.1%) and 1992 (girls: 8.8%/boys: 2.9%). However, the difference between the findings for 2010 (girls: 13.4%/boys: 5.5%) and the results for 2002 was not statistically significant (10). Mojtabai et al. (11) studied the development of the 12-month prevalences of major depression diagnoses among US adolescents and reported higher prevalence rates for 2014 (11.3%) compared to 2005 (8.7%). By contrast, no increase in the 12-month prevalence of major depression diagnoses between 2000 and 2014 was found among Canadian adolescents. However, the prevalence of self-reported mood disorder diagnoses established by health professionals showed an increase from 2003 to 2014 (12). In summary, the available data on the question whether internationally the numbers of adolescents with symptoms of depression has increased are inconclusive.

When interpreting prevalence estimates in population samples it should be taken into account that depressive symptoms (especially due to variations in the severity of the conditions) do not necessarily equate with need for treatment (13), which can be expected to be lower. However, given that adolescent depression (at least of longer duration) is a strong predictor of mental health problems in adult life (14), it makes sense from a public health perspective to assess the scale on which adolescents are already affected by depressive symptoms to be able to offer early preventative or interventional measures or to reduce or increase the extent of such offerings on the basis of epidemiological findings.

The aims of our study were to determine:

• The current point prevalence of depressive symptoms among adolescents in Germany and
• Which sociodemographic and psychosocial features are associated with depressive symptoms in adolescents.

Methods

Data collection

Data were collected by means of a telephone survey conducted by a market research and opinion polling company (Forsa) in August and September 2017. A total of 1001 adolescents were included in the survey (a random sample of 1698 12- to 17-year-olds were invited to participate in the survey; consequently, the participation rate was 59.0%). In preparation, households were selected from the ADM Telephone Master Sample by multi-stage systematic random selection and then contacted and asked whether 12- to 17-year-olds were living in the household and whether they agreed to being contacted again by phone. The average time required to conduct the telephone interview and collect the data was almost 17 minutes and the respondents were children and adolescents from Germany (age range 12 to 17 years; in the following, these will be covered collectively under the term “adolescents”). In order to achieve representativeness for the age group 12 to 17 years in Germany, the market research company weighted the sample by sex, age and region (eastern federal states/western federal states) based on the official current population statistics of the German Federal Statistical Office (as of 31 December 2015) after completion of data collection.

Survey methods

The validated Depression Screener for Teenagers (DesTeen, [15]) was used with the recommended cut-off value of ≥14 (16, p. 1307) to collect data on depressive symptoms among adolescents. Data on the parameter “interpersonal trust” were collected using the Interpersonal Trust short scale (KUSIV3 [17]) and on the subjective body image using 3 questions from the body image questionnaire (18). The problematic use of social media (websites of social networks, messengers, etc.) was assessed using the Social Media Disorder Scale (SMDS) (19) and the problematic use of computer games using the Internet Gaming Disorder Scale (IGDS) (20). Family functioning was measured using the Family APGAR questionnaire (21). Key sociodemographic characteristics, such as sex and age of the adolescents, and academic achievement were also recorded.

Data analysis

In order to estimate the point prevalence of depressive symptoms in adolescents, relative frequencies with 95% confidence intervals [95% CI] were calculated, first for the weighted total sample and then stratified according to sex (the weight variable was adjusted for each subgroup). In 13 of the 1001 surveyed persons (1.3% of the sample), there were altogether 16 missing values in the DesTeen items, making it impossible to calculate the respective questionnaire’s sum score. In view of the low absolute number (16 of a total of 14 014 entries or 0.1%) the missing values were not imputed. All statistical analyses on depressive symptoms were performed without the 13 respondents with missing DesTeen sum scores, i.e. based on 988 cases. For comparisons of groups (for example, boys vs. girls) the chi-square test was used; for independent samples we used the t-test. Associations between the dependent variable (two categories: depressive symptoms not present versus present) and the independent variables (sex, age, academic achievement, interpersonal trust, body image, problematic use of social media or computer games, as well as family functioning) were evaluated using logistic regressions (forced entry method, also in the weighted data set). First, bivariate associations between depressive symptoms and each of the various characteristics were calculated, followed by a multivariable regression analysis with all independent variables included in one model. The statistical software package SPSS version 22.0 (IBM Inc., Armonk, NY, USA) was used for all calculations. In line with internationally accepted conventions, the level of statistical significance was set at α = 0.05.

Results

Table 1 lists the sociodemographic und psychosocial characteristics of the 988 12- to 17-year-olds included in the analysis. As expected (given the age range of the sample), most of the adolescents still attended school at the time of the survey (926 cases [93.7%]). A total of 45 (4.6%) surveyed adolescents were in vocational training, 6 (0.6%) were in military service or working as volunteers in the German federal voluntary service, 4 (0.4%) reported to be unemployed or looking for work, 1 (0.1%) was a student at a university/universtity of applied sciences, and another 6 (0.6%) respondents selected the answer option “Other“. The 62 adolescents (6.3%) who were no longer attending school were asked for supplementary information about their highest educational attainment. One person (0.1% of the sample) had not obtained a school leaving certificate, 1 (0.1%) had a special-school leaving certificate (“Förderschulabschluss”), 17 (1.7%) a lower secondary school-leaving certificate (“Hauptschulabschluss”), 35 (3.6%) a secondary school-leaving certificate (“Realschulabschluss“), 7 (0.7%) a university of applied sciences or general university entrance qualification (“Fachhochschulreife”/”Abitur”), and 1 (0.1%) had acquired another school-leaving certification (with these results, the age of the respondents at the time of the survey and the percentage of the sample still attending school are to be taken into account).

Table 1

Sociodemographic and psychosocial characteristics of the total sample and of adolescents without and with depressive symptoms

Enlarge All figures

Altogether 81 adolescents screened positive for depressive symptoms (sum score of the DesTeen ≥ 14). Accordingly, the estimated prevalence of depressive symptoms among adolescents in Germany is 8.2% (point prevalence, [95% CI: 6.5; 9.9]). Girls (11.6% [8.8; 14.4]) reported depressive symptoms more frequently than boys (5.0% [3.1; 6.9]) (χ² = 14.12, df = 1, p<0.001). Respondents who screened positive were statistically significantly older than those with negative screening results (t = 2.61, df = 986, p = 0.009).

Bivariate regression analyses showed statistically significant associations between depressive symptoms and (Table 2):

• Female sex
• Older age
• Poorer performance at school
• Lower interpersonal trust
• More negative body image
• More problematic use of social media or computer games, as well as
• Lower satisfaction with family relations.

Table 2

Association between sociodemographic and psychosocial aspects and depressive symptoms in a representative German sample of adolescents (all associations were first evaluated in bivariate models and in the next step in a full model)

Enlarge All figures

In a full model, comprising all independent variables, statistically significant associations were found between depressive symptoms and female sex, older age, more negative body image, more problematic use of social media, and lower satisfaction with family relations.

In this multivariable regression analysis, about one third of the variation was explained by being a member of one of the two groups (adolescents without versus adolescents with depressive symptoms) (Nagelkerke’s R² = 0.35, χ² = 120.20, p<0.001).

Discussion

In this study, the prevalence of depressive symptoms was determined in a representative sample of German adolescents. The estimated prevalence revealed that a substantial percentage of the sample (8.2%)—equivalent to about one in twelve adolescents in Germany—experienced depressive symptoms (point prevalence). Significant differences were found between the sexes, with about one in nine girls being affected but only one in 20 boys. Compared to the last available findings for Germany, reported by Bettge et al. (7), which were based on a representative sample, the percentage of male adolescents with depressive symptoms [2017: 5.0% vs. 2008: 4.7% (7)] remained almost unchanged, while a slight increase in the estimated prevalence was observed among female adolescents [2017: 11.6% vs. 2008: 9.7% (7)]. However, the use of different screening instruments in the two studies limited the comparability of the results. Besides the two distinct questionnaires with different cut-off values, the main difference between the studies is that Bettge et al. (7) collected additional data on the degree of impairment resulting from the psychological problems and combined this aspect with a positive screening result for depressive symptoms. By contrast, our study made exclusive use of the screening instrument DesTeen (which, however, was validated during its development against a structured diagnostic interview; see eMethods). While depressive symptoms are indicative of subjective burden and burden of disease in those affected (potentially including their relatives), they cannot be equated with the diagnosis of depression. Based on their analysis of data from a German statutory health insurance (Gmünder Ersatzkasse), Hoffmann et al. (22) reported that in 2009 3.1% of the 12- to 18-year-old insured persons were diagnosed with depression in an outpatient setting (with this administrative prevalence, however, it should be noted that most likely not all affected adolescents received outpatient treatment or were diagnosed with depression even if they suffered from the disorder).

Internationally, the results have been inconsistent, with some studies indicating an increase in depressive symptoms among adolescents (11) and others not (12). It is conceivable that there is no uniform international trend and a more differentiated analysis is required taking additional country-specific aspects into account. For example, the study of Torikka et al. (9) showed that the socioeconomic status of the family of origin plays a key role. The findings of our study indicate that in Germany depressive symptoms are not rarer among adolescents than among adults, because in the German Health Interview and Examination Survey for Adults (DEGS1) a point prevalence of 8.1% was reported for a representative sample of 18- to 79-year-olds (23). However, when comparing the findings for adults with those for adolescents the methodological limitation has to be taken into account that in DEGS1 (23) another screening instrument (PHQ-9, [24]) with a different cut-off value (≥ 10) was used.

Several of the identified correlates of depressive symptoms among adolescents (female sex, older age, more negative body image, and poorer family functioning) are backed by sound empirical evidence (for example [6, 25, 26]). Especially the association between problematic use of social media and depressive symptoms among adolescents is a new field of research (27) and was demonstrated for Germany for the first time, whereas the associations between problematic computer game behavior and depression had already been empirically observed before (28). Furthermore, relationships between the various associated factors, such as, for example, use of social media by adolescents and impact on their body image, are of great interest and should be further evaluated in future longitudinal studies with regard to their effect on the development of depressive symptoms.

This study has a number of limitations. Data on depressive symptoms were collected using a screening questionnaire and not by a structured clinical interview which would have allowed more precise data acquisition and establishment of diagnosis. However, larger epidemiological studies often use questionnaires instead of interviews (e.g., also in [7] and [23]) and the survey instrument used was validated against a structured diagnostic interview (16). The prevalence estimates are based on self-assessments of adolescents which are considered more reliable compared to external assessments (for example by parents) (29); however, it would have facilitated the assessment of clinical significance if data on the impairment by depressive symptoms would also have been collected. In addition to the correlates studied, there are other relevant factors which were not included in the survey to save resources, for example, comorbid mental disorders or chronic somatic conditions. Furthermore, the survey did not cover the socioeconomic background of the family of origin or their migration background. Thus, it cannot be ruled out that these aspects are also of relevance to depressive symptoms in adolescents. While the chosen design of a cross-sectional survey in a representative sample is well suited for estimating prevalence (30), it does not allow to infer causal relationships. For example, performance at school can drop as a result of mental stress, but, on the other hand, it cannot be excluded that psychopathological stress is promoted by, for example, declining performance at school.

Our study produced several important new findings with regard to depressive symptoms. A substantial percentage of 12- to 17-year-olds in Germany is affected by depressive symptoms. In view of the increased risk of chronification of the symptoms into adulthood (31), it appears relevant to supplement or advance existing strategies of primary and secondary prevention for adolescents in the future. Here, the identified correlates (especially the particularly relevant individual psychosocial characteristics negative body image and problematic use of social media) may be of help; however, these should be further assessed—ideally in longitudinal studies—with regard to their predictive significance for the development of depressive symptoms.

Funding
This study was conducted with financial support from the DAK-Gesundheit health insurance.

Conflict of interest statement
The authors declare that no conflict of interest exists.

Manuscript received on 14 January 2018, revised version accepted on 22 May 2018

Translated from the original German by Ralf Thoene, MD.

Corresponding author
PD Dr. phil. Dipl.-Psych. Lutz Wartberg
Deutsches Zentrum für Suchtfragen des Kindes- und Jugendalters (DZSKJ)
Universitätsklinikum Hamburg-Eppendorf
Martinistr. 52, 20246 Hamburg, Germany
lwartberg@uke.de

►Supplementary material
For eReferences please refer to:
www.aerzteblatt-international.de/ref3318

eMethods:
www.aerzteblatt-international.de/18m0459