1997 behavioral characteristics of dsm-iv adhd subtypes ina scholl - gaub, carlson

Upload: donileite

Post on 03-Jun-2018

218 views

Category:

Documents


0 download

TRANSCRIPT

  • 8/12/2019 1997 Behavioral Characteristics of DSM-IV ADHD Subtypes Ina Scholl - Gaub, Carlson

    1/9

    Journal

    of

    Abnormal Child Psychology, Vol. 25, No. 2, 1997,pp. 103-111

    Behavioral

    Characteristics

    of

    DSM-IV

    ADHD

    Subtypes

    in a School-Based Population

    Miranda

    Gaub1

    and Caryn L.

    Carlson

    1,2

    Received

    August 8,1995; revision received February 21,1996;

    accepted

    February 27,

    1996

    From anethnically diverse sampleof2,744 school children,221attention

    deficit

    hyperactivity

    disorder

    (ADHD) [123 (4.5%) predominantly inattentive (IA), 47

    (1.7%)

    predominantly hy-

    peractive/impulsive

    (HI), and 51(1.9%) combined type (C)] were

    identified

    using teacher

    ratings on a

    Diagnostic

    andStatisticalManualofMentalDisorders(4th ed.) (DSM-IV) symptom

    checklist. Subjects were compared to 221 controls on teacher ratings of behavioral, academic,

    and social

    functioning.

    Theresults revealed relatively independent areasof impairmentfor

    eachdiagnostic group.

    The IA

    children were impaired

    in all

    areas,

    but

    were rated

    as

    dis-

    playing

    more appropriate behavior andfewerexternalizing problems than HI or C children.

    The HI

    group displayed externalizing

    and

    social problems,

    but was

    rated

    as no

    different

    than controls

    in

    learning

    or

    internalizing problems.

    The C

    groupdemonstrated severe

    and

    pervasive

    difficulties across domains. These findings support

    the

    validity

    of the

    DSM-IV

    ADHD subtypes;

    all

    ADHD groups demonstrated impairment relative

    to

    controls,

    but

    show

    different patternsofbehavioral characteristics.

    KEY WORDS: ADHD; ADHD subtypes; behavioral characteristics; impairment; teacher ratings; ethnic

    minority.

    Since its inception, attention

    deficit

    hyperactivity

    disorder

    (ADHD)

    has

    been known

    bymany

    names,

    such asminimal brain dysfunction, hyperkinesis,hy-

    peractivity,

    and

    attention

    deficit disorder with

    (ADD/H) or without (ADD/WO) hyperactiviry. The

    changing

    nomenclature reflects disagreement regard-

    ingthe diagnostic necessity of the three core charac-

    teristics: impulsivity, inattention, and motor excess.

    Various

    editionsof the

    DiagnosticandStatisticalMan-

    ual

    of

    MentalDisorders (DSM) have implemented

    varioussubtyping systems which cluster the three core

    characteristicsin anumberofways.Thethird edition

    [DSM-III; American Psychiatric Association (APA),

    1The

    UniversityofTexasat Austin, Austin, Texas78712.

    2Addressallcorrespondence to CarynL.Carlson, Departmentof

    Psychology,

    Mezes 330, The University of Texas at Austin,

    Austin,Texas78712.

    1 3

    CWl-OttW/WOMWOSllSM)

    C

    1997Plenum

    PublishingCorporation

    1980] subtyped ADHD children usingabidimensional

    approach such that a child was diagnosed as ADD/H

    orADD/WO.The revised third edition (DSM-III-R;

    APA,1987)

    implemented a unidimensional approach

    which

    included children with varying degrees

    of hy-

    peractivity under

    the

    diagnostic category

    of

    attention

    deficit hyperactivitydisorder.

    The fourth

    edition (DSM-

    IV;

    APA, 1994)

    has

    returned

    to a

    bidimensionalsystem,

    clustering hyperactivity and impulsivity symptoms into

    one

    dimension,

    and

    separating

    it from the

    inattention

    dimension. Based

    on

    this system,

    the

    DSM-IVpre-

    sents an ADHD diagnosis with three subtypes:

    pre-

    dominantly inattentive (IA), predominantly

    hyperactive/impulsive

    (HI),and combinedtype(C),

    Children displaying symptoms of both inatten-

    tion

    and

    hyperactivity

    (i.e.,

    ADD/H

    and

    ADHD)

    un-

    der the

    various

    DSM

    systems have

    consistently

    demonstrated significant

    difficulties

    in adjustment,

  • 8/12/2019 1997 Behavioral Characteristics of DSM-IV ADHD Subtypes Ina Scholl - Gaub, Carlson

    2/9

    1 4

    socialfunctioning,and internalizing and externalizing

    behavior

    (Barkley, 1990; Hinshaw, 1994; Whalen,

    1989).The

    pervasiveness

    of

    such associated

    difficul-

    ties has been demonstrated primarily within clinic-

    referred

    samples.

    Although behavior problems

    associated with ADHD have been explored in popu-

    lation-based samples

    to a

    lesserextent, they have

    also

    been documented

    (August,

    Ostrander, &

    Bloomquist, 1992; Szatmari,

    Offord,

    & Boyle,1989).

    Research

    has

    demonstrated that children display-

    ing both inattention and hyperactivity differ from

    thosewhodisplay inattentionin theabsenceofmotor

    excess. This latter

    form

    of

    ADHD,

    first

    formallyiden-

    tified

    inDSM-III

    as

    ADD/WO,

    has

    been less

    widely

    studied.

    In a

    recent review

    of the

    ADD/WO litera-

    ture, Lahey, Carlson, and Frick (in press) concluded

    that, as compared to ADD/H children, ADD/WO

    children are more internalizing but less antisocial, less

    rejected

    by

    peers,

    and less

    externalizing.

    The

    DSM-IV

    C and IA

    subtypes were intended

    to be

    congruent with past diagnoses

    of

    ADD/H

    and

    ADD/WO,

    respectively. However, DSM-IV diagnos-

    tic

    criteria deviate somewhatfrom those

    of

    previous

    DSMversions, interms of numbersand clustersof

    symptoms.Thus,

    it

    remains

    to be

    seen whether cur-

    rent DSM

    criteria identify

    children

    similar

    to those

    identified

    in previous research.

    The HIsubtype, introduced tocaptureagroup

    of

    childrenwhowere perceived byclinicians in the

    field trials to have a clinically significant disorder

    (Lahey et al., 1994), has no previous diagnostic coun-

    terpart. The

    validity

    of the HI subtype has been ques-

    tioned by those who have suggested that these

    children, who were identified primarily among pre-

    schoolers

    in the field

    trials,

    may

    eventually display

    inattention symptoms

    and

    qualify

    for a C

    diagnosis

    (Barkley,

    1997).

    The

    DSM-IV

    field trials (Lahey et al., 1994)

    found

    that, among clinic-referred children, the DSM-

    IV ADHD subtypes showed

    different

    patterns of as-

    sociated impairment. C and HI groups were rated as

    more globally impaired than the IA group, while C

    and

    IA

    children

    had

    more academic problems than

    the HI group. Teachers rated C children as less

    liked

    and more disliked than HI children, while IA chil-

    dren

    did not differ

    from

    the

    other

    twosubtypeson

    like

    or

    dislike

    scores.

    Apart

    from

    the field

    trials,

    descriptive informa-

    tion about the DSM-IV diagnostic subtypes is scarce.

    McBurnett, Pfiffner, Swanson, Ottolini, & Tamm

    (1995) used parent

    and

    teacher ratings

    on a

    DSM-

    III-R diagnostic checklist of 520 child referrals to an

    ADHD clinic to retrospectively classify them into

    DSM-IV subtypes. A comparison of the behavioral

    characteristics of the three ADHD subtypes indicated

    that, consistent

    with

    Lahey et al. (1994), HI children

    were more academically successful than C or IA chil-

    dren.HI

    children

    did not differ from C

    children

    on

    ratings of peer dislike, but both of these groups were

    more disliked than the IAgroup. In addition, both

    the HI and C groups received higher ratings than the

    IA group on measures of disruptive behavior.

    Twostudies examined behavioral correlates

    of

    the DSM-IV subtypes in nonreferred

    samples

    (Baumgaertel, Wolraich,

    &

    Dietrich, 1995; Wolraich,

    Hannah, Pinnock, Baumgaertel, & Brown, inpress).

    Wolraich

    et al. obtained teacher ratings for

    8,258

    children in grades K-5 in a middle Tennessee county.

    An overall ADHD prevalence

    rate

    of 11.4% was ob-

    tained, with ratesfor LA, C, and HI of 5.4%, 3.6%,

    and

    2.4%, respectively. Baumgaertel

    et al.

    (1995)

    obtained teacher

    ratings

    for

    1,077 children

    in

    Grades

    1 to 4 inRegensburg, Germany.An overall

    prevalence rate of 17% was obtained, with rates for

    IA, C, and HI of 9%, 4.8%, and 3.9%, respectively.

    Consistent with the results of Lahey et al. (1994)

    and

    McBuraet et al. (1995), the C and HI groups

    displayed behavioral problems,

    while the C and LA

    groups

    were

    associated with academic problems.

    Wolraichet al. also

    found

    the proportion of children

    displaying anxiety/depression

    was

    lower

    for the HI

    (9.2%)

    than

    for the C

    (29.3%)

    or IA

    (21.9%)

    groups.

    The goals of the present study were to enhance

    the existing literature by (1) using a large-scale,

    population-based sample to prevent contamination

    bypossible referral bias; (2) exploring behavioral cor-

    relatesof the three

    DSM-IV-diagnosed subtypes

    of

    ADHD;

    and (3)attemptingtoevaluatethe extentto

    which

    the current DSM-IV system identifies subtypes

    comparable

    to

    those identified using previous diag-

    nostic systems.

    METHOD

    Data

    for the present study were obtained under

    the auspices of a larger community service program

    entitled

    the Schoolof the Future Project.This pro-

    gram, implemented by the Hogg Foundation, pro-

    vides school-based mental health services

    to low

    income communities. Since 1990, this program

    has

    Gaub and

    Carlson

  • 8/12/2019 1997 Behavioral Characteristics of DSM-IV ADHD Subtypes Ina Scholl - Gaub, Carlson

    3/9

    ADHD

    Subtypes

    1 5

    Table I.

    Demographic Characteristics

    of

    Population

    Grade

    Overall

    Gender K 1 2 3 4 5

    totals

    Boys

    Girls

    Totals

    255

    263

    518

    19%

    288

    249

    537

    19%

    261

    235

    496

    18%

    202

    178

    380

    14%

    218

    184

    402

    15%

    193

    218

    411

    15%

    1,417

    52%

    1,327

    48%

    2,744

    100

    been in effect in some elementary and secondary

    schoolsin several Texas cities.Yearlyevaluations, in-

    cluding

    teacherratings, have been used to assess the

    effectiveness ofservice deliveryin these schools.

    Data were collected for allchildren for whom

    parental consent

    was

    obtained (approximately

    96%

    of

    eligible

    children).

    The 3% of

    children

    of unknown

    or

    other ethnicity were excluded, leaving

    a

    sample

    of 2744children

    in

    nine elementary schools

    who had

    adequate

    data available for the 1993-1994 school

    year.Demographic characteristicsfor thesampleare

    presented

    in

    Table

    I. Children were predominantly

    from low-socioeconomic-status (low-SES) back-

    grounds.Basedonpaternal occupation, availablefor

    30 of the sample, meanSESrating (using total la-

    bor force comparisons) on the revised Duncan So-

    cioeconomic Index (Stevens

    &

    Featherman, 1981)

    was

    23. The

    ethnic composition

    was 76%

    Hispanic,

    16%

    African American, and 8% Caucasian.

    Subjects

    For

    thecurrent study,221children, representing

    8.0

    of the total population, met the criteria for

    ADHD

    (via teacher reports)

    and

    were included

    in

    analyses. The 221 ADHD children were compared

    to 221

    non-ADHD subjects selected from among

    nondiagnosedchildrenandmatchedforgender, age,

    grade,and

    ethnicity. Demographic characteristics

    of

    thegroups are reported inTable II.

    Measures

    Teacher evaluationsforeach subject consistedof

    theTeacher's Report Form (TRF; Achenbach,1991),

    a

    teacher-completed DSM-IV-based diagnostic

    checklist for

    ADHD

    and

    oppositional

    defiant

    disor-

    der (ODD), and three Likert-type scale questions re-

    garding social functioning

    [adapted

    from

    a

    questionnaire developed by Dishion (1990) used in

    the

    DSM-IV

    field

    trials].

    TRF.

    The TRF (Achenbach, 1991) is a widely

    used, standardized tool for the assessment ofchild-

    hood functioning and impairmentinbehavioral and

    emotional realms. All analyses used raw scores rather

    than T-scores, since the latter are scaled

    differently

    for each gender.

    SNAP-IV

    (Swanson

    Carlson, 1994). This diag-

    nostic checklist

    for

    ADHD

    and ODD

    [Swanson,

    No-

    lan, and Pelham Checklist-IV and the DSM-IIIR

    Disruptive Behavior Disorder Rating Scale (Pelham,

    Gnagy,

    Greenslade, & Milich, 1992)] consists of 26

    questions thatclosely parallel

    in

    wording

    the

    diagnos-

    tic

    symptoms

    forboth ADHD and ODD astheyap-

    pear in the DSM-IV The instructions ask the teacher

    toindicate, foreach question, whichof the following

    fourchoices best describes

    the

    child:

    not at

    all, just

    a

    little,

    quite

    a

    bit,

    orvery

    much.

    Social

    Functioning.

    Three

    questions, adapted

    from Dishion (1990), were included to assess the

    teacher's

    perception of the child's level of social

    functioning.

    Teachers estimated

    the

    proportion

    of the

    child's peers that like/accepted, dislike/rejected,

    and

    ignoredhim/her based

    on a

    5-point Likert-type scale

    ranging from

    1(very

    few/less than25 )

    to 5

    (almost

    all/more

    than

    75 ).

    Procedure

    Data

    were collected during April 1994

    for the

    1993-1994

    school year. Subjects were excluded from

    the original sample if their TRF

    scores

    had more

    than eight unanswered items

    totalor

    more than

    three

    unanswereditems on anyscab.This sample was then

    screened for those children who met DSM-IV

    teacher

    rating criteria

    for the

    three subtypes

    of

    ADHD. For diagnostic purposes, items endorsed as

    very

    muchwere tallied

    as present

    symptoms,

    and

  • 8/12/2019 1997 Behavioral Characteristics of DSM-IV ADHD Subtypes Ina Scholl - Gaub, Carlson

    4/9

    1 6

    Gaub andCarlson

    Table II.

    Group Demographic Characteristics

    Hyperactive/

    Demographic Combined Inattentive

    impulsive

    Controls

    characteristic (n = 51) (n = 123) (n = 47) (n =221)

    Age

    [mean (SD)]

    Gender

    ratio (Male:Female)

    Ethnicity

    withingroup:Hispanic

    African American

    Caucasian

    7.6years

    (1.6)

    2.8:1

    70%

    24%

    6%

    7.6

    years (1.9)

    2.3:1

    79%

    15%

    6%

    7.5

    years (1.6)

    4.1:1

    57%

    30%

    13%

    7.6

    years (1.7)

    2.6:1

    77%

    15%

    8%

    students

    were assigned ADHD subtype diagnoses ac-

    cording to DSM-IV criteria.

    Toform a control group, a nondiagnosed child

    wasmatched to each ADHD child. These nondiag-

    nosed

    controls (NC) were sought within the same

    classroom

    to

    adjust

    for

    potential rater

    differences.

    If

    a

    nondiagnosed child (of the same gender and eth-

    nicity)

    was not in the ADHD child's classroom, a

    match

    was

    chosen from

    the

    next classroom

    (on an

    alphabetical list) of the same grade within that

    school.Onrare occasions, amatchhad to bechosen

    from

    a

    same grade classroom

    at a

    different elemen-

    taryschool. Approximately

    94% of the

    matches were

    from the

    same classroom,

    5%

    werefrom

    a

    different

    classroom

    in the

    same school,

    and

    fewer than

    1%

    werefrom adifferent school.

    Analyses

    compared the three ADHD subtypes

    and

    the NC group on 19 variables

    (four

    TRF adjust-

    ment

    questions, eleven TRF behavior scales, three

    sociometric questions,and ODD

    symptom

    rating).To

    maintain a focus on diagnostic group, data for the

    current study were collapsed across ethnicity and age.

    Gender was not included as an independent variable

    inthe current

    study

    since a separate report of gender

    effects

    is in preparation.

    RESULTS

    Therewere no significant differences among

    ADHD subgroups in age

    [F(2,

    215) = .05,p = .96],

    gender \yf(df = 2) = 2.0,p = .37], or ethnic group

    composition[%2(4f

    = 4) =

    8.8,p =.07].

    The

    number

    of children

    in the

    sample obtaining

    the C, LA, and

    HI diagnoses were

    51

    (1.9%),

    123

    (4.5%),

    and 47

    (1.7%),respectively.

    Using a 0- to 3-point scoring system on the DSM

    checklist,

    the

    followingscores were obtained

    for the

    total sample: LAsymptom total (M = 6.82, SD =

    7.79),

    HI

    symptomtotal

    (M = 4.22, SD =

    6.29),

    and

    ODD symptom total (M= 2.94, SD = 5.35).

    Behavioral Variables

    One-wayanalysesofvariance (ANOVAs) were

    executed

    for

    each

    of the

    nineteen dependent vari-

    ables. Maineffects werefurtherevaluated using the

    Tukey

    (Tukey, 1972) test to explore group differ-

    ences.Cell

    sires,

    means and standard deviations, F

    andp values, andpost hoc comparisons for the be-

    havior

    ratings are reported in

    Table

    III.

    On the

    adjustment

    variables, the three ADHD

    groups were

    all

    rated

    as

    significantly more impaired

    than the NC group on measures of Hard Working,

    Appropriate Behavior, and Happy. Surprisingly, on

    the fourth

    adjustment

    variable, Learning, the HI

    group did not

    differ

    from

    the NC

    group, while

    the

    C and

    LA

    groups received significantly poorer ratings

    than the NC and HI groups. Although more im-

    paired than the NCgroup, the HI group received

    higher ratings of Hard Working than the othertwo

    ADHD groups, and higher ratings of Happy than the

    Cgroup. However,theLAgroup obtained higherAp-

    propriate Behavior ratings than the other two diag-

    nosed groups.

    On the three sociometric rating variables and

    the TRF Social Problems scale, all three of the di-

    agnosed

    groups were rated

    as

    having poorer social

    functioning than the NCgroup. The Cgroup re-

    ceived higher ratingson the Peer Dislike than the

    LAgroup and higher ratings on the TRF Social Prob-

    lems variables than either the HI orLAgroups, which

    did not

    differ

    significantly

    from

    each other oneither

    variable. On the Peer Like variable, the

    LA

    group was

    rated as liked by more peers than the C group, with

    no significant

    differences

    betweeneitherof these two

    groups

    and the HI

    group.

    On the

    Peer Ignore vari-

    able,

    there

    were no differences among groups.

  • 8/12/2019 1997 Behavioral Characteristics of DSM-IV ADHD Subtypes Ina Scholl - Gaub, Carlson

    5/9

    ADHD Subtypes

    1 7

    Table

    DDLGroup

    Comparisons

    on

    Behavioral Ratings

    a

    Significant groupdifferences based

    ADHD ADHD ADHD onTukey/wrthoctests

    Variable C IA HI NC F-ratio

    (p C

    Disliked

    by

    2.96 2.27 2.65 1.50 35.62*

    C, ffl and IA > NQ OIA

    Peers (1.34) (1.29) (1.23) (0.86)

    Peers

    Neutral 112 2.22 1.98 1.44 18.5s C, HI. and IA>NC

    Toward (1.21) (1.17) (1.18) (0.80)

    ODD

    symptom 16.09 7.13 15.02 3.65 71.0*

    C and

    HI>IA>NC

    ratiag (6.64) (7.53) (6.74) (5.92)

    Hardworking 1.67 1.68 3.26 4.32 116.9* NC>HI>C

    and IA

    (0.97) (0.98) (1.34) (1.64)

    Appropriate 1.55 2.45 1.67 4.12 67.2* NC>IA>Cand HI

    Behavior (0.87) (1.48) (1.15) (1.77)

    Learning 1.84 1.82 3.67 4.17 86.4* NC andHI>C and IA

    (1.03) (1.08) (1.49) (1.67)

    Happy

    2.53 3.08 3.42 4.32 30.8* NOC,

    HI, and IA;

    (1.26) (1.35) (1.69) (1.54) HI>C

    Withdrawn

    4.46 5.82 2.45 2.27 29.3* C andIA>HIand NC

    (3.65) (4.52) (2.19) (2.86)

    Somatic 1.78 1.12 0.67 0.53 7.2* C>HIand NC;IA>NC

    Complaints (Z97) (2.13) (1.60) (1.24)

    Anxious/

    8.94 5.15 5.81 3.24 18.2* C>NIandIA>NC

    Depressed

    (6.35) (4.93) (5.75) (4.62)

    Social 9.79 6.12 6.36 2.32

    56.2*

    C>HI andIA>NC

    Problems

    (5.02) (4.62) (4.45) (3.35)

    Thought 1.80 1.16 1.79 0.35 15.9* C, HI, and IA>NC

    Problems (2.54) (2.01) (Z47) (1.05)

    Attention 29.89 25.8 18.98 8.03 218.4* C>IA>HI>NC

    Problems (4.25) (6.16) (6.85) (7.90)

    Delinquency 7.43 4.83 6.59 2.04 51.2* C andHI>IA>NC

    (3.87) (3.63) (4.13) (2.87)

    Aggressive 31.22 14.76 31,95 8.12 88.5*

    C and

    HI>IA>NC

    Behavior (11.12) (12.82) (10.31)

    (10.72)

    Internalizing 14.35 11.57 8.36 5.85 21.2* C>HI

    and NC;

    IA>NC

    Behavior (10.98) (8.84) (8.04) (7.06)

    Externalizing

    38.52 19.61 37.84 10.25 89.5* C and

    Hl>IA>NC

    Behavior (13.42) (15.47) (13.13) (13.07

    Total

    Problem 93.94 64.61 75.34 27.40 121.5*

    C> HI and

    IA>NC

    Behavior (27.00) (26.93) (27.25) (26.54)

    aADHD

    =

    attention

    deficit

    hyperactivitydisorder;

    C =

    combined type;

    IA =

    predominantlyinattentive;

    HI =

    hyperactive/impulsive;

    NC =nondiagnosedcontrols. Mean

    values

    forADHDC,ADHD IA,

    ADHD

    HI, and NCgroupsare followed bystandard deviationsinparentheses.

    bp

  • 8/12/2019 1997 Behavioral Characteristics of DSM-IV ADHD Subtypes Ina Scholl - Gaub, Carlson

    6/9

    1 8

    Gaub andCarlson

    Table IV.PercentageofChildreninADHD

    Subtypes Classified

    asImpaireda

    Impairment

    C IA HI

    variable

    (a =51) (n = 123) (n = 47)

    Social(Peer Like

    or

    Dislike)

    82 59 53

    Behavioral

    (Appropriate Behavior)

    90 58 80

    Academic(Learning) 82 76 23

    NotImpairedin any of the

    three

    domains 2 11 4

    aADHD = attention deficit hyperactivity disorder; C = combined type; IA =

    predominantlyinattentive;HI =hyperactive/impulsive.

    On all TRF externalizing variables (Aggressive

    Behavior, Delinquency,

    and

    Externalizing Behavior)

    and

    the ODD symptom rating, the three diagnosed

    groups received higher scores than

    the NC

    group.

    No

    differences

    emerged between

    the C and HI

    groups,

    with

    both

    receiving

    significantly

    higher ratings than

    the IA

    group

    on all

    variables. Thus,

    on

    externalizing

    variables, the C and HI children are rated as most

    deviant, while

    the IA

    children obtain lower ratings.

    The

    pattern

    of

    group differences

    for the TRF in-

    ternalizing variables

    was

    more complex than those

    in

    other domains.

    On all

    four internalizing variables

    (Withdrawn, Somatic Complaints, Anxious/De-

    pressed, and Internalizing Behavior), the C and IA

    groups were rated

    as

    significantly more impaired than

    the NC group. HI children received higher ratings

    than NC children only on the Anxious/Depressed

    variable; on all other internalizing variables, the HI

    group

    was

    rated

    as no

    different than

    the NC

    group.

    The C group was rated as having significantly more

    problems than

    the HI group on allfourvariables. The

    IAchildren werenot significantly

    different

    from the

    other two diagnostic groups on Somatic Complaints

    or

    Internalizing.

    The IA

    group

    was

    rated

    as

    more

    Withdrawnthan the HI group, but the groups did not

    differ on the other threeinternalizing variables.

    On the Thought Problems scale,the three diag-

    nosed groups were rated

    as

    significantlymore deviant

    than the NC group, with the C group rated assignifi-

    cantly more impaired than the HI and IA groups,

    which

    did not

    differ from each other.

    On the

    Atten-

    tion Problems scale, allthree diagnosed groups re-

    ceived

    significantly

    higher ratings than

    the NC

    group,

    with

    the C

    group receiving significantly higher ratings

    than the IA group. The HI group received signifi-

    cantlylower ratings than either the C or the IA group.

    On the Total Problem scale the C groupwas

    ratedby

    teachers

    as

    having significantly higher scores

    than anyother group. No differences emerged be-

    tween the HI and IA groups, both of which were

    rated as

    having significantly

    moreproblemsthan the

    NC

    group.

    Impairment

    Criteria

    To

    approximate

    the

    impairment criteria required

    by

    DSM-IV, NCscoreswere used to calculate objec-

    tive

    criteria

    on various ratingitems.A child was con-

    sidered impaired if he/she scored 1SD or greater

    from the NC

    mean

    (in the

    deviant direction)

    in the

    following domains: academic (TRF Learning), be-

    havioral (TRF Appropriate Behavior), social (Peer

    Dislike

    or Peer Like). Table IV shows the proportion

    of

    children

    in

    each subtype

    ratedas

    unpaired

    in

    each

    domain, and the proportion who did not meet im-

    pairment criteria in anydomain.

    As indicated, the C group was most pervasively

    impaired, with percentages

    of

    children rated

    as im-

    paired

    socially, behaviorally, and academically of

    82%, 90%,

    and

    82%, respectively. Only

    one C

    child

    (2%) did not show impairment in any domain.

    AmongLAchildren, academic impairmentwasmost

    common(76%), with moderate rates of social (59%)

    and behavioral (58%) impairment.Few LAchildren

    (11%) were rated as nonimpaired in anyarea.Chil-

    dren in the HI group were most likely to show be-

    havioral impairment (80%), with moderate rates

    of

    social impairment (53%)

    and

    relatively

    lowratesof

    academic (23%) impairment.Very few HI children

    (4%) were rated as unimpaired in any domain.

    DISCUSSION

    Prevalence

    The present findings in this nonreferred popu-

    lation indicate that prevalence ratios varied

    across

    the

    three ADHD subtypes, with

    the

    following rates;

  • 8/12/2019 1997 Behavioral Characteristics of DSM-IV ADHD Subtypes Ina Scholl - Gaub, Carlson

    7/9

    ADHD Subtypes

    1.9 for C, 4.5%for LA, and 1.7% for HI. Thepro-

    portionof the population identified as C(1.9%)was

    slightly lower than prevalence estimates of ADHD

    using

    previous

    diagnostic criteria, which are generally

    cited

    in the 2-3% range (APA,

    1987).

    While the cur-

    rent

    findingscouldmerely reflectthe stringent crite-

    riaforsymptom presence (i.e.,a very muchraring),

    it ispossible

    that changes

    in

    diagnostic criteria have

    decreased

    the prevalence of the C subtype. If, as sug-

    gestedby the field

    trials (Lahey

    et

    al., 1994), most

    children meeting criteria

    for the HI

    subtype would

    likely have been diagnosed byclinicians ashaving

    ADHD(according

    to

    DSM-III

    and

    DSM-C

    III-R

    cri-

    teria),the current results showing a 3.6% prevalence

    of

    these

    two

    subtypes combined (1.9%

    for C and

    1.7

    for HI) appear consistent with previous preva-

    lence

    estimates.

    The

    rate

    of IA was also higher than previous

    ADD/WO

    prevalence rates

    of approximately 3%

    (Szatmari

    et al.,

    1989).

    This was

    likely

    due to the

    change in diagnostic criteria resulting in more chil-

    dren

    qualifyingfor an LAdiagnosis; this conclusion

    isconsistent with the DSM-IV field trials (Lahey et

    al., 1994) findingthatthe majorityof the new cases

    identified by DSM-IV were IA children.

    Thehigher rates

    of all three subtypesfound in

    previous

    population-based studies of DSM-IV crite-

    ria

    (Baumgaertel

    et al., 1995; Wolraich et al., in

    press)werelikely

    due to the more lenient

    cutoff

    cri-

    teria employed in those studies, bothof which em-

    ployed

    slightly

    differently

    phrasingforanchor points,

    butcounted symptoms

    as

    present

    if

    either

    of the two

    mostextremeratings (i.e.,often or very

    much)

    were

    endorsed.

    Despite

    some differences in relative pro-

    portions of the subtypes, both previous studies and

    thecurrent studyfound that

    LA

    wasmost prevalent,

    and

    HIleast prevalent, in nonreferred populations.

    An interesting

    difference between current re-

    sults

    and

    those from

    research using clinic-referred

    samples(Laheyetal., 1994;

    McBurnett

    et al., 1995)

    involvesthe relative

    ratios

    of the

    three

    ADHD sub-

    types. In the

    current study,

    LA was

    identified twice

    as

    often as C(C:IA ratio of1:2.4);conversely, both

    the field trials (Lahey et al., 1994) andMcBurnett

    et al.(1995) found C to be much more prevalent

    than

    LA

    (C:IA ratios of 2.1:1 and3.5:1,respectively).

    In addition, while the current study found nearly

    identical ratesof

    C

    and

    HI C: Hl

    ratio

    =

    1.1:1),

    both

    previous studies (Laheyetal., 1994; McBurnett

    et al., 1995) found much higher CM ratios (3.0:1

    and4.3:1,respectively).Thesediscrepancies

    in

    preva-

    1 9

    lence

    ratios likely reflected

    the

    nature

    of the

    samples

    (population-based

    vs.

    clinic-referred).

    Relatedly,

    Wolraichet al. (in press) found that C children were

    two

    to four times more likely than

    LA

    or HI children

    to be

    referred

    to a

    clinic.

    Behavioral

    Characteristics

    The distinctive pattern of impairment by subtype

    found in the current study clearly indicates that the

    DSM-IV diagnostic system distinguishes three groups

    of

    children that are all impaired, relative to nondi-

    agnosed controls, and that can be differentiated from

    each other based on differentpatternsof difficulties.

    The C

    subtype

    was

    associated with

    the

    most perva-

    sive

    pattern ofdifficulties,with severe ratingsof im-

    pairment found

    in all

    major domains

    of

    functioning.

    For variables includingAnxious/Depressed, Social

    Problems, Attention Problems,

    and

    Total Problem

    Behavior, the C group was rated as more unpaired

    than

    any of the

    other groups.

    As

    noted earlier, this

    pattern

    of

    pervasive deficits displayed

    by C

    children

    is consistent with that shown in children diagnosed

    using previous ADHD

    criteria.

    In the DSM-IVfield

    trials sample, Lahey

    et al.

    (1994)also demonstrated

    that C children experienced extensivedifficulties,and

    were significantly impaired

    in all

    assessed areas

    in-

    cluding social, academic, and global functioning.

    The IA

    children, like

    the C

    children, were rated

    by

    teachers

    as significantly

    impaired

    in all

    major

    do-

    mains

    of functioning. Nonetheless, the LAchildren

    demonstrated areas of advantage relative to the other

    two diagnosed groups. Specifically,

    the LA

    children

    were perceived as displaying more appropriate behav-

    ior and

    less externalizing behavior than

    C or HI

    chil-

    dren.

    Previous research comparing ADD/WO and

    ADD/H children found that ADD/WO children

    showed

    more internalizing behavior, but less external-

    izing

    and

    antisocial behavior,

    and

    lower

    ratesof

    peer

    rejection, than those

    with

    ADD/H (Lahey

    et

    al.,

    in

    press).This pattern

    is

    consistent with

    the

    present

    find-

    ings, with IA children receiving significantly lower

    TRF ratings on Aggression, Delinquency and Exter-

    nalizingscales, lower

    ODD

    symptomscores,lower

    Peer Dislike ratings, and higher Peer Like ratings than

    C children. Surprisingly, however, LA and C children

    were rated as no different fromeach other (although

    both received

    significantly

    higher

    scores

    than NC chil-

    dren) on three internalizing variables (Withdrawn, So-

    matic Complaints

    and Internalizing Behavior).

  • 8/12/2019 1997 Behavioral Characteristics of DSM-IV ADHD Subtypes Ina Scholl - Gaub, Carlson

    8/9

    11

    Furthermore, although they did differ on theAnx-

    ious/Depressed rating, this difference

    was in the un-

    expected direction; the C children were rated as more

    Anxious/Depressed than

    the IA

    children. This

    finding

    parallels those of Wolraich et al. (in

    press),

    who also

    found cooccurring anxiety/depression to be more

    common

    among

    C

    (29.3%) than

    IA

    (21.9%)children.

    While

    this unexpected result may be attributable

    to the

    revised diagnostic criteria,

    it

    seems unlikely

    that the changes would produce such a strong effect

    A closer examination of past research, as reviewed

    byLahey, Carlson,

    and

    Frick

    (in

    press),suggests that

    the epidemiological (school-based) studies available

    (Lahey,

    Schaughency, Strauss, & Frame, 1984; Pel-

    ham,

    Atkins, Murphy,

    &

    White, 1981)

    did not find

    differences in

    anxiety/depression between

    the

    ADD/H

    and

    ADD/WO subgroups. Rather, these dif-

    ferences were evident primarily in studies using

    clinic-referred samples

    (e.g.,

    Lahey, Schaughency,

    Hynd,

    Carlson, & Nieves,1987).There may be a re-

    ferralbias operating such that those

    IA

    children

    with

    highlevelsofanxietyanddepressionaremore likely

    to be referred to clinics. Thus, IA (or ADD/WO)

    childrenwho are includedin clinic-referred popula-

    tions may be more anxious/depressed than their

    counterparts in the overall population, resultingin

    this

    discrepancy across studies.

    The HI

    group

    in the

    current study demonstrated

    a

    pattern

    of

    impairment that

    was

    quitedifferent than

    that of the C and IA groups. While the HI group

    received significantlypoorer ratings than the NC

    group in the

    socialfunctioning,externalizing, atten-

    tion, andthought problems domains, theydid notdif-

    fe r significantly from

    NC

    children

    on

    several

    variables, including

    the

    Learning

    and

    most internal-

    izingsubscales (Withdrawn, Somatic Complaints,

    and

    Internalizing).

    In

    addition,

    the HI

    group

    was

    rated

    assignificantlymore hard working than

    the IA and

    C

    groups

    and

    significantlyhappier than

    the C

    group.

    Thus, consistent with existing research (Baumgaertel

    et

    al., 1995; Laheyetal., 1994; McBurnettet al,1995;

    Wolraich et al., inpress),HI children appear to be

    characterized byexternalizingand peer relationship

    problems, but do not demonstrate internalizing prob-

    lems

    or academic impairment. Overall, the results

    from this

    and

    other research indicate that

    the new

    subtypeof HI is a valid and useful addition to the

    DSMdiagnostic system; HI children are significantly

    impaired

    in some areas of functioning, but the pat-

    tern of behavioral problems is distinctly different

    from that

    found

    among

    IA and C

    children.

    Gaub andCarlson

    The younger age of onset for HI reported in the

    field trials (5.65 years) (Lahey et al., 1994) led to

    speculation that, rather than representing

    a

    separate

    ADHD subtype, younger, Hi-diagnosed children

    might

    eventually display inattentive symptoms

    and

    qualify

    for C diagnoses (Barkley,

    1997).

    In the

    McBurnettet al.(1995)sample, the HI subtype was

    also significantly younger than the C or IA subtypes;

    however, the mean HI age (7.25 years) was not as

    young

    as

    that reported

    in the field

    trials. Further-

    more,

    in the

    current study,

    no

    significant

    age differ-

    enceswere

    found

    among the C(M= 7.6 years), IA

    (M = 7.6

    years),

    or HI

    (M =

    7.5

    years) groups. Since

    disruptive,hyperactive behavior would likely lead to

    referral

    at an

    earlier

    age than inattentive, less dis-

    ruptive

    behavior, the younger ages for HI children

    in

    studies using clinic samples (Lahey et. al, 1994;

    McBurnett

    et al., 1995) mayreflect referral patterns

    rather than true differences across subtypes in age

    of onset.Thus,

    the

    current

    findings of

    different

    as-

    sociated impairment

    for the HI

    subtype, along with

    evidence that this subtype

    was

    found

    in all age

    groups, support the validity of HI as a distinct sub-

    type of ADHD rather than a precursor to C.

    Limitations

    There

    are

    several limitations

    of the

    current study.

    The

    exclusive

    use of

    teacher ratings

    on a

    symptom

    checklist

    to

    assign diagnoses

    is

    problematic, since nei-

    ther the age of onset of symptomsnor information

    regarding impairment

    or

    cross-situationality

    was ob-

    tained.In an attempt to address this limitation within

    the parameters of the available data, the strictest pos-

    siblecut-off scores were implementedfor theSNAP-

    IV The low

    proportion

    of

    children

    (2% of C, 11% of

    IA, and 4% of HI) who

    were

    not

    rated

    as

    unpaired

    in

    any

    realm suggests that

    the

    current study appro-

    priately avoided identifying subthreshold cases. The

    study

    is

    also limited

    by its

    reliance

    on

    teachers

    for

    both diagnostic and dependent variable ratings, al-

    though the finding of

    different

    patterns ofbehavior

    problems across subtypes provided evidence

    of the as-

    sociation

    between teacher perceptions of diagnostic

    symptoms

    anddomain-specific

    dysfunction

    .

    Another

    limitation

    of the present study is the

    extent

    to

    which theseresults

    are

    generalizable

    to the

    overall population since the current sample consisted

    primarily

    of lower-SES Hispanicsubjects. However,

    the consistency of the obtained pattern of results

  • 8/12/2019 1997 Behavioral Characteristics of DSM-IV ADHD Subtypes Ina Scholl - Gaub, Carlson

    9/9

    ADHD

    Subtypes

    111

    (particularlyregarding group differences

    in

    areas

    of

    functioning) with findings

    from

    previous

    research

    suggests that ADHD children

    from

    both minority

    andnonminorityethnic backgrounds share common

    behavioral

    characteristics.

    Thisand other research supports the concurrent

    validity of the DSM-IV ADHD subtypes. The com-

    parability of the current results and those of the

    other two population-based studies (Baumgaertel et

    al., 1995; Wolraich et ah, in press), which used pri-

    marily

    Hispanic,German, and Caucasian subjects, re-

    spectively, establishes the cross-cultural congruency

    of

    behaviors associated withADHD. Future research

    should work toward examining the etiological and

    predictivevalidity of the current diagnostic system by

    exploring potential subtype differences in causes,

    outcomes,

    and

    treatment responsiveness.

    ACKNOWLEDGMENTS

    Thisstudywas made possible through collabora-

    tion

    with

    the

    Hogg Foundation School

    of the

    Future

    Project (SOF).The authors thank Wayne Holtzman,

    Ph.D.,

    Special Consultant to the Hogg Foundation,

    as well as Scott Keir,

    Ph.D.,

    SOF

    Director

    of Re-

    search,

    and Pam

    Diamond, Ph.D.,

    SOF

    Senior

    Re-

    searchAssociate,

    for their enthusiastic support. Also,

    thanks

    to

    Anne Anderson, Joey Martin,

    and

    Scott

    Davis

    fortheir

    manyhours

    of

    data entry.

    This

    research

    was partially supported by an

    NIMHFIRST grant, MH49827 awardedto the sec-

    ond

    author.

    REFERENCES

    Achenbach,

    T. M. (1991).Manualfor the teacher'sReport

    Form

    and

    1991

    Profile.

    Builington:Universityof Vermont,Depart-

    ment

    ofPsychiatry.

    AmericanPsychiatric Association(1980).Diagnostic

    and

    statistical

    manual of

    mental

    disorders

    (3rd ed.). Washington,DC:

    Author.

    AmericanPsychiatric Association(1987).Diagnosticandstatistical

    manualof

    mental

    disorders (3rd ed.,

    rev.).

    'Washington,

    DO

    Author.

    AmericanPsychiatric Association(1994).Diagnostic

    and

    statistical

    manual of mentaldisorders (4th ed.). 'Washington, DC:

    Author.

    August,

    G. J.,

    Ostrander,

    R., &

    Bloomquist,

    M. J.

    (1992).Atten-

    tion deficit

    disorder:

    An

    epidemiological screening method.

    American

    Journal

    of

    Orthopsychiatry,

    62,387-3%.

    Barkley, R. A. (1990).

    Attention

    deficit

    hyperactivity disorder:A

    handbook for diagnosis and treatment. New

    York:

    Guilford

    Press.

    Barkley,

    R.

    (1997).

    Behavioral inhibition, sustained attention, and

    executive functioning: Constructing a

    unifying

    theoryof

    ADHD.Psychological

    Bulletin,

    121,65-94.

    Baumgaertel, A.,Wolraich,M., &Dietrich,M.(1995).Compari-

    son of

    diagnostic criteria

    for

    attention deficit disorders

    in a

    German

    elementary school sample.Journal of the American

    Academy of Child and Adolescent Psychiatry, 34, 629-638.

    Dishion,

    T.

    (1990).

    The

    peer context

    of

    troublesome child

    and

    ado-

    lescent behavior. In O. E. Leone(Ed.),Understanding troubled

    and troublingyouth.

    Newbury

    Park,CA ,

    Hinshaw,S P.

    1994).

    Attentiondeficits an d

    hyperactivity

    in

    children.

    Thousand

    Oaks,

    CA:Sage.

    Lahey,B. B.,Applegate,B.,McBurnett,K.,Biederman,J, Green-

    hill,

    L, Hynd, G. W,

    Barkley,

    R. A., Newcorn, J., Jensen, P.,

    Richters, J., Garfinkel, B.,Kerdyk,L., Frick, P.J., Ollendick,

    T, Perez, D., Hart, E.,Waldman,I., &Shaffer, D.(1994).

    DSM-IV

    fieldtrialsforattention deficit/hyperactivitydisorder

    in childrena nd adolescents.American

    Journal

    of Psychiatry,

    152,

    1673-1685.

    Lahey,B. B, Carlson, C. L,, & Frick, P. J. (in press).Attention

    deficitdisorders:

    A

    review

    of

    research relevant

    to

    diagnostic

    classification.

    In T. A.Widiger,A. J.Frances,W.Davis,& M.

    First(Eds.),DSM-IV

    sourcebook

    (Vol. IV). Washington,DC:

    AmericanPsychiatric Press.

    Lahey,

    B. B., Shaughency, E. A., Hynd, G. W., Carlson, C. L, &

    Nieves,

    N.

    (1987).Attention deficit disorder with

    and

    without

    hyperactivity. Comparison of behavioral characteristics of

    clinic-referred

    children.Journal of theAmerican

    Academy

    of

    Child and Adolescent Psychiatry, 26, 718-723.

    Lahey,B. B.,Shaughency,E.A.,Straus, C. C, &Frame,C. L.

    (1984).Are attentiondeficit disorders withand withouthy-

    peractivity

    similaror dissimilar disorders?

    Journal of the

    AmericanAcademy

    of

    Childand AdolescentPsychiatry, 23,302-

    309.

    McBurnett, K.,Pfiffner, L. J., Swanson, J. M., Ottolini, Y. L., &

    Tamm,

    L.(1995, May).

    Clinical

    correlatesof children retrospec-

    tively

    classified by

    DSM-IV attention

    deficit/hyperactivity

    disorder

    subtypes: Contrast with field trials impairment patterns.Paper

    presented

    at the

    International Conference

    on

    Research

    and

    Practice inADD, Jerusalem, Israel.

    Pelham,

    W. E.,

    Atkins,

    M. S.,

    Murphy,

    H. A., &

    White,

    K. S.

    (1981).Operationalization and validation of attention

    deficit

    disorders. Paper presentedat the annual meetingof the As-

    sociation

    for the Advancement of

    Behavior Therapy,Toronto,

    Canada.

    Pelham, W. E., Gnagy,E. M., Greenslade, K. E., & Milich, R.

    (1992).Teacher ratingsof DSM-III-R symptomsof the dis-

    ruptive

    behavior disorders.

    Journal

    of the Am erican

    Academy

    of Child and Adolescent Psychiatry, 31,210-218.

    Stevens,

    G., &

    Featherman,

    D. L.(1981).Arevised

    socioeconomic

    in-

    dex of occupationalstatus.

    Social

    Science

    Research,

    10,364-395.

    Swanson, J., & Carlson, C L. (1994). DSM-IV

    rating scale

    for

    ADHD

    and ODD. Unpublished manuscript.

    Swanson, J. M.

    (1991).

    School-based

    assessments

    and

    interventions

    for

    ADD students. Irvine, CA : K. C.

    Publishing.

    Szatmari,

    P.,Offord, D. R., &Boyle,M. H. (1989). Ontario

    child

    health

    study: Prevalence

    of

    attention deficitdisorder hyperac-

    tivity.Journal of ChildPsychology and

    Psychiatry,

    30, 219-230.

    Tukey,

    J. W.

    (1972).

    Exploring

    data

    analysis.

    Redding, Massachu-

    setts: Addison-Wesley.

    Whalen,C. K.(1989).Attention

    deficit

    hyperactivity disorders.In

    X H.Ollendick& M.Hersen

    (Eds.),Handbook

    of child

    psy-

    che-pathology(2nd ed, pp.131-169).New York: Plenum Press.

    Wolraich,

    M. L., Hannah, J. N., Pinnock, T Y., Baumgaertel, A,

    Brown,J. (inpress).Comparisonofdiagnostic criteriafor at-

    tention deficit hyperactivitydisorderin acounty-widesample.

    Journal of the

    American Academy

    of Child and Adolescent Psy-

    chiatry.