link to page 1 link to page 1 link to page 2 link to page 2 link to page 3
FOI 23/24- 1400
DOCUMENT 9
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 Atkins
Endorsed by (EL1 or above): Shannon Atkins
Cleared by: Felicity Fallon
1. 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
Page 83 of 125
link to page 4 link to page 6 link to page 19 link to page 20
FOI 23/24- 1400
5.
Influence of DSM-5 ASD criteria on the prevalence of ASD ......................................... 4
6.
Literature Summary ...................................................................................................... 6
7.
References ................................................................................................................. 19
8.
Version control ............................................................................................................ 20
2. 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 smal or unrepresentative samples. That
being said, there is wide-spread consensus in the literature around the above findings.
3. 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
Page 84 of 125
FOI 23/24- 1400
older sibling with ASD (Hansen et al 2019). Girault et al (2020) also note that a sibling is even
more likely to get a diagnosis of ASD if there are multiple people in the family with ASD.
Family members are also more likely to have more autistic traits (short of an ASD diagnosis) if
someone in the family is diagnosed with ASD (Girault et al 2020; Page et al 2016). Girault et al
also notes that a person with ASD getting a higher score on the Social Communication
Questionnaire results in an increased chance of their sibling getting a diagnosis of ASD
(Girault et al 2020).
4. 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).
Page 85 of 125
FOI 23/24- 1400
Hausman-Kedem et al looked at a group of 87 participants who had been diagnosed with ASD
from psychologists or physicians in the community. They had predominately single-disciplinary
diagnoses. Hausman-Kedem et al found the diagnoses did not hold up in 23% of cases when
compared with best practice clinical estimates (2018, p.6). They also find that results of Autism
Diagnostic Observation Schedule-2 (ADOS-2) substantially agrees with final best practice
clinical estimates (2018, p.7). While the support for ADOS-2 is backed up by other studies, the
discrepancy between community diagnoses and best practice clinical estimates is complicated
by the participants’ having DSM-IV diagnoses and the researchers using updated DSM-5
categories.
In their 2018 systematic review, Wigham et al found some support for diagnostic measures
such as ADOS for adults, though they note that accuracy increases when multiple
questionnaires and measures are used. They also observe that difficulties arise when
distinguishing between ASD and some mental health conditions such as schizophrenia (p.15).
While there is better evidence to support tools used to diagnose ASD in children (Whigham,
2018, p.1), Randall et al found reason to be cautious about results supporting accuracy of
diagnostic tools (2018, p.3). According to the evidence obtainable, ADOS scored highest for
sensitivity and all tools assessed had similar results for specificity (p.2).
Further investigation will be required to provide a fuller picture of the overall accuracy of DSM-
5 diagnoses and of tools based on DSM-5 diagnostic criteria. Despite some lack of confidence
in the evidence, there is agreement in the literature that use of a variety of tools from a multi-
disciplinary team gives the highest chance of correctly diagnosing a person with ASD.
5. 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
FOI 23/24- 1400
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.
Page 87 of 125
FOI 23/24- 1400
6. Literature Summary
Author /
Title
Source
Aim / Objective
Methods
Results
Quality of
Date
Research
Inter-rater reliability of DSM-5 ASD Diagnoses
1
Mazurek et al
Factors associated Autism, 23(2):468-
To evaluate the use of
Descriptive
Higher severity
The study is based
2019
with DSM-5 severity 476
these severity ratings
quantitative study of ratings in both
on a large sample
level ratings for
for social
248 children and
domains were
that appears
autism spectrum
communication and
adolescents with
associated with
representative in
disorder
repetitive behaviour
DSM-5 diagnoses.
younger age, lower terms of gender and
domains and to
All participants
intelligence
functioning. No
examine their relation
received a non-
quotient, and
significant bias was
to other measures of
standardized
greater Autism
detected however
severity and clinical
diagnostic clinical
Diagnostic
the clinicians
features.
interview,
Observation
undertaking the
standardized
Schedule–Second
assessments are
observation using
Edition domain-
specialists in ASD
the Autism
specific symptom
diagnosis and may
Diagnostic
severity. Greater
not be
Observation
restricted and
representative of
Schedule–Second
repetitive behavior
clinicians in the
Edition (ADOS-2),
severity was
community.
cognitive
associated with
assessment, and
higher parent-
assessment of
reported
behavioral
stereotyped
functioning.
behaviours.
Participants were
Severity ratings
assessed by a
were not associated
psychologist,
with emotional or
physician or multi-
behavioural
disciplinary team.
problems. Strong
associations
between
Page 88 of 125
FOI 23/24- 1400
Author /
Title
Source
Aim / Objective
Methods
Results
Quality of
Date
Research
intelligence quotient
and DSM-5 severity
ratings in both
domains suggest
that clinicians may
be including
cognitive
functioning in their
overall
determination of
severity.
2
Hausman-
Accuracy of
Journal of
To compare community 87 participants
23% of participants The sample size is
Kedem et al
Reported
Psychopathology
diagnoses of Autism
(85% male, average with a reported
small and males are
2018
Community
and Behaviour
Spectrum Disorder
age 7.4 years), with community
over-represented.
Diagnosis of Autism Assessment. 40(3): (ASD) reported by
reported community diagnosis of ASD
Consensus
Spectrum Disorder
367–375.
parents to consensus
diagnosis of ASD
were classified as
diagnoses made
diagnoses made using
were evaluated
non-spectrum
using DSM-5
standardized tools plus using the Autism
based on our
criteria were
clinical observation.
Diagnostic
consensus
compared to
Observation
diagnosis.
community
Schedule) (ADOS-
Participants
diagnoses using
2), Differential
enrolled with
DSM-IV criteria.
Ability Scale (DAS-
community
Results may reflect
II), and Vineland
diagnosis of PDD-
changes in criteria
Adaptive Behaviour NOS were
as well as
Scales (VABS-II).
significantly more
differences between
Detailed
likely to be
community
developmental and classified as non-
diagnosis and
medical history was spectrum on the
consensus
obtained from all
study consensus
diagnosis.
participants.
diagnosis than
Diagnosis was
Participants with
based on clinical
Autism or Asperger.
consensus of at
This study shows
Page 89 of 125
FOI 23/24- 1400
Author /
Title
Source
Aim / Objective
Methods
Results
Quality of
Date
Research
least two expert
suboptimal
clinicians, using test agreement between
results, clinical
community
observations, and
diagnoses of ASD
parent report.
and consensus
diagnosis using
standardized
instruments.
3
Wigham et al
Psychometric
Autism, 23(2): 287-
Systematic review of
Systematic review
Limited evidence for Design of included
2019
properties of
305
research evidence on
accuracy of
studies were case–
questionnaires and
structured
structured
control, cross
diagnostic
questionnaires and
questionnaires.
sectional or
measures for
diagnostic measures for
Sensitivity and
retrospective,
autism spectrum
adults with Autism
specificity of
making comparison
disorders in adults:
published since 2014.
structured
of results difficult.
A systematic review
questionnaires were Case-control
best for individuals
studies are at risk of
with previously
bias which limits to
confirmed ASD and relevance of the
reduced in
reviews results.
participants referred However, the
for diagnostic
authors point out
assessments, with
that both stronger
discrimination of
and weaker studies
ASD from mental
agreed on the poor
health conditions
psychometric
especially limited.
properties of the
For adults with
tools investigated.
intel ectual
disability, diagnostic
accuracy increased
when a combination
of structured
questionnaires were
Page 90 of 125
FOI 23/24- 1400
Author /
Title
Source
Aim / Objective
Methods
Results
Quality of
Date
Research
used. In mental
health settings, the
use of a single
structured
questionnaire is
unlikely to
accurately identify
adults without
autism spectrum
disorder or
differentiate autism
spectrum disorder
from mental health
conditions.
4
Randall et al
Diagnostic tests for Cochrane Database To identify which
Systematic review
ADOS scored a
Studies reviewed
autism spectrum
of Systematic
diagnostic tools,
summary sensitivity showed some risk
disorder (ASD) in
Reviews
including updated
of 0.94 and a
of bias though
preschool children
versions, most
summary specificity studies at high risk
accurately diagnose
of 0.80. When
of bias were
ASD in preschool
compared with
excluded. Overall,
children when
other assessed
authors advice to
compared with multi-
tools, ADOS scored interpret results with
disciplinary team
highest for
caution due to
clinical judgement. To
sensitivity and all
sample sizes of
identify how the best of
tools had similar
included studies
the interview tools
results for
and potential
compare with CARS,
specificity.
conflicts of interest.
then how CARS
compares with ADOS:
which ASD diagnostic
tool - among ADOS,
ADI-R, CARS, DISCO,
GARS, and 3di - has
the best diagnostic test
Page 91 of 125
FOI 23/24- 1400
Author /
Title
Source
Aim / Objective
Methods
Results
Quality of
Date
Research
accuracy?; is the
diagnostic test
accuracy of any one
test sufficient for that
test to be suitable as a
sole assessment tool
for preschool children?;
is there any
combination of tests
that, if offered in
sequence, would
provide suitable
diagnostic test
accuracy and enhance
test efficiency?; if data
are available, does the
combination of an
interview tool with a
structured observation
test have better
diagnostic test
accuracy (i.e. fewer
false-positives and
fewer false-negatives)
than either test alone?
Frequency of ASD diagnoses in families
1
Bai et al
Inherited Risk for
Biological
Review data on
Quantitative
1.55% of children in The sample is large
Sept 2020
Autism Through
Psychiatry; 88:480– frequency of ASD
correlational study
the cohort were
(847,732 children in
Maternal and
487
among family members using data from the diagnosed with
total and 13,103
Paternal Lineage
Swedish National
ASD. Among their
diagnosed with
Patient Register
maternal /paternal
ASD) and so results
and the Multi-
aunts and uncles
are robust.
Generation Register 0.24% and 0.18%
However the
Page 92 of 125
FOI 23/24- 1400
Author /
Title
Source
Aim / Objective
Methods
Results
Quality of
Date
Research
for a cohort of
were diagnosed
sample is drawn
children born
with ASD,
entirely from
between 2003 and
respectively.
Swedish national
2012. Researchers Offspring of
registers and so
compared
mothers with a
may not be wholly
frequency of ASD
sibling(s) diagnosed applicable to other
diagnosis with
with ASD had
national contexts
family relations and higher rates of ASD (depending on
sex in a group of
than the general
variation in
847,732 children.
population (relative diagnostic habits).
risk, 3.05; 95%
Also, diagnoses are
confidence interval, made using ICD
2.52–3.64). These
8,9, and 10. Results
findings establish a may be different
robust general
using DSM-5
estimate of ASD
diagnoses.
transmission risk for
siblings of
individuals affected
by ASD, the first
ever reported. Our
findings do not
suggest female
protective factors as
the principal
mechanism
underlying the male
sex bias in ASD.
2
Girault et al
Quantitative trait
Journal of
To investigate how
Compared 385
Older siblings’
The study uses a
2020
variation in ASD
Neurodevelopment
quantitative variation
pairs of toddlers
scores on the
substantial sample
probands
al Disorder 12:5
in ASD traits and
and their older
Social
of 385 sibling pairs.
and toddler sibling
broader developmental siblings using data
Communication
However, the study
outcomes at 24
domains in older
from the Infant
Questionnaire
uses DSM-IV to
months
siblings with ASD
Brain Imagining
predicts whether
diagnose toddlers
Page 93 of 125
FOI 23/24- 1400
Author /
Title
Source
Aim / Objective
Methods
Results
Quality of
Date
Research
(probands) may inform
Study. Toddlers and younger siblings will as the IBIS data
outcomes in
older siblings were
receive an ASD
was gathered
their younger siblings.
each assessed
diagnosis. There is
before the DSM-5 a
using age
large variation in
was released. The
appropriate
autistic traits
study did not
diagnostic and
exhibited by
consider if this
adaptive behaviour siblings. However,
would have an
assessment tools to the severity of
impact on results.
determine presence autistic traits in the
Also, while age-
of ASD and autistic older sibling
appropriate clinical
traits.
predicts severity in
tools were used in
the younger sibling. the assessment of
the subjects, the
different tools casts
some doubts on the
comparison
between older and
younger siblings.
3
Hansen et al
Recurrence risk of
Journal of the
To estimate ASD
International
Research found an Very large sample
2019
autism in siblings
American Academy recurrence risk among
population-based
8.4-fold increase in of almost 9 million
and cousins: a
of Child and
siblings and cousins by cohort study of
the risk of ASD
children (with
multi-national
Adolescent
varying degree of
children born 1998– following an older
29,998 cases of
population-based
Psychiatry, 58(9):
relatedness and by
2007. Follow up
sibling with ASD
ASD and 33,769
study
866–875
sex
2011–2015.
and a 17.4-fold
cases of childhood
Subjects were
increase in the risk
autism). Measures
monitored for an
of Childhood Autism both shared genetic
ASD diagnosis in
(CA) following an
and non-genetic
their older siblings
older sibling with
factors. There was
or cousins
CA. A 2-fold
missing parental
(exposure) and for
increase in the risk
information in only a
their own ASD
for cousin
small proportion of
diagnosis
recurrence was
the sample. Results
(outcome). The
observed for both
are robust.
relative recurrence
disorders.
Page 94 of 125
FOI 23/24- 1400
Author /
Title
Source
Aim / Objective
Methods
Results
Quality of
Date
Research
risk was estimated
Researchers also
for different sibling- found a significant
and cousin-pairs,
difference in sibling
for each site
ASD recurrence risk
separately and
by sex.
combined, and by
sex
4
Sandin et al
The Heritability of
Journal of the
To calculate the
Sample of
On one model
The sample is very
2017
Autism Spectrum
American Medical
heritability of ASD by
3,557,446 pairs of
comparing on
large but taken from
Disorder
Association;
reanalysing a previous
siblings was
heritability and non- only Swedish
318(12): 1182-1184 data set.
examined for
shared
sample and so may
presence of ASD.
environmental
not be wholly
Total of 14,516
factors, heritability
applicable to other
children were
was estimated at
national contexts.
diagnosed with
0.87. On a model
Frequency of ASD
ASD. Liability
with all 4 factors,
diagnoses in the
threshold models
heritability was
sample (<0.5%) is
were used to
0.69. Using only
far below incidence
identify additive and twins in the sample, in the general
non-additive genetic heritability was
population (1-2%).
factors, shared and 0.87. The
This study focusses
non-shared
heritability of ASD is on heritability and
environmental
high and the risk of may not reflect
factors.
ASD increased with other familial
increasing genetic
factors.
relatedness.
5
Page et al
Quantitative autistic Molecular Autism
To fill a gap in the
Researchers
Measured
Small sample
2016
trait measurements 7:39
literature by
examined QAT
correlations
relative to these
index background
investigating the
scores in siblings
(between children
types of studies and
genetic risk for ASD
relationship of
and parents of 83
with ASD and i) first while the study
in Hispanic families
quantitative autistic
Hispanic children
degree relative, ii)
depends on a less
traits (QAT) to liability
with ASD, and 64
unaffected first
heterogeneous
of ASD in an example
non-ASD controls,
degree relatives in
sample than other
using the Social
ASD affected
studies (Hispanics),
Page 95 of 125
FOI 23/24- 1400
Author /
Title
Source
Aim / Objective
Methods
Results
Quality of
Date
Research
non-Caucasian
Responsiveness
families and iii)
it also i) restricts its
population.
Scale-2.
spouses) supported conclusions to this
previous studies of
population group;
non-Hispanic
and ii) shows how
populations.
the results for this
group coincide with
studies of other
population groups.
6
Tick et al
Heritability of
Journal of Child
To assess the evidence Systematic review
ASD heritability
Results are robust.
2016
autism spectrum
Psychology and
of environment and
and meta-analysis
estimates were 64–
The review contains
disorders: a meta-
Psychiatry 57:5;
genetic factors in the
of all ASD twin
91%. Shared
a meta-analysis of
analysis of twin
585-595
aetiology of ASD
studies.
environmental
twin studies, which
studies
effects became
are the standard for
significant as the
heritability studies.
prevalence
Authors have also
rate decreased from explained
5–1%: 07–35%.
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
Quantitative autism Molecular Autism
To establish the
Researchers
Non-ASD children
2015
symptom patterns
extent to which family
analysed data from manifested elevated
recapitulate
transmission pattern
5515 siblings (2657 ASD symptom
differential
non-ASD and 2858 burden when they
Page 96 of 125
FOI 23/24- 1400
Author /
Title
Source
Aim / Objective
Methods
Results
Quality of
Date
Research
mechanisms of
and sex modulate ASD ASD). Autism
were members of
genetic
trait aggregation
symptom levels
multiple incidence
transmission in
were measured
families—this effect
single and multiple
using the Social
was accentuated for
incidence families
Responsiveness
male children in
Scale (SRS) and by female ASD-
computing DSM-5
containing
symptom scores
families—or when
based on items
they had a history
from the SRS and
of language delay
Social
with autistic
Communication
qualities of speech.
Questionnaire.
Recurrence risk for
ASD was higher for
children from
female ASD-
containing families
than for children
from male-only
families
Influence of DSM-5 ASD criteria on the prevalence of ASD
1
Kulage et al
How has DSM-5
Journal of Autism
To 1) determine the
Systematic review
Using a random
The study is of high
2019
Affected Autism
and Developmental change in frequency of using PRISMA
effects model, the
quality as a
Diagnosis? A 5-
Disorders
ASD diagnosis in the
guidelines.
pooled proportion
systematic review
Year Follow-Up
first five years after
Qualitative and
suggests a 20.8%
and meta-analysis,
Systematic
publication of the
quantitative meta-
reduction in ASD
although the
Literature Review
revised DSM-5 ASD
analysis of 33
diagnoses. Pooled
underlying data has
and Meta-analysis
criteria; (2) identify the
published articles.
effects suggest
a moderate risk of
DSM-IV-TR autism
statistically
bias stemming from
subtypes most affected
significant
lack of masking of
by the new criteria; and
reductions in ASD
raters to results of
(3) assess the potential
diagnoses of 10.1% the references
of an alternative
for those with AD
standard, DSM-IV-
Page 97 of 125
FOI 23/24- 1400
Author /
Title
Source
Aim / Objective
Methods
Results
Quality of
Date
Research
diagnosis of SCD for
and 23.3% for those TR diagnosis, and
individuals who meet
with Asperger’s
failure to assess
DSM-IV-TR but
Disorder when
interrater
not DSM-5 ASD
DSM-5 criteria were agreement in
diagnostic criteria
applied. The
classification of
reduction in
DSM-5 diagnoses.
diagnoses for PDD- Findings should be
NOS was not
interpreted with
statistically
caution however
significant. Less
this study does
than one-third
represent the most
[28.8%] of those
comprehensive
who met DSM-IV-
exploration of the
TR ASD diagnostic data available.
criteria but not
DSM-5 would meet
SCD diagnostic
criteria.
2
Baio et al
Prevalence of
Centre for Disease
To determine ASD
The first phase
For 2014, the
Sample size is
2018
Autism Spectrum
Control and
prevalence in 11
involves review and overall prevalence
adequate to draw
Disorder Among
Prevention –
communities in the
abstraction of
of ASD among the
conclusion about
Children Aged 8
Morbidity and
United States.
comprehensive
11 ADDM sites was estimated
Years — Autism
Mortality Weekly
evaluations that
16.8 per 1,000 (one prevalence in the
and Developmental Report –
were completed by
in 59) children aged age and
Disabilities
Surveil ance
professional service 8 years.
communities
Monitoring Network, Summaries 67(6)
providers in the
studied. However,
11 Sites, United
community. In the
ADDM study is
States, 2014
second phase of
sometimes used as
the study, all
an estimate of
abstracted
prevalence for the
information is
entire United
reviewed
States. Samples
systematically by
chosen are not
experienced
representative of
Page 98 of 125
FOI 23/24- 1400
Author /
Title
Source
Aim / Objective
Methods
Results
Quality of
Date
Research
clinicians to
the entire US. Did
determine ASD
not specify whether
case status.
raters were aware
of the findings of
other raters.
3
Taheri, Perry,
A Further
Journal on
To determine whether
File review of 22
Only 55% of the
Reassessments
Factor
Examination of the
Developmental
children and
children and
sample met the
were completed
2014
DSM-5 Autism
Disability 20(1)
adolescents diagnosed adolescents
DSM-5 criteria for
using DSM-5
Spectrum Disorder
with Autistic Disorder or previously
ASD; this included
checklist rather than
Criteria in Practice
PDD-NOS on DSM-IV
diagnosed under
69% of those who
clinical diagnoses.
criteria would continue
DSM-IV criteria.
had an original
Children diagnosed
to meet DSM-5 ASD
Records were then
DSM-IV-TR
with Aspergers
criteria. To replicate
reassessed using
diagnosis of AD,
were excluded.
and extend the findings DSM-5 criteria.
and only 17% (one
Small sample size
of an earlier paper in a
child) with an
although study
different sample of
original diagnosis of intentional y worked
older individuals with
PDD-NOS.
as an extension of a
lower cognitive and
previous study with
adaptive skills
an adequate
sample size.
Although masking
of participants
occurred, the study
did not specify
whether raters were
aware of the
findings of other
raters.
4
Hiller, Young
Sex Differences in
Journal of Abnormal To explore sex
Quantitative
While no sex
Adequate sample
and Weber
Autism Spectrum
Child Psychology
differences in the
descriptive study of differences were
size and reported
2014
Disorder based on
behavioural
138 children with
found in the broad
inter-rater reliability
DSM-5 Criteria:
presentation of girls
ASD. Diagnoses
social criteria
between clinicians
Evidence from
and boys diagnosed
were provided by
presented in the
but study did not
two clinicians and
DSM-IV-TR or
specify whether
Page 99 of 125
FOI 23/24- 1400
Author /
Title
Source
Aim / Objective
Methods
Results
Quality of
Date
Research
Clinician and
with high-functioning
then statistical
DSM-5, numerous
raters were aware
Teacher
ASD.
analyses were
differences were
of the findings of
Reporting
applied.
evident in how boys other raters.
and girls came to
meet each criterion.
5
Young and Rodi Redefining Autism
Journal of Autism
To compare overlap of
223 subjects who
Of the 210
Adequate sample
2014
Spectrum Disorder
and Developmental DSM-IV pervasive
were either referred participants in the
size and reported
Using DSM-5: The
Disorders 44:758–
development delay
for a DSM-IV
present study who
inter-rater reliability
Implications of the
765
diagnoses and DSM-5
diagnosis and did
met DSM-IV TR
between clinicians
Proposed DSM-5
autism diagnoses.
not receive one, or
criteria for a PDD
but study did not
Criteria for Autism
who received a
only 57.1 % met
specify whether
Spectrum Disorders
DSM-IV diagnoses
DSM-5 criteria for
raters were aware
were reassessed
autism spectrum
of the findings of
using DSM-5
disorder when
other raters. DSM-5
criteria.
criteria were applied diagnoses were
concurrently during completed by one
diagnostic
or two clinicians
assessment
and so did not meet
best practice
guidelines for
clinical
assessments.
Page 100 of 125
FOI 23/24- 1400
7. References
Autism Spectrum Australia (2018)
Autism prevalence rate up by an estimated 40% to 1 in 70
people. Autismspectrum.org.au. Available at: https://www.autismspectrum.org.au/news/autism-
prevalence-rate-up-by-an-estimated-40-to-1-in-70-people-11-07-2018 (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 Surveil ance Summaries, 67(6), pp. 1–23
Centre for Disease Control and Prevention (2020)
Autism prevalence studies data table,
Cdc.gov. Available at: https://www.cdc.gov/ncbddd/autism/data/autism-data-table.html
(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.
Page 101 of 125
FOI 23/24- 1400
Mazurek, M. O.
et al. (2019) “Factors associated with DSM-5 severity level ratings for autism
spectrum disorder,”
Autism: the international journal of research and practice, 23(2), pp. 468–
476. doi: 10.1177/1362361318755318.
Page, J.
et al. (2016) “Quantitative autistic trait measurements index background genetic risk
for ASD in Hispanic families,”
Molecular autism, 7(1). doi: 10.1186/s13229-016-0100-1.
Randall, M. et al. (2018) “Diagnostic tests for autism spectrum disorder (ASD) in preschool
children,”
Cochrane database of systematic reviews, 7(7), p. CD009044. doi:
10.1002/14651858.CD009044.pub2
Sandin, S.
et al. (2017) “The heritability of autism spectrum disorder,”
JAMA: the journal of the
American Medical Association, 318(12), pp. 1182–1184.
Taheri, A, Perry, A, and Factor, D C. (2014) “A Further Examination of the DSM-5 Autism
Spectrum Disorder Criteria in Practice,”
Journal of Developmental Disabilities, 20(1), pp.116-
121
Taylor, M. J.
et al. (2020) “Etiology of autism spectrum disorders and autistic traits over time,”
JAMA psychiatry (Chicago, Ill.), 77(9), pp. 936–943
Tick, B.
et al. (2016) “Heritability of autism spectrum disorders: a meta-analysis of twin
studies,”
Journal of child psychology and psychiatry, and allied disciplines, 57(5), pp. 585–595.
Wigham, S.
et al. (2019) “Psychometric properties of questionnaires and diagnostic measures
for autism spectrum disorders in adults: A systematic review,”
Autism: the international journal
of research and practice, 23(2), pp. 287–305. doi: 10.1177/1362361317748245.
Young, R. L. and Rodi, M. L. (2014) “Redefining autism spectrum disorder using DSM-5: the
implications of the proposed DSM-5 criteria for autism spectrum disorders,”
Journal of autism
and developmental disorders, 44(4), pp. 758–765. doi: 10.1007/s10803-013-1927-3.
8. Version control
Version Amended
Brief Description of Change
Status
Date
by
0.1
AHR908 Literature review on the incidence and
Draft
10-12-21
reliability of ASD diagnoses using DSM-5
criteria.
1.0
FFM634 Final
Completed
10-12-21
Page 102 of 125
