An integrative literature review of birth cohort and time period trends in adolescent depression in the United States

Summary Abstract

The aim of this literature review is to examine evidence of time trends and birth cohort effects in depressive disorders and symptoms among US adolescents in peer-reviewed articles from January 2004 to April 2022. We conducted an integrative systematic literature review. Three reviewers participated at different stages of article review. Of the 2,234 articles identified in three databases (Pubmed, ProQuest Central, Ebscohost), 10 met inclusion criteria (i.e., adolescent aged United States populations, included information about birth cohort and survey year, focused on depressive symptoms/disorders). All 10 articles observed increases in depressive symptoms and disorders in adolescents across recent survey years with increases observed between 1991 and 2020. Of the 3 articles that assessed birth cohort trends, birth cohort trends were less prominent than time period trends. Proposed explanations for increases included social media, economic-related reasons, changes in mental health screening and diagnosis, declining mental health stigma, increased treatment, and, in more recent years, the COVID-19 pandemic. Multiple cross-sectional surveys and cohort studies documented rising prevalence of depressive symptoms and disorder among adolescents from 1991-2020. Mechanisms driving this increase are still unknown. Research to identify these mechanisms is needed to inform depression screening and intervention efforts for adolescents.

INTRODUCTION

Since the late 2010s, there have been growing concerns related to adolescent mental health, including impacts of social media use on face-to-face interaction,^1–3 cyber-bullying,⁴ and negative influence on self-esteem.⁵ In addition to changes in societal trends, such as social media use, more recently concern has emerged of the adverse effects of the COVID-19 pandemic on adolescent mental health.⁶ Thus, it is an important time to summarize recent trends in adolescent depression in the United States (US), including both pre-pandemic and pandemic years.

Evidence from some epidemiological studies suggests that the prevalence of depression among adolescents in the US increased across the first two decades of the 21^st Century.^7–10 This increase in depression occurred even as other mental health problems became less prevalent among adolescents, including parent-assessed severe psychological impairment¹¹ and externalizing symptoms such as conduct problems.¹² Social trends might be selectively affecting internalizing conditions, even though internalizing conditions are correlated with externalizing conditions and share some risk factors.^{13, 14} Increases in adolescent depression have been observed from several cross-sectional data sources across a range of different measurements of depression. For example, in one national sample, adolescent past-year major depressive episodes (MDE) increased from 8.8% in 2005 to 15.7% in 2019 and past-year prevalence of MDE with severe impairment increased from 5.5% in 2006 to 11.1% in 2019.¹⁰ A recent Morbidity and Mortality Weekly Report highlighted the importance of the growing public health concerns related to adolescent mental health in the US, finding high rates of depression-related indicators (i.e., symptoms, episodes, and disorders) across multiple national surveillance systems during 2018-2019.¹⁵ Prevalence of seriously considering suicide among adolescents in the US increased from 13.8% in 2009 to 17.2% in 2017.¹⁶ From 1999-2014, deaths by suicide among adolescents have increased from 1.9 to 2.6 per 100,000 among people ages 10-14 and 16.8 to 18.2 among older adolescents and young adults ages 15-24.¹⁷

The drivers of these observed increases in adolescent depression can be divided into two categories referred to by demographers as “time period effects” and “cohort effects”.^{9, 18} These two categories of trends have different implications for understanding the driving factors.¹⁹ First, a change in environment could be influencing the risk of depression for all adolescents in the US regardless of age (time period effects).^18–21 Changes in mental health prevalence can also occur for specific cohorts as they emerge through the life course, commonly referred to as birth cohort effects, indicative of shared cumulative experiences for individuals born during a defined time period.^18–21 In order to identify time period and birth cohort effects, measurements of mixed age groups at different time points are needed. Research assessing participants of the same age from distinct birth cohorts can elucidate drivers of changes in depression prevalence. The implications of time period effects and birth cohort effects are distinct, as changes in environment could be transient (i.e., changes from a societal event or shifting political landscape that are not sustained over time) and birth cohort effects might forecast rising depression prevalence in other age groups as these birth cohorts age into young adulthood (i.e., drivers that alter depression risk for specific birth cohorts uniquely such as the rise of youth social media use).

Prior integrative reviews have examined trends in depressive symptoms across period and cohort effects. Prior research has synthesized evidence on birth cohort trends in child (ages less than 12 years old) and adolescent (ages 12 to 17 years old) populations. A meta-analysis of studies prior to 2004 found no relationship between birth cohort and depression prevalence.²² Other reviews have examined changes in prevalence of mental health treatment, but not change in underlying symptoms.²³ Reviews of the literature in more recent years have been scant, despite growing evidence of increases in depression^7–9 and other mental health outcomes.^{16, 17} In 2014, a review was conducted on broad mental health problems in children and adolescents throughout the world,²⁴ documenting increasing internalizing symptoms in populations from the United States and other countries in recent versus prior birth cohorts, specific to adolescent populations. However, this review was not restricted to depression and included other mental health outcomes.²⁴

The aim of this literature review was to summarize recent trends in adolescent depressive disorders and symptoms in the United States. To answer this question, we summarized evidence from studies of time period effects and birth cohort effects in adolescent depressive disorders and symptoms in published peer-reviewed articles from January of 2004 to April of 2022. A secondary objective was to synthesize proposed explanations for observed adolescent depression trends to inform future research studying potential mechanisms of adolescent depression trends.

METHODS

Databases and searches

This integrative literature review incorporated components of a systematic literature review and a scoping review, allowing for flexibility of methodology while including the rigor of a systematic literature review. Three online databases were searched: Pubmed, ProQuest Central, and Ebscohost. ProQuest Central was restricted to relevant databases: Psychology Database, Public Health Database, Social Science Database. Ebscohost was also limited to just relevant databases related to this topic area: APA PsycInfo Database, SocINDEX with Full Text Database, Social Sciences Full Text Database. These databases were searched on April 6, 2022. All search terms are outlined in Appendix Table 1; search terms included combinations of search terms relating to adolescent populations, the United States, depressive symptoms or disorders, and birth cohort and time trends.

Inclusion and exclusion criteria

The inclusion and exclusion criteria are outlined in Table 1. Inclusion criteria were peer-reviewed English-language articles reporting original research focused on the United States published between January 2004 and April 2022. Articles also needed to assess associations between birth cohort or time trends and depression symptoms or disorders and focus on depression being assessed during adolescent ages, defined as age 12 to 17 years. Inclusion of articles was based on age of depression (ages 12-17) assessment, and thus a wide range of birth cohorts were included. Information must have been available about participants’ year of birth or age at interview and survey year. Psychiatric diagnoses of depressive disorders using an established taxonomy and structured or semi-structured psychiatric interview, or non-diagnostic depressive symptoms, were also required. Studies outside the United States, focused on adults, related to other mental health or other outcomes, and qualitative or non-empirical studies were excluded. Studies of age trends in depression over a study period that did not explicitly reference time period at each age and/or birth cohort were excluded. The rationale for this decision was to only include studies that clearly assessed time trends or birth cohort trends, and not age trends alone. It is important to note that articles could have looked at a combination of period and cohort effects rather than isolating these effects using methods such as hierarchical age, period, and cohort models.

Table 1:

Summary of inclusion and exclusion criteria for literature review on birth cohort and time trends in adolescent depression in the United States

Inclusion criteria

Exclusion criteria

•Original research papers   •Peer-reviewed   •Published between 2004 and April 2022   •Written in English   •Focused in the United States   •Shows associations between birth cohort or time trends and depression symptoms or disorders   •Participants aged 12-17 at time of interview   •Information is available about participants’ year of birth/age at interview and survey year   •Formal psychiatric diagnoses of depressive disorders using an established taxonomy and structure or semi-structured psychiatric interview OR generic depressive symptoms are assessed

•Outside the United States   •Focus on adults and not children   •Focused on younger children <12 years old only   •Outcome is related to another psychiatric outcome (e.g., PTSD, suicide) and not depression   •Outcome is related to depression treatment but not depressive symptoms/disorders   •Other irrelevant outcome   •Qualitative or other non-empirical study   •Does not include 2 out of these 3 factors: age, birth cohort, and study year

Utilizing Covidence software,²⁵ an initial title and abstract screening was performed for all articles identified from the search terms. Relevant articles were then assessed with a full-text review and two independent reviewers screened half of these articles each with two total reviewers per article. Review agreement was needed before articles were included. During the full text review, 87.1% (115/132 articles) of article inclusion/exclusion decisions were agreed upon. Reconciliation was also done to come to a consensus on exclusion reasons, discussing the inclusion criteria again in the context of each article. Following the full text review, two independent reviewers and the lead author (MSA) abstracted each article. Any discrepancies in the abstraction details were discussed and reconciled. At every stage of article selection, articles were assigned to each reviewer randomly to avoid potential bias.

Data extraction, synthesis, and analysis

For every study included, an abstraction form was utilized to collect relevant details. Data pertaining to general article information (title, publication year), methods (study objective, study design, study population and data source, age of participants, total analytic sample, main exposure, main outcome including details about measurement, analytic method details), results (quantitative results and descriptive summary of main findings), limitations/main sources of bias, and conclusions/implications (including whether articles noted potential explanations for observed trends) were extracted. This review was approved for registration with Prospero, an international registration system for literature reviews, on March 16, 2022 (CRD42022310592).

RESULTS

Literature search

Of the 2,234 articles identified in the 3 databases of interest, 12 were excluded due to duplicate articles (Figure 1). After duplicates were removed, 2,222 article titles and abstracts were reviewed and 2,090 of these articles were excluded for not meeting the inclusion criteria at this stage. After removing irrelevant articles from the title and abstract stage, there were 132 articles for which the full text was reviewed. Of these 132 articles, 125 were excluded. The most common reason for exclusion was because they did not include explicit information about 1) age and 2) birth cohort and/or study year, preventing ascertainment of birth cohort (k=67). The second most frequent exclusion reason was that articles focused on adults and not adolescent populations (k=22). Ten articles were excluded for focusing on an outcome related to another psychiatric outcome (e.g., PTSD) and not depression. Seven articles were excluded for referring to populations outside the United States. Six articles were excluded for focusing on children younger than adolescent age (<12 years old) only. Five articles were excluded for referring to other irrelevant outcomes (i.e., not depressive disorders or symptoms or another psychiatric outcome). Five articles were qualitative or non-empirical studies (e.g., literature reviews) and three articles were focused on depression treatment and not depressive symptoms or disorders. Three articles that were cited by the 7 included articles and met inclusion criteria were additionally included. Thus, there were 10 articles total included.

Figure 1:

Flowchart of literature review on birth cohort and time trends in adolescent depression in the United States

Study design of included articles

Seven of the ten included articles studied changes in depressive symptoms^{1, 9, 26–30} and three studied major depressive episodes^{7, 31, 32} as the main outcome of interest. Eight of the ten articles were repeated cross-sectional studies and the only two cohort studies were both focused on COVID-19 and included short time periods of follow-up.^{29, 30} Studies had methodological differences in design. For example, there were some distinctions in type of depressive symptom measured (e.g., somatic depressive symptoms²⁶ versus feelings of sadness or hopelessness in the prior two weeks²⁸). The data sources included primarily large national studies including the Monitoring the Future study, the National Survey on Drug Use and Health, and the Youth Risk Behavior Survey. The only studies including smaller study populations that were not nationally representative were those related to the COVID-19 pandemic.^{29, 30} Of the ten articles identified, three assessed both birth cohort effects and time period effects whereas the other seven assessed time period effects alone.

Increases in adolescent depression

Time-period effects:

All studies found increases in depressive symptoms and episodes in recent years starting in the 2010s (Table 3). For example, mean depressive symptoms measured by the Bentler Medical and Psychological Functioning Inventory depression scale increased from 7.67 in 2012 to 9.18 in 2018.⁹ Past-year major depressive episodes also increased from 8.7% in 2005 to 11.3% in 2014.⁷ During the COVID-19 pandemic, Revised Child Anxiety and Depression Scales DSM-based depression symptom scores increased from 45.10 2018-Feb 2020 to 50.95 in Spring 2020.²⁹

Table 3:

Increases in depressive symptoms observed in select included articles for literature review on birth cohort and time trends in adolescent depression in the United States

Study	Depression measure	Years	Age ranges (years)	Birth cohort ranges	Prevalence or Mean Score (SD)
Mojtabai, 2016	12-month major depressive episodes	2005 2014	12-17	1988-2002	8.7% 11.3%
Twenge, 2017	Depressive symptoms; included six items from the Bentler Medical and Psychological Functioning Inventory depression scale. Prevalence is of high depressive symptoms with an item mean score of 3+	1991-1994 2010 2015	13-18	1973-2002	16.96% 16.13% 21.48%
Weinberger, 2018	12-month major depressive episodes	2005 2015	12-17	1988-2003	8.69% 12.66%
Twenge, 2019	12-month major depressive episodes	2005 2017	12-17	1988-2005	8.7% 13.2%
Keyes, 2019	Depressive symptoms; included four items from the Bentler Medical and Psychological Functioning Inventory depression scale	2012 2018	13-18	1994-2005	7.67 (3.85) 9.18 (4.34)
Keyes, 2020	Depressive symptoms; included four items from the Bentler Medical and Psychological Functioning Inventory depression scale	2014 2018	13-18	1996-2005	8.02 (3.78) 8.86 (3.86)
Pontes, 2020	Depressive symptoms; survey item that measured feelings of sadness or hopeless-ness that lasted at least 2 weeks during the past 12 months with a dichotomous yes/no response	2009 2017	14-18	1991-2003	340 per 1000 female students 408 per 1000 female students
Breaux, 2021	Depressive symptoms; Self-report utilizing the Revised Child Anxiety and Depression Scales (RCADS) to assess DSM-based depression symptoms	2018-Feb 2020 Spring 2020 Summer 2020	15-17	2001-2005	45.10 (13.04) 50.95 (15.01) 46.19 (13.70)

Note: This table includes data from 8 of the 10 included articles. The 2 other included articles did not explicitly display descriptive data by year. Unless otherwise indicated (i.e., Pontes article), these estimates were calculated among the total sample and were not gender-specific. Birth cohorts were derived by getting the maximum range of survey year-age and articles did not always explicitly list the birth cohorts included.

Birth cohort effects:

Of the three articles that assessed birth cohort effects, two articles reported few birth cohort effects and found that trends were primarily driven by time period effects.^{9, 32} However, there were some birth cohort effects that emerged from sub-group analyses considering gender and race/ethnicity. For example, girls born in 2001 had higher depressive symptom scores compared with those born in 1990.⁹ The third article reporting birth cohort effects found birth cohort effects in somatic depressive symptoms with those born in 1982-1999 being more likely to experience somatic depressive symptoms compared with those born in 1961-1981.²⁶

Proposed explanations

It was not feasible to systematically examine statistical mediators of the observed increases in adolescent depression since studies reviewed did not test mediation. However, several studies speculated as to the mechanisms that might explain the observed trends. These hypothesized mechanisms explanations are summarized in Figure 2. These studies did not empirically test proposed explanations, but rather conjectured about potential explanations for the observed trends observed in the studies.

Figure 2:

Frequency of proposed explanations for increases in adolescent depression over time in included articles

Note: More than one explanation could be selected for each study.

The leading hypothesized mechanisms offered for the increases in prevalence of depression among adolescents were social media, economic-related reasons, and changes to mental health screening or diagnosis. Other explanations were social disruptions related to the COVID-19 pandemic, changes to mental health treatment systems, the opioid epidemic, rises in parental monitoring and supervision, and changes to mental health stigma. A broader group of other more infrequently noted explanations included changes to binge drinking trends that are related to depression trends, overall increases in stress generally, decreases in adult-related activities (e.g., receipt of driver’s licenses, working for pay, dating habits), declining age of menarche, decreases in sleep quality, and increases in other suicide-related correlates. These proposed explanations were described in relation to time period effects (i.e., drivers increasing adolescent depression time trends rather than impacting specific birth cohorts alone), though it is likely that some of the explanations also could have cohort specific impacts on depression (e.g., social media use).

Study quality and strength of evidence

There were various sources of bias in the included articles. The main source of bias identified across studies was measurement error or information bias, particularly for studies in which the depression outcome measure included non-specific depressive symptoms. For example, one study assessed broad somatic depressive symptoms,²⁶ such as whether an adolescent reported having a sore throat or headache in the prior 30 days, which could represent distinct trends and be influenced by different mechanisms than more specific depression measures, such as major depressive episodes or a depression diagnosis. Other studies did not clearly specify methods to account for confounding¹ or confounding control methods were insufficient (e.g., did not include birth cohort not exposed to COVID during adolescence as a historic control).²⁹ Lastly, the two studies focused on the impact of the COVID-19 pandemic on adolescent depression may not generalize to other populations and did not include nationally representative study populations, as one study focused on a predominantly White population in select regions of the US²⁹ and the other studied a homogenous sample restricted to New York State.³⁰

DISCUSSION

This literature review assessed birth cohort and time trends of depressive symptoms and disorders in adolescent populations in the US after 2004. Overall, adolescent depressive symptoms and disorders are increasing in recent years and birth cohorts. Several hypothesized mechanisms have been proposed as contributing to these increases in time period and birth cohort effects including social media, economic-related factors, changes to mental health screening and diagnosis, and the opioid epidemic; however, there was little consensus across studies about the main driver of these trends and studies lacked rigorous research designs that would have enabled identification of causal effects of drivers. These studies did not always speculate on how these drivers would differentially impact time period and birth cohort effects, and it is likely they could contribute to changes in both.

This study was the first to our knowledge to synthesize research published after 2004 that studied the relationship between birth cohort/time period depressive symptoms or disorders in adolescent populations. Increases in depression in recent years have been concentrated in adolescents and previously there was no review of papers focused specifically on this age group. Other reviews that include childhood depression trends have utilized a different search strategy, studying broad mental health outcomes instead of focusing on depression specifically, and synthesizing studies of younger children in addition to adolescents.²⁴ As the time trends in depression have been observed to be distinct between younger children less than 12 years of age and adolescent children ages 12 to 17,²⁴ it is important to narrow the focus to adolescents. This review of 10 articles found similar results to an international review of 19 articles that included some studies with adolescent populations²⁴ noting increases in depression for recent adolescent populations.

We acknowledge limitations. While some studies identified in this review used advanced methodological techniques to assess trends controlling for external confounding factors, some of the included studies were descriptive. The search strategy did not exclude descriptive studies to allow for synthesis of a broader range of studies. However, it is important to note that some of the descriptive studies may not have controlled for factors that influenced both the risk of depression and belonging to a specific birth cohort. One of these factors is age, as depression risk differs by age and age is closely linked to birth cohort (i.e., age = survey year - birth cohort). Studies that only assessed trends in age were excluded. As a result of the narrow age group of interest, the risk of confounding due to age was minimized. As the identification strategy did not specify specific sub-groups of interest, this review also did not summarize results based on sub-group differences, yet it is important to acknowledge that there are important differences in depression risk and trends by various sub-populations (e.g., sex and race/ethnicity).⁹ The objective of this review was to identify the overall population-level change in depression trends and further synthesis is needed to sufficiently synthesize literature on sub-group differences and search terms related to sub-groups were also not included in this study.

Evidence supports the claim that depressive symptoms in adolescents are increasing, consistent with a prior 2014 literature review focused on a broader array of mental health outcomes and international study populations.²⁴ Yet, there is still little consensus about the mechanisms driving these increases. The main mechanisms for these increases mentioned in this literature review included social media, economic-related factors, changes to mental health screening and diagnosis, and the opioid epidemic. Many of these mechanisms, particularly social media,^33–36 have been empirically explored. Increases in social media use have been hypothesized to a major driver of observed increases in poor adolescent mental health in the past decade.^1–5 Recent research using robust confounding control methods did not find that time spent on social media use was a risk factor for depressive symptoms.^{36, 37} Recent literature has also assessed changes in adolescent time use³⁸ and changes in parental supervision/monitoring¹ as potential drivers of increasing trends in depression since the early 2010s,⁹ but further research is needed to unpack how changes in parental monitoring have influenced adolescent mental health, including depression, and whether effects were more focused on time period effects or birth cohort effects. There have been further studies published examining potential social drivers of increases in adolescent depressive symptoms and disorders in US populations. For example, 2005-2018 Monitoring the Future study results indicate that depressive symptoms have increased since 2010, particularly among adolescent females with liberal political views.³⁹ Other recent research has assessed the influence of sleep patterns on internalizing symptoms, including depressive symptoms, finding that frequency of sleep can partially explain patterns in internalizing symptoms over time.⁴⁰ Lastly, research has explored relationships between religious beliefs and service attendance and depressive symptoms finding that declines in religious beliefs may explain some of the increases in adolescent depressive symptoms over the past two decades.⁴¹

Future directions for research using rigorous identification methods, such as age, period, and cohort models that independently assess age, period, and cohort effects, should include further empirical exploration of the proposed mechanisms contributing to the increase in adolescent depression. There is a trend of increasing adolescent depression which has worsened and is receiving attention because of the COVID-19 pandemic. In this context, it is essential to clarify secular time period trends and birth cohort changes in depression in the recent past, as this could elucidate whether specific birth cohorts may need targeted additional depression intervention and treatment efforts. If time period trends in adolescent depression are more dominant than birth cohort effects, then depression intervention and treatment efforts should be focused across multiple birth cohorts. Adolescents, while experiencing less risk for physical health problems during the COVID-19 pandemic with fewer infections and deaths compared to older individuals,⁴² are showing the greatest vulnerability to adverse mental health consequences associated with the pandemic.⁶ Increases in various adolescent mental health outcomes related to depression, such as suicide attempts, have been documented during the pandemic.⁴³ Future research should also consider the effects of economic factors on adolescent depression, as negative economic impacts of the COVID-19 pandemic, such as predicted recessions, are hypothesized to contribute to increased risk of poor adolescent mental health.⁶

This literature review synthesized evidence on birth cohort and time trends of depressive symptoms and disorders in adolescent US populations since 2004. This review identified 10 studies and we found that adolescent depression is increasing in recent years and for recent birth cohorts. While various societal explanations for these increased have been proposed, there is no single clear, empirically tested and supported mechanism that has been identified as driving this increase. Given heightened concerns about adolescent mental health during the COVID-19 pandemic, better understanding these mechanisms and the interplay between them in the context the pandemic will continue to be a priority for the years to come.

Table 2:

Summary of included articles (n=10) for literature review on birth cohort and time trends in adolescent depression in the United States

First author last name, year	Title of article	Depression-related research question	Study population and data source	Study design	Exposure measure	Outcome measure	Main conclusions	Limitations and Biases
Twenge, 2015	Time period and birth cohort differences in depressive symptoms in the US, 1982-2013	To analyze differences in somatic symptoms of depression (12th graders, 1982-2012)	12th grade students included in the Monitoring the Future Surveys (1982-2012; N =70,838); 79-83% response rate	Repeated cross-sectional study	Both birth cohort and time period	Depressive symptoms, past 30-day somatic symptoms assessed in the CES-D and other depression scales	Increases observed in the 2010s with adolescents reporting more somatic symptoms of depression than adolescents in the 1980s, including trouble remembering things, difficulty thinking or concentrating, and trouble sleeping	Measurement error/information bias: focuses on psychosomatic symptoms of depression that are very non-specific
Mojtabai, 2016*	National trends in the prevalence and treatment of depression in adolescents and young adults	To examine national trends in 12-month prevalence of major depressive episodes (MDEs) in adolescents and young adults overall and in different sociodemographic groups, as well as trends in depression treatment between 2005 and 2014	Participants ages 12-25 included in the National Surveys on Drug Use and Health (2005 to 2014; N= 172,495 adolescents aged 12 to 17 and 178,755 adults aged 18 to 25); annual mean response rate of 65.2%	Repeated cross-sectional study	Time period (i.e., survey year trends are presented)	Major depressive episodes; Lifetime and 12-month MDE were assessed using a structured interview based on the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition criteria. Participants were next asked whether they had experienced an episode in the past year. Questions were adapted from the depression section of the NCS-Replication	The prevalence of adolescent 12-month MDE was stable over the 2005 to 2011 period but gradually increased from 8.7% (2005) to 11.3% (2014) corresponding to a 37% increase in odds (P < .001)	Measurement error/information bias: depressive episodes based on self-report by adolescents and does not include clinical assessment or assessment from parents or teachers
Twenge, 2017*	Increases in depressive symptoms, suicide-related outcomes, and suicide rates among U.S. adolescents after 2010 and links to increased new media screen time	To determine whether the prevalence of depressive symptoms, suicide related outcomes, and suicide deaths has increased and to examine possible causes behind trends, focusing on shifts in adolescents	8th and 10th grade students Monitoring the Future (1991-2015; N=388,275); 9th, 10th, 11th, and 12th grade students Youth Risk Behavior Survey (1991-2015; N=118,545); 79-83% response rate	Repeated cross-sectional study	Time period (i.e., survey year trends are presented)	Depressive symptoms; included six items from the Bentler Medical and Psychological Functioning Inventory depression scale	Between 2009/2010 and 2015, 33% more adolescents had high levels of depressive symptoms (3+ symptoms on average; 16.13% in 2010, 21.48% in 2015) Compared with those born 1961-1981, those born in 1982-1999 were more likely to experience depressive symptoms.	Measurement error/information bias: omits depressive symptoms such as insomnia and anergia Confounding control method unclear
Weinberger, 2018	Trends in depression prevalence in the USA from 2005 to 2015: widening disparities in vulnerable groups	To estimate trends in the prevalence of major depression in the US population from 2005 to 2015 overall and by demographic subgroups	Participants ages 12+ National Survey on Drug Use and Health (2005-2015; N= 607,520); annual mean response rate of 65.2%	Repeated cross-sectional study	Time period (i.e., survey year trends are presented)	Major depressive episodes; Lifetime and 12-month MDE were assessed using a structured interview based on the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition criteria. Participants were next asked whether they had experienced an episode in the past year. Questions were adapted from the depression section of the NCS-Replication.	2005: ~9% prevalence of adolescent MDE 2015: ~13% prevalence of adolescent MDE A year by age interaction was observed, illustrating that the increase in depression among adolescents occurred at a faster rate than the increases in other age groups [aOR=1.05, 95% CI=1.04, 1.06 for time trend among ages 12-17]	Measurement error/information bias: depressive episodes based on self-report by adolescents and does not include clinical assessment or assessment from parents or teachers
Twenge, 2019*	Age, period, and cohort trends in mood disorder indicators and suicide-related outcomes in a nationally representative dataset, 2005-2017	To explore trends in psychological distress, past-year MDEs, and suicide-related outcomes	Participants ages 12+ National Survey on Drug Use and Health (2005-2017; N=212,913 for 12-17 years only); annual mean response rate of 65.2%	Repeated cross-sectional study	Both birth cohort and time period	Major depressive episodes; 12-month MDE was assessed using a structured interview based on the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition criteria. Participants were next asked whether they had experienced an episode in the past year. Questions were adapted from the depression section of the NCS-Replication	Rates of past-year adolescent MDE increased 52% 2005-2017 (from 8.7% to 13.2%). Age, period, cohort analyses found that the trend among adolescents was driven by time period, with MDE increasing from 8.8% in 2006 to 14.8% in 2017 (a 68% increase). Majority of the increase occurred from 2010 (9.1%) to 2017 (14.8%), as adolescent MDE increased 63%.	Measurement error/information bias: depressive episodes based on self-report by adolescents and does not include clinical assessment or assessment from parents or teachers
Keyes, 2019	Recent increases in depressive symptoms among US adolescents: trends from 1991 to 2018	To estimate age, period, and cohort effects in depressive symptoms among US nationally representative samples of school attending adolescents from 1991 to 2018	8th, 10th, and 12th grade students Monitoring the Future (1991-2018; N=512,283); 85.0% to 87.3% response rate with an annual average of 86.5%	Repeated cross-sectional study	Both birth cohort and time period	Depressive symptoms; included four items from the Bentler Medical and Psychological Functioning Inventory depression scale; All items included in Twenge, 2017 except “I feel that I can’t do anything right” and “I feel that my life is not very useful”	Starting in 2012, depressive symptoms increased annually through 2018, from 7.67 (SD=3.85) in 2012 to 9.18 in 2018 (SD=4.34) Few birth cohort effects were observed overall, but cohort effects were observed in some sub-groups. For example, girls born in 2001 had higher depressive symptoms than those born in 1990.	Measurement error/information bias: omits depressive symptoms such as insomnia and anergia
Keyes, 2020	Diverging trends in the relationship between binge drinking and depressive symptoms among adolescents in the U.S. from 1991 through 2018	To examine time trends in the strength of the relationship between binge drinking and depressive symptoms among nationally representative samples of adolescents from 1991 to 2018 in the MTF data, including examination of subgroup differences by sex, race/ethnicity, and socioeconomic status	12th grade students Monitoring the Future (1991-2018; N=68,670); 85.0% to 87.3% response rate with an annual average of 86.5%	Repeated cross-sectional study	Time period (i.e., survey year trends are presented)	Depressive symptoms; included four items from the Bentler Medical and Psychological Functioning Inventory depression scale; All items included in Twenge, 2017 except “I feel that I can’t do anything right” and “I feel that my life is not very useful”	Depressive symptoms were stable from 1991 to 2013, then mean depressive symptoms increased from 8.02 (SD: 3.78) in 2014 to 8.86 (SD: 3.86) in 2018	Measurement error/information bias: omits depressive symptoms such as insomnia and anergia Temporality unclear: potential bidirectional relationship between binge drinking and depressive symptoms (note- results shown only focus on depression trends alone)
Pontes, 2020	Trends in depressive symptoms and suicidality: Youth Risk Behavior Survey 2009-2017	To investigate whether gender moderates the linear time trend for the rate of depressive symptoms, suicidal ideation, and suicide attempts among high school students within the United States	High school students in the Youth Risk Behavior Survey (YRBS) data (2009 to 2017: 2009, 2011, 2013, 2015, 2017; N= 31,175); 67%-90% response rate	Repeated cross-sectional study	Time period (i.e., survey year trends are presented)	Depressive symptoms; survey item that measured feelings of sadness or hopeless-ness that lasted at least 2 weeks during the past 12 months with a dichotomous yes/no response	Depressive symptoms increased from 2009 (rate females=340 per 1000 students, males=192 per 1000 students) to 2017 (rate females=408 per 1000 students, males=211 per 1000 students)	Measurement error/information bias: self-reported single measure from YRBS for depressive symptoms
Breaux, 2021	Prospective impact of COVID-19 on mental health functioning in adolescents with and without ADHD: protective role of emotion regulation abilities	To examine changes in and predictors of adolescent mental health from before to during the COVID-19 pandemic in the Southeastern and Midwestern United States	8th grade adolescents both with and without ADHD from 2 sites in Southeastern and Midwestern United States (September 2018- August 2020; N=238); 90.8% response rate	Cohort study	Time period (i.e., survey year trends are presented)	Depressive symptoms; Self-report utilizing the Revised Child Anxiety and Depression Scales (RCADS) to assess DSM-based depression symptoms	Pre-COVID [September 2018-February 2020], M (SD)= 45.10 (13.04), Spring 2020 [May 15-June 14, 2020], M (SD)= 50.95 (15.01), Summer 2020 [July 1-August 5, 2020], M (SD)= 46.19 (13.70), F=20.40, p<.001, d=0.75 The change across timepoints was quadratic with spring 2020 depression scores being elevated compared with both pre-COVID-19 and summer 2020	Random error/small sample size (<n=500) Confounding: did not include birth cohort not exposed to COVID during adolescence as control Selection bias: adolescents identified as predominantly White (82%)
Hawes, 2021	Trajectories of depression, anxiety and pandemic experiences; A longitudinal study of youth in New York during the Spring-Summer of 2020	This study aimed to use multilevel growth modeling to capture the trajectory of symptoms and pandemic experiences across Spring-Summer 2020 and to explore predictors of depression symptoms across the study period	Adolescents and young adults in New York, pre-pandemic= July 2016-2019; during pandemic= March 27th and July 14th, 2020; Age 15 at the 5th wave of data collection pre-pandemic (N= 532 adolescents and young adults); response rate not provided	Cohort study	Time period (i.e., survey year trends are presented)	Depressive symptoms; Symptoms of depression assessed using the Child Depression Inventory and the Screen for Child Anxiety Related Disorders; Children’s Depression Inventory (CDI), a 27-item self- report questionnaire designed to assess symptoms of depression occurring in the past 2 weeks in youth ages 7-17	Depressive symptoms initially increased linearly (positive slope effects), t(822) =4.54, p <.001, with variability in the rate of linear change (95% CI =0.44-2.13), followed by a rapid decline in depression (negative curve effects), t(818) =−5.56, p <.001.	Selection bias: Homogenous sample with limited generalizability to other more nationally representative populations (only in New York and not nationally representative) and attrition analyses suggest non-participants differ significantly from participants

Additional notes: Twenge, 2015 outcome measure: “These next questions concern your health. During the LAST 30 DAYS, on how many days (if any) did you have the following problems or symptoms?” with the following items: “Headache,” “Sore throat or hoarse voice,” “Trouble with sinus congestion, runny nose, or sneezing,” “Coughing spells,” “Chest colds,” “Coughing up phlegm or blood,” “Shortness of breath when you were not exercising,” “Wheezing or gasping,” “Trouble remembering things,” “Difficulty thinking or concentrating,” “Trouble learning new things,” “Trouble sleeping,” “Trouble getting started in the morning,” “Stayed home most or all of a day because you were not feeling well. For each item, response choices were 1 = None, 2 = 1 day, 3 = 2 days, 4 = 3-5 days, 5 = 6-9 days, 6 = 10-19 days, 7 = 20+ days.”

Twenge, 2017 outcome measure: ““Life often seems meaningless,” “I enjoy life as much as anyone” (reverse scored), “The future often seems hopeless,” “I feel that I can’t do anything right,” “I feel that my life is not very useful,” and “It feels good to be alive” (reverse scored). Response choices ranged from 1 (dis-agree) to 5 (agree).”

Abbreviations: CES-D= Center for Epidemiologic Studies Depression Scale.

^*Indicates articles that were added due to being cited by included articles.

Implications and Contribution.

This literature review summarizes recent birth cohort and time trends of adolescent depression in the United States, illustrating increases from 1991 to 2020 observed in all included articles. Even as adolescent depression trends are rising, mechanisms driving these trends remain unclear.