Diagnoses of Autism Spectrum Disorder using the DSM-5

The content of this document is OFFICIAL.

Please note: The research and literature reviews collated by our TAB Research Team are not to be shared external to the Branch. These are for internal TAB use only and are intended to assist our advisors with their reasonable and necessary decision-making. Delegates have access to a wide variety of comprehensive guidance material. If Delegates require further information on access or planning matters they are to call the TAPS line for advice. The Research Team are unable to ensure that the information listed below provides an accurate & up-to-date snapshot of these matters

Research questions:

1. What is the accuracy of Autism Spectrum Disorder diagnoses using the DSM 5, particulary for ASD levels 2 and 3 and particularly focussing on the interrater reliability of single discipline assessments?
2. What is the incidence of ASD diagnosis among family members? How likely is it that multiple siblings in a family will all have Autism Spectrum Disorder?
3. How has the rate of diagnosis of ASD changed since the publication of the DSM 5 diagnostic criteria?

Date: 10/12/2021 Requestor: Shannon s22(1)(a)(ii) - irrelevant Endorsed by (EL1 or above): Shannon s22(1)(a)(ii) - irrelevant Cleared by: Felicity s22(1)(a)(ii) - irrelevant

Contents

Diagnoses of Autism Spectrum Disorder using the DSM-5 ……………………………………………….. 1

1. Contents ……………………………………………………………………………………………………….. 1
2. Summary ………………………………………………………………………………………………………. 2
3. Frequency of ASD diagnoses in families ……………………………………………………………. 2
4. Accuracy and inter-rater reliability of ASD diagnoses using DSM-5………………………………. 3

Summary

This literature review addresses questions relating to the prevalence of Autism Spectrum Disorder (ASD). Findings include:

• ASD is strongly genetic. If someone has a family with ASD it is more likely that they will be diagnosed with ASD and it is more likely they will display autistic traits even if they don’t meet the threshold for a diagnosis.
• DSM-5 diagnoses of ASD are overall more accurate than DSM-IV diagnoses. A true positive diagnosis is more likely if multiple assessment tools are used in the context of a multi-disciplinary team.
• The changes to DSM-5 ASD criteria likely reduced the frequency of ASD diagnoses, although prevalence continues to rise as a result of other factors.

These findings are provisional and may be altered with further research. Evidence supporting the high heritability of ASD is strong. Evidence is less reliable for prevalence estimates and accuracy of diagnoses. There is significant effort to understand the prevalence of ASD worldwide and to understand the effect of changes to the DSM-5 criteria. However, current studies are often marred by bias, lack of controls and small or unrepresentative samples. That being said, there is wide-spread consensus in the literature around the above findings.

Frequency of ASD diagnoses in families

Estimations of heritability of ASD range from 0.64 – 0.91, with some consensus emerging in the range 0.80 – 0.87 (Bai et al 2020; Sandin et al 2017; Tick et al 2016). High heritability means that for any two people, the more genes they share with each other, the more likely it is that they will share the highly heritable trait (Downes & Matthews, 2020). The closer the genetic relationship between a person with ASD and their relative, the more likely the relative will also have ASD. The literature notes recurrence rates of 80% for identical twins and 20% for non-identical siblings (Bai et al 2020; Girault et al 2020). This is supported by population-based studies showing the likelihood of a person having ASD is increased if they have a family member with ASD (Girault et al 2020; Bai et al, 2020; Hansen et al 2019). One study predicts a 2-fold increase in likelihood of ASD diagnosis if you have a cousin with ASD and an 8-fold increase in likelihood of ASD diagnosis if you have an

Accuracy and inter-rater reliability of ASD diagnoses using DSM-5

While I was able to locate information establishing inter-rater reliability of DSM-5 ASD diagnoses, this should be treated with caution. The results do not come from studies that explicitly set out to study the accuracy of DSM-5 diagnoses. Studies examining other features of ASD or ASD diagnostic practices will often use inter-rater reliability to ensure study quality.

In their study of ASD prevalence, Baio et al found 92.3% inter-rater agreement on presence or absence of ASD using DSM-5 criteria (2018, p.7). Taheri et al secured 100% inter-rater agreement for overall diagnosis and between 70% and 100% agreement on individual criteria (2014, p.118). In their study of gender differences in ASD diagnosis, Hiller, Young and Weber found substantial inter-rater agreement with Cohen’s kappa scores of between 0.75 and 0.93 (2014, pp.4-5). Young and Rodi also secured strong inter-rater agreement for overall diagnosis with Cohen’s kappa score of 0.91 (2014, p.761). These results demonstrate potential for high inter-rater agreement with DSM-5 ASD diagnoses, with somewhat lower agreement in individual criteria. They do not speak to accuracy of severity ratings (i.e., requiring support, requiring substantial support, requiring very substantial support).

Mazurek et al (2019) looked at use of severity ratings among clinicians. They found that assessment of severity levels of social communication and restrictive, repetitive behaviours using DSM-5 criteria largely agrees with other assessment tools as well as parental assessment of severity (p.7). However, they do point out a strong link between intelligence and severity ratings, which may mean that clinicians are conflating ASD symptoms with difficulties related to intellectual disability. Mazurek et al suggest that clinicians may be having difficulty:

“determining whether to assign ratings based on ASD symptom severity alone (more consistent with text examples) or based largely on need for support (more consistent with the level descriptors). If clinicians adhere to the latter interpretation, there may be greater potential for conflation of intellectual and symptom-related impairment. This poses problems for both inter-rater reliability and construct validity” (p.7).

Mazurek et al are also unaware of any studies looking at the inter-rater reliability of severity level assessments (p.8).

Influence of DSM-5 ASD criteria on the prevalence of ASD

Autism prevalence rates are increasing (Taylor et al, 2020; Chiarotti & Venerossi, 2020; CDC 2020). The Autism and Developmental Disabilities Monitoring network (ADDM) estimates prevalence at 1 in 44 in sample United States communities (CDC 2020; Maener et al, 2021; Baio et al, 2018). Autism Spectrum Australia estimates prevalence at 1 in 70 in Australia (Autism Spectrum Australia, 2018). The reasons for the increase are likely to be complex and the exact proportion of the increase that is attributable to different factors is still a matter for debate. Kulage et al suggest:

“parental awareness and acceptance, less stigmatization, better trained clinicians, more thorough data collection methods, and even increasing genetic tendencies could be contributing factors. In addition, comorbid diagnoses are now allowable for ASD under DSM-5, enabling clinicians to give multiple comorbid diagnoses of intellectual disability, ASD, and ADHD, which could also explain why ASD rates have continued to rise since publication of the DSM-5” (Kulage et al, 2019, p.19).

Page 86 of 125

Estimates of ASD Prevalence

Estimates of ASD prevalence are rising despite tightening diagnostic criteria in the current addition of the DSM-5. Since before publication of the DSM-5 there was concern about what the changes to ASD diagnostic criteria would do to ASD prevalence rates and especially whether people who failed to meet the new criteria would no longer be eligible for support (Kulage et al, 2019).

Kulage et al published a systematic review of the literature looking at the effect of the changes to ASD diagnostic criteria between the DSM-IV-TR and the DSM-5. They found that approximately 1 in 5 people who would have received a diagnosis in DSM-IV-TR would not have received a diagnosis in the DSM-V. Further, only 28.8 percent of those who no longer meet ASD criteria would go on to meet diagnostic criteria for Social Communication Disorder (SCD) (Kulage et al, 2019, p.19). This means roughly 14% of people who met diagnostic criteria under DSM-IV no longer meet criteria for ASD or SCD. It is unclear what proportion of those people would go on to meet other diagnostic criteria and what proportion would remain below threshold for any DSM-5 diagnosis.

According to this review, DSM-5 is contributing to a reduction in ASD diagnoses while the overall prevalence estimates continue to rise.

Literature Summary

| Author / Date | Title | Source | Aim / Objective | Methods | Results | Quality of Research | | — | — | — | — | — | — | — | | 3 | Hansen et al 2019 | Recurrence risk of autism in siblings and cousins: a multi-national population-based study | Journal of the American Academy of Child and Adolescent Psychiatry, 58(9): 866-875 | To estimate ASD recurrence risk among siblings and cousins by varying degree of relatedness and by sex | International population-based cohort study of children born 1998–2007. Follow up 2011–2015. Subjects were monitored for an ASD diagnosis in their older siblings or cousins (exposure) and for their own ASD diagnosis (outcome). The relative recurrence | Research found an 8.4-fold increase in the risk of ASD following an older sibling with ASD and a 17.4-fold increase in the risk of Childhood Autism (CA) following an older sibling with CA. A 2-fold increase in the risk for cousin recurrence was observed for both disorders.

Very large sample of almost 9 million children (with 29,998 cases of ASD and 33,769 cases of childhood autism). Measures both shared genetic and non-genetic factors. There was missing parental information in only a small proportion of the sample. Results are robust.

| Author / Date & Title & Source & Aim / Objective & Methods & Results & Quality of Research | — | — | — | — | — | — | — | & & & non-Caucasian population. & Responsiveness Scale-2. & families and iii) spouses) supported previous studies of non-Hispanic populations. & it also i) restricts its conclusions to this population group; and ii) shows how the results for this group coincide with studies of other population groups. | 6 | Tick et al 2016 | Heritability of autism spectrum disorders: a meta-analysis of twin studies | Journal of Child Psychology and Psychiatry 57:5; 585-595 | To assess the evidence of environment and genetic factors in the aetiology of ASD & Systematic review and meta-analysis of all ASD twin studies. & ASD heritability estimates were 64—91%. Shared environmental effects became significant as the prevalence rate decreased from 5–1%: 07–35%. Results are robust. The review contains a meta-analysis of twin studies, which are the standard for heritability studies. Authors have also explained discrepancy between the results of the meta-analysis and previous studies, namely, an over-estimation of the significance of environmental factors was due to some previous studies’ overinclusion of non-identical twins in the samples. | 7 | Frazier et al 2015 | Quantitative autism symptom patterns recapitulate differential & Molecular Autism & To establish the extent to which family transmission pattern & Researchers analysed data from 5515 siblings (2657 non-ASD and 2858 & Non-ASD children manifested elevated ASD symptom burden when they

References

• Autism Spectrum Australia (2018) “Autism prevalence rate up by an estimated 40% to 1 in 70 people.” Autismspectrum.org.au (Accessed: December 9, 2021).

• Bai, D. et al. (2020) “Inherited risk for autism through maternal and paternal lineage,” Biological psychiatry, 88(6), pp. 480–487.

• Baio, J. et al. (2018) “Prevalence of autism spectrum disorder among children aged 8 years - autism and Developmental Disabilities Monitoring Network, 11 sites, United States, 2014,” MMWR Surveillance Summaries, 67(6), pp. 1–23.

• Centre for Disease Control and Prevention (2020) “Autism prevalence studies data table.” Cdc.gov (Accessed: December 9, 2021).

• Chiarotti, F. and Venerosi, A. (2020) “Epidemiology of autism spectrum disorders: A review of worldwide prevalence estimates since 2014,” Brain sciences, 10(5).

• Downes, S. M. and Matthews, L. (2020) “Heritability,” The Stanford Encyclopedia of Philosophy. Spring 2020. Edited by E. N. Zalta. Metaphysics Research Lab, Stanford University.

• Girault, J. B. et al. (2020) “Quantitative trait variation in ASD probands and toddler sibling outcomes at 24 months,” Journal of neurodevelopmental disorders, 12(1), p. 5.

• Hansen, S. N. et al. (2019) “Recurrence risk of autism in siblings and cousins: A multinational, population-based study,” Journal of the American Academy of Child and Adolescent Psychiatry, 58(9), pp. 866–875.

• Hausman-Kedem, M. et al. (2018) “Accuracy of reported community diagnosis of Autism Spectrum Disorder,” Journal of psychopathology and behavioural assessment, 40(3), pp. 367–375. doi: 10.1007/s10862-018-9642-1.

• Hiller, R. M., Young, R. L. and Weber, N. (2014) “Sex differences in autism spectrum disorder based on DSM-5 criteria: evidence from clinician and teacher reporting,” Journal of abnormal child psychology, 42(8), pp. 1381–1393. doi: 10.1007/s10802-014-9881-x.

• Kulage, K. M. et al. (2020) “How has DSM-5 affected autism diagnosis? A 5-year follow-up systematic literature review and meta-analysis,” Journal of autism and developmental disorders, 50(6), pp. 2102–2127.

• Maenner, M. J. et al. (2021) “Prevalence and characteristics of Autism spectrum disorder among children aged 8 years - Autism and Developmental Disabilities Monitoring Network, 11 sites, United States, 2018,” MMWR Surveillance Summaries, 70(11), pp. 1–16.

8. Version control

Page 102 of 125