RESEARCH DOSSIER
Developmental Psychopathology and Clinical Trajectories
Abstract. Developmental psychopathology examines the emergence, organization, continuity, transformation, remission, recurrence, and consequences of mental disorders as people develop within changing biological and social contexts. This integrative research monograph analyzes the conceptual history of the field, age-related incidence and diagnosis, homotypic and heterotypic continuity, dynamic risk processes, clinical course heterogeneity, prospective cohort evidence, and methods for estimating transitions and causal pathways. Landmark investigations from the Dunedin and Great Smoky Mountains cohorts are examined alongside cross-national epidemiology, dimensional and transdiagnostic models, trajectory-analysis methodology, current classification guidance, and recent youth follow-up studies. The central synthesis is that a clinical diagnosis observed at one age represents a time-localized state within a potentially changing developmental course, and that symptom persistence, functional impairment, diagnostic succession, and recovery require separate measurement. The evidence supports longitudinal, multi-informant and developmentally sensitive assessment, while substantial uncertainties remain concerning causal mechanisms, transportability, the meaning of statistical trajectory classes, and durable intervention effects. This is a critical narrative synthesis of identified scientific literature, rather than a preregistered systematic review, meta-analysis, or original clinical investigation.
Keywords: developmental psychopathology; clinical trajectories; age of onset; developmental transitions; homotypic continuity; heterotypic continuity; prospective cohort; diagnostic succession; remission and recurrence; longitudinal epidemiology; mental health; developmental cascades; p factor; survival analysis; clinical staging; measurement invariance
The Research Object and Its Boundaries
Developmental psychopathology studies how clinically significant patterns of emotion, cognition, behavior, and interpersonal functioning arise and change during development. Its central unit of explanation is a developing person observed over time in a changing environment. A clinical trajectory includes the timing of first symptoms, the emergence of clinically significant impairment, movement among symptom states and diagnoses, periods of recovery, possible recurrence, and consequences for later participation and functioning. The field addresses both individual continuity and population-level variation. A trajectory is therefore a temporally ordered set of observations interpreted within a defined measurement and diagnostic framework.
This dossier concentrates on the developmental course of psychopathology: clinical onset, continuation, substitution, transformation, remission, recurrence, and the measurement of those processes. It includes risk and protective mechanisms insofar as they explain clinical transitions. A separate dossier on early adversity examines exposure types and causal consequences in depth; a dossier on genetic and environmental liability examines familial and genomic architectures; and the broader Developmental Trajectories dossier studies changes in psychological capacities regardless of clinical status. The distinctive question here is when and how a developing person crosses between clinically meaningful states and what can be inferred from those changes.
The field's modern conceptual architecture was articulated in Cicchetti, D. (1984) and critically elaborated by Rutter, M., & Sroufe, L. A. (2000). Their work treats continuity, discontinuity, causal processes, and developmentally organized functioning as connected problems. A classification obtained from a single encounter has clinical value, yet its developmental interpretation requires observations before and after that encounter, information about the surrounding ecology, and explicit assumptions about change.
Evidence Review and Interpretive Method
The method of this monograph is critical integrative narrative review. The evidence corpus comprises classic theoretical formulations, prospective population cohorts, clinically recruited longitudinal samples, diagnostic epidemiology, methodological investigations, a recent meta-analysis, and official clinical guidance. The selection prioritizes studies that can identify change within persons, distinguish ages and transitions, measure disorder histories, or evaluate the methodological validity of trajectory claims. It includes studies capable of challenging broad continuity narratives. Bibliographic records and direct URLs are documented in References. No exhaustive database search, prospective protocol, double-screening process, or quantitative pooling was conducted for this monograph.
Evidence is interpreted according to design. A birth cohort can establish temporal ordering between measured childhood exposures and adult outcomes but does not by itself establish that an exposure caused the outcome. Repeated diagnostic interviews can estimate observed transitions under the chosen criteria, whereas a retrospective lifetime survey relies on recall and is susceptible to forgotten episodes. Clinical recruitment offers rich symptom measurement but typically changes the distribution of severity relative to the community. A randomized intervention can identify the effects of a particular offered treatment under its trial conditions without automatically demonstrating prevention of disorders decades later.
A recurrent distinction guides the synthesis: description concerns what was observed; prediction concerns conditional probability in a specified population; explanation proposes mechanisms; and causal inference requires defensible counterfactual assumptions or random assignment. Hernán, M. A., & Robins, J. M. (2016) demonstrate why causal questions in observational data benefit from specifying the intervention and target trial. These distinctions prevent a correlated early marker from being mistaken for an inevitable developmental cause.
The Historical Formation of Developmental Psychopathology
The field emerged through the convergence of developmental psychology, child psychiatry, clinical psychology, attachment research, epidemiology, and developmental biology. Earlier clinical traditions often placed symptoms within age-specific categories, while developmental investigators asked how competence, relationships, regulation, and vulnerability evolved. The intellectual integration redirected attention toward the developmental production of clinical outcomes and toward ordinary adaptive systems whose disturbance could contribute to disorder. Cicchetti, D. (1984) is an explicit early statement of the emerging field, rather than the beginning of every contributing tradition.
An important historical movement was from single-factor accounts toward organized processes spanning person and environment. The Minnesota longitudinal tradition, documented by Sroufe, L. A., Egeland, B., Carlson, E. A., & Collins, W. A. (2005), studied patterns of adaptation from childhood into adulthood while embedding behavior within relationships and changing social opportunities. Rutter, M., & Sroufe, L. A. (2000) subsequently articulated unresolved challenges concerning heterotypic continuity, comorbidity, measurement, sex differences, nature–nurture relations, and treatment mechanisms. These remain live scientific questions.
Longitudinal population studies transformed the historical program into empirically testable course descriptions. The New Zealand Dunedin Study and the US Great Smoky Mountains Study repeatedly assessed psychiatric functioning rather than relying exclusively on a single adult recollection. Advances in latent-variable analysis, event-history models, and clinical interviewing supplied methods for examining age-specific changes. The historical significance of a model, however, is different from its contemporary evidential status: a foundational taxonomy remains a hypothesis whose calibration and generalizability require evaluation against later cohorts.
Developmental Principles: Equifinality and Multifinality
Developmental reasoning begins with the possibility that multiple pathways yield similar outcomes. Equifinality describes convergence on a comparable clinical phenotype from different developmental histories. Multifinality describes divergence of later outcomes among individuals who share an early exposure, symptom, or vulnerability. Cicchetti, D., & Rogosch, F. A. (1996) established these principles as central organizing ideas. They are useful because a diagnosis at a particular age neither specifies a unique causal history nor determines a single future.
A child with persistent irritability, for example, may later show a range of internalizing, externalizing, or relatively compensated trajectories, depending on other features and contexts. Conversely, adults diagnosed with depression may reach that clinical outcome through diverse constellations of genetic vulnerability, relational changes, recurrent symptoms, environmental stress, and age-linked developmental transitions. These examples express the logical structure of the concepts; the probability of any specific pathway must be estimated from data rather than inferred from the principle alone.
Developmental change also involves reorganization: an earlier functional difficulty may manifest differently when the tasks demanded by the social environment change. Problems with emotion regulation may become visible as school difficulties, peer conflict, substance-related behavior, or later relational instability without preserving a single symptom label. Developmental mechanisms therefore have to specify what persists—behavioral appearance, regulatory process, underlying liability, impairment, or exposure—not simply say that 'the disorder continued.'
Defining Clinical Course and Its Distinct States
Clinical course is a longitudinal construct built from repeated classifications of symptoms, impairment, duration, and context. Onset may refer to the first subthreshold symptom, first threshold-level episode, first documented diagnosis, first treatment encounter, or the beginning of functional disruption. These dates often differ. An analysis that calls an administrative first contact the onset of illness changes its estimand and may systematically delay the apparent age of onset among people who have less access to care.
Persistence refers to continuing symptoms or disorder across a specified observation interval. Chronicity commonly implies extended duration, but its threshold depends on diagnosis and study protocol. Remission is reduction below a defined symptom threshold; recovery is often distinguished by longer duration or restoration of functioning. Recurrence concerns return after remission, whereas relapse may describe symptom deterioration during a continuing episode or early recovery. These words cannot be interpreted consistently without prespecified observation periods and operational definitions.
Diagnostic transitions add a separate dimension. Someone can remain impaired while moving from one disorder category to another, or can cease to meet diagnostic criteria while retaining important social or cognitive consequences. A valid trajectory analysis therefore keeps at least four dimensions apart: diagnostic state, symptom severity, functional impairment, and exposure to care. Caspi, A., Houts, R. M., Ambler, A., et al. (2020) provide a powerful demonstration of changing diagnostic histories across decades.
Homotypic and Heterotypic Continuity
Homotypic continuity means that the same disorder category or comparable symptom pattern is observed across time. Heterotypic continuity refers to temporal association between different disorder categories or symptom presentations. Both are conditional constructs: their magnitude changes when age, assessment instrument, diagnostic thresholds, informants, interval length, and baseline comorbidity change. Developmental continuity may be detectable even when repeated diagnoses differ, but the interpretation then depends on whether a common mechanism or liability has independent supporting evidence.
In the Dunedin birth cohort, Kim-Cohen, J., Caspi, A., Moffitt, T. E., Harrington, H., Milne, B. J., (2003) evaluated juvenile diagnoses in 1,037 cohort members with assessments extending to age 26. The authors reported that 73.9% of adults meeting their interview-based case definition had received a psychiatric diagnosis before age 18. The estimate belongs to this cohort, definitions, and assessment history; it should not be converted into a universal claim that every adult disorder begins in childhood. The investigators identified both same-category and cross-category antecedents.
A three-cohort analysis by Copeland, W. E., Adair, C. E., Smetanin, P., et al. (2013) used nearly 20,000 observations of 3,722 participants across ages 9 to 30. Childhood or adolescent disorder was associated with substantially higher subsequent disorder odds; both homotypic and heterotypic transitions appeared across categories. These data establish the clinical relevance of an earlier disorder without implying that a particular diagnostic switch is biologically preordained. The most defensible conclusion is probabilistic continuity amid substantial heterogeneity.
Onset, Incidence, Prevalence, and Cumulative Risk
An age-of-onset curve describes the distribution of first recognized episodes among people who develop a particular disorder under a defined assessment procedure. Point prevalence, period prevalence, incidence, lifetime prevalence, and projected morbid risk answer different questions. Incidence requires identifying new cases within a risk set; retrospective age-of-onset distributions reconstruct reported first events; cumulative risk models the probability of developing a disorder by a specified age. These measures should remain distinct when describing adolescence as an interval of elevated clinical transition.
In the US National Comorbidity Survey Replication, Kessler, R. C., Berglund, P., Demler, O., Jin, R., Merikangas, K. R., & (2005) used diagnostic interviews among 9,282 adult respondents to describe lifetime prevalence and onset distributions of selected DSM-IV disorder groups. The methods were well suited to broad adult epidemiology but relied on retrospective reporting. As such, the analysis indicates age-patterned disorder histories as recalled in adulthood, not direct annual clinical observation during childhood.
The cross-national World Mental Health analysis by McGrath and colleagues (2023)00193-1) included 156,331 adult respondents across 29 countries. The estimated median age of first onset among the 13 surveyed disorders was 19 years for male and 20 for female respondents, with the conditional probability of first onset peaking near age 15. The design extends geographic coverage while retaining dependence on retrospective interviews, defined disorder coverage, and modeled morbid-risk assumptions. It supports sustained service capacity across young developmental periods without making any specific age an inevitable disease threshold.
Childhood, Adolescence, and Emerging Adulthood
Clinical expression changes when language, executive capacities, peer relationships, schooling, autonomy, and identity tasks change. A preschool child may communicate distress primarily through behavior and bodily complaints, whereas an adolescent may describe complex cognitions, self-evaluation, mood states, or obsessive thoughts. The apparent developmental rise of a disorder can reflect real changes in liability, changing exposure conditions, improved capacity to report symptoms, revised thresholds, or assessment methods calibrated to an older age.
Adolescence brings simultaneous changes in educational placement, peer ecology, sleep organization, puberty, reward-related behavior, and opportunities for substance exposure. These contextual changes can influence risk and can also make existing difficulties newly impairing. Costello, E. J., Copeland, W., & Angold, A. (2011) examined trends across the adolescent years and emphasized the importance of age-specific patterning. An analytic account must examine the relative timing of transitions instead of treating all change as a uniform developmental effect.
Emerging adulthood introduces new role demands, including employment, independent living, higher education, partnerships, and changing access to services. An adolescent diagnosis can remit while later social or affective impairment remains; a new adult presentation may arise without earlier clinically recorded symptoms. The transition between pediatric and adult care is also an administrative event that can create apparent discontinuities in records. Follow-up studies need harmonized assessment across those service boundaries.
Internalizing Pathways: Anxiety and Depression
Internalizing psychopathology encompasses varied symptoms of fear, anxiety, low mood, anhedonia, and associated cognitive and bodily distress. Anxiety presentations can begin early and change with cognitive maturity, while many depressive presentations become more common during adolescence. These are descriptive population tendencies, not a claim that any individual path is fixed. Differences in age, sex, stress exposure, access to treatment, and informant reports materially affect observed rates.
The Copeland, W. E., Shanahan, L., Costello, E. J., & Angold, A. (2009) found patterns of continuity and switching between childhood and young-adult disorders. The importance of a childhood anxiety diagnosis therefore extends beyond predicting an identical future diagnosis; it can indicate future vulnerability in a broader internalizing domain. At the same time, adolescent depression may emerge in people without earlier diagnosable depression, and many affected youths recover from particular episodes.
When analyzing the course of depression, researchers should distinguish first incidence, recurrent episodes, residual symptoms, concurrent anxiety, and long-term functional burden. The international National Institute for Health and Care Excellence. (2019) addresses depression in children and young people and emphasizes developmentally appropriate assessment and treatment planning. Clinical guidance is evidence-informed policy within its jurisdiction, while the long-run causal effects of every early treatment choice require separate experimental follow-up. A single observed improvement is not an estimate of lifelong prevention.
Externalizing Pathways and Antisocial Development
Externalizing trajectories include problems involving aggression, disruptive behavior, disinhibition, oppositionality, and rule violations. Their developmental organization is heterogeneous: behavior that is rare and serious at one age may be common or less diagnostically informative at another. The distinction between early-onset persistent problems and adolescent-limited conduct was prominently formalized by Moffitt, T. E. (1993). This taxonomy supplied hypotheses about developmental timing, social contexts, neuropsychological vulnerabilities, and cumulative transactions.
The historical model does not justify assigning an individual child to a permanent life-course type on the basis of an early observation. Persistent antisocial outcomes vary by measurement definitions, social opportunities, interventions, and cohort characteristics; adolescence-limited behavior can still entail consequential harm or impairment. The model's value lies in proposing contrasts that can be empirically checked with repeated assessments rather than in turning statistical differences into immutable personal categories.
Copeland, W. E., Adair, C. E., Smetanin, P., et al. (2013) observed substantial continuity for behavioral disorders, particularly attention-deficit/hyperactivity disorder, across examined periods. Yet externalizing symptom clusters may change with age and may be associated with later substance-related disorders. Developmental research must measure conduct outcomes, executive functioning, environmental opportunity, and service exposure separately, while preserving the distinction between population prediction and causal mechanisms.
Neurodevelopmental Conditions and Longitudinal Clinical Needs
Neurodevelopmental disorders are often recognized during early development but may first become clinically evident when environmental demands exceed available supports. Attention-deficit/hyperactivity disorder, autism, intellectual developmental conditions, and developmental language difficulties involve different diagnostic and functional architectures. For this dossier, their relevance lies in timing of recognition, persistence or changing expression, comorbidity, and clinical transitions; their full diagnostic science belongs in condition-specific research dossiers.
The same symptom score can signify different challenges in preschool, secondary school, and adulthood because required tasks, social contexts, and available accommodations differ. A child's later improvement may reflect learning, changing environments, compensatory strategies, medication or therapy, measurement changes, or a combination. A delayed diagnosis can reflect earlier missed recognition rather than new biological onset. These possibilities should be distinguished with developmental history and functional assessment.
The World Health Organization. (2024) provide a current international diagnostic framework. Classification supplies a common clinical vocabulary; it does not itself establish an individual's causal developmental pathway. Longitudinal comparisons must account for revisions to diagnostic definitions and for changing norms of neurodevelopmental recognition across cohorts and jurisdictions.
Bipolar Spectrum, Psychosis, and Diagnostic Reclassification
Longitudinal work on bipolar spectrum conditions offers a strong example of fluctuating symptom burden and the difference between categorical diagnosis and clinical time course. In a clinically recruited sample, Birmaher, B., Axelson, D., Goldstein, B., et al. (2009) followed 413 youths aged 7–17 at intake for approximately four years. The investigators reconstructed week-by-week symptomatic status using longitudinal interval assessments to study recovery and recurrence. Clinical referral sampling and retrospective interval ratings shape the generalizability and precision of this evidence.
Substantial mood fluctuation within an established diagnosis does not imply that every adolescent fluctuation is bipolar disorder. Developmentally informed assessment differentiates sustained or episodic mood syndromes, background temperament, substance effects, sleep loss, trauma-related arousal, medication effects, and other clinical possibilities. A child or adolescent who presents with depression may receive a different diagnosis later; such reclassification should be understood through repeated assessments and explicit criteria, rather than assumed to reflect diagnostic error at every earlier encounter.
Psychosis-related trajectories require similar caution about broad nonspecific symptoms, rare outcomes, distress, functional impairment, and developmental stage. The long-run course depends on case definition, ascertainment, treatment context, and associated conditions. Clinical staging models can organize questions about early signs and progression, but a putative early marker needs prospective predictive validation before it becomes a justification for high-stakes labeling or intervention.
Comorbidity as a Developmental Process
Comorbidity means that symptoms or disorders meet operational thresholds concurrently or successively, with the relevant time windows stated. Cross-sectional comorbidity and sequential comorbidity have different explanations. Shared etiological liability, correlated exposures, reciprocal reinforcement, overlapping diagnostic criteria, and sequential changes in symptom organization can all produce association between disorder categories. The Angold, A., Costello, E. J., & Erkanli, A. (1999) showed the significance of disorder co-occurrence in childhood and adolescence.
The developmental question asks whether earlier symptoms change the risk of later symptoms beyond shared background factors and baseline comorbidity. For example, anxiety might predict later depression because of common liability, because prolonged avoidance restricts rewarding opportunities, because environmental adversity influences both, or because a combination operates. These explanations require different measurement and analytic strategies. A transition matrix alone cannot identify which mechanism produced the observed conditional probabilities.
In the Dunedin Study, Caspi, A., Houts, R. M., Ambler, A., et al. (2020) found that a person's cumulative psychiatric history frequently included disorders from different categories over decades. This strengthens the case for integrated longitudinal clinical research. It also complicates statements such as 'a childhood condition becomes an adult condition': what continues may be general liability or impairment, while the diagnoses recorded at successive times are distinct.
Dimensional Structure and the General Psychopathology Factor
Dimensional models describe variation in severity and covariance among mental health symptoms, potentially across conventional diagnostic boundaries. The general psychopathology factor, commonly called p, is a statistical construct introduced to summarize common liability or covariance. Caspi, A., Houts, R. M., Belsky, D. W., et al. (2014) analyzed Dunedin longitudinal data and reported that a general factor accounted for substantial common variation beyond broader internalizing, externalizing, and thought-disorder structures. A factor is defined by the measures, model, cohort, and identification constraints used to estimate it.
Caspi, A., & Moffitt, T. E. (2018) explored the implications of p for comorbidity, severity, and developmental psychiatric histories. The construct can integrate observations across changing diagnoses, but it does not by itself identify one biological cause, one disorder, or one effective intervention. Measurement overlap, informant effects, and modeling assumptions can all affect estimated general-factor strength.
A Harris, J. L., Swanson, B., & Petersen, I. T. (2024) assembled 65 studies yielding 110 effect-size estimates and reported that the general factor explained approximately 56% of reliable symptom variance in their selected models, with no significant age moderation across the studied ages. That is an aggregate statement about factor-model covariance, not evidence that every individual follows one common clinical pathway. Future longitudinal work must connect dimensional liabilities with observed state transitions and clinically meaningful functioning.
Developmental Cascades and Reciprocal Transactions
Developmental cascades occur when functioning in one domain contributes over time to changes in another domain, potentially altering subsequent developmental opportunities. For example, school difficulties and peer exclusion may become mutually reinforcing with emotion or behavior problems; these associations can then influence subsequent educational and relational contexts. The term 'cascade' is scientifically useful only when specific directed relationships and appropriate time ordering are demonstrated. A simple correlation among domains measured once is insufficient.
Reciprocal transactions refer to repeated interactions between a developing person's characteristics and surrounding environments. A youth's behavior influences parental, peer, and institutional responses, while those responses can modify later behavior. The Minnesota tradition, summarized in Sroufe, L. A., Egeland, B., Carlson, E. A., & Collins, W. A. (2005), emphasized person–environment organization across developmental periods. Temporal measurement allows hypotheses about accumulating vulnerability to be tested rather than assumed.
Causal analyses are challenging because time-varying exposures can be both outcomes of earlier symptoms and predictors of later symptoms. Standard regression may introduce bias if it conditions on variables that are consequences of prior exposures. A stronger study specifies timing, a directed causal model, confounding assumptions, the target intervention, and sensitivity analyses. Clinical trajectory research gains explanatory power when it connects observed sequential states to independently measured transactional mechanisms.
Early Adversity, Timing, and Clinical Vulnerability
Early adversity is relevant to clinical trajectories because duration, timing, severity, and accumulation of difficult experiences can correlate with later disorder onset and persistence. At the same time, children exposed to comparable adversity display diverse outcomes. This dossier considers adversity as a time-varying potential risk exposure, with full exposure science allocated to the independent Early Adversity and Life-Course Developmental Risk research object. The developmental question concerns whether different exposure sequences help explain later clinical transitions.
A robust analysis must examine familial and community contexts, socioeconomic inequalities, parental mental health, the timing of assessment, and factors that may jointly influence exposure and outcome. Observational associations are susceptible to confounding and measurement error. Earlier clinical symptoms may themselves change exposure likelihood by affecting relationships or school placement. Prospective exposure records and designs using sibling comparisons or other genetically informed methods can improve inference, although no single design eliminates all causal uncertainty.
The inference that an adverse experience raises population-level risk does not authorize prognosis for a particular child. Effective prevention can target modifiable conditions when evidence and proportionality support intervention, but a developmental theory must preserve multifinality and the possibility of adaptive change. The Masten, A. S. (2001) shows why protective processes and ordinary adaptive systems deserve positive, observable definitions within course research.
Resilience, Compensation, and Recovery
Resilience is typically inferred from relatively favorable adaptation despite meaningful adversity or elevated risk. Its operationalization requires defining exposure, outcome, time scale, and the comparison group. It can describe continuing functioning, recovery after a decline, or successful reorganization under altered conditions; these are not interchangeable outcomes. Masten, A. S. (2001) emphasized protective processes embedded in ordinary adaptive systems such as supportive relationships, learning opportunities, and effective caregiving.
Clinical recovery must likewise be measured across several dimensions. A person may no longer meet symptom criteria while experiencing reduced educational participation or ongoing interpersonal challenges. Another person may continue to experience some symptoms but function well with accommodations and effective support. Studies that measure symptoms alone may misrepresent recovery, especially when participants move between life stages with very different everyday demands. A Schaefer, J. D., Caspi, A., Belsky, D. W., et al. (2017) investigated enduring mental health as an outcome across a substantial observation history; this approach shows why one symptom-free interval does not define an entire life course.
Recovery trajectories are not simply mirror images of disorder onset. Subsequent episodes may be influenced by residual symptoms, ongoing stressors, access to care, pharmacological exposures, social support, and evolving individual capacity. A useful course study measures both symptom change and the conditions under which participation improves or deteriorates. Resilience research contributes explanatory variables to clinical trajectories, rather than replacing diagnosis with an imprecise judgment that a person is 'doing well.'
Family, Peers, School, and Social Ecologies
Development occurs within social institutions whose demands and resources change over time. Parent–child relationships, extended families, peers, schools, neighborhoods, online communities, and healthcare systems may influence symptom identification, access to intervention, and ongoing functioning. Such contexts can act as exposures, moderators, mediators, and measurement environments. A diagnosis is often recorded when challenges become visible to caregivers or teachers, which means surveillance and referral patterns can influence observed disorder incidence.
Multi-informant assessment is essential because children may behave differently across contexts and because informants observe different domains. Parent reports may capture persistent home difficulties, teacher reports may detect classroom impairment, and youth self-reports may better reveal private anxiety, hopelessness, and intrusive thoughts. Differences among reports can contain meaningful contextual information rather than merely measurement error.
School transitions and social role changes can alter both risk and the threshold for clinical impairment. The Patalay, P., & Fitzsimons, E. (2018) provides longitudinal evidence concerning change in mental ill-health and well-being from childhood to adolescence. Disentangling environmental transitions from age and cohort differences requires repeated observations, appropriate comparisons, and care with causal interpretations.
Puberty, Neurodevelopment, and Biological Time
Developmental age combines several partly distinct temporal processes: chronological age, pubertal maturation, neural and cognitive development, social transitions, and accumulated exposure history. These processes correlate but vary among individuals. A clinical change near puberty may coincide with hormonal changes, peer relationships, sleep shifts, or school transitions. The temporal overlap generates hypotheses about mechanisms; a physiological explanation requires independent measurement and a design that can distinguish it from social and contextual covariates.
Neural development supports changing capacities for attentional control, emotional regulation, learning, and social prediction. These processes are relevant to symptoms, yet no imaging marker alone identifies an individual's developmental psychiatric trajectory. Neurodevelopmental measurements can help generate mechanistic explanations only when reproducibility, sampling, analytic choices, and clinical prediction are carefully evaluated.
A useful developmental design assesses chronological and pubertal indicators separately and considers the consequences of using age as a proxy for maturity. It also studies differences in developmental tempo rather than assuming identical biological schedules. The World Mental Health findings of McGrath and colleagues (2023)00193-1) show why age of clinical onset matters epidemiologically; they do not identify puberty as the cause of those onset distributions. The distinction is fundamental to interdisciplinary interpretation.
Sex, Gender, Culture, and Social Inequality
The apparent distribution of clinical trajectories can vary by sex-related biology, gendered social experience, cultural meanings, socioeconomic position, and differential pathways to diagnosis. These influences often intersect. A population disparity might arise from differing exposures, expression and recognition of symptoms, care availability, reporting practices, or all of these. Biological and social variables should be specified at the level actually measured rather than collapsed into one explanatory label.
Cross-cultural comparison requires attention to whether symptom scales, diagnostic interviews, language translations, and impairment thresholds function similarly. A category standardized in one national setting may capture experiences differently in another. For this reason the geographic scope of McGrath and colleagues' 29-country analysis (2023)00193-1) is important, but differences among participating countries, survey periods, and interview practices remain methodological considerations.
Disability, migration, displacement, poverty, discrimination, and access to stable services can alter clinical opportunities and observed records. Clinical trajectory research should distinguish changing prevalence from changing documentation and changing service utilization. A scientifically credible account describes populations explicitly, examines potential measurement noninvariance, and avoids transferring a cohort-specific risk estimate indiscriminately to populations omitted or underrepresented in the original study.
Prospective Cohorts and the Dunedin Evidence
The Dunedin Multidisciplinary Health and Development Study has provided unusually extensive repeated assessments across childhood and adulthood. In Kim-Cohen, J., Caspi, A., Moffitt, T. E., Harrington, H., Milne, B. J., (2003), retrospective follow-back conducted within the prospectively assessed cohort indicated that many adult cases had recorded juvenile diagnoses. This is stronger than relying on adult retrospective recall alone, because earlier observations were obtained contemporaneously with developmental periods. Nevertheless, differences in available diagnoses, age windows, and case definitions still shape continuity estimates.
Caspi, A., Houts, R. M., Ambler, A., et al. (2020) assessed clinically impairing mental disorders at nine ages between 11 and 45 in a New Zealand birth cohort. Of 1,037 original participants, 1,013 had mental health data available. The study found frequent shifts among internalizing, externalizing, and thought-disorder diagnoses across four decades, supporting description of a person's history by onset, duration, and diversity of diagnoses. The strength of this design is within-person temporal coverage; an important limitation is the cohort's predominantly White regional composition and the historic period in which members developed.
The investigators' results also make a methodological point: repeated single-year or past-year prevalence estimates cannot simply be added to calculate lifetime burden, because the same people contribute across waves. Cumulative history depends on carefully linking cases across intervals. The Dunedin evidence is powerful for showing recurrent and changing clinical states, while explanations of why each transition occurred require distinct causal studies.
The Great Smoky Mountains and Cross-Cohort Evidence
The Great Smoky Mountains Study followed children and adolescents through key developmental transitions using repeated psychiatric assessments. Copeland, W. E., Shanahan, L., Costello, E. J., & Angold, A. (2009) analyzed childhood and adolescent disorders as predictors of young-adult conditions, documenting both within-category and cross-category predictive relationships. The study underscored that a childhood psychiatric history is relevant to later risk even when the later syndrome differs from the original presentation.
Copeland, W., Shanahan, L., Costello, E. J., & Angold, A. (2011) investigated cumulative psychiatric disorder prevalence by young adulthood. Cumulative measurement addresses the fact that disorders can appear for relatively brief intervals and may be missed in a single annual snapshot. Yet estimates depend on diagnostic breadth, follow-up coverage, interview schedules, attrition, and assumptions concerning unobserved periods.
An important extension used pooled evidence from several cohorts. Copeland, W. E., Adair, C. E., Smetanin, P., et al. (2013) examined 3,722 participants and nearly 20,000 observations, finding elevated later disorder likelihood after an earlier diagnosis, while showing variation across categories. Cross-cohort analysis can test whether patterns replicate beyond one population but creates new issues concerning harmonized criteria, different follow-up ages, and historical changes in clinical practice. Agreement on a broad conclusion is stronger evidence than exact equality of effect estimates across incompatible measures.
Newer Evidence from Preschool Through Late Adolescence
The importance of early and repeated direct observation is reinforced by Silver, J., Calentino, A. E., Lawhead, C., et al. (2026), who studied 609 families recruited when children were age three in a geographically defined area around Stony Brook, New York. Assessments at three-year intervals tracked depression, anxiety, attention-deficit/hyperactivity disorder, disruptive behavior, and substance use disorders through age 18 using semistructured interviews. The published study reported substantial continuity within and between diagnostic categories, with stronger heterotypic continuity within broad internalizing or externalizing domains than between them.
The study reported an estimated 71.8% cumulative prevalence of at least one assessed disorder by age 18 in this sample. This number is specific to the assessment schedule, included diagnoses, selected population, and study design. It should not be presented as a universal prevalence rate for all children, countries, or diagnostic systems. The exclusion of children with significant medical or developmental disorders and the English-language requirement are relevant to transportability.
The authors also observed age-varying profiles across disorder families. The important scientific contribution is the combination of repeated diagnostic ascertainment and developmental span from preschool to late adolescence. Continued investigation should examine whether these patterns replicate in wider demographic settings and how sensitivity to symptom thresholds, informant changes, and missing data affects estimates. The existence of an early disorder indicates elevated later risk, while individual outcomes remain diverse.
Cross-National Evidence and the Problem of Retrospection
Cross-national psychiatric epidemiology provides a broad setting for estimating the timing of disorder recognition and onset. In the World Mental Health survey analysis (McGrath et al., 2023)00193-1), adult interviews conducted across 29 countries supported the conclusion that many surveyed mental disorders have their first onset during childhood, adolescence, or early adulthood. The estimates are linked to 13 disorders, specific diagnostic instruments, and the included countries.
Retrospective onset information may be distorted when episodes are brief, early, stigmatized, or no longer salient. A person may remember a first episode differently after receiving a later diagnosis, and onset reports may cluster around culturally recognizable life events. Kessler, R. C., Berglund, P., Demler, O., Jin, R., Merikangas, K. R., & (2005) provided landmark US retrospective distributions, but prospective repeated interviews may reveal substantially more lifetime disorder histories. A disagreement between designs can therefore reveal ascertainment differences rather than a simple error in one dataset.
Comparative developmental inference should report sampling frame, survey years, age coverage, disorders included, recall window, functional thresholds, and the distinction between observed and projected cumulative risk. For policy, cross-national evidence supports services adapted to multiple developmental stages. It does not support treating a global median age as a local diagnostic threshold or a single universal schedule for clinical intervention.
Longitudinal Designs and Time Scales
Prospective cohort designs measure participants repeatedly, allowing observation of changes in symptoms and clinical classifications. Panel studies measure at discrete waves; intensive longitudinal designs sample more frequently; registry studies record health contacts and diagnoses at high volume; clinical follow-up studies often describe treatment-seeking patients. Each design captures a different part of the underlying process, and none automatically reconstructs continuous psychopathology between observations.
Temporal resolution matters. Annual interviews may detect episodes lasting months while missing short disturbances; weekly ratings can describe rapid symptom changes but may rely on retrospective reconstruction between interviews. Birmaher, B., Axelson, D., Goldstein, B., et al. (2009) illustrate detailed interval follow-up within a clinical cohort. An event that first appears between survey waves is interval-censored; assigning it to the next interview date without adjustment can systematically alter estimated onset distributions.
Developmental samples require particular attention to attrition, death, migration, institutionalization, and withdrawal from care. Participants with severe symptoms may be differentially difficult to recontact, biasing long-term estimates. Methods for handling missing observations must state whether loss is treated as random, conditionally ignorable, or potentially informative. Robustness checks should test alternative assumptions and account for differing observation schedules across participants and periods.
Growth Curves, Mixtures, and Trajectory Classes
Longitudinal growth models estimate average change and variation in individual change under chosen functional forms, such as linear, nonlinear, or piecewise trends. Growth-mixture and group-based trajectory models additionally approximate observed heterogeneity with latent classes. Their practical appeal is the identification of recurrent statistical patterns such as low-stable, increasing, decreasing, or persistent symptom levels. Nagin, D. S., & Odgers, C. L. (2010) explain the uses and interpretive hazards of these methods in clinical research.
A fitted latent class should not automatically be interpreted as a distinct natural developmental subtype. Estimated groups can depend on class number, starting values, distributional assumptions, outcome scale, measurement occasions, sample composition, and missing data. High posterior classification probabilities provide evidence about assignment under a fitted model, not proof that human populations contain immutable biological categories corresponding to each fitted line.
Clinical interpretation requires comparing alternate model specifications, reporting uncertainty, testing out-of-sample prediction, and showing whether observed patterns have independent prognostic or intervention relevance. Continuous random effects may fit nearly as well as a discrete typology. Researchers should preserve individual variability within each fitted group and avoid using group labels as deterministic clinical identities. Model simplicity, theoretical meaning, and predictive validation must be evaluated together.
Event-History, Multistate, and Transition Models
Survival and event-history models estimate the timing of transitions such as first onset, remission, recurrence, or service entry. A risk set must be defined precisely: people are at risk of a first onset only until it occurs, while recurrence requires entry into an appropriately defined remitted state. A hazard is a conditional event rate among people still at risk at a particular time, and it differs from cumulative incidence. Competing events and loss to follow-up alter the interpretation of both.
Multistate models can represent transitions among no diagnosed disorder, active disorder, remission, recurrent disorder, or death; other state structures can represent separate diagnostic families. Markov models assume that transition probabilities depend on the current state under their specification, while semi-Markov and history-dependent approaches can incorporate time spent in a state or previous events. Clinical histories with multiple concurrent diagnoses demand careful choices about how combinations are represented.
The scientific value of a multistate model is that it makes the transition of interest explicit. The challenge is that detailed state spaces require large samples and strong measurement consistency. Diagnoses observed only at interviews can conceal multiple events between waves, and a model can underestimate short episodes if intervals are long. Model results are therefore conditional on state definitions, assessment frequency, and untested assumptions regarding unobserved transitions.
Developmental Measurement and Diagnostic Invariance
A symptom scale can change meaning over development because linguistic comprehension, behavioral expectations, and contexts change. Measurement invariance examines whether the same underlying construct is represented comparably across ages, groups, or time. Putnick, D. L., & Bornstein, M. H. (2016) document conventions and reporting challenges for invariance analysis. When invariance fails, differences in mean scores or factor loadings may partly represent changing measurement rather than true change in clinical functioning.
In developmental psychiatry, different informants often assess the same person at different ages. Childhood caregiver ratings may be replaced by adolescent self-report, and adult clinical interviews may use revised criteria. Without bridging procedures, the resulting time series can conflate informant, method, and developmental effects. Studies should examine item functioning, criterion thresholds, differential reporting, and the feasibility of linking constructs across instruments.
Diagnostic categories also evolve. A longitudinal study that reclassifies all earlier data using one contemporary framework preserves analytic comparability but may lose historical clinical meaning; a study that retains historical diagnoses preserves source fidelity but introduces classification changes. Both approaches have a place if explicitly documented. A robust synthesis should distinguish continuity of symptoms, continuity of clinical impairment, and continuity under a particular diagnostic convention.
Multi-Informant Assessment and Clinical Phenomenology
Clinical trajectory measurement begins with systematic developmental history, including early milestones, symptom chronology, functional changes, health conditions, treatment exposure, family context, school or work transitions, and relevant sources of corroboration. The temporal order of symptoms may alter differential diagnosis, yet early recollection can be unreliable and dependent on current beliefs. Whenever possible, contemporaneous records and repeated assessments supplement retrospective narrative.
Interviews may be structured for comparability or semistructured to permit clinical clarification. Informants supply distinct evidence about settings. A caregiver may observe routines and external behavior, a young person may describe internal distress, a teacher may assess classroom performance, and clinical records may capture severe episodes or treatment effects. Disagreement should be investigated for environmental or measurement explanations rather than erased by averaging by default.
Current classification documents such as World Health Organization. (2024) guide the identification and description of disorders. They coexist with longitudinal assessment, which asks whether the same classification still applies later and what clinically meaningful changes occurred. The patient's evolving capacities and priorities should inform assessment, with safeguards for privacy and communication accessibility. This monograph describes research and assessment principles rather than offering personal diagnostic determinations.
Longitudinal Causal Inference and Confounding
Longitudinal data can establish that an exposure preceded an outcome, but causal inference also requires addressing common causes of exposure and outcome, measurement error, selection, and changes in exposure influenced by earlier symptoms. For example, treatment initiation depends on clinical severity, access, family preference, and clinician judgment. Comparing treated with untreated youths in routine observational data can therefore yield misleading estimates of treatment effects.
The Hernán, M. A., & Robins, J. M. (2016) offers a disciplined way to specify eligibility, treatment strategies, assignment, time zero, outcome, follow-up, and estimand. This approach is particularly useful when studying longitudinal changes in clinical care. It also clarifies when the evidence cannot identify a causal effect because required confounders were unmeasured or the intervention was poorly defined.
Genetic liability and shared familial context further complicate environmental interpretation. Sibling and twin comparisons, genetically informed designs, natural experiments, randomized preventive interventions, and sensitivity analyses can contribute complementary evidence, each with identifiable assumptions. A credible developmental psychopathology explanation ideally converges across designs and shows why an earlier process changes later clinical risk rather than merely predicting it. Causal claims should be proportionate to that evidence.
Clinical Staging, Prognosis, and Decision Thresholds
Clinical staging seeks to locate a person's observed presentation within a proposed course, often using severity, recurrence, functional impairment, and treatment history. Staging can help organize longitudinal observations, yet developmental psychopathology calls for caution when stages are presented as a necessary sequence. A person may experience early symptoms followed by recovery, several intermittent episodes, or changing symptom families. Prognostic validity must be demonstrated prospectively.
Risk prediction generally combines variables measured at baseline to estimate a future event over a specified horizon. Prediction improves clinical decisions only when models are calibrated, validated in new populations, clinically useful at actionable thresholds, and compatible with ethically acceptable consequences. A highly discriminating score may still be inappropriate if its false-positive burden leads to unnecessary stigma or intervention. Age-specific base rates and varying measurement should be reported.
The distinction between a person's current clinical state and predicted long-term trajectory should remain explicit. Caspi, A., Houts, R. M., Ambler, A., et al. (2020) showed the diversity of psychiatric histories even among individuals with substantial illness burden. Developmental assessment therefore supports dynamic re-evaluation rather than assigning an enduring identity based on a single early episode or a single model-derived score.
Treatment, Remission, and the Scope of Clinical Trials
Treatment studies assess whether an intervention improves specified outcomes under particular conditions. In youth care, relevant outcomes include symptoms, functioning, distress, educational engagement, adverse events, recurrence, and family participation. An effect on short-term symptom scores does not necessarily show reduced lifetime incidence of all later psychiatric diagnoses. Follow-up periods, treatment crossover, concurrent services, and retention influence the strength of long-run conclusions.
The National Institute for Health and Care Excellence. (2019) emphasizes assessment that accounts for developmental level, preferences, comorbidity, and context. Its recommendations are clinical guidance for the populations and jurisdiction specified by NICE; they are cited to illustrate the necessity of age-sensitive clinical evidence, not to prescribe treatment to an individual reader. Questions about a particular medication, dose, or regulatory approval belong to specialist clinical and pharmacological reviews.
Clinical trajectories can change after treatment, but the mechanism of change may involve specific therapeutic skills, common treatment factors, altered social contexts, time, spontaneous remission, or combinations. Randomization, appropriate comparison groups, valid outcomes, and sufficiently long follow-up help distinguish competing explanations. A complete developmental treatment science examines whether improvements persist through new developmental demands rather than assuming that success at one age transfers automatically to another.
Prevention and Timing of Interventions
Prevention can be universal, selective, or indicated, depending on the population and intervention target. In developmental psychopathology, intervention timing matters because the same program may operate differently before symptoms emerge, during a first clinical episode, after recurrence, or amid longstanding impairment. A preventive hypothesis requires specifying both the process intended to change and the clinically meaningful outcome to be prevented.
Evidence from developmental cohorts supports identifying periods of elevated onset probability and groups with higher observed subsequent risk. The McGrath and colleagues (2023) analysis00193-1) places much first-onset risk early in the lifespan for the surveyed disorders. However, allocating services based solely on aggregate age distributions overlooks individuals with later-onset conditions and can miss structural inequalities in access.
A promising preventive intervention should demonstrate efficacy or effectiveness under the relevant conditions, report harms, and measure durability. An intervention that improves school behavior may yield valuable immediate benefits even if the evidence is insufficient to claim prevention of adult psychiatric disorder. Developmental trajectory science can help specify those distinct endpoints. Future research should assess both symptom and participation outcomes, monitor equity, and consider whether interventions alter exposure to later risks.
The Artificial Era and Changing Developmental Ecologies
Human development increasingly occurs in environments containing artificial agents, adaptive information systems, automated educational interfaces, and algorithmically mediated communication. The Artificial Era introduces new conditions for attention, social learning, help-seeking, emotional disclosure, and the distribution of cognitive work. These are potentially relevant exposures and relational settings for developmental psychopathology, particularly when they interact with existing vulnerabilities and protective processes.
The appropriate research question is observable: whether particular patterns of human–artificial interaction alter symptom detection, clinical trajectories, access to support, or developmental functioning under defined conditions. Studies would need to separate platform type, interaction duration, user characteristics, prior mental health, human relationships, and institutional context. A person who uses an artificial system extensively may already differ in distress or access to care; observed association alone cannot establish benefit or harm.
A longitudinal framework can compare developmental transitions before and after exposure while addressing confounding, privacy, selection, platform change, and measurement drift. The classification of artificial systems within the philosophical architecture of Aisentica is a separate theoretical matter; this psychology dossier makes no empirical claim about machine consciousness or clinical status. The field's direct contribution is to establish testable human developmental and clinical consequences of changing artificial environments.
Critical Examination of Replication and Generalizability
The strongest developmental findings recur across study traditions: earlier disorder histories often predict later clinical difficulties, symptoms can change diagnostic expression, and population averages conceal substantial heterogeneity. Yet exact estimates differ because cohorts vary in ethnicity, geography, birth period, clinical access, symptom instruments, and ascertainment. The Dunedin Study's intensive follow-up is a major advantage, while its relatively narrow birthplace and historical cohort limit universal projection.
Prospective diagnostic assessments, cross-national retrospective surveys, and clinical treatment cohorts provide complementary but noninterchangeable evidence. The Silver, J., Calentino, A. E., Lawhead, C., et al. (2026) adds direct observation from preschool to late adolescence, but its recruitment and exclusions require explicit reporting. A broad age-of-onset distribution from retrospective international surveys does not establish the same cumulative diagnostic prevalence as a local cohort assessed directly every few years.
Replication should include diagnostic harmonization, disclosure of analytic decisions, missing-data sensitivity analyses, comparison of latent versus continuous developmental structures, and external validation of clinical prediction. Results that remain stable across instruments and cohorts carry stronger weight than a precise but sample-dependent trajectory class. Scientific disagreement can be productive when it clarifies which outcome, population, or time scale each study actually measures.
Unresolved Scientific Problems
A central unresolved question is how to identify the processes that explain cross-diagnostic continuity. Shared symptom covariance may reflect common liabilities, causal interactions among symptoms, family and social conditions, developmental timing, or multiple overlapping mechanisms. The p-factor provides a useful statistical summary in many datasets, but its ontological interpretation and mechanistic specificity remain open. Diagnostic continuity also changes when definitions and instruments evolve, making long-run measurement equivalence a major challenge.
A second problem concerns the integration of clinical and developmental timescales. Ordinary psychiatric interviews often observe individuals at intervals longer than many episodes. Symptoms may fluctuate rapidly during a day while risk changes over years. Models connecting momentary experience to life-course disorder histories require compatible measurement, valid interpolation, and transparent assumptions. A model that fits mean symptom trends may be inadequate for predicting abrupt crises or rare clinical transitions.
A third challenge is intervention translation. Knowledge that a childhood disorder predicts later difficulties does not identify which preventive action will change that future risk. Randomized or plausibly causal intervention evaluations must test durable outcomes, adverse effects, equity, and implementation in real settings. Cross-cultural and geographically diverse cohorts, reliable multi-informant instruments, and explicit treatment-history recording are necessary to test whether theoretical pathways represent broad developmental processes.
Integrated Scientific Synthesis
Developmental psychopathology is most coherent when the clinical state observed at time t is treated as an outcome of prior developing processes and as a potential influence on later environments, adaptations, and symptoms. Its complete description combines an onset history, current symptom organization, impairment, contextual conditions, and subsequent transitions. The distinction between state, course, liability, and mechanism provides a disciplined framework for integrating diverse literatures.
Evidence from Kim-Cohen, J., Caspi, A., Moffitt, T. E., Harrington, H., Milne, B. J., (2003), Copeland, W. E., Adair, C. E., Smetanin, P., et al. (2013), and Caspi, A., Houts, R. M., Ambler, A., et al. (2020) establishes that prior psychiatric difficulties often carry prognostic information, and that successive disorders can differ in clinical presentation. McGrath and colleagues (2023)00193-1) locate substantial onset risk in youth across surveyed countries; Harris, J. L., Swanson, B., & Petersen, I. T. (2024) demonstrate the importance of general covariance in child and adolescent psychopathology models; and Silver, J., Calentino, A. E., Lawhead, C., et al. (2026) contribute additional prospective evidence of both forms of continuity.
Together, these results support a developmental clinical science built from repeated, age-sensitive observation, analytically explicit transitions, context-aware explanations, and cautious causal inference. They also place limits on simple diagnostic forecasts: no one clinical label, latent class, or risk score can substitute for measuring developmental change. The most productive research architecture connects clinical states to processes and institutions across time while permitting recovery, reorganization, and plural developmental outcomes.
Conclusion
Developmental psychopathology explains clinical phenomena through their temporal organization. The scientific object is the course by which psychological symptoms become clinically consequential, persist, transform into different presentations, remit, recur, or interact with later developmental opportunities. An adequate account identifies the measured clinical state, the duration and timing of observations, the developing person's context, and the evidential basis for any proposed mechanism.
The evidence reviewed here supports three durable conclusions. First, mental disorder histories frequently begin before adulthood, and early disorder carries information about later probability. Second, continuity occurs both within and across diagnostic categories, while the timing, severity, and eventual outcome vary substantially among people. Third, longitudinal inference depends on measurement consistency, appropriate temporal design, accounting for confounding and attrition, and independent validation of predictive and causal claims.
The research program ahead is to connect clinical histories to manipulable and testable developmental mechanisms, explain individual differences without reifying statistical trajectory classes, evaluate durable prevention and treatment effects, and extend culturally valid observations across rapidly changing human and artificial ecologies. Such work positions developmental psychopathology as a core science of clinical time: how disorders emerge, move, and sometimes recede across the developing life course.
References
- Angold, A., Costello, E. J., & Erkanli, A. (1999). Comorbidity. Journal of Child Psychology and Psychiatry, 40(1), 57–87. https://doi.org/10.1111/1469-7610.00424
- Birmaher, B., Axelson, D., Goldstein, B., et al. (2009). Four-year longitudinal course of children and adolescents with bipolar spectrum disorders: The Course and Outcome of Bipolar Youth (COBY) study. American Journal of Psychiatry, 166(7), 795–804. https://doi.org/10.1176/appi.ajp.2009.08101569
- Caspi, A., Houts, R. M., Belsky, D. W., et al. (2014). The p factor: One general psychopathology factor in the structure of psychiatric disorders? Clinical Psychological Science, 2(2), 119–137. https://doi.org/10.1177/2167702613497473
- Caspi, A., Houts, R. M., Ambler, A., et al. (2020). Longitudinal assessment of mental health disorders and comorbidities across 4 decades among participants in the Dunedin birth cohort study. JAMA Network Open, 3(4), e203221. https://doi.org/10.1001/jamanetworkopen.2020.3221
- Caspi, A., & Moffitt, T. E. (2018). All for one and one for all: Mental disorders in one dimension. American Journal of Psychiatry, 175(9), 831–844. https://doi.org/10.1176/appi.ajp.2018.17121383
- Cicchetti, D. (1984). The emergence of developmental psychopathology. Child Development, 55(1), 1–7. https://pubmed.ncbi.nlm.nih.gov/6705613/
- Cicchetti, D., & Rogosch, F. A. (1996). Equifinality and multifinality in developmental psychopathology. Development and Psychopathology, 8(4), 597–600. https://doi.org/10.1017/S0954579400007318
- Copeland, W. E., Shanahan, L., Costello, E. J., & Angold, A. (2009). Childhood and adolescent psychiatric disorders as predictors of young adult disorders. Archives of General Psychiatry, 66(7), 764–772. https://doi.org/10.1001/archgenpsychiatry.2009.85
- Copeland, W., Shanahan, L., Costello, E. J., & Angold, A. (2011). Cumulative prevalence of psychiatric disorders by young adulthood: A prospective cohort analysis from the Great Smoky Mountains Study. Journal of the American Academy of Child & Adolescent Psychiatry, 50(3), 252–261. https://doi.org/10.1016/j.jaac.2010.12.014
- Copeland, W. E., Adair, C. E., Smetanin, P., et al. (2013). Diagnostic transitions from childhood to adolescence to early adulthood. Journal of Child Psychology and Psychiatry, 54(7), 791–799. https://doi.org/10.1111/jcpp.12062
- Costello, E. J., Copeland, W., & Angold, A. (2011). Trends in psychopathology across the adolescent years: What changes when children become adolescents, and when adolescents become adults? Journal of Child Psychology and Psychiatry, 52(10), 1015–1025. https://doi.org/10.1111/j.1469-7610.2011.02446.x
- Harris, J. L., Swanson, B., & Petersen, I. T. (2024). A developmentally informed systematic review and meta-analysis of the strength of general psychopathology in childhood and adolescence. Clinical Child and Family Psychology Review, 27(1), 130–164. https://doi.org/10.1007/s10567-023-00464-1
- Hernán, M. A., & Robins, J. M. (2016). Using big data to emulate a target trial when a randomized trial is not available. American Journal of Epidemiology, 183(8), 758–764. https://doi.org/10.1093/aje/kwv254
- Kessler, R. C., Berglund, P., Demler, O., Jin, R., Merikangas, K. R., & Walters, E. E. (2005). Lifetime prevalence and age-of-onset distributions of DSM-IV disorders in the National Comorbidity Survey Replication. Archives of General Psychiatry, 62(6), 593–602. https://doi.org/10.1001/archpsyc.62.6.593
- Kim-Cohen, J., Caspi, A., Moffitt, T. E., Harrington, H., Milne, B. J., & Poulton, R. (2003). Prior juvenile diagnoses in adults with mental disorder: Developmental follow-back of a prospective-longitudinal cohort. Archives of General Psychiatry, 60(7), 709–717. https://doi.org/10.1001/archpsyc.60.7.709
- Masten, A. S. (2001). Ordinary magic: Resilience processes in development. American Psychologist, 56(3), 227–238. https://doi.org/10.1037/0003-066X.56.3.227
- McGrath, J. J., Al-Hamzawi, A., Alonso, J., et al. (2023). Age of onset and cumulative risk of mental disorders: A cross-national analysis of population surveys from 29 countries. The Lancet Psychiatry, 10(9), 668–681. https://doi.org/10.1016/S2215-0366(23)00193-1
- Moffitt, T. E. (1993). Adolescence-limited and life-course-persistent antisocial behavior: A developmental taxonomy. Psychological Review, 100(4), 674–701. https://doi.org/10.1037/0033-295X.100.4.674
- Nagin, D. S., & Odgers, C. L. (2010). Group-based trajectory modeling in clinical research. Annual Review of Clinical Psychology, 6, 109–138. https://doi.org/10.1146/annurev.clinpsy.121208.131413
- National Institute for Health and Care Excellence. (2019; reviewed 2024). Depression in children and young people: Identification and management (NG134). https://www.nice.org.uk/guidance/ng134
- Patalay, P., & Fitzsimons, E. (2018). Development and predictors of mental ill-health and wellbeing from childhood to adolescence. Social Psychiatry and Psychiatric Epidemiology, 53(12), 1311–1323. https://doi.org/10.1007/s00127-018-1604-0
- Putnick, D. L., & Bornstein, M. H. (2016). Measurement invariance conventions and reporting: The state of the art and future directions for psychological research. Developmental Review, 41, 71–90. https://doi.org/10.1016/j.dr.2016.06.004
- Rutter, M., & Sroufe, L. A. (2000). Developmental psychopathology: Concepts and challenges. Development and Psychopathology, 12(3), 265–296. https://doi.org/10.1017/S0954579400003023
- Schaefer, J. D., Caspi, A., Belsky, D. W., et al. (2017). Enduring mental health: Prevalence and prediction. Journal of Abnormal Psychology, 126(2), 212–224. https://doi.org/10.1037/abn0000232
- Silver, J., Calentino, A. E., Lawhead, C., et al. (2026). The development and course of youth psychopathology: A longitudinal study of prevalence and continuity from early childhood through late adolescence. European Child & Adolescent Psychiatry, 35(10), 3347–3359. https://doi.org/10.1007/s00787-026-03087-3
- Sroufe, L. A., Egeland, B., Carlson, E. A., & Collins, W. A. (2005). The development of the person: The Minnesota study of risk and adaptation from birth to adulthood. Guilford Press. https://www.guilford.com/books/The-Development-of-the-Person/Sroufe-Egeland-Carlson-Collins/9781606232491
- World Health Organization. (2024). Clinical descriptions and diagnostic requirements for ICD-11 mental, behavioural and neurodevelopmental disorders. World Health Organization. https://www.who.int/publications/i/item/9789240077263