Asperger syndrome (AS) is a sub-group on the autistic spectrum (Baron-Cohen, 1995; Frith, 1991; Wing, 1981, 1988) and is diagnosed on the basis of DSM-IV criteria (APA, 1994). The DSM-IV criteria for autism and AS both require patients to demonstrate the same number of impairments in social interaction and to demonstrate obsessions or repetitive behaviour. However, for autism, but not for AS, qualitative impairments in communication must be evident. In addition, for AS, there must be no significant general delay in language (defined by the use of single words at the age of 2 years and communicative phrases at the age of 3 years) or in cognitive development. For autism, this restriction is not stated, so that an autism diagnosis can be made in an individual of any IQ or language level. A final difference between the DSM-IV criteria for autism and AS is that the AS diagnosis specifies that the disturbances must cause clinically significant impairment in social, occupational or other important areas of functioning. (Bizarrely, this criterion is not specified for autism.) These similarities and differences between the DSM-IV criteria for autism and AS are summarised in Table I.

Table I. Similarities and Differences between the DSM-IV Diagnostic Criteria for Autism and Asperger Syndrome (AS)

These diagnostic criteria for AS are unsatisfactory, at two levels (Howlin, 2000). First, no communication abnormalities are included as symptoms, despite the difficulties in the domain of pragmatics (Baron-Cohen, O’Riordan, Jones, Stone, & Plaisted, 1999). Secondly, the number of required criteria are so few that there is a risk of over-diagnosis. In this paper we describe a new diagnostic instrument, the Adult Asperger Assessment (AAA), designed to assess adults for AS and high functioning autism (HFA). The AAA uses more stringent diagnostic criteria than DSM IV.

A new instrument for assessing AS in adults is needed because at present nothing exists for this purpose. This is because it is only recently that clinicians have really become aware of AS. The condition was first described more than 60 years ago by Hans Asperger (Asperger, 1944), but the attention of the English-speaking scientific and medical world was only drawn to AS by the publication of articles by Lorna Wing (Wing, 1981, 1988). Apart from a small number of research studies in the 1980s (Tantam, 1988a, 1988b, 1988c; Tantam, Monaghan, Nicholson, & Stirling, 1989), this condition still went mostly undiagnosed. The publication of the first book on the subject in English in the early 1990s changed this situation significantly (Frith, 1991). Today, there are hundreds of books on AS, and a wealth of research debating its relationship to autism and characterizing its cognitive profile, epidemiology, and biology (Baron-Cohen, 2000; Baron-Cohen, Jolliffe, Mortimore, & Robertson, 1997; Ghaziuddin, Tsai, & Ghaziuddin, 1992; Klin, Volkmar, Sparrow, Cicchetti, & Rourke, 1995; Szatamari, Bartolucci, Bremmer, Bond, & Rich, 1989; Szatmari, Archer, Fisman, Streiner, & Wilson, 1995; Szatmari, Bartolucci, & Bremner, 1989; Szatmari, Tuff, Finlayson, & Bartolucci, 1990). Family pedigrees of AS implicate heritability (Gillberg, 1991). There are also numerous first person and practical accounts (Attwood, 1997; Sainsbury, 2000; Willey, 1999).

Due to this relatively recent increase in awareness of AS, professionals are now alert to the possibility of AS in primary school age children. Current estimates of its prevalence are as high as 1 in 200 children (Ehlers & Gillberg, 1993; Scott, Baron-Cohen, Bolton, & Brayne, 2002). But for individuals who were born before 1980, a new diagnostic dilemma has arisen. What instruments should be used to diagnose AS in adults who have reached this point in development without being recognised as having it?

Currently available diagnostic instruments for autism spectrum conditions, such as the ADI-R (Autism Diagnostic Interview) (Le Couteur et al., 1989; Lord, Rutter, & Le Couteur, 1994), the ADOS-G (Autism Diagnostic Observation Schedule) and the CARS (Childhood Autism Rating Scale) (Schopler, Reichler, & Renner, 1986), are not age-appropriate for adults with AS. Moreover, as mentioned above, the DSM-IV criteria for AS are currently a cause for concern, not least because of how few symptoms an individual needs to manifest to warrant a diagnosis. For this reason, we describe new criteria and a new method for diagnosis of AS or HFA in adults.

The Adult Asperger Assessment (AAA)

The distinction between AS and HFA is simplified in the AAA so that those patients who meet criteria and who did not have a language delay are diagnosed with AS, whereas patients who meet criteria and did have a language delay are diagnosed with HFA.

The Adult Asperger Assessment (AAA) is shown in Appendix A. It comprises 4 sections each describing a group of symptoms (A–D), and then a final section (E), describing 5 key prerequisites. The AAA incorporates all the symptoms from the DSM-IV diagnosis of Asperger’s Disorder as well as additional relevant symptoms. The additional symptoms reflect the more stringent approach to diagnosis taken in the AAA compared with DSM-IV. This conservative design was employed so as to err on the side of under-diagnosis.

Section A of the AAA is ‘Qualitative impairment in social interaction’. This includes 4 symptoms from the DSM-IV AS criteria, and 1 extra symptom. The wording of symptom A3 has been slightly adapted from DSM-IV for the AAA. Section B is “Restricted, repetitive and stereotyped patterns of behaviour, interests and activities”. This section also has 4 symptoms from the DSM-IV AS criteria, and 1 extra symptom. Section C, “Qualitative impairments in verbal or non-verbal communication”, and Section D, “Impairments in imagination”, are not included in the DSM-IV AS diagnosis but form part of the DSM-IV autistic disorder diagnosis. The AAA retains these additional sections for the diagnosis of AS. None of the 5 symptoms in Section C appear in DSM-IV and are worded to be appropriate for diagnosing adults with AS. They refer to the often-noted difficulties with the pragmatic (rather than syntactic or phonological) aspects of language and communications. (Baron-Cohen, 1988; Tager-Flusberg, 1993). In Section D, 1 of the symptoms is from the DSM-IV autistic disorder diagnosis and the other 2 are new. Any symptoms not found in DSM-IV are asterisked on the AAA.

To meet criteria for a DSM-IV diagnosis of AS, patients need to have 2 or more symptoms from Section A (A1–A4 of the AAA) and one or more symptoms from Section B (B1–B4 of the AAA) . They also need to meet the pre-requisites in Sections E of the AAA. In our view, the DSM-IV criteria for AS are too lax, and so the AAA criteria have been designed to be more stringent. Therefore, it follows that anyone who meets the AAA criteria will also meet the DSM-IV criteria. On the AAA, patients must have at least 3 out of 5 of the symptoms in each of Sections A–C, 1 out of 3 symptoms from Section D. This means the patient has to score on at least 10 symptoms (max = 18). They must also meet all 5 pre-requisites in Sections E. Note that this includes showing no current or post evidence of psychosis, which the AAA, like DSM-IV, uses as an exclusion criterion for AS (ICD-10 allows for comorbid AS and schizophrenia).

Before the clinical interview, patients are asked to complete the AQ (Autism Spectrum Quotient: Baron-Cohen, Wheelwright, Skinner, Martin, & Clubley, 2001) and the EQ (Empathy Quotient: Baron-Cohen & Wheelwright, 2004) as screening questionnaires. These are described next.

The AQ comprises 50 questions, made up of 10 questions assessing 5 different areas (see Appendix B): social skill (items 1, 11, 13, 15, 22, 36, 44, 45, 47, 48); attention switching (items 2, 4, 10, 16, 25, 32, 34, 37, 43, 46); attention to detail (items 5, 6, 9, 12, 19, 23, 28, 29, 30, 49); communication (items 7, 17, 18, 26, 27, 31, 33, 35, 38, 39); imagination (items 3, 8, 14, 20, 21, 24, 40, 41, 42, 50). Individuals score in the range 0–50. In our previous study we compared n = 58 adults with Asperger syndrome (AS) or high-functioning autism (HFA); with n = 174 randomly selected controls. The adults with AS/HFA had a mean AQ score of 35.8 (SD = 6.5), significantly higher than controls (x = 16.4, SD = 6.3). Eighty percent of the adults with AS/HFA scored 32+, vs. 2% of controls. The AQ strongly predicts an AS diagnosis in a clinic sample (Woodbury-Smith, Robinson, Wheelwright, & Baron-Cohen, 2005)

The EQ comprises 60 questions, 40 assessing empathy and 20 filler (control) items (see Appendix C). Filler items are 2, 3, 5, 7, 9, 13, 16, 17, 20, 23, 24, 30, 31, 33, 40, 45, 47, 51, 53 and 56. It has a maximum score of 80 and a minimum of zero. In our previous study using this instrument we employed the EQ with n = 53 adults with AS or high-functioning autism (HFA). Their mean EQ score was 20.4 (SD = 11.6), which was significantly lower than n = 53 age and sex-matched controls (mean EQ = 42.1, SD = 10.6). 80% of the adults with AS/HFA scored equal to or less than 30 out of 80, compared to only 10% of controls.

The AAA template is a Microsoft Excel document. The patient’s response to each item on the AQ and EQ is entered and a macro is then run to score the AQ and EQ. The scores from these questionnaires are automatically entered on to the front sheet of the AAA. In addition, items from the AQ and EQ, which the patient has endorsed, provide examples of the symptoms in Sections A–D, so these are automatically entered into the appropriate sections of the AAA. The AAA in Appendix A shows a hypothetical patient (Max Asperger) who has the maximum AQ score (=50) and the minimum EQ score (=0). This AAA therefore includes all possible examples of symptoms from the AQ and EQ. Note that the AQ and EQ do not provide examples for all the symptoms.

During the clinical interview, the clinician then seeks to validate the symptom examples provided by the AQ and EQ by gathering examples from the patient and their relative/informant and checks the other symptoms and prerequisites. In our national clinic, CLASS (Cambridge Lifespan Asperger Syndrome Service), patients are only seen if they can provide an informant who can supply details of their developmental history. Following this interview, the AAA is completed by entering a 1 in the YES or NO box for each symptom and prerequisite as appropriate. There is space for the clinician to include additional examples of symptoms and comments. The number of symptoms and prerequisites met are automatically totalled on the front sheet of the AAA. The AAA is completed by the clinician entering their diagnosis, if any.

Validation study

Sample

Forty two patients in series attending the Cambridge Lifespan Asperger Syndrome Service (CLASS), a national diagnostic clinic for adults referred with suspected AS, were assessed using the AAA. The mean age of the sample was 34.1 years (SD = 10.6 years) and the sex ratio was 36:6 (or 9:1) (male:female). The occupational levels of the patients were mixed, varying from university researchers to unemployed and unskilled workers. Twelve of them were married or in long-term relationships, 3 were divorced or separated, and 27 were single. Nine (or 21.4%) had a history of violence. Finally 12 (28.6%) were living alone.

Method

The AAA was administered by a team comprising either a consultant clinical psychologist (SBC) or consultant psychiatrist (MWS) and a clinical psychologist (JR) in the team. Two professionals were involved in every assessment. Each patient was accompanied by at least one parent as an informant. In the case of those patients with a partner, the partner was also invited to the assessment to act as an additional informant. Each area in the AAA was probed, in order to collect a range of anecdotal examples of a specific kind, from the patient’s life, either from self-report or via the informant(s). Each AAA interview took on average 3 hours, including collecting early developmental history information about the patient, information about the patient’s educational history, occupational history, medical history, family factors, and to allow time for feedback on the diagnosis. The team of two clinicians filled in the AAA independently and where there was disagreement, this item was conservatively not scored.

Each patient was therefore assigned a total AAA score, a DSM-IV score, an AQ score, an EQ score, and a AAA diagnosis. The AQ and EQ were sent by post in advance of the clinic visit, and filled in by the patient alone.

Results

Of the 42 patients assessed, 31 (28 male, 3 female) were diagnosed with AS, 3 (all male) were diagnosed with HFA and 8 (5 male, 3 female) did not receive an autism spectrum diagnosis. Of the 8 (all male) who did not receive an AAA diagnosis, 3 met DSM-IV criteria for AS.

The mean AQ and EQ scores for the patients with and without a diagnosis of autism spectrum are shown in Table II. As would be predicted, the diagnosed group score significantly higher on the AQ, and significantly lower on the EQ, than the patients without a diagnosis (t = 3.1, p = .004 and t = −2.5, p = .015 respectively). The AS group’s scores on the AQ are well above the usual cut-off of 32, and below the usual cut-off on the EQ of 30. Table II also shows the mean AAA score for each of the two groups, the AS spectrum group scoring well above the minimum cut-off (10/18) whilst the non-AS group scored well below this.

Table II. Mean AQ and EQ Scores (and SD’s) of Patients with and without Final Autism Spectrum Diagnosis

Discussion

The study reported here introduces a method for the diagnosis and assessment of adults with AS. It is easy to use and links with two screening instruments, the AQ (Autism Spectrum Quotient) (Baron-Cohen et al., 2001) and the EQ (Baron-Cohen & Wheelwright, 2004) thus making maximum use of information that can be collected prior to the clinic visit. Results confirm that the AAA diagnosis is more conservative than the DSM-IV diagnosis. Results also replicate earlier findings that adults with a clinical diagnosis of AS score above 32 on the AQ, and below 30 on the EQ. It is recommended that clinicians may wish to adopt the AAA in order to maintain a stricter definition of AS.

There is some debate over whether “impairments in imagination” should be diagnostic criteria for AS, since cases of individuals with AS who are gifted at drawing, film-making, and poetry are well-documented (e.g. Myers et al., 2005). It should be noted that in the AAA, “impairments in imagination” carry less weight in that only one symptom out of the minimum of 10 is required in this domain. Future studies could examine the value of including this one symptom or not.

We report this instrument at this stage of its development, because of the need for a quantitative, clinical method. However, we recognise that in future it will be useful to compare its performance relative to other approaches, such as an adapted form of the ADI-R (Le Couteur et al., 1989). Equally important will be a test of this and other instruments’ power to distinguish between AS and the milder manifestations described as the “broader phenotype” in family genetic studies of autism (Bailey et al., 1995)

Future studies will need to test the sensitivity and specificity of the AAA, as well as the linked screening instruments, by inclusion of psychiatric control groups. It is expected that in the medium term there will be little or no clinical need for the AAA, if most or all of the actual cases of AS are identified when they should be, namely in early childhood. Alternatively, there may be value in adapting the AAA for use in childhood if it is more conservative than alternative methods. But in the interim, whilst undiagnosed cases of AS exist in the adult population, both screening and diagnostic instruments of this kind will be needed. Finally, it is hoped that the extended and more conservative criteria for AS adopted by the AAA will contribute to the debate on what should constitute useful criteria for defining this syndrome.