Abstract
Background: The prevalence of diagnosed mental disorders has risen substantially over the past two decades, prompting debate about whether this reflects a true increase in psychopathology, an artifact of measurement and incentive structures, or both. Discussions of this trend are frequently reduced to a single explanatory factor — commonly, pharmaceutical industry influence — while other well-evidenced contributors receive comparatively little attention.
Objective: This review proposes a four-category framework for organizing the causes of rising mental health diagnoses — biological/neurodevelopmental, psychological/relational, social/environmental, and measurement/systemic — and evaluates the evidentiary basis for each. It further considers how these categories interact and translates the resulting synthesis into pragmatic, category-specific recommendations.
Methods: This is a narrative synthesis drawing on peer-reviewed epidemiological studies, meta-analyses, global burden of disease data, and health-policy research identified through targeted literature search. It is not a systematic review and does not apply formal quality-scoring or meta-analytic pooling across categories.
Results: Evidence supports genuine contributions from all four categories. Heritability of major psychiatric disorders ranges from approximately 0.32 for major depressive disorder to 0.67 for schizophrenia. Relationship-based stressors, particularly intimate partner violence, show consistent, dose-responsive associations with depression and anxiety. Social and environmental exposures — sleep disruption, digital engagement, economic precarity, and the COVID-19 pandemic — show convergent, if largely correlational, evidence of impact. Measurement and systemic factors, including diagnostic criteria expansion, reimbursement-linked incentives, documented financial ties between diagnostic-manual panels and pharmaceutical firms, and direct-to-consumer advertising's demonstrated causal effect on prescribing, contribute an inflationary component that is distinct from — but compounds — true prevalence change. Population growth alone accounts for the majority of the raw increase in global disorder counts.
Conclusions: Rising diagnosis rates are best explained as the additive and interacting product of genuine increases in risk exposure and measurement-driven inflation, rather than either alone. Effective response requires category-matched interventions: biological and clinical infrastructure investment, relationship- and sleep-focused psychological intervention, social and digital-environment policy, and structural reform of diagnostic governance and pharmaceutical marketing practice.
Keywords: mental health epidemiology; diagnostic inflation; intimate partner violence; heritability; direct-to-consumer advertising; DSM; global burden of disease
1. Introduction
Rates of diagnosed mental disorder have risen across most high-income countries over the past two decades, a trend that has intensified public debate about its causes (Nature/Molecular Psychiatry, 2026). One common explanation attributes the rise substantially to the commercial interests of the pharmaceutical industry, operating through an analogy to the statistical normal distribution: if psychological health were distributed like IQ, the argument goes, only a small tail of the population should qualify as impaired, just as only a small tail qualifies as intellectually gifted. This analogy is intuitively appealing but does not survive close inspection. IQ is a single, deliberately unidimensional construct, psychometrically normed to produce a Gaussian distribution. Mental health is not a single variable; it comprises dozens of partially independent dimensions — mood regulation, anxiety proneness, psychotic symptomatology, personality functioning, substance use — most of which are not normed to any particular distributional shape, and diagnosis typically reflects a functional-impairment threshold rather than a position in a ranked distribution. The analogy, while rhetorically useful, cannot substitute for an evidentiary account of what is actually driving the trend.
A more productive approach treats the question empirically and structurally. This review organizes the causal landscape into four categories — biological/neurodevelopmental, psychological/relational, social/environmental, and measurement/systemic — and evaluates the evidence for each. The measurement/systemic category is where pharmaceutical-industry incentives properly belong; treating it as one mechanism among several, rather than the master explanation, allows its real but bounded contribution to be assessed against comparably rigorous evidence for the other categories, including one that is frequently underweighted in public discussion: relationship and attachment-related stress, particularly intimate partner violence.
2. Methods
This is a narrative literature synthesis rather than a systematic review. Sources were identified through targeted searches of peer-reviewed journals, preprint and open-access repositories, and health-policy literature, prioritizing meta-analyses, large cohort and registry studies, and Global Burden of Disease (GBD) modeling where available. Because the four categories draw on distinct methodological traditions (behavioral genetics, clinical epidemiology, social epidemiology, and health-policy research), no attempt is made to pool effect sizes across categories; instead, each is reported using the effect metric native to its own literature (heritability coefficients, odds ratios, prevalence percentages, or qualitative policy findings), and the discussion section addresses how findings of different types can be integrated into a single interpretive framework.
3. Biological and Neurodevelopmental Factors
Genetic contribution to major psychiatric disorders is well established but highly variable across diagnoses. Registry-based heritability estimates derived from the Danish national population (n > 2.6 million) place the heritability of liability to schizophrenia at 0.67, bipolar disorder at 0.62, and major depressive disorder at 0.32 (Wray & Gottesman, 2012), a pattern broadly consistent with twin-study estimates ranging from roughly 0.35 for major depression to over 0.60 for schizophrenia (Polderman et al., cited in Cai et al.). Multi-condition sibling and molecular-genotype analyses corroborate this ordering, with sibling-based heritability estimates spanning 0.30 for major depression to 0.80 for ADHD, and measured-genotype estimates lower but directionally consistent. This variability matters for the framework proposed here: disorders at the high-heritability end of the spectrum are unlikely to show large true-prevalence swings over a period as short as two decades, whereas disorders with lower heritability and larger environmental variance components — including major depression and many anxiety disorders — are the categories most plausibly influenced by the social and relational factors discussed in Sections 4 and 5.
Neurodevelopmental timing constitutes a second biological contributor. Adolescence is characterized by asynchronous maturation between limbic structures, which mature relatively early, and the prefrontal cortex, which continues developing into the mid-twenties, producing a developmental window of heightened vulnerability to the onset of mood, anxiety, and psychotic disorders (Molecular Psychiatry, 2026). This mismatch does not by itself explain a rising trend, since it is a stable feature of human development, but it may interact multiplicatively with newer environmental exposures — for example, concentrating the mental-health impact of social media use or academic pressure into a developmental period already characterized by heightened neural reactivity to social evaluation.
Substance exposure is a third relevant factor, most extensively studied in relation to cannabis and psychosis. Systematic review evidence indicates that tetrahydrocannabinol (THC) is consistently associated with increased risk of psychotic symptoms and schizophrenia onset, while cannabidiol (CBD) has been investigated for potential antipsychotic and therapeutic properties, producing a genuinely bidirectional evidence base rather than a uniformly harmful one (PMC, 2025). Rising rates of cannabis use and potency in many jurisdictions following legalization are therefore a plausible, if disorder-specific, contributor to rising psychotic-spectrum diagnoses, distinct from the broader mood and anxiety trends driven more heavily by non-biological factors.
Finally, physical-health comorbidity operates bidirectionally with mental health: cardiovascular disease and mental disorders are mutually aggravating, with mental illness complicating cardiovascular risk management and cardiovascular disease elevating psychiatric symptom burden (epidemiological review literature). This comorbidity pathway means that biological and lifestyle-medicine trends unrelated to psychiatry per se — diet, urbanization, sedentary behavior — may indirectly inflate psychiatric caseloads.
4. Psychological and Relational Factors
Relationship-based stress, and intimate partner violence (IPV) specifically, constitutes one of the most consistently documented psychological contributors to depression and anxiety, and is frequently underweighted relative to biological and social-media explanations in public discussion. A meta-analysis of longitudinal studies encompassing more than 36,000 participants found that IPV exposure was associated with substantially increased incident depression, with weighted mean odds ratios for the association between violence and mental health problems ranging from 3.55 to 5.62 across outcomes, and a dose-response relationship between violence severity and both depression and PTSD symptoms (Golding, 1999; Devries et al., 2013). More recent systematic reviews focused specifically on psychological (non-physical) IPV, synthesizing 194 studies, found strong associations with PTSD, depression, and anxiety, with coercive control most strongly linked to PTSD and emotional or isolating abuse most strongly linked to depression (Dokkedahl et al., 2022). Contemporary population-based studies continue to confirm this pattern across diverse settings, including nationally representative surveys from Zambia, Mozambique, and Lesotho, and meta-analyses in sexual-minority populations reporting pooled adjusted odds ratios of 1.71 for depressive symptoms and 1.89 for anxiety among men exposed to IPV (Hong et al.). The umbrella review evidence base, while methodologically variable in quality, is consistent in direction: IPV is a genuine and substantial driver of common mental disorder, not a peripheral one (Pubmed umbrella review, 2025).
Global Burden of Disease decomposition analyses independently corroborate the relational-risk category at a population level, identifying childhood sexual abuse, bullying, and intimate partner violence among the leading specific risk factors contributing to the burden of anxiety and depressive disorders in individuals under 24 (Frontiers/GBD 2021 analysis). This places relationship and attachment-related trauma on comparable epidemiological footing with more commonly cited social factors such as academic pressure, despite receiving considerably less attention in popular discourse about the causes of rising diagnosis rates.
Sleep disruption functions as a second psychological/behavioral factor of particular importance because it operates transdiagnostically — that is, as a risk factor shared across depression, anxiety, and other conditions rather than being specific to any one diagnosis. Recent reviews of youth mental health trends identify declining sleep quality as a central, cross-cutting contributor to rising rates of multiple disorders simultaneously, which makes it a high-leverage target for intervention precisely because it does not require disorder-specific programming (Molecular Psychiatry, 2026; Youth Futures Foundation, 2025).
5. Social and Environmental Factors
Digital engagement and its correlates — cyberbullying, social comparison, and sleep displacement — are among the most frequently cited environmental contributors to the recent rise in youth anxiety and depression, though the literature is more consistently correlational than causal, and some evidence points to sleep disruption as a partial mediator of digital engagement's effects rather than a fully independent pathway (Molecular Psychiatry, 2026).
Economic and academic pressure represents a second social contributor with substantial supporting evidence. Cross-sectional and repeated cross-sectional studies of economically disadvantaged young adults document a marked worsening of depressive and anxiety symptoms, alongside rising suicidal ideation, between 2020 and 2024–2025, with financial hardship, social isolation, and unequal access to psychological support identified as structural drivers (Macalli et al., 2025). Family instability, parental unemployment, and household financial strain show similar associations in adolescent populations more broadly.
The COVID-19 pandemic constitutes the clearest example of an acute, dateable shock producing a measurable step-change in mental health indicators, distinguishable from slower-moving structural trends. Cross-national analysis reported a 25 to 30 percent rise in youth depression and anxiety across eleven high-income countries during the pandemic period, while a meta-analysis of 29 studies found a twofold increase in adolescent anxiety risk during lockdown periods specifically (Racine et al., 2021, cited in Frontiers/GBD 2021 analysis). Importantly, GBD decomposition analysis finds that population growth — not increased per-capita risk — accounts for the large majority of the raw increase in global mental-disorder prevalence and disability-adjusted life years (84.9 percent and 57.9 percent respectively), a finding that substantially qualifies any narrative built solely on rising individual-level risk (Frontiers, GBD 2021 study).
6. Measurement and Systemic Artifacts
This category captures mechanisms that inflate diagnosed prevalence independent of any change in true underlying psychopathology, and it is where pharmaceutical-industry incentives are most defensibly situated. Three specific, well-documented mechanisms are relevant.
6.1 Diagnostic criteria expansion and industry ties in manual revision
Successive editions of the Diagnostic and Statistical Manual of Mental Disorders (DSM) have broadened diagnostic thresholds for numerous conditions, a process that mechanically increases the pool of people eligible for diagnosis without requiring any change in population health. Independent analyses of financial disclosures have found that a majority of DSM panel members reported financial relationships with pharmaceutical companies at rates that did not meaningfully decline between DSM-IV and DSM-5 despite new disclosure requirements, with 57 percent of DSM-5 work-group members reporting industry ties in one analysis (Cosgrove & Krimsky, 2012) and a subsequent BMJ-published study finding DSM-5 and DSM-5-TR panel members collectively received approximately $14.2 million in industry funding, concentrated most heavily in panels covering conditions for which pharmacological treatment is the first-line intervention. A related analysis of clinical-trial principal investigators found that 61 percent of DSM Task Force members and 27 percent of Work Group members reported financial conflicts of interest with manufacturers of drugs under review, with a subset holding simultaneous decision-making authority over the diagnostic revision itself (Cosgrove et al., cited study). These findings do not establish that any specific diagnostic threshold was altered for commercial rather than clinical reasons, but they establish a documented, structural conflict-of-interest pathway through which such influence could operate, and they support disclosure and governance reform as a legitimate, evidence-based policy target.
6.2 Reimbursement-linked diagnostic incentives
Independent of pharmaceutical influence, health-financing structures that require a billable diagnostic code to authorize reimbursement create incentives for clinicians to assign a diagnosis in borderline or subthreshold presentations, contributing to diagnostic inflation through an entirely separate administrative mechanism (Psychology Today/health-policy commentary, 2025).
6.3 Direct-to-consumer pharmaceutical advertising
Direct-to-consumer (DTC) prescription drug advertising, legal in essentially only the United States and New Zealand among developed nations, has a demonstrated causal effect on prescribing behavior. A randomized controlled trial published in JAMA used standardized patients (actors) presenting with either major depressive disorder or subclinical adjustment-disorder symptoms to unannounced physicians, varying whether the patient made a brand-specific medication request, a general request, or no request. Among patients presenting with adjustment-disorder symptoms — for whom antidepressant medication is not indicated by clinical guidelines — a brand-specific request tied to an advertisement resulted in a prescription 55 percent of the time, compared to 10 percent when no request was made (Kravitz et al., 2005). Evidence synthesis grading this literature concludes that DTC advertising is associated with both appropriate increases in treatment-guideline fidelity for genuine major depression and inappropriate increases in prescribing for subclinical presentations, indicating a real but bidirectional effect rather than a uniformly harmful one; a separate economic analysis estimated that DTC advertising accounted for roughly 12 percent of the growth in prescription drug sales between 1999 and 2000 (Donohue, cited KFF-published analysis).
Together, these three mechanisms constitute a coherent measurement/systemic category distinct from genuine changes in population mental health: they operate through diagnostic-manual governance, health-system financing design, and marketing-driven demand generation respectively, and each has an identifiable, targetable policy lever, addressed in Section 8.
7. How the Categories Interact
The four categories are not independent; each is capable of amplifying or masking the others. Biological vulnerability establishes a baseline sensitivity to environmental stressors — the adolescent neurodevelopmental mismatch described in Section 3 plausibly concentrates the psychological impact of both relational stress (Section 4) and digital/social exposures (Section 5) into a developmentally sensitive window, producing a larger population-level effect than either pathway would generate alone. Measurement and systemic factors (Section 6) then interact multiplicatively with genuine increases: a population genuinely experiencing more relational and economic stress, evaluated by a diagnostic apparatus with progressively lower thresholds and reimbursement structures that reward diagnosis, will show a diagnosed-prevalence increase larger than the true-prevalence increase alone would produce. This compounding relationship explains why single-cause narratives — pharmaceutical industry interest alone, or social media alone, or population growth alone — each capture a real but partial mechanism, and why the GBD finding that population growth explains the majority of the raw prevalence increase does not contradict, but rather sits alongside, genuine per-capita increases in specific stressor-linked conditions such as anxiety and eating disorders (Frontiers, 2025).
8. Pragmatic Recommendations
Recommendations are organized by category to ensure that interventions are matched to the mechanism they target; a single generic response (e.g., "expand access to therapy") is necessary but insufficient given the structural, not merely clinical, nature of several contributing factors.
8.1 Biological/neurodevelopmental
● Prioritize early neurodevelopmental screening during adolescence, when the limbic-prefrontal maturation mismatch confers heightened risk, rather than waiting for full diagnostic threshold presentation.
● Integrate psychiatric and cardiometabolic care given documented bidirectional comorbidity, rather than treating them in separate clinical silos.
● Monitor cannabis potency and psychosis-risk indicators specifically in jurisdictions undergoing legalization, since this is a disorder-specific rather than general risk pathway.
8.2 Psychological/relational
● Fund and scale routine IPV screening in primary care and mental health intake settings, given odds ratios for depression and anxiety comparable to or exceeding many biological risk factors currently screened as a matter of course.
● Treat sleep intervention (e.g., structured sleep-hygiene programs, cognitive behavioral therapy for insomnia) as a transdiagnostic, low-cost, high-leverage population intervention rather than a condition-specific afterthought.
● Expand attachment-informed early relationship and family intervention programs, particularly for adolescents and young adults, given the concentration of relational risk factors in this developmental window.
8.3 Social/environmental
● Address economic precarity directly (income support, housing stability, student financial aid) as a mental-health intervention, not merely an economic one, given the strength of financial-hardship associations in recent cohort data.
● Pursue evidence-based, proportionate regulation of youth social-media design features implicated in sleep displacement and social comparison, while avoiding overstatement of causality given the largely correlational evidence base.
● Build pandemic-style shock-response protocols into public mental health infrastructure, given the demonstrated speed and magnitude of COVID-19's measurable impact.
8.4 Measurement/systemic
● Require independent, industry-free composition of diagnostic-manual revision panels, particularly for disorder categories where pharmacological treatment is first-line, extending beyond disclosure toward structural exclusion of conflicted members from decision-making authority.
● Reform reimbursement structures so that access to care does not require assignment of a formal diagnostic code, reducing the administrative incentive toward diagnostic inflation in borderline presentations.
● Restrict or more tightly regulate direct-to-consumer pharmaceutical advertising, following the near-universal international norm outside the United States and New Zealand, given the randomized-trial-level causal evidence of inappropriate prescribing it generates.
● Publish routine, disaggregated prevalence-vs-incidence reporting in national mental health statistics so that policymakers and the public can distinguish true increases from measurement-driven ones.
9. Limitations
This review is a narrative rather than systematic synthesis and does not apply formal risk-of-bias scoring or meta-analytic pooling; effect sizes cited are drawn directly from the primary or secondary sources referenced and vary considerably in methodological rigor, particularly in the social/environmental category, where much of the evidence remains correlational. The four-category framework is an organizational heuristic rather than an empirically derived taxonomy, and some phenomena (e.g., sleep disruption) plausibly span more than one category. Finally, most of the underlying epidemiological and policy literature is drawn from high-income countries, particularly the United States and Western Europe, and the relative weight of these categories may differ substantially in other health-system and cultural contexts.
10. Conclusion
The rise in diagnosed mental disorder over the past two decades is not adequately explained by any single mechanism. Genuine biological vulnerability, relational and attachment-based trauma, social and economic stressors, and measurement-driven inflation each carry independent, evidence-supported weight, and they compound rather than compete with one another. Framing the discussion around a single villain — whether pharmaceutical industry interest or social media — forecloses the category-matched policy response that the evidence actually supports. A structural response requires simultaneous investment in early biological and psychological intervention, serious attention to relationship-based trauma as a mainstream rather than peripheral risk factor, social and economic policy addressing precarity, and governance reform of the diagnostic and marketing systems that shape how, and how often, distress becomes a diagnosis.
References
Cosgrove, L., & Krimsky, S. (2012). A comparison of DSM-IV and DSM-5 panel members' financial associations with industry: A pernicious problem persists. PLoS Medicine, 9(3), e1001190. https://doi.org/10.1371/journal.pmed.1001190
Devries, K. M., Mak, J. Y., Bacchus, L. J., Child, J. C., Falder, G., Petzold, M., Astbury, J., & Watts, C. H. (2013). Intimate partner violence and incident depressive symptoms and suicide attempts: A systematic review of longitudinal studies. PLoS Medicine, 10(5), e1001439.
Dokkedahl, S. B., Kirubakaran, R., Bech-Hansen, D., Kristensen, T. R., & Elklit, A. (2022). The psychological subtype of intimate partner violence and its effect on mental health: A systematic review with meta-analyses. Systematic Reviews, 11, 163.
Golding, J. M. (1999). Intimate partner violence as a risk factor for mental disorders: A meta-analysis. Journal of Family Violence, 14(2), 99-132.
Hong, C., Wang, Y., Wang, Y., Pushpanadh, S., Stephenson, R., Keum, B. T. H., Goldbach, J. T., Graham, S. M., & Holloway, I. W. Associations between intimate partner violence and mental health outcomes among sexual minority men: A systematic review and meta-analysis.
Kravitz, R. L., Epstein, R. M., Feldman, M. D., Franz, C. E., Azari, R., Wilkes, M. S., Hinton, L., & Franks, P. (2005). Influence of patients' requests for direct-to-consumer advertised antidepressants: A randomized controlled trial. JAMA, 293(16), 1995-2002.
Macalli, M., et al. (2025). Post-pandemic changes in anxiety and depression symptom networks among socioeconomically disadvantaged young adults: A repeated cross-sectional study.
Molecular Psychiatry. (2026). Young minds in distress: Exploring the global rise in youth mental health diagnoses. Nature Publishing Group.
Frontiers in Public Health. (2025). Global burden and trends of major mental disorders in individuals under 24 years of age from 1990 to 2021, with projections to 2050: Insights from the Global Burden of Disease Study 2021.
Wray, N. R., & Gottesman, I. I. (2012). Using summary data from the Danish national registers to estimate heritabilities for schizophrenia, bipolar disorder, and major depressive disorder. Frontiers in Genetics, 3, 118.
Youth Futures Foundation. (2025). Understanding drivers of recent trends in young people's mental health.
Additional sources: PMC systematic review of cannabis use and schizophrenia (2025); Psychology Today, Should We Be Concerned About the Trends in Mental Illness? (2025); Donohue, J. Effects of direct-to-consumer advertising on medication choice: The case of antidepressants; community.the-hospitalist.org, Does DTC advertising affect physician prescribing habits?