Are the Kids (Sort of) Alright, After All?

In recent years, there has been a growing focus on a perceived mental health crisis among adolescents, as embodied by the alarming rise in reported cases of suicidal behaviors and depression. The Centers for Disease Control and Prevention (CDC), for example, reports that between 2013 and 2023, the number of US adolescents reporting “feelings of persistent sadness or hopelessness” increased from 30% to a staggering 40%.¹

Understandably, this disconcerting trend has generated much concern among academics and policymakers. Mental health experts have pointed to a range of factors, from rising academic pressures due to the increasing importance of education in the labor market to the pervasive influence of social media. Indeed, some trends show the rise in adolescent depressive symptoms starting shortly after 2007, when, coincidently or not, the first iPhone model was released. One prominent voice in this discussion, the social psychologist Jonathan Haidt, has argued that the rise of social media usage plays a large role in declining mental health, particularly among teenage girls. Haidt’s and colleagues’ research emphasizes a correlation between increased screen time, social media use, and rates of anxiety, depression, and suicidal ideation in adolescents.² Their findings suggest that the heightened exposure to curated online identities and the pressures of social comparison can lead to deteriorating mental health, especially in vulnerable teens. The problem with studies such as these, however, is that it’s nearly impossible to sort out cause and effect: Does online media exposure depress the moods of teens?  Or do teens who are falling into despair fall into more screen time? Economists Luca Braghieri, Ro’ee Levy, and Alexey Makarin leveraged the quasi-random roll-out of Facebook across college campuses and found more modest effects when social media is exogenous to individual mood: The rise of social media may explain one-fifth of the increase in depressive symptoms among adolescents—leaving the lion’s share unaccounted for.³

While Haidt’s and others’ work has been provocative in pointing to social media as a key factor in the mental health crisis, another important element often overlooked is the role of reporting rules in shaping the trends documented by the CDC and others. Research by Adriana Corredor-Waldron and Janet Currie sheds light on this aspect, revealing that changes in how suicidal behaviors are reported and classified may have a significant impact on the observed rates of these behaviors.⁴ Their findings suggest that much of the reported increase in teen suicide-related hospital visits between 2008 and 2019 may be due to shifts in reporting and diagnostic criteria rather than a dramatic rise in underlying behaviors.⁴ This raises a critical question: How much of what we believe about trends in mental health—whether driven by social media or other factors—reflects genuine changes versus changes in how we identify and report these behaviors?

Two specific reporting rule changes stand out. First, in 2011, the US Preventive Services Task Force recommended that adolescent girls be routinely screened for depression—this is when rates for girls start rising most quickly, diverging from those of boys. This recommendation was later incorporated into the Affordable Care Act, leading to a broader adoption of mental health screenings as part of routine adolescent health care. As a result, many more teens—particularly girls—began undergoing mental health assessments as part of their regular doctor visits. This increased screening led to more diagnoses of depression and suicidal ideation, which in turn contributed to the rise in suicide-related hospital visits.

The second major change occurred in 2016 when the International Classification of Diseases (ICD-10) updated its guidelines to encourage clinicians to record suicidal ideation as a secondary diagnosis, even when the primary diagnosis was a different mental health condition, such as depression. This adjustment meant that doctors were more likely to record suicidal thoughts in a patient’s medical records, even if the patient was primarily being treated for anxiety, depression, or another disorder. The study found that while suicide-related hospital visits increased significantly during this period, there was no corresponding increase in completed suicides. This suggests that the rise in hospital visits may have more to do with changes in how suicidal behaviors are recorded than with an actual increase in suicidal actions among teens.

These shifts in reporting practices are significant because they highlight how changes in diagnostic guidelines can inflate the appearance of an epidemic. When more doctors are encouraged to screen for depression and record suicidal ideation, the numbers of reported cases naturally increase—not because more teens are suddenly developing mental health problems, but because we are doing a better job of identifying them.  Moreover, the increased focus on suicide-related behaviors also led to a phenomenon known as “diagnostic drift,” where certain behaviors are more readily classified as symptoms of suicidal ideation due to heightened awareness and the broader criteria introduced by the ICD-10. This means that behaviors that might not have been labeled as suicidal in previous years are now classified as such, contributing to the rise in reported suicide-related incidents.

The lesson here is clear: reporting rules play a critical role in shaping the data we rely on to understand health trends. As reporting standards evolve, they can dramatically influence the observed rates of diseases and behaviors.

This issue of reporting rules shaping trends in health outcomes is not unique to mental health. Similar patterns have been observed in other areas of medicine, where changes in diagnostic criteria or reporting practices have led to apparent increases in disease prevalence.

One well-known example is the rise in autism diagnoses. The introduction of the DSM-5 in 2013, which redefined and broadened the criteria for autism spectrum disorder (ASD), led to a significant increase in the number of children diagnosed with the condition. Between 2000 and 2016, the prevalence of autism in the United States rose from about 1 in 150 children to 1 in 54. While some of this increase may be attributed to a genuine rise in autism rates, much of it likely stems from the changes in how autism is diagnosed.

Similarly, the expansion of criteria for diagnosing depression and anxiety has contributed to rising rates of these disorders. The DSM-5, for example, lowered the diagnostic thresholds for anxiety disorders and redefined depressive disorders to include new categories such as persistent depressive disorder and premenstrual dysphoric disorder. These changes have likely contributed to the increase in reported cases of depression and anxiety over the past decade.

Even in physical health, changes in reporting rules can dramatically alter the perceived prevalence of diseases. Take heart disease as an example. Over the years, the definition of what constitutes high blood pressure has been lowered several times, resulting in millions more people being classified as hypertensive. While this has improved early detection and treatment, it also has led to an apparent rise in the prevalence of the condition, even though the underlying rate of heart disease may not have changed as dramatically.

Understanding how reporting and classification rules affect observed trends in health is critical for both policymakers and the public. When changes in diagnostic criteria or reporting practices lead to spikes in the observed rates of disease, it can create the impression of a growing crisis, even if the underlying problem remains relatively stable. This can lead to a misallocation of resources, with public health efforts focusing on conditions that appear to be increasing, while other equally important issues receive less attention.

Moreover, changes in reporting rules can disproportionately affect certain populations. For instance, the research by Corredor-Waldron and Currie shows that, after the introduction of new screening guidelines for adolescent depression, girls were more likely than boys to be diagnosed with suicidal thoughts. This suggests that changes in reporting practices may not only inflate overall numbers but also create disparities between different demographic groups​.

As we continue to grapple with the challenges of adolescent mental health, it is crucial to approach reported trends with caution. Rising numbers of diagnoses may not always reflect an actual increase in disease but rather improvements in detection and changes in how conditions are reported. This does not diminish the seriousness of the teen mental health crisis, but it does underscore the importance of understanding the role that reporting rules play in shaping our perceptions of health trends.

At the same time, we must advocate for policies that ensure consistent and accurate reporting across all populations. Expanding access to mental health screenings and care is essential, but so is ensuring that changes in diagnostic practices do not disproportionately affect certain groups or inflate trends in ways that mislead public health efforts. As we move forward, public health officials and researchers would be wise to adopt a policy of parallel trends. That is, when possible, two trends should be reported: Those using the old classification scheme alongside data obtained using the new approach. For example, rates of hypertension are easily adjusted to show how trends look under different standards, retrospectively and prospectively, since we have the information we need: raw blood pressure readings from medical records. Other times, we may have to satisfy ourselves with simply marking when major policy or classification shifts occur—akin to the marker showing the break in presidential polling when President Biden dropped out of the race. Such may be the case for understanding trends in adolescent depression. What is certain is that it is health policy malpractice for the CDC or anyone else to report trend data collected under different regimes without highlighting that fact. Good policy always starts with good data.

References

1. Centers for Disease Control and Prevention (CDC), 2023. Youth Risk Behavior Survey Data Summary & Trends Report: 2013-2023.  Available at: https://www.cdc.gov/yrbs/dstr/index.html [Accessed 25 September 2024].
2. Twenge, J.M., Haidt, J., Lozano, J. and Cummins, K.M., 2022. Specification curve analysis shows that social media use is linked to poor mental health, especially among girls. Acta psychologica, 224, p.103512.
3. Braghieri, L., Levy, R.E. and Makarin, A., 2022. Social media and mental health. American Economic Review, 112(11), pp.3660-3693.
4. Corredor-Waldron, A. and Currie, J., 2023. The Long-Term Impact of Medicaid and the Affordable Care Act on Teen Suicide-Related Behaviors. Journal of Human Resources. 58(2), pp. 423-454.

About the Author