1. Bibliographic Information

1.1. Title

Trends in generalised anxiety disorders and symptoms in primary care: UK population-based cohort study

1.2. Authors

The authors of the paper are April Slee, Irwin Nazareth, Nick Freemantle, and Laura Horsfall. They are affiliated with University College London (UCL), specifically the Department of Primary Care and Population Health and the Comprehensive Clinical Trials Unit. Their affiliations indicate strong expertise in primary care research, epidemiology, biostatistics, and clinical trials, which are all central to this study's design and interpretation.

1.3. Journal/Conference

The paper was published in the British Journal of Psychiatry (BJP), the flagship journal of the Royal College of Psychiatrists in the UK. The BJP is a leading international peer-reviewed journal covering all branches of psychiatry, with a high impact factor and a strong reputation for publishing significant clinical and epidemiological research. Publication in this journal signifies that the work is considered important and methodologically sound by experts in the field.

1.4. Publication Year

2020. The paper was first received in April 2020 and accepted for publication in July 2020.

1.5. Abstract

The study investigates temporal trends in the recording of generalised anxiety and related disorders in UK primary care between 1998 and 2018. Using data from 795 general practices from The Health Improvement Network (THIN) database, the researchers calculated annual incidence rates of generalised anxiety diagnoses and symptoms. They found a significant increase in the recording of generalised anxiety, particularly among men and women aged 18–24 between 2014 and 2018. A composite measure of anxiety and depression also showed increases in younger populations. In terms of treatment, approximately half of patients who were not previously on medication received a prescription for anxiety within a year of diagnosis, with selective serotonin reuptake inhibitors (SSRIs) being the most common choice. Conversely, benzodiazepine prescribing showed a steady decline. The study concludes that there has been a recent and substantial rise in primary care consultations for anxiety and depression, concentrated among younger individuals, especially women.

1.6. Original Source Link

The original source link provided is a local file path: /files/papers/690566510b2d130ab3e047df/paper.pdf.

The paper is officially published and is an Open Access article available under the Creative Commons Attribution license. The official publication can be accessed via its Digital Object Identifier (DOI): https://doi.org/10.1192/bjp.2020.159.

2. Executive Summary

2.1. Background & Motivation

The core problem this paper addresses is the lack of up-to-date, large-scale evidence on the trends of generalised anxiety disorder (GAD) in UK primary care settings. GAD is a common and debilitating mental health condition that places a significant burden on individuals (through reduced functioning) and the healthcare system (through frequent consultations). While most mental health conditions are managed in primary care in the UK, there was limited research examining trends after 2010.

The authors identified a gap in understanding how the incidence of GAD has changed over time, particularly in different age and gender groups. This is important because societal changes, economic pressures, and shifts in clinical practice could all influence how many people present with and are diagnosed with anxiety. The paper's innovative approach is to leverage a massive, longitudinal electronic health record database (THIN) to conduct a 20-year retrospective analysis (1998-2018), providing a high-resolution picture of recent trends in both diagnosis and initial pharmacological treatment.

2.2. Main Contributions / Findings

The paper's main findings provide crucial insights into the changing landscape of mental health in the UK:

1. A Sharp, Recent Increase in Anxiety in Young Adults: The most significant finding is a dramatic increase in the recorded incidence of GAD and its symptoms between 2014 and 2018. This increase was not uniform across the population but was heavily concentrated in young people aged 18–24, and was most pronounced in young women.

2. Validation of a Genuine Rise in Mental Health Issues: By using a composite measure of "any anxiety or depression," the study cleverly tested whether the rise in anxiety was simply a re-labeling of depression diagnoses. For the youngest age groups, the composite rate also rose sharply, suggesting a genuine increase in mental health consultations, not just a diagnostic shift.

3. Shifting Prescription Patterns Aligning with Clinical Guidelines: The analysis of pharmacological treatments revealed that prescribing practices have evolved to better align with the 2011 National Institute for Health and Care Excellence (NICE) guidelines. The use of selective serotonin reuptake inhibitors (SSRIs) as a first-line treatment has steadily increased, while the prescription of benzodiazepines, which are associated with dependence, has consistently declined.

4. Identification of a Major Public Health Concern: By quantifying the scale and demographic focus of this trend, the study highlights a significant and emerging public health issue. The findings signal a need for greater awareness among clinicians and call for more resources to be directed toward mental health services for young adults.

3. Prerequisite Knowledge & Related Work

3.1. Foundational Concepts

To fully understand this paper, a novice reader should be familiar with the following concepts:

• Generalised Anxiety Disorder (GAD): A mental health condition characterized by persistent, excessive, and uncontrollable worry about a variety of things. As the paper notes, this worry becomes counterproductive and debilitating, affecting physical, emotional, and social functioning. It is distinct from the normal anxiety people experience in response to stressful situations.

• Primary Care & General Practitioners (GPs): In the UK's National Health Service (NHS), primary care is the first point of contact for patients. General Practitioners (GPs) are family doctors who provide comprehensive care and manage a wide range of health problems, including mental health. Most common mental disorders like GAD are identified and managed at this level.

• Epidemiology: The scientific study of the distribution (who, where, when) and determinants (causes, risk factors) of health-related states and events in specified populations. This study is an example of descriptive epidemiology, as it describes the trend of a disorder over time.

• Cohort Study: A type of observational research where a group of individuals (a cohort) who share a common characteristic are followed over a period of time. In this study, the cohort consists of patients registered in specific UK general practices. Researchers track them to determine the incidence of new GAD diagnoses.

• Incidence Rate: A measure of the frequency with which new cases of a disease or condition occur in a population over a specified period. It is a key metric in epidemiology for tracking the emergence and spread of health issues.

• Person-Years at Risk (PYAR): A measurement used in cohort studies to calculate incidence rates. It combines the number of people in the study and the amount of time each person is followed. For example, 100 people followed for one year contribute 100 person-years; similarly, 50 people followed for two years also contribute 100 person-years. This method accurately accounts for participants entering and leaving the study at different times.

• The Health Improvement Network (THIN) Database: A very large database of anonymized longitudinal patient records collected from UK primary care practices that use Vision software. It contains clinical information such as diagnoses, symptoms, prescriptions, and referrals, making it a powerful resource for epidemiological research.

• Read Codes: A comprehensive, hierarchical system of clinical codes used in UK primary care to record every aspect of a patient's care, including symptoms, diagnoses, tests, and procedures. Researchers use specific Read codes to identify patients with the conditions they want to study.

• British National Formulary (BNF): The standard UK reference guide for prescribers, providing information on medicines and their uses. The paper uses BNF chapter classifications (e.g., chapter 4.3.3 for SSRIs) to identify drug classes based on their intended therapeutic purpose.

• Key Drug Classes for Anxiety:

  • Selective Serotonin Reuptake Inhibitors (SSRIs): A class of antidepressants that work by increasing the level of the neurotransmitter serotonin in the brain. They are recommended as a first-line pharmacological treatment for GAD.
  • Serotonin-Norepinephrine Reuptake Inhibitors (SNRIs): Another class of antidepressants that increase levels of both serotonin and norepinephrine.
  • Tricyclic Antidepressants (TCAs): An older class of antidepressants. Their use has declined due to a higher side-effect burden compared to SSRIs.
  • Benzodiazepines: A class of sedative drugs that provide rapid relief from anxiety symptoms. However, they are associated with a high risk of dependence, tolerance, and withdrawal symptoms, so their long-term use is discouraged.

3.2. Previous Works

The authors build upon several key pieces of prior research and clinical guidance:

• Walters et al. (2012): This is the most direct predecessor to the current study. The Walters paper used the same THIN database to examine trends in anxiety diagnoses and symptoms from 1998 to 2008. The current study explicitly aims to extend and update these findings, confirming trends from the first decade and, more importantly, analyzing the subsequent decade (2008-2018).

• NICE Guidelines (2011): The National Institute for Health and Care Excellence (NICE) provides evidence-based recommendations for health and care in the UK. Their 2011 guideline for GAD was a landmark publication that shaped clinical practice. Key recommendations included:

  • Offering psychological interventions (like cognitive-behavioural therapy) first.
  • If medication is used, offering an SSRI (specifically sertraline) as the first-line choice.
  • Warning against the routine prescription of benzodiazepines and advising they only be used for short-term crisis management due to the risk of addiction. This guideline provides the benchmark against which the paper evaluates the observed trends in prescribing.

• Research on Comorbidity: The paper cites work (e.g., Judd et al., 1998) establishing the high rate of comorbidity between GAD and major depressive disorder. This foundational knowledge justifies the authors' decision to analyze a composite outcome of anxiety and depression, as the symptoms overlap and diagnostic practices may vary.

3.3. Technological Evolution

The methodology of this paper reflects a major evolution in public health research: the move from traditional, small-scale surveys to the analysis of "big data" from electronic health records (EHRs). In the past, estimating disease prevalence or incidence required resource-intensive surveys of a sample of the population. The advent of large, standardized, longitudinal databases like THIN allows researchers to:

1. Analyze a much larger and more representative sample of the population (millions of patients).

2. Track trends over very long periods (decades).

3. Achieve high granularity, breaking down trends by detailed age groups, gender, and geographical location.

4. Link diagnostic data with prescription data to study real-world treatment patterns.

   This study sits at the modern frontier of pharmacoepidemiology and mental health surveillance, using computational analysis of real-world data to identify public health trends in near real-time.

3.4. Differentiation Analysis

Compared to previous work, this paper's key innovations are:

• Timeliness and Recency: Its analysis extends to 2018, capturing a very recent and dramatic shift in GAD incidence that earlier studies could not have seen.
• Focus on Young Adults: While other studies might have reported overall trends, this paper's stratified analysis successfully pinpoints young adults (18-24) as the group most affected by the recent surge.
• Methodological Rigor: The use of a composite anxiety/depression outcome is a sophisticated sensitivity analysis that strengthens the conclusion that the trend in young people is a genuine increase in mental ill-health, not just a change in diagnostic labels.
• Integration of Diagnosis and Treatment: By simultaneously analyzing diagnostic trends and prescribing patterns, the paper provides a holistic view of how primary care is responding to this evolving mental health challenge.

4. Methodology

4.1. Principles

The core principle of this study is to conduct a retrospective population-based cohort study. The researchers used historical, anonymized patient data from a large primary care database (THIN) to identify a cohort of eligible patients. They then followed this cohort over a 20-year period (1998–2018) to calculate the annual rate of new (incident) diagnoses of generalised anxiety. To analyze these trends while accounting for influential factors, they employed a statistical model known as a Poisson mixed effects model, which is specifically designed for analyzing count-based rates over time while controlling for variations between different individuals and GP practices.

4.2. Core Methodology In-depth

The methodology can be broken down into the following integrated steps:

4.2.1. Data Source and Patient Selection

1. Data Source: The study utilized The Health Improvement Network (THIN) database, containing records from 795 UK general practices. This database is considered representative of the UK population in terms of demographics and disease prevalence.
2. Cohort Definition: A cohort of patients was selected based on specific criteria to ensure data quality and that only new cases of anxiety were counted.
   • Inclusion Criteria:
     • Patients registered with one of the 795 practices between January 1, 1998, and December 31, 2018.
     • Aged 18 years or older.
     • Registered with their practice for at least 12 months before cohort entry (this is a "washout period" to ensure any recorded diagnosis is new and not a historical one from a previous practice).
     • Registered at a practice meeting quality standards for data recording.
   • Exclusion Criteria:
     • Patients with less than 12 months of valid follow-up data.
     • Crucially, any patient with a recorded diagnosis of generalised anxiety, depression, or mixed anxiety/depression before their cohort entry date was excluded. This ensures the study measures incidence (new cases) and not prevalence (all existing cases).
3. Follow-up Period: Each patient was followed from their cohort entry date until the earliest of: their first GAD diagnosis, leaving the practice, death, or the study end date (December 31, 2018).

4.2.2. Diagnosis and Prescription Identification

1. Diagnoses: Diagnoses were identified using clinical Read codes. The authors used pre-existing, validated lists of Read codes to identify:
   • Generalised anxiety disorder diagnoses.
   • Generalised anxiety symptoms.
   • Depression diagnoses and symptoms.
   • Mixed anxiety/depression states. The primary outcome was a recording of either a GAD diagnosis or GAD symptoms. A secondary, composite outcome included any of the above diagnoses.
2. Prescriptions: Pharmacological treatments were identified and classified by drug class according to the British National Formulary (BNF) coding system. This approach captures the GP's intended use for the prescription. For example, benzodiazepines were only included if coded under the 'Anxiolytics' section to exclude use for other conditions like alcohol withdrawal.

4.2.3. Statistical Analysis

1. Incidence Rate Calculation: The fundamental metric was the annual incidence rate. For each year, this was calculated as: $ \text{Incidence Rate per 1000 PYAR} = \frac{\text{Number of new cases in that year}}{\text{Total Person-Years at Risk (PYAR) in that year}} \times 1000 $ Where:

   • Number of new cases: The count of patients receiving their first-ever diagnosis of GAD in a given year.
   • Total Person-Years at Risk (PYAR): The sum of the time each person in the cohort was followed and remained "at risk" (i.e., had not yet been diagnosed) during that year.

2. Poisson Mixed Effects Model: To formally model the trends, the authors used a Poisson mixed effects model. This type of model is ideal for analyzing count data (like the number of diagnoses) that occur over a certain exposure time (PYAR).

   • Poisson Component: At its core, a Poisson model assumes that the number of events (diagnoses) follows a Poisson distribution. It models the logarithm of the expected rate of events as a linear function of predictor variables.
   • Mixed Effects Component: This term means the model includes both fixed effects and random effects.
     • Fixed Effects: These are the variables of primary interest whose impact the researchers want to measure directly. In this study, they were: age group, gender, time (as a linear variable), and the interactions between them (e.g., age group × time).
     • Random Effects: These account for sources of variability that are not of primary interest but could influence the results, such as differences between GP practices. By including the practice as a random effect, the model acknowledges that patients from the same practice may be more similar to each other than to patients from other practices, thus correcting for potential clustering in the data. The model also included a random effects smoother for time, allowing it to capture complex, non-linear trends over time that might differ between practices.
   • Model Formulation: The model essentially predicts the rate of diagnosis. The structure is: $ \log(\text{rate}{ijt}) = \log\left(\frac{\text{count}{ijt}}{\text{PYAR}{ijt}}\right) = \text{Fixed Effects}{ijt} + \text{Random Effects}_{jt} $ Where:
     • $\text{rate}_{ijt}$ is the diagnosis rate for age-gender group $i$ in practice $j$ at time $t$.
     • $\text{count}_{ijt}$ is the number of diagnoses.
     • $\text{PYAR}_{ijt}$ is the person-years at risk.
     • $\text{Fixed Effects}_{ijt}$ includes terms for age, gender, time, and their interactions.
     • $\text{Random Effects}_{jt}$ includes terms for practice-level variation and non-linear time trends within practices. (Note: The paper states the "log of the number of diagnoses" was used as an offset, which is likely a typo. In rate modeling, the standard offset is the logarithm of the exposure, which is log(PYAR). This is the standard and correct approach for a Poisson model of rates, and what was almost certainly implemented.)

3. Calculation of Rate Differences: The authors used a SAS macro (%NLEstimate) that employs the delta method. The delta method is a statistical technique used to estimate the variance of a function of one or more random variables. Here, it was used to calculate the standard errors and confidence intervals for the model-estimated rates for each year, and more importantly, for the difference in rates between 2014 and 2018. This allowed them to formally test whether the change over that period was statistically significant.

5. Experimental Setup

5.1. Datasets

The sole dataset used was The Health Improvement Network (THIN) database.

• Source: Anonymized, longitudinal electronic health records from 795 general practices across the UK. It is licensed by the data provider IQVIA.
• Scale and Characteristics: The study cohort comprised 6,630,040 patients, contributing a total of 52,695,827 person-years at risk (PYAR) over the 20-year period (1998–2018). Within this cohort, 400,667 incident cases of generalised anxiety (diagnoses or symptoms) were identified.
• Data Example: A data sample for one patient would look like a timeline of records, including their date of birth, gender, date of registration with the practice, and a sequence of dated entries with Read codes (e.g., 'Anxiousness - symptom') and prescription codes (e.g., for 'Sertraline 50mg tablets').
• Rationale for Choice: The THIN database was chosen for its large scale, its representativeness of the UK population, and its longitudinal nature, which allows for the tracking of patients over time to accurately calculate incidence rates and study long-term trends.

5.2. Evaluation Metrics

The primary evaluation metrics used to quantify the trends were:

5.2.1. Incidence Rate

• Conceptual Definition: The incidence rate measures the number of new cases of a condition that appear in a population over a specific time period. It quantifies the risk of developing the condition for individuals in that population. In this study, it is expressed as the number of new GAD diagnoses per 1,000 person-years at risk.
• Mathematical Formula: $ \text{Incidence Rate} = \left( \frac{\text{Number of new cases during a period}}{\text{Total person-time at risk during that period}} \right) \times 1000 $
• Symbol Explanation:
  • Number of new cases: The count of patients who received their first recorded GAD diagnosis or symptom code within the specified year.
  • Total person-time at risk: The sum of time (in years) that all individuals in the cohort were followed and remained disease-free (at risk). This is the PYAR.

5.2.2. Rate Difference

• Conceptual Definition: This metric measures the absolute change in the incidence rate between two points in time (in this case, between 2014 and 2018). A positive value indicates an increase in incidence, while a negative value indicates a decrease. It is used to quantify the magnitude of the recent trend.
• Mathematical Formula: $ \text{Rate Difference} = \text{Incidence Rate}{2018} - \text{Incidence Rate}{2014} $
• Symbol Explanation:
  • $\text{Incidence Rate}_{2018}$: The model-estimated incidence rate for the year 2018.
  • $\text{Incidence Rate}_{2014}$: The model-estimated incidence rate for the year 2014. The study reports this difference along with a 95% confidence interval to indicate the statistical significance of the change.

5.3. Baselines

This study is an epidemiological trend analysis, so it does not use "baseline models" in the machine learning sense. Instead, the comparisons are made:

• Over Time: The incidence rate in recent years (2014–2018) is compared to rates in earlier years. The year 2014 serves as a baseline for measuring the most recent changes.
• Across Demographic Groups: Rates are stratified and compared across different age groups (e.g., 18–24, 25–34, etc.) and between men and women. Each group effectively serves as a comparison for the others.

6. Results & Analysis

6.1. Core Results Analysis

The experimental results reveal several clear and significant trends in the primary care management of generalised anxiety in the UK.

6.1.1. Trend in Generalised Anxiety (GAD) Recording

The central finding is a sharp and recent increase in recorded GAD.

As shown in Figure 1 from the paper, the incidence rates of GAD were relatively stable or slightly increasing for most groups until around 2014. After 2014, there is a dramatic upward trend, which is most pronounced in younger age groups.

Figure 1: Incidence of generalised anxiety by age and gender from 1998 to 2018. The left panels show the crude (raw) rates, while the right panels show the model-fitted values, which smooth out year-to-year noise.

• For women aged 18–24, the incidence rate increased from 17.06 per 1,000 PYAR in 2014 to 23.33 per 1,000 PYAR in 2018.

• For men aged 18–24, the rate increased from 8.59 to 11.65 per 1,000 PYAR over the same period.

• In contrast, rates for patients aged 55 and older remained stable or showed slight decreases, indicating the phenomenon is specific to younger cohorts.

  Figure 2(a) visualizes these rate differences between 2014 and 2018, highlighting the magnitude of the change.

  Figure 2: Model-based rate differences between 2014 and 2018. Panel (a) shows the change for generalised anxiety only. Panel (b) shows the change for the composite of anxiety or depression.

The chart clearly shows that the increase in GAD incidence for women aged 18–24 is by far the largest, approximately double the increase seen in the next youngest group of women and in young men. The confidence intervals for these younger groups are well clear of zero, indicating a statistically significant increase. For those aged 55 and over, the confidence intervals overlap with zero, showing no significant change.

6.1.2. Trend in Composite Anxiety and Depression Recording

To investigate if the rise in anxiety was simply due to GPs re-classifying depression as anxiety, the authors analyzed a composite outcome of any anxiety or depression diagnosis.

Figure 3: Incidence of any anxiety or depression diagnosis by age and gender from 1998 to 2018. The pattern of a recent sharp increase in the youngest cohorts is still clearly visible.

The results (Figures 2b and 3) show:

• For young people (18–24), the composite rate also increased significantly, suggesting a genuine rise in mental health consultations, not just a relabeling of diagnoses.

• For women aged 45 and older, the composite rate actually decreased. When viewed alongside the stable GAD rates for this group, this suggests a potential decrease in depression diagnoses for older women.

• For men aged 35 and older, the composite rate remained largely unchanged.

  This sensitivity analysis provides strong evidence that the surge in mental health issues among young adults is a real phenomenon captured in primary care records.

6.1.3. Trends in Pharmacological Treatment

The analysis of prescribing patterns for patients with a new GAD diagnosis reveals a clear shift towards practices aligned with NICE guidelines.

Figure 4: Initial pharmacological treatment within one year of a generalised anxiety diagnosis. The left panel shows "drug-naïve" patients (no psychotropic medication in the year prior), and the right panel shows all patients.

Key findings from Figure 4 include:

• Proportion Treated: About half of drug-naïve patients receive a prescription for an anxiety medication within one year of diagnosis.

• Decline of Benzodiazepines: The prescription rate for benzodiazepines has fallen steadily and dramatically, especially after 2008. By 2018, they represent a very small fraction of initial treatments.

• Rise of SSRIs: The use of SSRIs has increased consistently over the 20-year period, becoming the most common first-line pharmacological treatment for GAD, in line with NICE recommendations.

• Decline of TCAs: The use of older tricyclic antidepressants (TCAs) has also steadily decreased over time.

  These results strongly suggest that GPs have adapted their prescribing habits in response to evidence and clinical guidelines, favoring safer, more effective long-term treatments over drugs with high addiction potential.

6.2. Ablation Studies / Parameter Analysis

This study is not a machine learning paper, so it does not contain "ablation studies" in the traditional sense. However, the analysis of a composite outcome (any anxiety or depression) serves a similar purpose to a sensitivity analysis. By running the same analysis on this broader definition, the authors "ablated" the effect of potential diagnostic shifting between anxiety and depression. The fact that the core finding of a surge in young people held up in this analysis strengthened the paper's main conclusion.

7. Conclusion & Reflections

7.1. Conclusion Summary

The study concludes that there has been a substantial and recent increase in primary care consultations for generalised anxiety and depression, particularly from 2014 to 2018. This increase is disproportionately concentrated within younger people (aged 18–24) and especially young women. Simultaneously, prescribing practices for GAD have evolved significantly, with a marked decline in benzodiazepine use and a rise in the use of SSRIs, reflecting adherence to modern clinical guidelines. The findings point to a significant and growing public health challenge concerning the mental well-being of young adults in the UK.

7.2. Limitations & Future Work

The authors acknowledge several limitations:

• Underestimation of True Incidence: The study is based on diagnoses recorded in primary care. It does not capture individuals with GAD who do not seek help from their GP, meaning the true incidence in the population is likely higher.

• Ambiguity of "Recording" Trends: An increase in recorded diagnoses could be driven by multiple factors: a true increase in the underlying condition, increased health-seeking behavior by patients (due to reduced stigma), or improved recognition and recording by GPs. The study cannot definitively separate these factors, though the concentration of the trend in a specific demographic makes a simple change in GP behavior less likely to be the sole explanation.

• Scope of Analysis: The analysis was restricted to generalised anxiety and did not examine other anxiety disorders like panic disorder or OCD. It also did not account for comorbid substance misuse, which is often poorly recorded in primary care data.

  The paper does not explicitly outline future work, but the findings strongly imply the need for research to understand the causes behind this surge in anxiety among young people (e.g., the role of social media, economic precarity, academic pressure) and to ensure healthcare services are adequately resourced to meet this growing demand.

7.3. Personal Insights & Critique

This is a high-impact, well-executed piece of epidemiological research with significant public health implications.

Strengths:

• Scale and Data Quality: The use of the THIN database, covering over 6 million patients, provides immense statistical power and generalisability.
• Methodological Rigor: The use of a mixed-effects model to account for clustering by practice and a non-linear time smoother is statistically sound. The sensitivity analysis using a composite outcome was particularly insightful for strengthening the main conclusion.
• Clarity and Importance of Findings: The paper identifies a clear, dramatic, and demographically-specific trend that is of immediate relevance to clinicians, policymakers, and the public.

Critique and Areas for Reflection:

• Interpretation of "Increased Recording": While the authors are careful, there is a risk that the findings are interpreted solely as a negative "mental health crisis." An alternative, more optimistic component of the story could be that reduced stigma and greater mental health literacy are empowering more young people to seek help, which is a positive development. The truth is likely a combination of both.
• Causality is Speculative: The discussion section links the findings to potential drivers like the 2008 economic downturn, austerity, and social media use. While these are plausible and important hypotheses for future research, the study's design is descriptive and cannot establish causality.
• The "Why" Remains Unanswered: The paper masterfully answers "what," "who," and "when." The crucial question of "why" this is happening remains open and is the most urgent follow-up question.

Inspiration and Application: This study is a powerful demonstration of how large-scale, real-world data from electronic health records can be harnessed for public health surveillance. The methodology serves as a blueprint for tracking other chronic and mental health conditions. The findings themselves act as an early warning system, highlighting an urgent need to bolster mental health support systems for young adults and to investigate the societal factors driving this trend. It underscores that mental health is not static and that continuous monitoring is essential for responsive and effective healthcare planning.