Changes in incidence of hospital diagnoses of mental disorders among children and adolescents over the past 3 decades Authors: Pernille B Hansen, Andreas Moses Appel, Elisabeth Wreford Andersen, Soren Dalsgaard, Michael E Benros Published: 2026 ABSTRACT AbstractBackgroundWHO considers mental disorders a significant disease burden among children and adolescents. However, there is a lack of population-based studies on the time trends of mental disorder diagnosis in this group.ObjectiveTo investigate changes in the incidence of clinical diagnoses of mental disorders among all Danish children and adolescents in the past 3 decades.MethodsIn this retrospective cohort study, we included all Danish-born individuals from 1976. Follow-up started in 1994 until first diagnosis of a mental disorder, 18th birthday, emigration from Denmark, death or end of study on 30 June 2023. Sex-specific cumulative incidences of diagnosis by age 18 were calculated in 5-year birth cohorts from 1976–1980 to 2001–2005 using the Kaplan-Meier estimator. Poisson regression was used to estimate the age-specific and sex-specific incidence rates.FindingsFrom 1994 to 2023, a total of 193 354 children and adolescents received a mental disorder diagnosis. Over 3 decades the cumulative incidence of any mental disorder diagnosis before age 18 increased from 1.46% (95% CI 1.40% to 1.52%) to 14.56% (95% CI 14.36% to 14.76%) among females and from 1.22% (95% CI 1.16% to 1.27%) to 14.77% (95% CI 14.57% to 14.96%) among males. A lower mean age at first diagnosis accounted for an estimated 15–22% of the observed incidence in the most recent years.ConclusionsThis study demonstrates a large increase in mental disorder diagnoses. Observed rates were in part accounted for by decreasing age at first diagnosis. While some of the remaining increase may reflect a true rise in the underlying prevalence of mental disorders, changes in diagnostic practice and greater awareness of mental disorders are probable contributors.Clinical implicationsEstimates of specific diagnostic rates are essential for appropriate healthcare planning and resource allocation to accommodate the increasing number of mental disorder diagnoses among children and adolescents. WHAT IS ALREADY KNOWN ON THIS TOPIC While prevalence of mental disorders among children and adolescents in screening studies have been stable during the past 30 years, psychiatric service use indicates an increasing trend. A review of the literature shows a need for comprehensive, up-to-date studies on the general and diagnosis-specific time trends of mental disorders among children and adolescents. While prevalence of mental disorders among children and adolescents in screening studies have been stable during the past 30 years, psychiatric service use indicates an increasing trend. A review of the literature shows a need for comprehensive, up-to-date studies on the general and diagnosis-specific time trends of mental disorders among children and adolescents. WHAT THIS STUDY ADDS In Denmark during the past 3 decades, there was a continuously increasing cumulative incidence of any mental disorder diagnoses from less than 2% to now more than 14%, which coincides with the underlying prevalence found in screening studies. Younger age at first diagnosis accounted for a substantial part of the observed incidence rate. In Denmark during the past 3 decades, there was a continuously increasing cumulative incidence of any mental disorder diagnoses from less than 2% to now more than 14%, which coincides with the underlying prevalence found in screening studies. Younger age at first diagnosis accounted for a substantial part of the observed incidence rate. HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY Knowledge about the magnitude of increase in prevalence and the diagnosis-specific trends is essential for optimising the allocation of clinical resources for assessment and treatment of mental disorders. Knowledge about the magnitude of increase in prevalence and the diagnosis-specific trends is essential for optimising the allocation of clinical resources for assessment and treatment of mental disorders. Background Mental health problems among children and adolescents have gained increased attention in recent years,1–5 and mental and behavioural disorders are now considered one of the main causes of disease burden among adolescents.6 Mental disorders commonly have an onset during childhood and adolescence7 with many persisting into adulthood,8 and throughout the lifespan they are associated with significant functional impairments, such as poorer academic outcomes,9 development of other psychiatric comorbidities,10 11 criminality12 and premature death.13 Thus, high quality data on the time trends of mental disorder diagnoses is of utmost importance for relevant healthcare planning and appropriate allocation of resources for this vulnerable age group. Despite the increasing concern, meta-analyses of representative, population-based screening studies have estimated the prevalence of children and adolescents fulfilling criteria of any mental disorder diagnosis to be relatively stable over the last four decades and around 13.4–15.5%.14 15 While the estimated occurrence of underlying disease has been relatively constant for most mental disorder diagnoses,16–18 an increasing number of youths receive a clinical diagnosis.2 3 19 20 A large observational retrospective study of administrative healthcare data on 4.1 million children aged 4–17 in the USA showed an increase of 34.6% in the prevalence of diagnosed mental disorders among children and adolescents from 2012 to 2018.2 A Danish register-based study compared age-specific incidence rates of mental disorder diagnoses from ages 0 to 80 years across birth cohorts and calendar periods between 2004 and 2021 and found higher rates among younger cohorts and more recent periods.20 However, data on the age-standardised incidence rates among children and adolescents going further back and including the full spectrum of mental disorders remain limited. To our knowledge, the present study represents the most comprehensive, up-to-date, nationwide examination of temporal changes in the incidence of clinical diagnoses of mental disorders among children and adolescents. Unlike previous studies, we also estimated how much the diagnostic trend could be attributed to a change in mean age of first-time diagnosis. Additionally, we explored potential changes in the associations between incidence rates of psychiatric diagnoses and important, well-established risk factors for mental disorders. Finally, we compared diagnostic trends of psychiatric and general medical conditions during the study period. Methods Study population In this retrospective cohort study, we included all children born in Denmark from 1 January 1976 to 30 June 2023. Using the unique personal identification number assigned to all individuals living in Denmark, person-level data was linked across the nationwide Danish registers. Information on all Danish residents was obtained from the Danish Civil Registration System.21 This is a nationwide registry established in 1969, containing information on all Danish residents, including sex assigned at birth, date of birth, migration and death, and links to identify family members. Diagnoses of mental disorders were identified using the Danish National Patient Register (NPR) and the Psychiatric Central Research Register (PCRR).22 23 From 1969, the PCRR includes data on all psychiatric admissions, and from 1977, the NPR includes information on all somatic admissions. In 1995, the PCRR was integrated into the NPR and data on all outpatient and emergency room contacts in Danish hospitals was added. For analyses of cumulative incidence of mental disorder diagnoses in childhood/adolescence, individuals were followed from birth. For analyses of incidence rates, the start of follow-up was from the introduction of the International Classification of Diseases V.10 (ICD-10) on 1 January 1994 or the date of birth, whichever came last, meaning that individuals in the study population could start follow-up after birth. Individuals diagnosed before the start of follow-up were excluded. Exclusions were based on inpatient data. In all analyses, individuals were followed up until the first diagnosis of a mental disorder (main or secondary), 18th birthday, emigration from Denmark, 30 June 2023 or death. For details of data sources and follow-up periods, see online supplemental eFigure 1. Outcomes and covariates The primary objective was to describe time trends in the incidence of diagnoses of mental disorders among children and adolescents, taking age and sex into account. The primary outcome of interest was any mental disorder. Secondary outcomes were the following diagnostic groups and specific diagnoses of mental disorders: substance use disorders, schizophrenia spectrum disorder, schizophrenia, mood disorders, bipolar disorder, depression, anxiety disorders, anxiety, obsessive–compulsive disorder (OCD), eating disorders, anorexia nervosa, personality disorders, autism spectrum disorders (ASD), attention-deficit hyperactivity disorder (ADHD), oppositional defiant disorder/conduct disorder (ODD/CD), attachment disorders and tic disorders. The choice of diagnostic groups and specific diagnoses was inspired by another Danish register-based study estimating cumulative incidence and incidence rates of mental disorders among children and adolescents.24 Diagnoses were identified in the NPR and the PCRR. Until 1994, diagnoses were defined according to the ICD-8 and ICD-10 from 1994 onwards (see eTable 1).25 Many diagnoses of specific mental disorders have shown high validity in the Danish registers.26 From the registers, we also included data on the following established risk factors for mental disorders: maternal age at childbirth, ethnicity (Danish: native-born Dane with at least one parent being a native born Danish citizen; immigrant/descendant: not native-born/no parents who are native-born Danish citizens), urbanicity of birthplace and family structure were identified at study entry. Number of addresses since birth, maternal/paternal level of education, maternal/paternal income quintile (within age group), maternal/paternal work status, maternal/paternal previous mental disorder diagnosis and maternal/paternal Charlson Comorbidity Index (CCI) were time varying covariates. Statistical analysis The incidence rates (IR) of first psychiatric diagnosis per 10 000 person-years were estimated for each sex and age group. Cumulative incidences of first psychiatric diagnoses before age 18 were estimated in 5-year birth cohorts for females and males separately using the Kaplan-Meier estimator. The 5-year interval for birth cohorts was chosen, as annual analyses for some specific diagnoses would result in insufficient sample sizes to provide meaningful estimates. As outpatient contacts were only available from 1995, we performed a sensitivity analysis to estimate the influence of incomplete diagnostic history on the main results,27 restricting the cumulative incidence analyses to 2-year birth cohorts from 1995 onwards. Poisson regression was used to model the IR by age, year and sex. The rates depended on two timescales simultaneously, that is, attained age and calendar period. Thus, the follow-up time was split into short time intervals along both timescales. For both attained age and calendar period, restricted cubic splines were applied with five knots, placing the knots so that events were approximately evenly distributed in between. The main model contained the interaction between the splines for the two timescales. Our primary interest was the effect of calendar period, so age-standardised rates were also calculated from the models. We also estimated whether the association between risk factors of mental disorders and diagnostic rate changed during the study period, by adding each covariate to the main model one by one, as interactions with period and age. To estimate the effect of a possible shift in age at first diagnosis, we adapted methods from demographic studies on birth rates, in which decreasing birth rates were shown to be partly attributed to older ages at childbirth.28 The same approach can be used to decompose the observed diagnostic rate into the rate that would have been observed without the change in age at diagnosis (the age-shift adjusted rate) and the distortion introduced by the changing age at diagnosis. We report this distortion as the proportion by which the observed rate exceeds the age-shift adjusted. To mitigate the effect of missing outpatient data, we estimated age-shifts both in periods 1976–2023 and 1994–2023. A detailed description of this analysis is given in the supplement. Finally, to compare diagnostic trends of somatic and mental disorders we estimated the IR of all somatic diagnoses from 1994 to 2023. Somatic diagnoses were defined as all ICD-10 codes apart from the mental disorder diagnoses (ICD-10: F00-99). All analyses were carried out using R statistical software V.4.4.1. Findings All 2 829 649 individuals in Denmark below 18 years of age during the study period were followed from 1994 through 2023, resulting in 30 454 953 person-years of observation. Within the study population, 193 354 individuals received a diagnosis of a mental disorder before the age of 18 years, 88 456 (45.7%) females and 104 898 (54.3%) males. The characteristics of the full cohort are presented in online supplemental eTable 2. In the following, we describe the time trends in cumulative incidence and incidence rates. All cumulative incidences and incidence rates are presented in the online supplemental material and in an interactive html at: https://biological-and-precision-psychiatry.github.io/Time-Trends-Of-Child-Adolescent-Mental-Disorder-Diagnoses/supplementary_viewer.html Time trends of cumulative incidences The risk of receiving any mental disorder diagnosis before the age of 18 years increased considerably over the 3-decade study period (figure 1A,online supplemental eTable 3, github.com). Among people born between 1976 and 1980, the cumulative incidence of any mental disorder diagnosis was 1.46% (95% CI 1.40% to 1.52%) in females and 1.22% (95% CI 1.16% to 1.27%) in males. Among people born in the period 2001–2005, the cumulative incidence had increased to 14.56% (95% CI 14.36% to 14.76%) in females and 14.77% (95% CI 14.57% to 14.96%) in males (p value for difference across birth cohorts: p<0.0001). Incidence rates of any mental disorder diagnosis per 10 000 person-years before age 18 from 1994 to 2023. While the diagnostic time trends were comparable between males and females for all mental disorders combined, we found considerable differences regarding the specific diagnoses (online supplemental eTables 3–21, github.com). Cumulative incidence increased more among males for the diagnoses of ADHD, ASD, ODD/CD, tic disorders, attachment disorders and substance use disorders. All other diagnoses increased more among females. Across the sexes, the greatest absolute increases in cumulative incidence comparing the birth cohorts of 1976–1980 and 2001–2005 were in the diagnoses of ASD (females: from 0.03% (95% CI 0.02% to 0.03%) to 3.44% (95% CI 3.34% to 3.55%) (p<0.0001); males: from 0.12% (95% CI 0.11% to 0.14%) to 6.11% (95% CI 5.98% to 6.24%) (p<0.0001)), ADHD (females: from 0.05% (95% CI 0.04% to 0.06%) to 4.77% (95% CI 4.65% to 4.89%) (p<0.0001); males: from 0.12% (95% CI 0.10% to 0.14%) to 7.68% (95% CI 7.53% to 7.82%) (p<0.0001)), anxiety disorders (females: from 0.71% (95% CI 0.66% to 0.75%) to 9.20% (95% CI 9.04% to 9.37%) (p<0.0001); males: from 0.57% (95% CI 0.53% to 0.60%) to 5.62% (95% CI 5.49% to 5.74%) (p<0.0001)) and mood disorders (including depression) (females: from 0.10% (95% CI 0.09% to 0.12%) to 3.45% (95% CI 3.34% to 3.55%) (p<0.0001); males: from 0.05% (95% CI 0.04% to 0.06%) to 1.70% (95% CI 1.63% to 1.77%) (p<0.0001)). For most diagnoses, we found similarly low cumulative incidence at baseline for both females and males. However, during follow-up time trends diverged for most of the diagnoses. Only eating disorders, anorexia nervosa and personality disorders showed consistently higher cumulative incidence among females across all birth cohorts. For ADHD, ASD and anxiety, the cumulative incidence increased more for each younger birth cohort group. The same was observed for depression, apart from the two most recent birth cohort groups having similar cumulative incidences among females. Cumulative incidence of ADHD and ASD increased most among males, with 7.56 and 5.99 percentage points increase, respectively, compared with females, with 4.72 and 3.41 percentage points increase. Among females, anxiety disorders and mood disorders increased 8.49 and 3.35 percentage points, respectively, compared with 5.05 and 1.65 percentage points increase among males. All other diagnoses, except subdiagnoses of the before mentioned diagnostic groups, had cumulative incidence below 2% during the entire period for both sexes, most of them closer to 1%. Cumulative incidence of any mental disorder also increased significantly after outpatient diagnoses were included in the registers, from 10.49–10.73% among people born in 1995–1997 to 15.59–15.61% among people born in 2004–2005 (online supplemental eTable 4). Time trends of incidence rates We also observed large increases in the incidence rate of any mental disorder diagnosis, with a consistent increase from 12.7 per 10 000 person years in 1994 to 129.3 per 10 000 person years in 2023 (figure 1, online supplemental eTable 22, github.com). We found noticeable differences in the time trends of diagnosis-specific rates, particularly when comparing males and females (figure 2, online supplemental eTables 23-39, github.com). Among diagnostic groups, we observed the largest absolute increases in incidence rates for ASD, ADHD, anxiety disorders (specifically anxiety) and mood disorders (specifically depression). After many years where ADHD and ASD rates increased more among males, rates among females increased more rapidly in later years and came close to the male rates at the end of follow-up (figure 2, online supplemental eTables 35-36). Rates of anxiety disorders diagnoses were similar in both sexes until around 2011, where the rate stagnated among males but increased rapidly among females, apart from a small dip in anxiety in 2017–2020 (figure 2, online supplemental eTables 29 and 30). Diagnostic rates of mood disorders and depression were consistently higher among females, plateauing around 20 new cases per 10 000 person years from 2014 and onwards, while peaking at around 9–10 new cases per 10 000 person years among males in 2017 and decreasing thereafter (figure 2, online supplemental eTables 26, 28). In males, incidence rates of eating disorders (including anorexia nervosa) and personality disorders were relatively low and stable throughout the entire period, while in females, continuously increasing incidence rates of eating disorders were observed. (figure 2; online supplemental eTable 32–33). Decreasing incidence rates were observed in the latest decade of follow-up among both sexes for substance use disorders, schizophrenia, personality disorders, ODD/CD and attachment disorders, (figure 2; online supplemental eTables 23, 25, 34, 37–38) and among males also for anxiety disorders, anxiety and OCD (figure 2; Online supplemental eTable 29–31). Since the year 2015, rates have stabilised for tic disorders in both sexes (figure 2; online supplemental eTable 39). Age-standardised incidence rates of mental disorder diagnoses per 10 000 person-years before age 18 from 1994 to 2023—by specific diagnosis types. ADHD, attention-deficit hyperactivity disorder; ASD, autism spectrum disorders; CD, conduct disorder; OCD, obsessive–compulsive disorder; ODD, oppositional defiant disorder. Incidence rates of any mental disorder diagnosis across age groups In the entire population, there were noticeable differences in the age-stratified incidence rates for females and males (figure 1B,online supplemental eTable 40). Among males, the incidence rates of any mental disorder diagnosis peaked around 6–12 years of age. This age group also had the highest incidence rate when directly compared with the other age-groups (p<0.0001). The incidence rates among females peaked around ages 14–17. The diagnostic rate of ADHD was highest among younger males aged 6–11 years, whereas it was highest for females during their teenage years (online supplemental eFigure 160). Incidence rates of ASD diagnoses among males peaked among 6–8 year-olds, whereas among females, the incidence of ASD diagnoses peaked among 12–17 year-olds (online supplemental eFigure 149). The incidence of anxiety peaked among females in the late teens, and in males among 9–11 year-olds until 2015, whereafter the incidence was highest among 15–17 year-olds (online supplemental eFigure 94). In depression, incidence rates were considerably higher among 15–17 year-olds throughout the entire study period, particularly among females (online supplemental eFigure 72). For schizophrenia and personality disorder, the highest rates were also seen among 15–17 year-olds for both sexes (online supplemental eFigures 39, 138). Among males, OCD is primarily assigned from age 9 and onwards, while in females it is most common among those aged 15–17 (online supplemental eFigure 105). Anorexia nervosa was almost exclusively diagnosed in females from age 12, and from around 2015 the incidence rate among the 15–17 year-old stagnated, while it continued to increase among 12–14 year-olds (online supplemental eFigure 127). Diagnoses of ODD/CD, attachment disorders and tic disorders were primarily assigned from age 6 onwards (online supplemental eFigures 170, 181 and 191). For ODD/CD, this pattern was observed in both sexes, whereas for tic disorders and attachment disorders, it was observed only among females. In males, attachment disorders showed higher incidence rates at younger ages, while tic disorders peaked in 6–11 years age group. Trends across risk factors of mental disorders The following risk factors of mental disorders were found to be associated with diagnosis of a mental disorder before age 18 across the entire study period (all p<0.001): Maternal/paternal psychiatric diagnosis, lower maternal/paternal educational level, not living with both parents, maternal/paternal unemployment, higher maternal/paternal CCI, higher number of different addresses during childhood and being a native-born Dane (figure 3, online supplemental eFigures 2-12, online supplemental eTables 41–55, github.com). For males, being born outside a capital municipality was also associated with a higher rate of a mental health diagnosis (online supplemental eTable 43). Age-standardised incidence rates of any mental disorder diagnosis per 10 000 person-years before age 18 from 1994 to 2023—by risk factors. ‘Municipality of birthplace’ is equivalent to ‘Urbanicity of birthplace’ elsewhere in the paper and supplement. CCI, Charlson Comorbidity Index. For most risk factors, the age-stratified, univariate association to diagnosis of any mental disorder did not change considerably during the study period (online supplemental eTables 41–55). However, the rate ratio of mental disorder diagnoses decreased slightly for the following risk factors: not living with both parents, a higher number of previous addresses, higher maternal age at childbirth, lower maternal/paternal work status, a maternal/paternal psychiatric diagnosis and higher maternal/paternal CCI score. Conversely, the rate ratio increased slightly for the following risk factors: being an immigrant/descendant, lower maternal income and being born in a municipality outside the capital. It should also be noted that the p values for time trend in rate ratios were significant for all risk factors, except for the paternal income quintile among females. Patterns were similar across specific diagnoses, however, small numbers complicated interpretation in some cases (online supplemental eFigures 13–197). Estimating the effect of shift in age at first diagnosis of a mental disorder While the overall incidence rates and cumulative incidences of mental disorder diagnoses have increased since 1976, the mean age of first diagnosis for both females and males has decreased since 1990 (figure 4B). In the most recent period (2020–2023), the decreasing diagnostic age accounted for an estimated 21.6% (95% CI 13.6% to 25.8%) of the total rate among females and 14.9% (95% CI 4.6% to 21.4%) among males. Over the entire study period (1976–2023), the corresponding estimates were 16.3% (95% CI 14.4% to 16.7%) and 14.2% (95% CI 11.9% to 15.2%) (figure 4C). Notably, the age-shift-adjusted rate increased in a similar fashion to the observed rate (figure 4A). Estimating the effect on incidence rates of changing age at first mental disorder diagnosis. Rate difference: Observed minus age−shift−adjusted incidence rate, average across period; Share of observed rate (%): proportion of the observed rate attributable to age−shift, average across period. Comparison with time trends of somatic disorder diagnoses The incidence rate of mental disorder diagnoses per 10 000 person years for males and females increased from 14.9 to 129.9 and 10.4 to 129.1, respectively, corresponding to an increase between 772% and 1141% (online supplemental eFigure 198, online supplemental eTable 22). In comparison, the incidence rate of somatic disorder diagnoses increased only slightly from 1086.6 to 1161.4 among males and 972.7 to 1084.8 among females, corresponding to an increase of 7% and 12%, respectively (online supplemental eFigure 198, online supplemental eTable 57). Discussion This is the most comprehensive study to date, quantifying the time trends in mental disorder diagnoses between 1994 and 2023 in a national sample of children and adolescents. During the study period, we observed a substantial and continuous increase in the risk of receiving a mental disorder diagnosis before age 18 in both females and males. This overall trend is consistent with previous research indicating a rising burden of mental disorder diagnoses in the younger population.15 We found the largest absolute increases in diagnoses of ADHD and ASD in both sexes, with higher rates among males throughout the entire study period. However, diagnostic rates among females have increased more rapidly, particularly among teenagers, and are now close to the overall male rate for these disorders. This convergence in diagnosis rates is also reported by previous literature and could represent a heightened awareness of existing symptoms in the population and healthcare system.20 24 29 Among females, a notable increase was seen in anxiety and depression, particularly pronounced among those aged 15–17 years. While the increase was continuous for anxiety, stagnation was observed for depression in both cumulative incidence and incidence rates, especially among teenagers. Consistent with our findings, previous studies support an increasing trend in the diagnoses of ADHD, ASD, anxiety and depression,20 29–31 with the highest incidences of anxiety and depression observed among teenage females.20 31 However, a population-based study from the UK found that children aged 6–9 years had the highest rates of ADHD and ASD for both sexes between 2003 and 2018.31 The increasing diagnostic rate does not necessarily reflect a rise in the number of individuals affected by a psychiatric condition in the general population. Screening studies have found a stable underlying prevalence of around 13.4–15.5%.14 15 This aligns well with our findings of a cumulative incidence in the most recent birth cohorts around 14.6–14.8%. As the underlying prevalence seems to have remained stable over time, the increase in diagnostic rates may reflect that those who meet the criteria are now more frequently diagnosed and, as a result, given the opportunity for treatment, reducing the prior treatment gap. We estimated that a substantial part of the observed diagnostic rate, both recently and over the past decades, was attributable to decreasing age at first diagnosis. Nonetheless, the diagnostic trend, after adjusting for the age-shift, indicates a clear increase in the underlying rate, independent of the age-shift. In total, these trends are consistent with an increasing share and number of initial psychiatric assessments occurring within child and adolescent mental health services, straining the available resources in this sector. While the associations between known risk factors of mental disorders and diagnosis of mental disorders changed significantly during the study period, changes in rate ratios were generally small, reflecting a relatively stable relationship between the established risk factors and diagnostic rates. This stability is reflected in the stratified incidence rates, which generally show similar time trends of diagnosis across risk factor levels for the included mental disorders. Looking at trends in rate ratios and incidence rates, we did not find any clear indications that any single factor could explain the observed increase in diagnostic rates of mental disorders. From 1994 data on outpatient and emergency ward visits were included in the Danish registers, which likely have contributed to the increase observed after 1994, although similar increases were observed in other countries during the 1990s,32 33 indicating that other factors were also important, and that in the period prior to 1994, outpatient treatment in child and adolescent psychiatry was less common. Greater public awareness and reduced stigma may have contributed to an increased help-seeking behaviour for mental health treatment. In addition, improved evidence-based treatments and expanded psychiatric services reflecting increased political prioritisation of the field, along with updates to diagnostic criteria and education of healthcare providers, may have led to earlier and more frequent diagnoses. Since child and adolescent psychiatry was established as a medical specialty in Denmark, the specialty has expanded in both capacity and specialisation, with a notable increase in the number of specialists.34 This increased capacity has enabled the psychiatric assessment of more children and adolescents and moving closer to the clinical demand. Interestingly, we did not find a comparable relative increase in diagnoses of somatic disorders, which is important for future healthcare planning. Somatic diagnoses might also be less influenced by contextual and societal factors, which may contribute to their comparatively more stable time trends in diagnoses. Strengths and limitations This study has several strengths. First, it is the most comprehensive nationwide study focusing on age-standardised diagnostic time trends among children and adolescents to date. It is based on validated registers, which allowed the inclusion of the entire population of children and adolescents and diagnosis-specific analyses. Furthermore, it has complete follow-up across 3 decades in the analyses of incidence rates, and a 47-year follow-up in the analyses of cumulative incidence. All available data were included, with individuals contributing observation time solely during the study period; although follow-up time varies depending on time of birth, the inclusion of age and calendar period in the statistical models adjusts for this variation. Moreover, we provide the first estimate on how much the decline in mean age at first diagnosis may have accounted for to the observed incidence rate, based on an established demographic method.28 Although this estimate is subject to statistical uncertainty, it nonetheless provides a useful indication of the magnitude of the contribution of the younger age at first diagnosis. Among limitations, the study is exclusively based on registry data from inpatient and outpatient psychiatric records in public hospitals. Thus, patients who only received a diagnosis in primary care or from a private practising psychiatrist, and who did not have a hospital contact with the given diagnosis during the whole study period, could not be included in the study. This may particularly affect diagnoses which are often treated outside public hospital settings, such as ADHD, ASD, anxiety and depression. In Denmark children receiving ADHD medication with no hospital diagnosis of ADHD has increased substantially since the year 2000.29 Moreover, the antidepressant treatments that are initiated in secondary care35 is first captured, if they subsequently have a hospital contact. Thus, combined with a recent migration of psychiatrists from the public to the private sector, we may to some degree underestimate the increase in children and adolescents with a mental disorder diagnosis in Denmark. Outpatient data was only available from 1995 onwards, meaning that cumulative incidence from 1976 to 1994 was based on inpatient data alone. We found this a reasonable approach for estimating nationwide diagnostic incidence, as outpatient diagnosis prior to 1995 was rare due to very limited facilities, although a small amount of under-representation must be expected. Furthermore, a sensitivity analysis revealed a considerable increase in cumulative incidence from 1995 and onwards where complete both inpatient and outpatient was available. While evidence indicates good validity for many mental disorder diagnoses in the Danish hospital registers, not all diagnoses in the present study have been validated in the registers, and validation studies generally cover short periods of time.26 Furthermore, to mitigate uncertainty of estimates not all rare diagnoses are included, but these are encompassed in the broader diagnostic groups. It should be noted that the rates in these groups are in some cases mainly driven by one or two frequent diagnoses, for example, anorexia nervosa within the group of eating disorders. Another potential limitation concerns the generalisability of the findings, which may be limited to countries with similar healthcare systems. Danish register data on diagnoses reflect a tax-funded system with universal access and gatekeeping referral pathways, which may lead to higher detection and registration of mental disorders compared with settings with more restricted or unequal access to care. The estimation of age-shift effects assumes that the age distribution of first diagnoses shifts along the age axis without changing its shape. If the spread of ages at diagnosis also changed over time, the estimated contribution of the age-shift could be biased in either direction.36 The estimate is also sensitive to changes in diagnostic practices and registrations. However, age-shift effects in 1976–2023, which included pre-outpatient data and changes from ICD-8 to ICD-10, were similar to effects in 1994–2023 with only ICD-10 from 1994 and inpatient data from 1995. Clinical implications Our study provides important information on the temporal changes in the epidemiology of mental disorder diagnoses among children and adolescents, which is essential knowledge for future research on mental disorders and psychiatric healthcare planning. A substantial part of the currently observed incidence rate is accounted for by younger age at first diagnosis. At the same time, the age-shift-adjusted rate indicates that the underlying diagnostic rate has genuinely risen over the study period, suggesting that while earlier identification is a likely mechanism, other factors also contribute to the trend. Earlier diagnosis implies that more of the associated clinical burden will be carried by child and adolescent mental health services. The present findings highlight the importance of reorganising clinical services to ensure that they adequately address the needs of the population. Further research would be valuable to investigate whether earlier diagnosis contributes to improved outcomes. Of interest is also investigating whether the included risk factors of mental disorders are non-specific predictors of poor health by testing the similarity of associations across specific mental disorder diagnoses and somatic disorder diagnosis. Supplementary material