2015 Collishaw Annual Research Review Secular Trends in Child and Adolescent Mental.

Download as pdf or txt
Download as pdf or txt
You are on page 1of 24
 
Annual Research Review: Secular trends in child andadolescent mental health
Stephan Collishaw
Institute of Psychological Medicine and Clinical Neurosciences, MRC Centre for Neuropsychiatric Genetics andGenomics, Cardiff University, Cardiff, UK
Background:
 Child and adolescent mental health problems are common, associated with wide-ranging functionalimpairments, and show substantial continuities into adult life. It is therefore important to understand the extent towhich the prevalence of mental health problems has changed over time, and to identify reasons behind any trends inmental health.
 Scope and Methodology:
 This review evaluates evidence on whether the population prevalence of child and adolescent mental health problems has changed. The primary focus of the review is on epidemiologicalcross-cohort comparisons identified by a systematic search of the literature (using the Web of Knowledge database).
Findings:
 Clinical diagnosis and treatment of child and adolescent psychiatric disorders increased over recentdecades. Epidemiological comparisons of unselected population cohorts using equivalent assessments of mentalhealth have found little evidence of an increased rate of ADHD, but cross-cohort comparisons of rates of ASD arelacking at this time. Findings do suggest substantial secular change in emotional problems and antisocial behaviourinhigh-incomecountries,includingperiodsofincreaseanddecreaseinsymptomprevalence.Evidencefromlow-andmiddle-income countries is very limited. Possible explanations for trends in child and adolescent mental health arediscussed. The review also addresses how cross-cohort comparisons can provide valuable complementaryinformation on the aetiology of mental illness.
 Keywords:
 Time trends, secular change, depression, antisocial,psychopathology.
Introduction
Epidemiological surveys indicate that psychiatricdisorders in children and adolescents are common,affecting at least one in ten 5
 – 
16-year olds (Green,McGinnity, Meltzer, Ford, & Goodman, 2005; Melt-zer, Gatward, Goodman, & Ford, 2000). Theseinclude neurodevelopmental disorders such as aut-ism spectrum disorder (ASD) and attention-deficithyperactivity disorder (ADHD), mood and anxietydisorders, and conduct disorders. Emotional andbehavioural problems more broadly defined areconsiderably more common. Mental health problemshave wide-ranging detrimental impacts on youngpeople’s well-being, health and education. They arealso not transient phenomena (Maughan & Colli-shaw, 2015; Thapar, Collishaw, Pine, & Thapar,2012). Long-term prospective studies highlight con-tinuitiesbetweenchildandadultmentalhealth,withover half of disorders in young adulthood precededby a psychiatric disorder in childhood (Kim-Cohenet al., 2003), and conversely strong persistence andrecurrence of childhood psychiatric conditions intoadulthood (Birmaher et al., 2004). Psychiatric prob-lems account for a substantial and increasing world-wide burden of ill health (Murray et al., 2012).Psychopathology is associated with risk for suicide,the second or third leading cause of death inadolescents and young adults (Windfuhr et al.,2008); and with effects on chronic illness and earlymortality (Jokela, Ferrie, & Kivimaki, 2009; Mau-ghan, Stafford, Shah, & Kuh, 2014; Thapar et al.,2012).Thepreventionofchildandadolescentmentalhealth problems is therefore an important healthpriority (Institute of Medicine, 2009), and an accu-rate understanding of secular trends in child andadolescent psychopathology is needed to gauge thesuccess of attempts to reduce the burden of childpsychiatric problems.Despite efforts to better understand child andadolescent psychiatric conditions and how to pre-vent them, there are concerns among parents,teachers and health professionals that today’s chil-dren are more prone to mental health problems thanwere previous generations of children and youngpeople (Barnard, Potter, Broach, & Prior, 2002;Russell, Kelly, & Golding, 2010). Indeed, the past20 years has seen considerable increases in thediagnosis and treatment of child and adolescentpsychiatric disorders (Fombonne, 2009; Getahunet al., 2013; Ma, Lee, & Stafford, 2005; Smith,Larkin, & Southwick, 2008). As discussed below,there are a range of explanations for such trends,including increased help-seeking by parents and young people, improved screening and clinical rec-ognition in schools and primary care, medicalizationof feelings and behaviours previously considered asnormal (Dowrick & Frances, 2013), and a broaden-ing of diagnostic classifications of psychiatric disor-ders. It is also possible that there have been realchanges in the prevalence of childhood psychiatricconditions.
Conflict of interest statement: No conflicts declared.
©
 2014 Association for Child and Adolescent Mental Health.Published by John Wiley & Sons Ltd, 9600 Garsington Road, Oxford OX4 2DQ, UK and 350 Main St, Malden, MA 02148, USA
Journal of Child Psychology and Psychiatry 
 56:3 (2015), pp 370393 doi:10.1111/jcpp.12372
 
 The first aim of this review was to evaluateevidence on whether the
 population prevalence 
 of child and adolescent mental health problems haschanged. It is first necessary to consider thestrengths and weaknesses of different methodologi-cal designs that have been used to address thisquestion. Theseincludecross-sectionalstudies com-paring lifetime prevalence across different agecohorts, meta-analyses of epidemiological studiesusing psychiatric interviews, cross-cohort compari-sons in which symptom prevalence is directly com-pared, and time series tracking population suicideand offending rates. Using convergent evidenceacross these different study methods, but with aprimary focus on epidemiological cross-cohort com-parisons, this paper will review trends in the mostcommon types of child and adolescent psychopa-thology
 – 
 neurodevelopmental problems, antisocialbehaviour and emotional problems. In understand-ing these trends, it is important to consider possiblevariations between children and adolescents, boysand girls, different socioeconomic groups and acrossdifferent countries. Social change affecting risks formental health problems is unlikely to have impactedin the same way on these different groups. Thereview focuses on trends since the middle of the lastcentury, but recent fluctuations may be relativelyminor when compared to very long-term historicalchange in rates of mental ill health and violence(Pinker, 2011). It is also important to note from theoutset that epidemiological ‘like-for-like’ compari-sons of trends in population prevalence are at thisstage only available for a small number of high-income countries, and almost completely lacking inlow- and middle-income countries.Next, the review considers the reliability andvalidity of evidence about trends in child andadolescent mental health, and in particular, thepossibility that changes in reporting confoundobserved trends. A number of methods can helppoint to more robust conclusions. These include theconsideration of convergent evidence across multipleinformants and study methods, data about predic-tive validity where longitudinal data have beencollected across multiple cohorts, and evidence forspecificity of trends in particular aspects of childmental health.Cross-cohort comparisons provide an opportunityto address a number of other important questionsbeyond simply tracking trends in prevalence.A second aim was to consider possible explanationsfor any increases or decreases in the prevalence of child mental health problems. The past 50 years hasseen substantial societal changes which are likely tohave had far-reaching impacts on children’s livesand well-being. Cross-cohort comparisons withlinked data on risk factors and mental health aresparse, but where such data are available theyprovide valuable opportunities for understandingthe possible impact of changes in risk factors onchild psychopathology. A related issue is whethersocial inequalities in child mental health haveincreased or decreased. Addressing health inequal-ities in morbidity and mortality is recognized as animportant policy priority nationally and internation-ally (Marmot, 2010). However, the extent of socialinequalities affecting child and adolescent mentalhealth, and whether progress is being made inreducing such inequalities, is rarely considered.A full review of the myriad changes in children’s livesthat haveoccurred overrecent decadesis beyond thescope of this review. Instead, a number of illustrativeexamples are discussed to highlight the methodolog-ical and conceptualchallenges involved.Athirdaimwastohighlight exampleswherecross-cohort comparisons have provided novel insights onthe role of hypothesized risk factors. Sudden dis-continuities in the social environment can approxi-mate features of a ‘natural experiment(Thapar &Rutter, 2015). In addition, studies of secular changeforarangeofphenotypes(e.g.height,ageatpubertalmaturation, cognitive ability) have pointed to some-what different conclusions regarding the role of environmental and heritable factors when comparedto studies of within-population variation (Dickens &Flynn, 2001; Silventoinen et al., 2003). Throughout, this review aims to identify importantknowledgegapsandtoprovideguidanceonprioritiesfor future research.
Methodological considerations
Studies have used various designs to considerwhether the prevalence of child mental health prob-lems has changed across time. Apart from researchtracking rates of diagnosis and treatment in clinicalpractice, studies have also examined populationsuicide and crime rates, generational differences inlifetime rates of retrospectively reported mental illhealth, used meta-analytic methods to comparerates of interview-assessed disorder in epidemiolog-ical surveys conducted at different time points, andundertaken direct cross-cohort comparisons of epi-demiological studies using equivalent mental healthsymptom screens.Each of these approaches has pros and cons. Fourgeneral issues are important to consider. First, towhatextentisastudyabletocompare‘likewithlike’,bothintermsofsamplesandmeasurement?Second,to what extent does the coverage of a study extend tothe whole population rather than focusing onselected subgroups? A third challenge relates to thepossibility that the way informants report mentalhealth problems has changed, even when equivalentmeasures are used in different cohorts. Finally,linked data on hypothesized explanatory factors isrequired in order to test explanations for trends inmental health.More specifically, time trends studies can beevaluated according to the extent to which they
©
 2014 Association for Child and Adolescent Mental Health.doi:10.1111/jcpp.12372 ARR: Secular trends in child and adolescent mental health
 371
5
 
encompasses the following features: (a) comparisonof unselected representative epidemiological cohortsusing the same sampling frames; (b) equivalence inresponse and methods for dealing with selectivenonresponse; (c) use of equivalent measures of mental health symptoms or disorders; (d) the abilityto provide convergent data across multiple methodsor informants; (e) data on impairment as well assymptoms of psychopathology; (f) contemporaneousrather than retrospective assessment; (g) linked andcomparable data on hypothesized explanatory fac-tors; (h) longitudinal follow-up or data linkageallowing a test of the predictive validity of mentalhealthtrendswithrespecttoindependentlyassessedfunctional outcomes.Finally, it is important to consider distinctionsbetween age, period and cohort effects (Robertson &Boyle, 1998). These are typically confounded in moststudies, but have important implications for under-standing reasons behind changes in prevalence. It iswellknownthatpatternsofdiseasevarybyage.Theymight also vary for different periods of historicaltime, for example as a result of economic recession.Cohorts of individuals born at the same time mayalso share risk for disease independent of their ageand period of assessment. Comparisons of longitu-dinal birth cohorts at multiple points in times arerequired to distinguish between period and cohorteffects (Keyes et al., 2014), but very few such studieshavebeenundertaken.Inthispaper,theterm‘cross-cohort comparison’ refers to comparisons of repeatepidemiological samples assessed at different pointsin times. It is important to remember that anydifference could reflect either period effects or cohorteffects or some combination of the two.
Trends in child and adolescent mental health
 This section of the review focuses on trends in themost commonly occurring types of child and adoles-cent psychopathology. The main focus is on studiesthat meet the most important of the criteria outlinedabove, namely those in which contemporaneousrepresentative population samples have been com-pared using equivalent assessments of mentalhealth. Tables 1
 – 
3 provide overviews of key studies of children, adolescents and mixed age groups respec-tively that were included in this section of the review(see online supporting information for description of search strategy). In total, 21 studies were identifiedthat met relevant inclusion and exclusion criteria. Of these, the majority were undertaken in the UnitedKingdom (6), Finland (3), other Nordic countries (3)or the Netherlands (3). Five used samples of chil-dren, 11 of adolescents, and 5 included children andadolescents. Studies used various samplingapproaches, including via schools, private house-holds or follow-ups of birth cohorts. Most studiesused equivalent sampling approaches at each timepoint. Studies have typically reported high responserates (
>
80%) or addressed the possibility of selectivedrop-out analytically, for example by using cohort-specific sample weights derived from prior morecomplete data. All selected studies used validatedsymptom screens with little or no variation acrossassessments. One longer term trends study, wheresimilar but nonidentical assessments were com-pared used calibration methods to ensure compara-bility of comparisons across time (Collishaw,Maughan, Goodman, & Pickles, 2004). Not includedare studies which have used bespoke items toexamine trends in mental health-related phenom-ena, including self-reported feelings of depression oranxiety, self-harm or suicidality, delinquency andsubstanceuse.Itwasbeyondthescopeofthisreviewto fully consider these here, but they may alsoprovide valuable information on trends in child andadolescent mental health. One example is the cross-national World Health Organization Health Behav-iour among School-aged Children series of studies(see HBSC.org for details). Turning back to thestudies included here, data from multiple infor-mants, information about functional impact of symptoms, longitudinal validation of trends, infor-mation on whether trends differ by gender or soci-odemographic subgroup and linked data onpotential explanatory factors all provide importantstrengths.Moststudiesincludedonlyoneinformant,typically parent reports for younger children and youth self-reports in adolescence. One Finnish studyof 8-year-old children also included child-basedreports of symptoms (Sourander, Niemel
a, Santa-lahti, Helenius, & Piha, 2008). Impact of symptomson psychosocial functioning has rarely beenassessed, and only one study used independent dataon longitudinal outcomes to validate trends (Colli-shaw et al., 2004). Finally, while most studiesexamined whether trends differed by gender, only aminority of studies directly tested how far trends inmental health might be accounted for by trends inhypothesized explanatory factors.
Neurodevelopmental disorders
Perhaps, the most dramatic increases in child psy-chiatric diagnoses relate to changes in ADHD andautism spectrum disorders (ASD). In the 1960saround one in 2500 children were diagnosed withautism (Lotter, 1966); the incidence of autism spec-trum disorders has increased steadily to around 3
 – 
6children per 1000 children at the beginning of the21st Century (Baird et al., 2006; Rutter, 2005). There is some evidence of a continuing increase indiagnoses of autism spectrum disorders in the 21stCentury (Boyle et al., 2011; Russell, Rodgers, Ukou-munne, & Ford, 2014; but see also Taylor, Jick, &MacLaughlin, 2013). A recent systematic review of worldwide studies found a median prevalence of 62per 10,000 (Elsabbagh et al., 2012), and in some
©
 2014 Association for Child and Adolescent Mental Health.
372
 Stephan Collishaw J Child Psychol Psychiatry 2015; 56(3): 37093
5
 
      T    a      b      l    e     1
    S   u   m   m   a   r   y   o    f   c   r   o   s   s  -   c   o    h   o   r    t   c   o   m   p   a   r    i   s   o   n   s    t   u    d    i   e   s   o    f   c    h    i    l    d   r   e   n   u   s    i   n   g   m   e   n    t   a    l    h   e   a    l    t    h   s   y   m   p    t   o   m   s   c   r   e   e   n   s    L   e   a    d   a   u    t    h   o   r    (   y   e   a   r    )    C   o   u   n    t   r   y    S    t   u    d   y   y   e   a   r   s    A   g   e    (   y   e   a   r   s    )    S   a   m   p    l   e    N   s    R   e   s   p   o   n   s   e   r   a    t   e   s    (    %    )    I   n    f   o   r   m   a   n    t   s    M   e   a   s   u   r   e   s    O   u    t   c   o   m   e   s    T   r   e   n    d   s    b   y   s   u    b   g   r   o   u   p    E   x   p    l   a   n   a    t   o   r   y    f   a   c    t   o   r   s    K   e   y    fi   n    d    i   n   g   s    S   o   u   r   a   n    d   e   r    (    2    0    0    4 ,    2    0    0    8    )    S   a   n    t   a    l   a    h    t    i    (    2    0    0    5 ,    2    0    0    8    )    I    l   o    l   a    (    2    0    1    3    )    L   u   n    t   a   m   o    (    2    0    1    2    )    F    i   n    l   a   n    d    1    9    8    9    1    9    9    9    2    0    0    5    8    H    H    /    S    C    H    H    H    /    S    C    H    H    H    /    S    C    H    9    8    6    8    3    1    8    7    0    9    6    8    6    8    4    P   a   r   e   n    t    T   e   a   c    h   e   r    C    h    i    l    d    R   u    t    t   e   r    A    R   u    t    t   e   r    B    C    D    I    M   e   a   n   s    H    i   g    h   s   c   o   r   e    G   e   n    d   e   r    S    E    S    F   a   m    i    l   y    t   y   p   e    L    i    f   e   e   v   e   n    t   s    D   e   c   r   e   a   s   e    d   p   a   r   e   n    t  -   r   e   p   o   r    t   c   o   n    d   u   c    t   a   n    d    t   o    t   a    l   p   r   o    b    l   e   m    h    i   g    h   s   c   o   r   e   s    1    9    8    9
   –
    1    9    9    9    (    b   o   y   s    )    I   n   c   r   e   a   s   e    d   c    h    i    l    d  -   r   e   p   o   r    t   e    d    d   e   p   r   e   s   s    i   o   n    (   g    i   r    l   s    )    I   n   c   r   e   a   s   e    d   s   o   m   a    t    i   c   s   y   m   p    t   o   m   s    C    h   a   n   g   e   s    i   n    b   u    l    l   y    /   v    i   c    t    i   m    i   z   a    t    i   o   n    S   e    l    l   e   r   s    (    2    0    1    4    )    U    K    1    9    9    9    2    0    0    4    2    0    0    8    7    H    H    H    H    B    C    1 ,    0    3    3    6    4    8    1    3    8    5    7    8    3    7    6    7    2    P   a   r   e   n    t    T   e   a   c    h   e   r    S    D    Q    M   e   a   n   s    V   a   r    i   a   n   c   e    H    i   g    h   s   c   o   r   e    I   m   p   a   c    t    G   e   n    d   e   r    S    E    S    D   e   m   o   g   r   a   p    h    M   e   a   n   p   r   o    b    l   e   m   s ,   v   a   r    i   a   n   c   e ,   a   n    d    h    i   g    h   p   r   o    b    l   e   m   s   a    l    l   r   e    d   u   c   e    d    (   e   s   p   e   c    i   a    l    l   y    b   o   y   s   ;   g   e   n    d   e   r   c   o   n   v   e   r   g   e    d    )    I   m   p   a   c    t   c    h   a   n   g   e    d    (   e   s   p    b   o   y   s    )    T   r   e   n    d   s   n   o    t   e   x   p    l   a    i   n   e    d    b   y    d   e   m   o   g   r   a   p    h    i   c   c    h   a   n   g   e   s    T    i   c    k    (   v   a   n    d   e   r    E   n    d   e ,    K   o   o    t ,    &    V   e   r    h   u    l   s    t    2    0    0    7    )    N   e    t    h   e   r    l   a   n    d   s    1    9    8    9    2    0    0    3    2
   –
    3    H    H    3    9    4    2    7    9    9    0    8    0    P   a   r   e   n    t    C    B    C    L    M   e   a   n   s    H    i   g    h   s   c   o   r   e    G   e   n    d   e   r    S    E    S    N   o    S   m   a    l    l    d   r   o   p    i   n   m   e   a   n   p   r   o    b    l   e   m   s   a   n    d    h    i   g    h   s   c   o   r   e   s    M   a    t    i    j   a   s   e   v    i   c    h    (    2    0    1    4    )    B   r   a   z    i    l    1    9    9    3    2    0    0    4    4    B    C    B    C    6    3    4    3 ,    7    5    0    8    7    9    1    P   a   r   e   n    t    C    B    C    L    M   e   a   n   s    H    i   g    h   s   c   o   r   e    G   e   n    d   e   r    S    E    S    D   e   m   o   g   r   a   p    h ,   p   e   r    i   n   a    t   a    l ,   m   a    t   e   r   n   a    l   p   s   y   c    h    i   a    t   r    i   c    S   u    b   s    t   a   n    t    i   a    l    i   n   c   r   e   a   s   e   s    i   n   m   e   a   n   s   a   n    d   a    b   n   o   r   m   a    l   r   a   n   g   e   s   c   o   r   e   s ,   e   s   p   e   c    i   a    l    l   y   a   g   g   r   e   s   s    i   v   e    b   e    h   a   v    i   o   u   r ,    b   u    t   a    t    t   e   n    t    i   o   n   p   r   o    b    l   e   m   s    d   e   c   r   e   a   s   e    d   ;    S    i   m    i    l   a   r    t   r   e   n    d   s    b   y   g   e   n    d   e   r   ;    I   n   c   r   e   a   s   e    d   p   r   o    b    l   e   m   s   m   o   s    t   m   a   r    k   e    d    f   o   r    l   o   w    S    E    S .    I   n   c   r   e   a   s   e   s   n   o    t   e   x   p    l   a    i   n   e    d    b   y   e   x   p    l   a   n   a    t   o   r   y    f   a   c    t   o   r   s    M   c    A   r    d    l   e    (    2    0    0    3    )    U    K    1    9    7    3    1    9    9    4    7
   –
    9    S    C    H    S    C    H    5    1    5    1 ,    0    4    4    N   o    t   r   e   p   o   r    t   e    d    T   e   a   c    h   e   r    R   u    t    t   e   r    B    H    i   g    h   s   c   o   r   e    N   o    N   o    R   e    d   u   c   e    d   r   a    t   e   o    f    h    i   g    h   s   c   o   r   e   s    H    H ,   r   a   n    d   o   m   s   a   m   p    l   e   o    f   p   r    i   v   a    t   e    h   o   u   s   e    h   o    l    d   s   ;    B    C ,    b    i   r    t    h   c   o    h   o   r    t   ;    S    C    H ,   s   c    h   o   o    l  -    b   a   s   e    d   s   a   m   p    l    i   n   g .
©
 2014 Association for Child and Adolescent Mental Health.doi:10.1111/jcpp.12372 ARR: Secular trends in child and adolescent mental health
 373
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5

Reward Your Curiosity

Everything you want to read.
Anytime. Anywhere. Any device.
No Commitment. Cancel anytime.
576648e32a3d8b82ca71961b7a986505