Comparing rates of psychiatric and behavior disorders in adolescents and young adults with severe intellectual disability with and without autism

Journal of Autism and Developmental Disorders, Vol. 34, No. 2, April 2004 ( 2004) Comparing Rates of Psychiatric and Behavior Disorders
in Adolescents and Young Adults with Severe Intellectual
Disability with and without Autism
Elspeth A. Bradley,1,5 Jane A. Summers,2 Hayley L. Wood,3 and Susan E. Bryson4
Eight males and four females with an Autism Diagnostic Interview-Revised (ADI-R) diagnosisof autism (mean age of 16.3 years) and severe intellectual disability (IQ < 40) were individu-ally matched to controls on the basis of chronological age, gender, and nonverbal IQ. The de-pendent measure was the Diagnostic Assessment for the Severely Handicapped-II, which is usedto screen for psychiatric and behavior disorders in lower-functioning individuals. Participantswith autism showed significantly greater disturbances as measured by the Diagnostic Assess-ment for the Severely Handicapped-II total score and seven of 13 subscales. They also averaged5.25 clinically significant disturbances compared with 1.25 disturbances for participants with-out autism. Specific vulnerabilities to anxiety, mood, sleep, organic syndromes, and stereotypies/tics were found in the participants with comorbid autism.
KEY WORDS: Autism; intellectual disability; adolescence; behavior and psychiatric disorders.
INTRODUCTION
several factors, including the characteristics of theindividuals being evaluated (e.g., their level of intel- One of the most consistent findings to emerge in lectual disability, age, gender, medical or genetic recent years is the increased prevalence of psychiatric conditions, family history), their living arrangements and behavior disorders among individuals with intel- (e.g., community settings, clinical or institutional lectual disability6 compared with the general popula- settings), the sampling methods used in the study (e.g., tion. It has been estimated that anywhere from 10% to population-based approaches, clinic referrals), and the 70% of individuals present with such behavioral or psy- clinical criteria (psychiatric disorders only vs. psy- chiatric disturbance (Bregman, 1991). The variability chiatric, behavioral, or emotional disorders) and as- in findings can be understood as an interaction among sessment approaches (e.g., semistructured clinicalinterviews, file reviews, observations, rating scales) that were used for ascertainment and diagnosis (Borthwick- Surrey Place Centre and the Department of Psychiatry, Universityof Toronto, Toronto, Ontario, Canada.
2 Hamilton Health Sciences and McMaster University, Hamilton, There is tremendous diversity among individuals with intellectual disability in terms of functioning level, 3 The Hospital for Sick Children, Toronto, Ontario, Canada.
4 etiology of disabilities, and other variables likely to Department of Pediatrics and Psychology, Dalhousie University, affect mental health. Existing prevalence data pertain 5 Correspondence should be addressed to Dr. Elspeth Bradley, largely to undifferentiated populations in terms of these Biomedical Sciences and Research Division, Surrey Place Centre, variables. There is a growing trend, however, for re- 2 Surrey Place, Toronto, ON, Canada, M5S 2C2; e-mail: e.bradley@ searchers to study mental health disorders within smaller, more homogeneous subgroups. Recent ad- The term “intellectual disability” is used synonymously with mental vances in the field of genetics have paved the way retardation as defined in Diagnostic and Statistical Manual (4th ed.)(DSM-IV; American Psychiatric Association, 1994).
for the identification of psychiatric and behavioral 0162-3257/04/0400-0151/0 2004 Plenum Publishing Corporation Bradley, Summers, Wood, and Bryson
phenotypes in groups of individuals with a genetic eti- this subset of individuals with autism who also have ology for their intellectual disability. As a result, there severe intellectual disability, although there has been is mounting evidence that some syndromes (e.g., Down, increasing attention to these issues in recent years Williams, Prader-Willi, Deletion 22q11.2, and Fragile X) are associated with different rates and patterns of, Diagnosing psychiatric and behavior disorders in as well as vulnerabilities to, mental health disorders persons with severe cognitive and communicative im- (Dykens, Hodapp, & Finucane, 2000). More discover- pairments such as those with intellectual disability and ies of this type will help bring greater clarity and speci- autism poses formidable challenges (King, DeAntonio, ficity to the prevalence statistics.
McCracken, Forness, & Ackerland, 1994). Existing di- Autism and autism spectrum disorders make up agnostic classificatory systems (DSM-IV, American one of the largest diagnostic subgroups within the en- Psychiatric Association, 1994; ICD-10-CDDG, World tire population of individuals with intellectual disabil- Health Organization, 1992) rely heavily on descriptions ity (Nordin & Gillberg, 1996; Stromme & Diseth, of the subjective experiences of the individuals who are 2000). Autism is a behaviorally defined syndrome that being diagnosed. Applying these diagnostic approaches is characterized by abnormalities or impairments in the to persons who are unable to share their subjective ex- areas of communication and play, socialization, and periences because of cognitive and communication im- range of interests and activities, all with an onset before pairments and disabilities is problematic, and some 3 years of age (DSM-IV; American Psychiatric Asso- would argue that an alternative conceptualization for ciation, 1994). More rigorous diagnostic tools are now diagnosing mental health disorders is needed for this available (Autism Diagnostic Observation Schedule; severely impaired group (e.g., Reid, 1980). At present, Lord et al., 2000; Autism Diagnostic Interview- attempts are being made to modify these existing sys- Revised; Lord, Rutter, & Le Couteur, 1994), permitting tems so that they can be more appropriately applied in greater precision in defining who meets these social, these circumstances, for example, Diagnostic Criteria communication, and behavioral criteria across chrono- for Psychiatric Disorders for Use with Adults with logical age and level of functioning, and thus provid- Learning Disabilities/Mental Retardation (DC-LD; ing a more homogenous group in which to explore Royal College of Psychiatrists, 2001) and Practice mental health issues. An important question is whether Guidelines for the Assessment and Diagnosis of Mental a diagnosis of autism is associated with higher rates of Health Problems in Adults with Intellectual Disability psychiatric and behavior disorder, a finding that, if sub- (Deb, Matthews, Holt, & Bouras, 2001); however, stantiated, would have important implications for the serious concerns about both the reliability and the planning and delivery of effective treatment and allo- validity of these approaches in diagnosing mental cation of limited resources (Bryson, 1996) and would health disorders in persons with intellectual disabilities, also raise the issue as to why those with autism are more vulnerable to such disturbances. Studies of co- Partly in response to these aforementioned chal- morbid mental health disorders among individuals with lenges, the Diagnostic Assessment for the Severely autism/PDD are few (e.g., Lainhart & Folstein, 1994; Handicapped or DASH (which was later revised to be- Tsai, 1996) and often focus on the minority of higher- come the DASH-II) was developed to explore issues of functioning, verbal individuals (e.g., Ghaziuddin, psychiatric and behavior disorders specifically among Alessi, & Greden, 1995; Kim, Szatmari, Bryson, individuals with severe and profound intellectual dis- Streiner, & Wilson, 2000), who are most likely to be ability (Matson, 1995; Matson, Gardner, Coe, & able to report their symptoms. By contrast, there is an Sovner, 1991b). Items that appear in the rating scale enormous literature on the assessment and treatment of were derived from DSM-III-R as well as from prior re- behavioral deficits (e.g., communication, social, and search on maladaptive behaviors in lower-functioning play skills) and excesses (e.g., aggression, self-injury, individuals. The instrument consists of 84 items that stereotypies) in individuals with autism or autistic- are grouped into 13 subscales. According to its devel- like behavior (see Matson, Benavidez, Compton, opers, the first five subscales cover the “classic” forms Paclawskyj, & Baglio, 1996b, for a review of over of mental illness (i.e., anxiety, PDD/autism, mania, de- 250 studies in this area), many of whom also have se- pression, and schizophrenia), and the eight remaining vere intellectual disability and minimal or poor verbal subscales cover a range of aberrant or maladaptive skills. Because behavioral approaches tend to deem- behaviors (i.e., stereotypies/tics, self-injury, eating phasize diagnostic and etiological issues, even less is disorders, sleep disorders, sexual disorders, organic known about the prevalence of psychiatric disorder in syndromes, elimination disorders, and impulse control Comparing Rates of Psychiatric and Behavior Disorders in Adolescents and Young Adults
problems and other miscellaneous behaviors). A third- occurring psychiatric and aberrant behavior disorders.
party informant uses a 3-point scale to rate each item In a more narrowly focused study, Matson, Baglio, on the DASH-II on the dimensions of frequency, sever- Smiroldo, Hamilton, and Paclawskyj (1996a) found that ity, and duration. Scores range from 0 to 2 on frequency rates of stereotypies/tics, mania, impulse-control dis- (whether the behavior has occurred between 0 to more orders, and organic disorders were at least 25% higher than 10 times in the last 2 weeks), duration (whether among adults who met the DASH-II scoring criteria for the behavior has occurred anywhere from less than PDD/autism versus those who did not. In addition, 1 month to up to more than 1 year), and severity (the Matson et al. (1999), using the DASH-II, found that extent of the damage or disruption caused by the stereotypies/tics and impulse-control disorders were the behavior, ranging from no disruptions or damages, to two most common comorbid disorders occurring in a caused injury or property damage at least once). The group of institutionalized adults who had previously DASH-II has been shown to be a reliable instrument received a DSM-IV diagnosis of autism.
(Sevin, Matson, Williams, & Kirkpatrick-Sanchez, This study, focusing on lower-functioning indi- 1995), and several subscales have been validated by its viduals, was designed to rigorously test the hypothesis primary author as screening tools for specific disorders that rates of psychiatric and behavior disorders are (Matson et al., 1999; Matson & Smiroldo, 1997; higher among individuals with autism and severe in- Matson, Smiroldo, & Hastings, 1998). This instrument tellectual disability compared with individuals who do thus offers a systematic approach to documenting be- not have the additional diagnosis of autism. Participants haviors that are likely to be associated with psychiatric were grouped according to the presence or absence of and behavior disorders in persons with severe intellec- autism; the etiology of the autism or intellectual dis- tual disability, and it provides the opportunity to com- ability, where known, was not a defining group char- pare groups of individuals to determine whether such acteristic. The participants were drawn from a larger disorders are different between the groups. Although it epidemiological study on the prevalence of mental may appear that the different subscales described in this health disorders among adolescents and young adults instrument (such as “anxiety,” “mania,” “schizophre- with intellectual disability, and as such represented the nia,” and “sleep disorder”) infer psychiatric etiological total population in a defined geographic area rather than diagnoses, it should be noted that these subscale cate- an institutionalized or clinical sample. We used a men- gories are not synonymous with DSM or ICD clinical tal health screening tool (the DASH-II) that has gained psychiatric diagnoses. The validity of many or most of acceptance in clinical and research studies (and that, to these subscales in correctly diagnosing such underly- our knowledge, is the only instrument in existence at ing psychiatric conditions in low-functioning, nonver- the time of the study that was based on DSM nosology bal individuals remains to be determined (as it does for and had established psychometric properties) to study any other diagnostic approach currently available), and this should be facilitated in the future by the identifi-cation of objective biomedical markers for these Studies using the DASH have provided some in- Recruitment of Participants
formation on the rates and pattern of psychiatric andbehavior disorder among selected populations of The participants for this study were drawn from individuals with severe and profound intellectual the population of adolescents and young adults with in- disability, with and without an additional diagnosis of tellectual disability living in the Niagara Region of PDD/autism. They offer evidence of higher rates of such Southern Ontario, Canada (Bradley, Thompson, & disorder among adults with autism or autistic-like be- Bryson, 2002). At the time of the investigation, the havior. In two studies involving large groups of adults Niagara region had a population of around 400,000 living in institutions, DASH frequency-subscale scores people, with a mix of rural and urban lifestyles and were highest for elimination disorders, PDD/autism, socioeconomic circumstances reflecting the diversity mania, and stereotypies/tics (Cherry, Matson, & found in the other parts of the province (Statistics Paclawskyj, 1997; Matson et al., 1991b). In terms of Canada, 1996). To be eligible for inclusion in the larger “clinical significance” (as determined by the scoring study of mental health disorders, participants had to be criteria for the instrument), PDD/autism, mania, between the ages of 14 and 20 years as of June 1, 1994, impulse control disorders, organic syndromes, and and to have a confirmed diagnosis of intellectual stereotypies/tics were among the most commonly disability. Potential participants were recruited Bradley, Summers, Wood, and Bryson
primarily through approaching public and separate level (mental age of less than 1–2 years), a modified school boards, various parent groups, and agencies ADI (Bryson & Bradley, 2004) was developed that was serving persons with intellectual disability.
appropriate for these lower-functioning individuals.
Psychometric testing of all potential participants The ADI-R cut-off scores were not changed in the identified 171 individuals (“participants”) as having intellectual disability (full scale IQ ≤ 75). Theseparticipants underwent an assessment to determine Assessment of Psychopathology and
whether they met diagnostic criteria for autism. In Maladaptive Behavior
addition to completing the DASH-II, caregivers tookpart in an evaluation of the participant’s adaptive func- The DASH-II (Matson, 1995; Matson et al., tioning and provided some background medical and 1991b) was used to screen for behavioral and psychi- demographic information. The subset of individuals atric disorder. The DASH-II was administered as per with severe and profound intellectual disability was the protocol outlined in the manual. The interviewer read each item aloud to the informant, providing ex-planations as needed. The informant then recorded hisor her response for each item, consisting of a separate Assessment of Cognitive and Adaptive Functioning
rating score on the dimensions of frequency, duration, Because of the age range and functioning levels and severity. Frequency scores were tallied for the anx- of the participants, a variety of tests was used to assess iety, depression, mania, PDD/autism, schizophrenia, their cognitive functioning. Nonverbal IQ was obtained stereotypies/tics, organic syndromes, and impulse- from either the performance scale of the Wechsler Adult control disorders subscales, and the resultant scores Intelligence Scale-Revised (WAIS-R; Wechsler, 1981) were compared to the established cutoffs for clinical for participants who were 17 years of age and older or significance (Matson, 1995). Clinical significance was the Wechsler Intelligence Scale for Children-Revised considered to have been met when a participant’s score (WISC-R; Wechsler, 1974) for participants up to the reached or surpassed the cutoff score. For the re- age of 16 years 11 months; less capable individuals maining subscales (self-injury, elimination, eating, were administered the Merrill-Palmer Scale of Mental sleep, and sexual disorders), clinical significance was Tests (Stutsman, 1948), excluding verbal items.
based on at least one subscale item receiving a sever- Language ability (specifically, single-word receptive ity score of 1 or 2. DASH-II factor scores were com- vocabulary) was measured using the Peabody Picture puted using the factor structure and scoring criteria Vocabulary Test-Revised, Form L (Dunn & Dunn, that were outlined by Matson, Coe, Gardner and 1981). The Vineland Adaptive Behavior Scales-Survey Edition (VABS; Sparrow, Balla, & Cicchetti, 1984) wasused to assess the participants’ functioning in the areas Procedure
of communication (expressive, receptive, and written),socialization (interpersonal relations, play and leisure, The interviews and assessments took place in a lo- and coping skills), and daily living skills (personal, cation that was convenient for the participants and their domestic, and community). The maladaptive behavior caregivers; most often this was their own homes. Re- domain (Parts 1 and 2) from the VABS was also ad- search staff were trained by an experienced psycholo- ministered. Part 1 of the domain considers minor be- gist in the administration and scoring of the cognitive havioral issues, whereas part 2 describes more serious and adaptive assessments. Training in the administra- tion of the ADI-R interviews was provided by observ-ing and scoring teaching tapes with supervision frompersons trained in its use. All ADI-R interviews were Autism Diagnostic Interview-Revised
conducted by one of two research staff, both of whom The ADI-R (Lord et al., 1994) was used to assess met the recommended criterion of greater than 85% the presence of autism. The ADI-R is a semistructured interrater reliability, as per the criteria outlined by Lord, interview that is linked to ICD-10 and DSM-IV crite- Rutter, and Le Couteur (1994) regarding the use of the ria for autism. It yields separate scores in three ADI-R for research purposes. Two additional steps domains—communication, social interaction, and re- were taken to ensure agreement regarding the identifi- stricted, repetitive, and stereotyped behaviors. Because cation of autism in these lower-functioning individu- many of the participants were functioning at a very low als: first, all interviews were audiotaped, and interrater Comparing Rates of Psychiatric and Behavior Disorders in Adolescents and Young Adults
agreement between the two interviewers was checked presence of a major medical problem, motor impair- at regular intervals throughout the study. In addition, the audiotapes were independently reviewed by two of The groups were each composed of eight males the authors (E. Bradley and S. Bryson). Difficulties in and four females. The mean age of the participants was scoring individual items occasionally arose in associa- 16.33 years (SD = 2.2 years) for those with autism and tion with very low levels of functioning or where there 16.08 years (SD = 2.8 years) for those without autism.
were additional sensory or motor impairments. These To overcome the problem of participants who were difficulties were resolved by considering observations untestable on the nonverbal measures or who were made by research staff at the time of psychological tested with different instruments, standard scores from assessment, and through consensus. Research staff were the VABS are used (see Fombonne, 1992). These data also trained on the administration and scoring of the consist of the adaptive behavior composite (T) and DASH-II according to the instructions provided in scores in the domains of communication (C), social- ization (S), and daily living skills (DLS). For the group The ADI-R, VABS, and DASH-II were completed with autism, the mean standard scores are as follows: during a face-to-face interview with an informant who T = 20.33, C = 19.83, S = 20.17, and DLS = 22.42.
had interacted with the participant on a daily basis over The median standard score for each of the domains and at least the last 5 years. Almost all of the informants the total composite was less than 20 (see Fombonne, 1992). For the nonautistic group, the mean standard Fifty-seven participants achieved either a nonver- scores are as follows: T = 24.75, C = 22.75, S = 30.08, bal IQ < 40 or were untestable because of lack of abil- and DLS = 24.92. As for the group with autism, the ity on the particular measure of nonverbal performance.
median standard score for each of the domains and the Of these 57 participants, 20 (35%) met the criteria for total was less than 20. Comparability of the groups was a diagnosis of autism on the ADI-R. It was possible to assessed using a series of independent-samples t-tests individually match 12 of these individuals with autism for each of the measures. The only test to reach to one of the remaining 37 participants on the basis of significance was for the socialization standard scores gender, chronological age, and where available, per- [t(22) = −2.348, p < .05]. See Table I for additional formance IQ. Other factors were also taken into con- information regarding the characteristics of the sideration during the matching process, such as the Table I. Additional Characteristics of Participants
a Classification of gross motor function from Palisano et al. (1997).
b Categories are not mutually exclusive.
c Psychotropic drug classes and specific medications consisted of antidepressants (paxil, prozac), mood sta- bilizers (tegretol), neuroleptics (risperidol, neuleptil), and anxiolytics (buspar).
Bradley, Summers, Wood, and Bryson
Data Analysis
a nonparametric test (χ2). Third, a Pearson correlation Data were analyzed using SPSS Version 9.0 for coefficient was calculated between the total frequency Windows. Because of the finding that 42% of the par- scores on the DASH-II and the total raw scores from ticipants with autism were taking psychotropic the VABS maladaptive behavior scale.
medication, versus only 8% of the participants with-out autism, a multivariate analysis of covariance(MANCOVA) procedure with psychotropic medication as the covariate was used to rule out a potential con-found. Because the multivariate F values were signif- Table II shows the mean total frequency score on icant for group but not medication for both the the DASH-II along with the mean frequency scores for DASH-II frequency and factor scores, the covariate was each of the subscales, by group. Participants with dropped from subsequent analyses. Independent- autism showed significantly higher scores (indicative samples t-tests were used post hoc to investigate of greater disturbance) than participants without autism between-group differences on the DASH-II frequency for the total score and for seven of the 13 subscales.
and factor scores as well as the raw scores from the Specifically, participants with autism received signifi- VABS maladaptive behavior scale. Levene’s test for cantly higher scores on four of the five subscales that homogeneity of variance (Miller, 1996) was performed make up the group of psychiatric disorders and three on the data. In the event that this test was significant, of the eight subscales that make up the group of aber- equal variances were not assumed, and the t-value based on separate variances was reported. Throughout Rates for disorders that reached clinical signifi- the analyses, one-tailed tests of significance were used cance (as defined by the DASH-II scoring criteria) in light of predictions that the group with autism would were examined next. Fifty percent or greater of the receive higher scores on measures of psychiatric and participants with autism reached clinical significance behavior disorders. Second, the proportion of partici- for seven disorders compared with one disorder meet- pants in both groups who reached clinical significance ing DASH-II clinical criteria for the participants with- for each disorder on the DASH-II was compared using out autism (see Table III). The number of participants Table II. Mean Frequency Scores on the DASH-II Subscales
t value and p levela a One-tailed test.
b Frequency cutoff score ≥ 2.
c Frequency cutoff score > 4.
d Frequency cutoff score ≥ 6.
e Frequency cutoff score > 2.
f Severity cutoff score ≥ 1.
g Frequency cutoff score ≥ 8.
Comparing Rates of Psychiatric and Behavior Disorders in Adolescents and Young Adults
Table III. Percentage of Participants Reaching Clinical Significance on the DASH-II Subscalesa
a Using criteria from Matson (1995).
b df = 1; one-tailed test.
in each group who scored at or above the clinical cut- participants without autism, with 75% of this group off for each of the 13 subscales was compared using having one or no comorbid disorders.
a χ2 statistic. Significant differences were found on The factor scores by group are presented in seven of 13 subscales, with each time a greater pro- Table IV. Mean factor scores were significantly higher portion of participants with autism reaching clinical for participants with autism for social withdrawal, emo- significance than participants without autism. The tional lability, and sleep disorders.
PDD/autism subscale correctly classified 83% of the The group means for the VABS maladaptive be- individuals with autism and 100% of individuals with- havior domain were compared using an independent- out autism using the ADI-R diagnosis as the criterion samples t-test. The finding of greater behavioral for assignment into different groups (χ2 = 17.143, disturbance in the group of participants with autism was df = 1, p < .001). Participants with autism had on borne out for the total raw score [X = 18.08, SD = average 5.25 clinically significant disorders (exclud- 10.87 for the participants with autism, and X = 7.25, ing the diagnosis of PDD/autism) versus an average SD = 5.15 for the participants without autism, t(22) = of 1.25 disorders in participants without autism.
3.118, p < .01]. The total frequency scores on the On closer inspection, 50% of the participants with DASH-II were correlated with the VABS maladaptive autism had more than five clinically significant disor- domain raw scores, yielding a strong positive correla- ders, whereas the opposite trend prevailed for the tion (r = .868, p < .001).
Table IV. Comparison of DASH-II Factor Scores
t value and p levela Bradley, Summers, Wood, and Bryson
DISCUSSION
The PDD/autism subscale scores showed good dis- criminant validity when evaluated against the ADI-R— This study compared psychiatric and behavior considered to be the gold standard for assessing autism.
disorders in two groups, one with a diagnosis of autism, Only two participants with an ADI-R diagnosis of the other without, of adolescents and young adults with autism were misclassified, as their scores did not ex- severe intellectual disability who were drawn from the ceed the clinical cutoff for the autism/PDD scale, total population in this age range living in a defined whereas all 12 participants without autism were cor- geographic area. Our aim was to catalog a range of rectly classified. In contrast Matson, Smiroldo, and behaviors in the two groups and then to use a controlled Hastings (1998) reported that the DASH-II correctly comparative approach to establish which behaviors are classified 100% of institutionalized adults who met more specifically related to autism versus intellectual DSM-IV diagnostic criteria for autism and 89% of con- disability in general. It was hypothesized that the par- trols who had not received a clinical diagnosis. One of ticipants with an independently confirmed diagnosis of the participants with autism in this study who was autism would display higher rates of disorder relative incorrectly classified received a frequency score of four to nonautistic controls who were individually matched on the PDD/autism subscale, which is just below the for gender, chronological age, and performance IQ.
clinical cutoff. The other participant received a fre- Data were obtained using the DASH-II, an informant- quency score of one on the PDD/autism subscale; more- based instrument previously developed to screen for over, this individual also had the lowest total (i.e., psychiatric and behavior disorders in lower-functioning including all subscales) frequency score among all the individuals. The hypothesis of greater disturbance among participants with autism was borne out in a Group differences emerged in the rate and pattern straightforward manner in relation to the frequency of comorbid psychiatric and behavior disorders. On scores on the DASH-II; the total frequency score was average, the rate of comorbidity was four times higher almost three times higher for the participants with in the group with autism than in the nonautism group, autism compared to those without autism.
again confirming our hypothesis that in the larger group A similar pattern of results occurred when indi- of persons with intellectual disability, those with co- vidual items were grouped together to yield subscale existing autism contribute disproportionately to the and factor scores, with higher scores reflecting signif- overall prevalence rate of psychiatric and behavior dis- icantly greater disturbance for the participants with orders. The process of diagnosing mental health disor- autism on seven of 13 subscales and three of six fac- ders in individuals with severe intellectual disability is tors. Looking first at the subscale scores, the signifi- complex and challenging, in part because of differing cant findings were divided almost evenly between the views regarding the definition of such disorders; that group of “psychiatric disorders” (i.e., anxiety, is, whether to take a categorical approach (e.g., using PDD/autism, mania and depression, as defined in the DSM-IV diagnostic criteria) versus a more dimensional DASH-II) and the group of “aberrant behavior approach (e.g., using scores on a continuous rating scale disorders” (i.e., stereotypies/tics, sleep disorders, and that provide a quantitative measure of disturbance; see organic syndromes, again as defined in the DASH-II).
Brereton & Tonge, 2001, for further discussion of these The higher scores for the PDD/autism subscale are to concepts). These conceptual issues take on greater be expected because all participants in the autism group prominence in the case of individuals with autism and were selected because they met ADI-R criteria for severe intellectual disability, as evidenced by ongoing autism. Turning next to the factor scores, the largest debate regarding whether additional disturbance is significant difference was found for the social with- really a manifestation of the underlying autistic dis- drawal factor. This finding can be explained in relation order (American Academy of Child and Adolescent to the items that load on this factor (four of eight items come from the autism/PDD subscale) and the fact that Current clinical practice in diagnosing mental seven of eight items met the Matson et al. (1996a) health disorders in persons with severe intellectual criterion for a critical (i.e., ≥30%) between-group disability is heavily dependent on obtaining as much difference, whereas the eighth item approached signif- clinical data as possible from a variety of key informants icance (25% difference). The other two factor scores to to develop a longitudinal picture of behavior, an under- reach significance were emotional lability and sleep standing of current symptomatology and behavior in disorders. Both of these factors contained one critical relation to baseline or typical patterns of behavior, and item as opposed to numerous critical items for the an appreciation of the biopsychosocial circumstances of the individual and any changes in these. Establishing Comparing Rates of Psychiatric and Behavior Disorders in Adolescents and Young Adults
what is “normal” for a given individual is often diffi- without autism is not unexpected given the reported in- cult in light of competing behavioral concerns that can creasing prevalence of seizures with increasing intel- cloud the picture and the need to obtain input from lectual impairment (see Sillanpaa, 1999, for a review knowledgeable informants who may hold different sub- of this issue). However, our finding of a three- to four- jective views. The DASH-II, with its operational defi- fold increase in psychiatric and behavior disorders (as nitions of behavior and the scoring dimensions of measured by the DASH-II) in the autism group com- frequency, severity, and duration, helps to overcome pared with the group without autism, in the absence of some of these difficulties. However, it provides at best a parallel difference in seizure rate, indicates that this a “snapshot” of current behavior versus data regarding greater rate of disorders in the autism group is not the onset, duration, pattern, and magnitude of behavior changes, all of which, in addition to any changes in the Our study has a number of strengths and unique biopsychosocial circumstances of the individual, are features. These include its rigorous ascertainment meth- taken into consideration in clinical (e.g., according to ods and the fact that the study population was drawn DSM criteria) diagnostic formulations. Thus, statements from the total group of adolescents and young adults regarding the rates of clinical “disorders” based solely in a specific geographic area that had a mix of services on data from the DASH-II must be interpreted with cau- and supports for persons with intellectual disability, tion. Nonetheless, the instrument does seem to be a use- rather than being restricted to a clinical or institutional ful screening tool for detecting potential mental health sample. Some cautions are in order as well. Although disorders, and as such, can help to identify those indi- the sample size was relatively small (n = 24), and repli- viduals who require a more thorough and intensive clin- cation with larger numbers is desirable, it is important ical evaluation, particularly those who score above the to bear in mind that this sample was drawn from a much larger population base of around 400,000 people. Given Additional support in this study for the hypothe- the exploratory nature of the study, several statistical sis of greater disturbance among individuals with se- analyses were conducted that may increase the proba- vere intellectual disability and autism versus those bility of type I errors. Future studies should take a more without an additional diagnosis of autism was provided conservative approach and use procedures to guard by two sources: the higher scores by the former group on the maladaptive behavior scale from the Vineland, One of the most intriguing of the questions that arise and data regarding the prevalence of psychotropic from the findings is why autism is associated with higher medication use among the participants in the study.
rates of mental health disorders in lower-functioning Approximately two-thirds of the participants with individuals. The answer to this question likely reflects a autism who were receiving medication were taking psy- complex interaction among neurophysiological, bio- chotropic drugs, whereas only one of the participants chemical, genetic, and psychosocial factors and requires without autism was taking such medication. In the latter a more comprehensive evaluation using different diag- group, anticonvulsant medication use was twice as high nostic approaches. The current findings, using a screen- proportionately, a finding that was roughly in keeping ing instrument, point to specific vulnerability in persons with the number of participants with active seizure with autism in areas of mood, anxiety, sleep, organic syndromes, and stereotypies/tics, and as such may pro- Follow-up studies of children with autism have vide a starting point for more focused research inquiry.
shown that aggravation of symptoms or deterioration In the meantime, decisions about pharmacological treat- in behavior may occur in a half to a third of children ment for these conditions must continue to be based on around the time of puberty and early adolescence an individualized clinical approach.
(Gillberg & Schaumann, 1981; Gillberg & Steffenburg,1987; Kobayashi, Murata, & Yashinaga, 1992; Rutter,Greenfield, & Lockyer, 1967). Although the cause of ACKNOWLEDGMENTS
this deterioration is poorly understood, researchers havedocumented a peak onset of seizures during these ado- We thank all the young people and their families lescent years (with up to one-third having developed who participated in this study and who gave so gener- seizures by late adolescence), although not all children ously of their time. We have also greatly appreciated who showed deterioration developed seizures. In our the support provided by staff within the school and study, one-quarter of those with autism had seizures or developmental disability service systems in the Niagara a history of seizures compared with one-half of those Region. A special thanks to Ann Thompson, who has without autism. The high rate of seizures in the group provided meticulous assistance in so many aspects of Bradley, Summers, Wood, and Bryson
the study. This research was supported by a grant King, B. H., DeAntonio, C., McCracken, J. T., Forness, S. R., & (Bradley and Bryson) from Health Canada through the Ackerland, V. (1994). Psychiatric consultation in severe andprofound mental retardation. American Journal of Psychiatry, National Research and Development Program (Project Kobayashi, R., Murata, T., & Yashinaga, K. (1992). A follow-up study of 201 children with autism in Kyushu and Yamaguchi Areas,Japan. Journal of Autism and Developmental Disorders, 22,395–411.
REFERENCES
Lainhart, J. E., & Folstein, S. E. (1994). Affective disorders in people with autism: A review of published cases. Journal of Autism andDevelopmental Disorders, 24, 587–601.
American Academy of Child and Adolescent Psychiatry (1999).
Lord, C., Riei, S., Lambrecht, L., Cook, E. H., Leventhal, B. L., Practice parameters for the assessment and treatment of chil- DiLavore, P. C., Pickles, A., & Rutter, M. (2000). The Autism dren, adolescents, and adults with autism and other pervasive Diagnostic Observation Schedule-Generic: A standard measure developmental disorders. Journal of the American Academy of of social and communication deficits associated with the Child and Adolescent Psychiatry, 38, 32S–54S.
spectrum of autism. Journal of Autism and Developmental American Psychiatric Association (1994). Diagnostic and Statistical Manual of Mental Disorders. (4th ed.). Washington, DC: Author.
Lord, C., Rutter, M., & Le Couteur, A. (1994). Autism Diagnostic Borthwick-Duffy, S. A. (1994). Epidemiology and prevalence of Interview-Revised: A revised version of a diagnostic interview psychopathology in people with mental retardation. Journal of for caregivers of individuals with possible pervasive develop- Consulting and Clinical Psychology, 62, 17–27.
mental disorders. Journal of Autism and Developmental Disor- Bradley, E. A., Thompson, A., & Bryson, S. E. (2002). Mental retardation in teenagers: Prevalence data from the Niagara Matson, J. L. (1995). The Diagnostic Assessment for the Severely Region, Ontario, Canada. Canadian Journal of Psychiatry, 47, Handicapped-Revised (DASH-II). Baton Rouge, LA: Scientific Bregman, J. D. (1991). Current developments in the understanding Matson, J. L., Baglio, C. S., Smiroldo, B. B., Hamilton, M., & of mental retardation. Part II: Psychopathology. Journal of the Paclawskyj, T. (1996a). Characteristics of autism as assessed by American Academy of Child and Adolescent Psychiatry, 30, the Diagnostic Assessment for the Severely Handicapped-II (DASH-II). Research in Developmental Disabilities, 17, Brereton, A. V., & Tonge, B. J. (2001, September). Psychopathology in children and adolescents with autism compared to an epi- Matson, J. L., Benavidez, D. A., Compton, L. S., Paclawskyj, T., & daemiological sample of young people with intellectual dis- Baglio, C. (1996b). Behavioral treatment of autistic persons: A ability and syndrome groups. Paper presented at the Australian review of research from 1980 to the present. Research in National Autism Conference, Adelaide, Australia.
Developmental Disabilities, 17, 433–465.
Bryson, S. E. (1996). Brief report: Epidemiology of autism. Journal of Autism and Developmental Disorders, 26, 165–167.
Matson, J. L., Coe, D. A., Gardner, W. I., & Sovner, R. (1991a). A Bryson, S. E., & Bradley, E. A. (2004). Prevalence of autism among factor analytic study of the Diagnostic Assessment for the adolescents with intellectual disability. Manuscript in preparation.
Severely Handicapped. Journal of Nervous and Mental Disease, Cherry, K. E., Matson, J. L., & Paclawskyj, T. R. (1997). Psy- chopathology in older adults with severe and profound mental Matson, J. L., Gardner, W. I., Coe, D. A., & Sovner, R. (1991b). A scale retardation. American Journal on Mental Retardation, 101, for evaluating emotional disorders in severely and profoundly mentally retarded persons. British Journal of Psychiatry, 159, Deb, S., Matthews, T., Holt, G., & Bouras, N. (Eds.) (2001). Prac- tice guidelines for the assessment and diagnosis of mental health Matson, J. L., Rush, K. S., Hamilton, M., Anderson, S. J., Bamburg, problems in adults with intellectual disability. The European J. W., Baglio, C. S., Williams, D., & Kirkpatrick-Sanchez, S.
Association for Mental Health in Mental Retardation. Brighton: (1999). Characteristics of depression as assessed by the Diag- nostic Assessment for the Severely Handicapped-II (DASH-II).
Dunn, L. M., & Dunn, L. M. (1981). Manual for the Peabody Pic- Research in Developmental Disabilities, 20, 305–313.
ture Vocabulary Test-Revised, Forms L and M. Circle Pines, MN: Matson, J. L., & Smiroldo, B. B. (1997). Validity of the mania subscale of the Diagnostic Assessment for the Severely Handicapped-II Dykens, E. M., Hodapp, R. M., & Finucane, B. M. (2000). Genetics (DASH-II). Research in Developmental Disabilities, 18, and Mental Retardation Syndromes. Baltimore, MD: Paul Matson, J. L., Smiroldo, B. B., & Hastings, T. L. (1998). Validity of Fombonne, E. (1992). Diagnostic assessment in a sample of autistic the autism/pervasive developmental disorders subscale of the and developmentally impaired adolescents. Journal of Autism Diagnostic Assessment for the Severely Handicapped-II. Jour- and Developmental Disorders, 22, 563–581.
nal of Autism and Developmental Disorders, 28, 77–81.
Ghaziuddin, M., Alessi, N., & Greden, J. F. (1995). Life events and Miller, R. G. (1996). Beyond ANOVA: Basics of applied statistics depression in children with autism and pervasive developmen- (2nd ed.). London: Chapman and Hall.
tal disorders. Journal of Autism and Developmental Disorders, Nordin, V., & Gillberg, C. (1996). Autism spectrum disorders in children with physical or mental disability or both. II Screening Gillberg, C., & Schaumann, H. (1981). Infantile autism and puberty.
aspects. Developmental Medicine and Child Neurology, 38, Journal of Autism and Developmental Disorders, 11, 365–371.
Gillberg, C., & Steffenburg, S. (1987). Outcome and prognostic Palisano, R., Rosenbaum, P., Walter, S., Russell, D., Wood, E., & factors in infantile autism and similar conditions: A population- Galuppi, B. (1997). Gross motor function classification system based study of 46 cases followed through puberty. Journal of for cerebral palsy. Developmental Medicine and Child Neurol- Autism and Developmental Disorders, 17, 273–287.
Kim, J. A., Szatmari, P., Bryson, S. E., Streiner, D. L., & Wilson, Reid, A. H. (1980). Diagnosis of psychiatric disorder in the severely F. J. (2000). The prevalence of mood problems among children and profoundly mentally retarded. Journal of the Royal Society with autism and Asperger syndrome. Autism, 4, 117–132.
Comparing Rates of Psychiatric and Behavior Disorders in Adolescents and Young Adults
Royal College of Psychiatrists (2001). Diagnostic Criteria for Psy- Statistics Canada, CANSIM (1996). Population by age group. Ottawa: chiatric Disorders for Use with Adults with Learning Disabilities/ Mental Retardation (DC-LD). London: Gaskell Press.
Stromme, P., & Diseth, T. H. (2000). Prevalence of psychiatric Rutter, M., Greenfield, D., & Lockyer, L. (1967). A five-to-fifteen diagnoses in children with mental retardation: Data from a year follow up study of infantile psychosis: II. Social and be- population-based study. Developmental Medicine and Child havioural outcomes. British Journal of Psychiatry, 113, Stutsman, R. (1948). Guide for administering the Merrill-Palmer Sevin, J. A., Matson, J. L., Williams, D., & Kirkpatrick-Sanchez, S.
Scale of Mental Tests. New York: Harcourt, Brace, & World.
(1995). Reliability of emotional problems with the Diagnostic Tsai, L. Y. (1996). Brief report: Cormorbid psychiatric disorders Assessment for the Severely Handicapped (DASH). British of autistic disorder. Journal of Autism and Developmental Journal of Clinical Psychology, 34, 93–94.
Sillanpaa, M. (1999). Definitions and epidemiology. In M. Sillanpaa, Wechsler, D. (1974). Wechsler Intelligence Scale for Children- L. Gram, S. I. Johannessnen, & T. Tomson (Eds.), Epilepsy Revised. San Antonio, TX: Psychological Corporation.
and mental retardation (pp. 1–5). Petersfield, Hampshire: Wechsler, D. (1981). Wechsler Adult Intelligence Scale-Revised. San Antonio, TX: Psychological Corporation.
Sparrow, S. S., Balla, D. A., & Cicchetti, D. V. (1984). Vineland World Health Organization (1992). The ICD-10 Classification of Adaptive Behavior Scales: Survey Form Manual. Circle Pines, Mental and Behavioural Disorders—Clinical Descriptions and Diagnostic Guidelines. Geneva: Author.

Source: http://autism.medicine.dal.ca/research/publications/bryson/Bradley,%20Summers,%20Wood,%20&%20Bryson%20(2004).pdf