Discrepancy Between Parent Report and Clinician Observation of Symptoms in Children With Autism Spectrum Disorders

Wiggins and Robins (2007) also investigated the ADI-R, and ADOS in their use of diagnosis. They brought in 142 toddlers, 60 with no disorder, 42 with Autistic Disorder, and 30 with another spectrum disorder to see the validity of diagnosis. Using a double-blind study Wiggins and Robins had clinicians perform the ADI-R, ADOS, and the Child Autism Rating Scale (CARS), which is an observation measure, on the participants (2007). They found disagreement between the parent report of ADI-R with the CARS and ADOS, and when controlling for diagnosis of Autism alone rather than including other ASD diagnoses, they found that the ADI-R under-diagnosed according to the diagnoses that the toddlers came in with. When they removed the behavior domain there was much greater agreement between the measures, which shows that the behavior domain may need further analysis (Wiggins & Robins, 2007).

Stone, Hoffman, Lewis and Ousley have investigated parent report and clinician observation in their research on early recognition of autism (1994). Stone et. al used the Childhood Autism Rating Scale (CARS) and the Parent Interview for Autism (PIQ) as the parent assessments and evaluated the child directly using either the Bayley Scales of Infant Development or the Merrill Palmer Scale of Mental Tests, and a motor imitation task (1994). They directly compared the parent reports with the clinician reports of behaviors in the domain of Rutter's criteria for a clinical diagnosis and compared the agreement on items within each domain for children under the age of four (Rutter, 1987). They found that agreement was strongest for the absence of behaviors rather than the presence, and only three of the twenty six participants had an acceptable level of agreement between clinicians and parents on the items of abnormal social play, stereotyped body movements, and restricted interests (Stone et. al, 1994). Low agreement was found on the items of impaired imitation, lack of awareness of others, impaired peer friendships, no mode of communication, abnormal nonverbal communication, absence of imaginative play, and preoccupation with parts of objects (Stone et. al, 1994). Their study is incredibly relevant to diagnosis, but because the age range is limited, information is lacking on whether the parent's knowledge of the child's behaviors at different ages can affect their assessment on current behaviors. Additionally, since this study no one has performed a comparative analysis of these using updated measures such as the ADI-R and ADOS.

Stone et. al's (1994) research shows the discrepancy between parent report and clinician observation but more research is needed. Some research indicates that parents can be a reliable source of information regarding their child's development (Glascoe, 2003). Little research has indicated whether clinician report is more accurate than the parent's or if it is offering information that the parent cannot provide. The research of Lord et. al (2006), Bender et. al (2008), Sakora et. al (2008) and Tomanik et. al (2007), lends itself to the study of parent versus clinician assessment. Using the ADI-R and the ADOS these clinicians have found discrepancies between the different kinds of tests and further investigation should follow up by comparing the differences in the subtests and if the results are different because of the reporters or because of the tests. The subtests are categorized based on the DSM IV criteria for diagnosis, and questions are coded and included on the subtests in each assessment. The assessments include subtests on social skills, restricted interests, behavior, and communication. This may be valuable information for the healthcare provider, educator, and therapeutic intervention service providers who will need specific details on a child's diagnosis. It could provide information to help them determine if they should rely more heavily on a specific test or reporter or give them equal weight in the diagnostic process.

Research indicates that using both the ADI-R and the ADOS gives the most accurate results (Risi et. al, 2006; Lord, 2006). Because both have been shown to be reliable and valid, further investigation should be done using the subtests to isolate current behaviors, past behaviors, adaptive behavior, communication skill, imaginative play, and other components of the child's activities to determine whether parents or objective clinicians have a more accurate rating of the activities. The purpose of this study is to investigate if parent and clinician report differ on the sub-scores of communication abilities, stereotyped behavior and restricted interests for children with ASD. It is hypothesized that parents will over report symptoms in the social domain, and under-report symptoms in the domains of interests and communication compared to the clinician. This should also be done to investigate if there is a deficit in the DSM IV criteria that parents and clinicians do not report the comparable scores. This may lead to the inclusion of not only the ADOS and the ADI-R in an official diagnosis but other assessments that may be aimed towards one particular deficit area. By doing this more specific analysis, future research can lead to finding why there may be discrepancies. If results indicate a discrepancy, programs can be created to help parents learn how to observe and analyze their child's behavior and clinicians to gather full information from their assessments.

Methods

Twelve participants who had previously received diagnosis of autism spectrum disorder including Pervasive Developmental DisorderNot Otherwise Specified (PDDNOS), Autism, or Asperger's syndrome and their parents participatedin the study . Participants' parents provided verification of diagnosis from a physician or psychologist. The parents acted as informants and are the persons who are most familiar with the person with daily behaviors of ASD.

Inclusion/Exclusion Criteria: All participants were individuals with ASD, PDD-NOS, or Asperger's Syndrome who have an IQ of less than 100. Those with additional diagnosis of genetic neurodevelopmental disabilities or major mental or psychological disorders were excluded from this study. Eligibility was determined by the most recent IQ test results provided to the study, and their ability to communicate verbally being present,. Participants were between the ages of 7 and 40 years old, have aparent or guardian as an informant that is familiar with their behavior, have an IQ of 100 or below, are verbal and use English as their primary language. The age range for this study was 9 to 22 years old (mean age = 15.8) and Participants also had Performance IQ's that ranged from 45 to 117 with a mean of 86. All of the participants lived with the person acting as the informant.

Participants were recruited through the Autism Society of Greater Cleveland, "Walk Now" for Autism Speaks and Northeast Ohio schools and camps for children with ASD. Participants and informants signed the Consent and Assent forms before beginning any part of the study. They were compensated for their participation, travel, and parking when they traveled to Case Western Reserve University's Neurodevelopment Research Lab in Cleveland.

Procedure

Measures: The Autism Diagnostic Observation Schedule (ADOS) is the "gold standard" for assessing individuals with PDD-NOS and ASD for toddlers through adults (Lord, Rutter, DiLavore, Risi, 2006). It is a 30 to 45 minute standardized observation measure designed to assess autistic behaviors. Scoring yields cutoffs for fitting diagnosis of categories of ASD and for fitting diagnostic criteria for each domain. The ADOS involves the participant and a trained clinician and consists of four modules, which are divided based on developmental and language level. The developmental and language level of each participant was matched to the proper module, and only that module was administered to that participant. In this study Modules 2, 3 and 4 were administered given the verbal level of participants recruited. During the ADOS the clinician presented many opportunities for the participant to exhibit behaviors of interest in the diagnosis of ASD through "standard 'presses" for communication and social communication (Lord, 2006). These presses are used in the different portions of the test, and are similar to psychoeducational or developmental tests such as the Psychoeducational Profile by Shopler and Reichler (1980). The ADOS uses structured activities and materials that provide standard contexts for social interaction, communication, and any atypical behaviors. The presses refer to the "immediate environment that has direct implications for the subjects behavior" (Lord, 1989). The environment includes the toys, tools, and the clinician who attempts to interact during play with the child. The ADOS domain for language and communication, variable ADOS Communication, includes : overall level of nonechoed language, speech abnormalities associated with ASD (intonation, volume, rhythm, rate), immediate echolalia, stereotyped/idiosyncratic use of words or phrases, offering information, asking for information, reporting events, conversation, and descriptive, conventional, instrumental or informational gestures. The domain for reciprocal social interaction, variable ADOS Social, includes; unusual eye contact, facial expressions directed to others, language production and linked nonverbal communication, shared enjoyment in interaction, empathy/comments on others emotions, insight, quality of social overtures, quality of social response, amount of reciprocal social communication, and overall quality of rapport. The domain for stereotyped behaviors and restricted interests, variable ADOS Behaviors, includes; unusual sensory interest in play material/ person, hand and finger and other complex mannerisms, self-injurious behavior, excessive interest in or references to unusual or highly specific topics or objects or repetitive behaviors, and compulsions or rituals.

Scoring for most tasks in the ADOS are on a threepoint scale. Beginning with 0= within normal limits, to 1= infrequent or possible abnormality, to 2 = definite abnormality (Lord, 1989). The same aspect of a behavior cannot be coded as abnormal more than once but different aspects of that behavior can be coded more than once (Lord, 1989). An algorithm based on the tasks and items was used to determine the number of participants who meet diagnostic criteria in each domain. Lord et. al (1989) found the discriminant validity of the algorithm as quite good for social and communication criteria. Using intraclass correlations, Lord et. al (1989) assessed the interrater reliability of the algorithm and found no changes in classification with different clinicians as raters. The ADOS was also videotaped and another trained clinician watched the video and scored for reliability.

The Autism Diagnostic Interview-Revised (ADI-R) is a diagnostic measure used for individuals of all ages and administered in interview format to informant (Rutter, Le Couteur, Lord, 2005). The informant is asked questions about the participant's family and education, diagnosis, and medications, informant's concerns and introductory questions that help to create a general picture of the participant's behavior, early development and key developmental milestones, language history, communication and language functioning, social development and play, interests and behaviors, and general behaviors of clinical importance (aggression, self-injury, epileptic features). The ADIR is divided into three domains when interpreting results. These domains correspond to ASD diagnostic criteria in the DSM IV: Qualitative abnormalities in reciprocal social interaction ("A"), Qualitative abnormalities in communication ("B"), and restricted, repetitive, and stereotyped patterns of behavior ("C"), and if these signs of ASD were evident before the age of 36 months ("D")(Rutter, 2005). The interviewer obtains specific and detailed responses from the informant. There are nine codes for responses from the informant. These codes begin at one with behaviors not being present, progress to abnormal behaviors being present at varying degrees, and the question being nonapplicable. The algorithm for scoring conversion was created based on the maximum likelihood of ASD according to clinical consensus that participants would score above the cutoff. The Current Behavior Algorithm is used in scoring so as to get a direct comparison between the current behavior that is observed in the ADOS. The Current Behavior Algorithm does not have cutoffs, but can be compared with ADOS scores on items endorsed and when scores are normalized with the ADI-R using z-scores (Noterdaeme et. al, 2002).

The ADI-R variable ADI-R Social includes failure to use nonverbal behaviors to regulate social interaction, failure to develop peer relationships, lack of shared enjoyment, and lack of socioemotional reciprocity in the domain of qualitative abnormalities in reciprocal social interaction. Included in the domain of qualitative abnormalities in communication, variable ADI-R communication is; lack of, or delay in, spoken language and failure to compensate through gesture, relative failure to initiate or sustain conversational interchange, stereotypes, repetitive idiosyncratic speech, and lack of varied spontaneous make-believe or social initiative play. Included in domain of restricted, repetitive and stereotyped patterns of behavior, variable ADI-R Behavior is; encompassing preoccupation or circumscribed pattern of interest, apparently compulsive adherence to non-functional routines or rituals, stereotyped and repetitive motor mannerisms, and preoccupations with part of objects or non-functional elements of material.

The ADOS and ADI-R do not use the same metrics in their scoring, but have been reliably converted and compared (Lord et. al, 2006; Bender et. al, 2008; Sakora et. al, 2008; Tomanik et. al, 2007; Noterdaeme et. al CITE). The domains correspond with one another and the scores can be compared directly after the conversion. In this study z-scores were created to directly compare the ADI-R and the ADOS. Many of the items on the ADI-R and the ADOS are also directly comparable and can be used to examine percent agreement.

Procedure: Participants and informants came to the Neurodevelopment Research Lab at Case Western Reserve University on the day of their appointment. The procedures were explained to them and informed consent was obtained. Then the participant was administered the ADOS. Later, a trained professional or research assistant interviewed the informant by administering the ADI-R. Once both assessments were completed the participants were compensated for their time.Continued on Next Page »