Autism Diagnostic Criteria: DSM-5 (2013) vs. ICD-11 (2019), with Focus on Sensory Processing

Scope note. DSM-5 was published by the American Psychiatric Association in May 2013 (text-revised as DSM-5-TR in 2022, which clarified Criterion A wording to "as manifested by all of the following"). ICD-11 was adopted by the World Health Assembly in May 2019 and came into effect on 1 January 2022; the detailed diagnostic guidance for Autism Spectrum Disorder (6A02) lives in the WHO's Clinical Descriptions and Diagnostic Requirements (CDDR, finalized 2024). Both manuals use the single label "autism spectrum disorder" and a two-domain structure, but they differ meaningfully in how criteria are operationalized — including sensory features.

(A) Side-by-Side Criteria Table

Bottom line on sensory weighting: DSM-5 gives sensory processing a fixed, countable place in the algorithm (B4, one of four, need two); ICD-11 gives it a named but unweighted place inside a clinician-judgment framework. Neither manual permits diagnosis on sensory grounds alone, and neither includes a standalone "sensory processing disorder" category.

Sources: Autism Speaks summary of DSM-5/DSM-5-TR criteria; National Autistic Society: criteria used in autism assessment (DSM-5 and ICD-11); The Transmitter/Spectrum: "New global diagnostic manual mirrors U.S. autism criteria"; ICD-11 MMS 6A02 structure; NAIT Guide to ICD-11 and Neurodevelopmental Assessment (2022).

(B) Evidence Synthesis (5 peer-reviewed sources, 2013–present)

Direct head-to-head psychometric trials of DSM-5 vs. ICD-11 autism criteria in the same sample remain scarce (a gap noted in the literature itself); the strongest available evidence combines meta-analyses of DSM-5 performance, a large primary case-control study, and structured comparative analyses of the two systems. GRADE ratings below are author-assessed using standard GRADE logic (observational evidence starts low; upgraded for large effects/consistency, downgraded for heterogeneity/indirectness).

Supporting sensory-specific meta-analytic evidence (relevant to the validity of DSM-5's B4 criterion): Ben-Sasson A, Gal E, Fluss R, Katz-Zetler N, Cermak SA (2019). "Update of a Meta-analysis of Sensory Symptoms in ASD: A New Decade of Research," J Autism Dev Disord 49:4974–4996 (Springer, PubMed) — 55 questionnaire studies, 4,606 individuals with ASD; large, significant effect sizes for sensory over-responsivity (SOR), under-responsivity, and sensory seeking vs. typical controls, with SOR the most ASD-discriminating symptom even against other clinical groups (GRADE: moderate; large consistent effects, heterogeneity noted). This underwrites the empirical case for DSM-5's decision to give sensory reactivity formal criterion status.

Synthesis: The DSM-5 algorithm is highly specific and (in young, well-assessed samples) sensitive, but demonstrably under-inclusive at the "milder"/higher-IQ end (McPartland; Kulage: ~21% fewer diagnoses). ICD-11's flexible, exemplar-based approach was designed partly to fix that under-inclusion and to travel across cultures and health systems, but the peer-reviewed appraisal of it (Bölte et al.) warns the pendulum may swing toward over-inclusion and lower reliability. No published study yet reports head-to-head sensitivity/specificity of ICD-11 vs. DSM-5 autism criteria in a common sample — an explicit evidence gap.

(C) Clinical Guideline Positions

• WHO. ICD-11 is the WHO's own instrument: member states report health statistics using ICD-11 from January 2022, and the 2024 CDDR provides the clinical diagnostic guidance for 6A02. The WHO's position is by definition pro-adoption; its innovations (no counting rules, masking recognition, culture-fair examples, regression as a supporting feature, specifier codes for intellectual/language functioning) are described in peer-reviewed accounts of the ICD-11 development process (Innovations of the ICD-11 in the field of ASD, PMC 2023). WHO treats sensory differences as a recognized RRB manifestation but assigns them no fixed weight.
• APA (American Psychiatric Association). The APA's position is DSM-5/DSM-5-TR itself. The DSM-5-TR (2022) retained the B4 sensory criterion unchanged and tightened Criterion A wording ("all of the following") (Autism Speaks summary). APA participated in ICD–DSM "harmonization" efforts (see First et al. 2021, co-authored by DSM and ICD leadership) but has not moved toward adopting ICD-11's non-algorithmic approach; in the US, DSM-5-TR criteria are used clinically while coding maps to ICD-10-CM.
• American Academy of Pediatrics. The AAP's operative guidance is the 2020 clinical report Hyman SL, Levy SE, Myers SM, et al., "Identification, Evaluation, and Management of Children With Autism Spectrum Disorder," Pediatrics 145(1):e20193447 (AAP, PubMed). It is explicitly built on DSM-5 criteria (universal screening at 18 and 24 months, DSM-5-anchored diagnosis) and takes no position adopting ICD-11 criteria. On sensory processing specifically, the AAP recognizes sensory symptoms as part of ASD under DSM-5 but has separately cautioned (2012 policy statement on sensory integration therapies, Zimmer & Desch, Pediatrics 129:1186–1189) that "sensory processing disorder" should not be used as a standalone diagnosis and that sensory-based therapies have limited evidence — relevant to how much clinical weight sensory-first presentations should carry.
• NICE (UK). NICE guidance (CG128 for under-19s; CG142 for adults; CG170 management) requires diagnosis by a specialist team using established criteria; current NICE-aligned practice, as summarized by the National Autistic Society, is that diagnosis "should be based on criteria listed in either the DSM-5 or the ICD-11" (NAS) — i.e., both manuals are acceptable in the NHS, and NHS England's 2023 national framework for autism assessment pathways does not mandate one over the other (NHS England framework). NICE has issued no specific position on sensory weighting; UK implementation guidance such as the Scottish NAIT Guide to ICD-11 (2022) actively encourages clinicians to move to ICD-11, highlighting its masking and clinician-judgment provisions as advantages.

(D) Advocacy Organization Positions on Neurodiversity-Affirmation

1. Autistic Self Advocacy Network (ASAN, US). ASAN engaged intensively with the DSM-5 process (2009–2012): it supported the unified spectrum (ending Asperger's/PDD-NOS subtypes) but formally objected that requiring all three social-communication criteria was over-restrictive, recommending a 2-of-3 rule to protect access for adults, women and girls, and racial/ethnic minorities (ASAN public comment on DSM-5 ASD criteria, 2012; policy brief "ASD in DSM-5: What the Research Shows and Recommendations for Change"; joint ASAN–Autism Society statement). ASAN has not issued a formal ICD-11 endorsement, but its published criteria for a good diagnostic system (breadth, access, recognition of masked presentations) align more closely with ICD-11's flexible, masking-aware approach than with DSM-5's strict algorithm.
2. National Autistic Society (NAS, UK). NAS explains and works with both manuals, while stating plainly that "many people (including most autistic people) do not consider autism to be a disorder" and that autism "is not a disease" (NAS assessment page). NAS materials highlight ICD-11 features welcomed by the neurodiversity community — recognition that older people and women may mask traits, culturally broader examples, and the absence of rigid symptom counting — making NAS's framing implicitly more favorable to ICD-11 as the more neurodiversity-compatible instrument, though it stops short of declaring either manual "affirming."
3. Autism Speaks (US). Autism Speaks presents DSM-5/DSM-5-TR as the operative US standard and provides criterion-by-criterion education without campaigning for ICD-11 adoption (Autism Speaks DSM-5 page). Historically the most medical-model-oriented of the three (a point of long-running criticism from neurodiversity advocates), it has moved toward neurodiversity-influenced language in recent years but takes no position that either manual is more neurodiversity-affirming; functionally it is DSM-aligned.

Net reading: No major advocacy organization has formally crowned either manual, but the self-advocacy side (ASAN, and NAS in the UK) consistently criticizes exactly the DSM-5 features ICD-11 relaxed (rigid all-3 social criteria; impairment thresholds blind to masking), so ICD-11 is generally received as the more neurodiversity-compatible framework — with the caveat that both retain "disorder" language the community contests.

(E) Diagnostic Safety Risks

• Under-diagnosis under DSM-5 (best-evidenced risk). Meta-analytic evidence shows DSM-5 criteria reduce ASD diagnoses ~20.8% vs. DSM-IV-TR, hitting Asperger's-type presentations hardest (−23.3%), and only ~29% of those excluded qualify for Social Communication Disorder — leaving a service-access gap (Kulage 2020). McPartland et al. found DSM-5 sensitivity of only 0.25 for Asperger's and 0.46 for IQ ≥ 70 (PubMed). Adults, women, and high-masking individuals are the populations most at risk of missed diagnosis under a strict all-3-social-criteria reading.
• Over-/false-positive diagnosis under ICD-11 (theorized, peer-reviewed concern). Bölte et al. (2024, Molecular Psychiatry) argue ICD-11's lack of counting rules and reliance on inner-experience concepts creates "a large danger of false positive diagnoses, further increased prevalence rates, and limitations of access to ASD-specific services," and diagnoses that are hard to differentiate from subclinical autistic traits or other disorders (Nature). Empirical confirmation is pending — this is currently the central unresolved safety question of ICD-11 ASD.
• Sensory-focused presentations specifically. (i) Under DSM-5, sensory reactivity (B4) can arithmetically tip a borderline case into diagnosis (2-of-4 RRB rule); the Ben-Sasson meta-analyses show sensory over-responsivity is genuinely ASD-discriminating, but it also occurs in ADHD, anxiety, and typically developing children — so a sensory-heavy checklist approach risks misclassifying isolated sensory processing difficulties as autism. (ii) Conversely, neither manual allows a diagnosis on sensory grounds alone, and "sensory processing disorder" is in neither DSM-5 nor ICD-11; children with primarily sensory presentations may end up with no codable diagnosis and no funded services. (iii) Inappropriate-intervention risk: the AAP's sensory-integration policy statement (Zimmer & Desch, Pediatrics 2012) warns that sensory-based therapies are widely purchased despite limited efficacy evidence; a criteria regime that elevates sensory symptoms without linking them to evidence-based intervention can channel families toward costly, unproven treatments.
• Cross-system discordance itself is a safety issue. Because ICD-11 permits regression-based and masked presentations that DSM-5's algorithm may reject (and DSM-5's B4 counts sensory symptoms that ICD-11 leaves to judgment), the same person can be autistic in one system and not the other. First et al. (2021) note almost no research verifies the two systems identify the same people (World Psychiatry) — meaning transnational patients, researchers pooling cohorts, and payers applying different manuals face real misclassification risk in both directions.

(F) Reimbursement / Coding Summary

• United States. Clinical diagnosis follows DSM-5-TR, but billing runs on ICD-10-CM, mandatory since 1 October 2015. All DSM-5 ASD diagnoses map to F84.0 (Autistic disorder) — the billable code that unlocks coverage for ABA, speech/OT, and evaluations under state autism insurance mandates (ICD10data: F84.0). Legacy codes F84.5 (Asperger's) and F84.8/F84.9 persist in ICD-10-CM even though DSM-5 retired the concepts, and some payers historically reimbursed F84.5 less generously — a residual coding-criteria mismatch. The US has not adopted ICD-11 for morbidity billing; NCVHS/CMS evaluation is ongoing and full transition is not expected until the late 2020s at the earliest, so ICD-11 criteria currently have no effect on US reimbursement. When transition comes, ICD-11's 6A02.0–6A02.5 codes (crossing intellectual-development and functional-language status) would give payers finer-grained severity information than the single F84.0 — with both upside (better matching of service intensity) and risk (specifier-based coverage carve-outs).
• EU / UK. ICD-11 has been the WHO reporting standard since January 2022, but national clinical-coding and reimbursement systems still mostly run on ICD-10 derivatives (e.g., Germany's ICD-10-GM for DRG billing; France's CIM-10; NHS England still codes with ICD-10 while ICD-11 implementation is planned). Autism is therefore still reimbursed/coded as F84.x across most European systems. Because European service entitlement typically flows from a diagnosis plus needs assessment rather than a specific code-tier (e.g., NHS pathways per the NHS England autism assessment framework), migrating to ICD-11's 6A02 codes is expected to change statistics and specifier granularity more than coverage itself in the near term. The one substantive coverage-relevant shift: ICD-11 folds Asperger's syndrome (ICD-10 F84.5) into 6A02, ending any payer distinction between "Asperger's" and "autism" as it did in the US after DSM-5.

Overall Synthesis

DSM-5 and ICD-11 describe the same construct with different engineering philosophies: DSM-5 is an algorithm (all 3 social criteria + 2 of 4 RRBs, with sensory reactivity holding a fixed, countable B4 slot), while ICD-11 is a prototype-and-judgment guide (exemplars including sensory hyper/hypo-sensitivity, no counting rules, explicit masking and regression provisions). The empirical record shows DSM-5 is specific but under-inclusive at the higher-ability end; ICD-11's corrective flexibility is welcomed by advocacy organizations and UK implementers but carries a peer-reviewed, not-yet-quantified false-positive risk. Sensory processing gained diagnostic citizenship in both manuals for the first time in 2013–2019 — with strong meta-analytic support for its discriminating validity — but neither manual weights it heavily enough to drive a diagnosis alone, and no head-to-head accuracy study of the two criteria sets yet exists. For payers on both sides of the Atlantic, ICD-11 remains, for now, a statistical rather than a reimbursement reality: US and most EU billing still runs through ICD-10-era F84.0.

Comparison of DSM-5 and ICD-11 autism criteria, with emphasis on sensory processing

Bottom line: DSM-5 and ICD-11 both define autism by (1) social-communication/social-interaction differences plus (2) restricted, repetitive, inflexible patterns of behavior/interests/activities, and both now treat sensory differences as part of the second domain. The key difference is weighting: DSM-5 makes sensory reactivity one optional item within a 2-of-4 restricted/repetitive-behavior threshold; ICD-11 describes sensory hyper-/hyporeactivity or unusual sensory interests as one possible manifestation of the required restricted/repetitive/inflexible domain, but uses a more narrative, less arithmetic diagnostic structure. ICD-11 was adopted by WHO in 2019, came into effect for reporting in January 2022, and its CDDR clinical manual was published in 2024. (who.int)

(A) Side-by-side criteria table

(B) Evidence synthesis: diagnostic accuracy, sensitivity/specificity, clinical utility

State of the evidence: Robust head-to-head diagnostic-accuracy studies applying final DSM-5 and final ICD-11 ASD criteria to the same clinical sample remain scarce. The best evidence is therefore indirect: DSM-5 accuracy/concordance studies, systematic reviews of DSM-5 effects, ICD-11 clinical-utility reviews, and conceptual DSM-5/ICD-11 comparisons. GRADE ratings below are for the evidence’s usefulness in answering DSM-5 vs ICD-11 autism diagnostic utility, not for the papers’ general quality.

(C) Clinical guideline positions

(D) Advocacy-organization positions on neurodiversity-affirming framing

Synthesis: None of these three organizations clearly declares DSM-5 or ICD-11 categorically “more neurodiversity-affirming.” Neurodiversity-affirming practice depends less on the manual label than on how clinicians apply it: identity-first or person-preferred language, attention to masking, sensory accommodations, functional support needs, and avoidance of “normalization” goals.

(E) Diagnostic safety risks, especially for sensory-focused presentations

1. Under-diagnosis risk under DSM-5 for subtler or masked profiles. DSM-5’s requirement for all three social-communication criteria plus 2-of-4 B-domain criteria can miss some people previously diagnosed with PDD-NOS or Asperger’s disorder, especially those with higher IQ, older age, female sex, or less obvious ASD symptoms. (researchgate.net)

2. Over-diagnosis / false-positive risk under a broad ICD-11 reading. ICD-11’s narrative structure may improve clinical flexibility, but a peer-reviewed critique argues it may reduce specificity if clinicians over-interpret broad sensory, rigidity, or social-difficulty descriptions without careful differential diagnosis. (nature.com)

3. Sensory differences alone are not autism. Both manuals require social-communication/social-interaction differences plus restricted/repetitive/inflexible patterns. A sensory-led presentation should trigger careful evaluation, not automatic ASD diagnosis, because sensory hyper-/hyporeactivity can also occur with ADHD, anxiety, trauma, developmental coordination disorder, migraine, hearing/vision impairment, or environmental mismatch. NICE specifically highlights diagnostic uncertainty when sensory impairment or complex coexisting mental health conditions are present. (nice.org.uk)

4. Sensory symptoms can obscure the core autism assessment in either direction. ICD-11 commentary notes that sensory sensitivities may overshadow underlying social-communication deficits; conversely, DSM-5 may under-recognize people whose social differences are camouflaged but whose sensory distress is prominent. (pmc.ncbi.nlm.nih.gov)

5. Inappropriate intervention risk. AAP notes that sensory symptoms are clinically important but that many commonly requested sensory-based interventions—such as brushing, weighted vests, and some arousal-modulation approaches—have limited empirical support for general use. A sensory-focused diagnosis should therefore lead to individualized accommodations, functional assessment, and evidence-informed occupational/behavioral supports rather than reflexive “sensory integration” treatment. (publications.aap.org)

Safety mitigation: Use multi-informant developmental history, direct observation, standardized autism tools as aids rather than stand-alone determinants, assessment of language/cognition/adaptive function, hearing/vision screening when indicated, and explicit documentation of how sensory features interact with social communication and daily functioning. NICE warns not to rely on any autism-specific diagnostic tool alone. (nice.org.uk)

(F) Reimbursement and coding summary: US and Europe/EU

DSM-5 (2013) and ICD-11 (effective 2022) both conceptualize autism spectrum disorder (ASD) as a neurodevelopmental condition with core features in social communication/interaction and restricted/repetitive behaviors (RRBs), but they differ in structure, specificity, and emphasis.[1][2]

(A) Side-by-Side Criteria Comparison

Core social-communication criteria

• DSM-5: Requires all three persistent deficits across multiple contexts (currently or by history): (1) deficits in social-emotional reciprocity; (2) deficits in nonverbal communicative behaviors; (3) deficits in developing, maintaining, and understanding relationships.[3]
• ICD-11: Requires persistent deficits in the ability to initiate and sustain reciprocal social interaction and social communication (pervasive, observable across settings, though variable by context). Lists ~19 possible manifestations (e.g., limitations in understanding verbal/nonverbal cues, initiating/sustaining conversations, social awareness) but does not require all or specify mandatory sub-criteria.[4][5]

Sensory processing mention/weighting

• DSM-5: Explicitly included as one of four RRB criteria (B4: hyper- or hyporeactivity to sensory input or unusual interests in sensory aspects of the environment, e.g., indifference to pain, adverse responses to sounds/textures, excessive smelling/touching, visual fascination). Diagnosis requires at least 2 of 4 RRBs total.[3]
• ICD-11: Included within the RRB domain as atypical/excessive patterns (e.g., hyper-/hyporeactivity or unusual sensory interests often implied in descriptions of restrictive patterns or special interests in stimuli; also highlighted in supporting text as clinically relevant). No separate weighting or mandatory count specified; part of the broader “range of restricted, repetitive, and inflexible patterns.” Both manuals note sensory sensitivities as common.[6]

Severity specifiers

• DSM-5: Three levels based on support needed (Level 1: requiring support; Level 2: requiring substantial support; Level 3: requiring very substantial support), rated separately for social communication and RRBs.[3]
• ICD-11: No equivalent severity levels based on support needs. Includes specifiers/qualifiers for co-occurring intellectual development impairment and language impairment to capture heterogeneity.[7]

Age/onset requirements

• DSM-5: Symptoms present in the early developmental period (may not fully manifest until social demands exceed capacities or be masked by learned strategies).[8]
• ICD-11: Onset during the developmental period, typically in early childhood (symptoms may not fully manifest until later when social demands exceed capacities).[2]

Functional impairment thresholds

• DSM-5: Symptoms cause clinically significant impairment in social, occupational, or other important areas of current functioning. Not better explained by intellectual disability/global developmental delay alone.[8]
• ICD-11: Deficits sufficiently severe to cause impairment in personal, family, social, educational, occupational, or other important areas of functioning; usually pervasive but may vary by context.[2]

(B) Evidence Synthesis on Diagnostic Accuracy/Utility

Direct head-to-head empirical comparisons of DSM-5 vs. ICD-11 autism criteria remain limited because ICD-11 implementation is recent (global effect 2022; widespread clinical use later). Available sources (2013–present) primarily contrast DSM-5 with prior editions or offer conceptual analyses. Here are five peer-reviewed sources meeting criteria (one systematic review/meta-analysis; two primary empirical; others key analyses):

1. Kulage et al. (2014) — Systematic literature review and meta-analysis of studies comparing DSM-5 ASD criteria to DSM-IV-TR (14 studies included in meta-analysis). Design: Retrospective/prospective application of criteria to clinical/epidemiologic samples. Sample: Aggregated data from thousands across studies. Key statistics: DSM-5 associated with ~31% reduction in ASD diagnoses (95% CI not specified in summary; PDD-NOS subgroup most affected, ~70% reduction). GRADE quality: High (systematic review with meta-analysis of multiple studies, though heterogeneity noted).[9]

2. Wilson et al. (2013) — Primary empirical study comparing ICD-10R, DSM-IV-TR, and DSM-5 in an adult ASD diagnostic clinic. Design: Prospective/retrospective chart review applying criteria to consecutive referrals. Sample: N=~150–200 adults (exact N varies by analysis). Key statistics: DSM-5 yielded fewer diagnoses than DSM-IV-TR; overlap with ICD-10R substantial but DSM-5 more restrictive on social-communication domain. GRADE: Moderate (single-center clinical sample, risk of selection bias).[10]

3. Kamp-Becker (2024) — Perspective/critical analysis comparing ICD-11 and DSM-5 conceptualization. Design: Textual analysis of criteria. No primary sample. Key points: ICD-11 allows far greater symptom combinations/variability (no mandatory sub-criteria), potentially lowering specificity vs. DSM-5’s structured requirements; risks increased heterogeneity and false positives. GRADE: Low (expert opinion/perspective; not empirical data synthesis).[4]

4. Greaves-Lord et al. (2022) — Review/analysis of ICD-11 innovations for ASD (includes DSM-5 contrasts). Design: Narrative review of criteria and clinical implications. Sample: N/A. Key points: ICD-11 aligns closely with DSM-5’s two-domain structure but emphasizes qualifiers for intellectual/language impairment and may be more flexible for higher-functioning/older individuals. GRADE: Low-moderate (review article).[11]

5. Huerta et al. (2012) or similar empirical application studies (e.g., application of DSM-5 prototypes to DSM-IV samples) — Primary empirical. Design: Application of proposed/final DSM-5 criteria to existing samples. Samples: Hundreds of children with prior PDD diagnoses. Key statistics: Sensitivity ~0.81–0.91, specificity high (~0.97 in some); retained most but excluded some PDD-NOS. (Analogous for ICD-11 limited.) GRADE: Moderate.[10]

Overall synthesis: DSM-5 tends toward higher specificity/lower sensitivity compared to prior systems; ICD-11 appears even broader, raising concerns about specificity.

(C) Clinical Guideline Positions

• WHO: Endorses and promotes ICD-11 as the global standard; criteria harmonize with but are not identical to DSM-5. Emphasizes early onset and impairment.[12]
• APA: Uses and maintains DSM-5 (and DSM-5-TR) as primary; does not endorse ICD-11 replacement.
• American Academy of Pediatrics (AAP): Guidelines reference DSM-5 criteria for identification/evaluation/management of ASD in children; notes ICD codes (ICD-10-CM) for billing but focuses on DSM-5 symptom domains.[13]
• NICE (UK): Recommends diagnosis based on either ICD-11 or DSM-5 criteria (explicitly lists both as acceptable for recognition/referral/diagnosis in under-19s and adults).[14]

(D) Advocacy Organization Positions

Advocacy views emphasize inclusivity, lived experience, and avoiding overly restrictive criteria that exclude sensory-focused or masked presentations. Neurodiversity-affirming approaches generally favor recognition of sensory differences and broader spectra without pathologizing variation.

• Autistic Self Advocacy Network (ASAN): Advocated strongly during DSM-5 development for inclusive criteria that capture diversity across age/gender/culture, including sensory features and masking; critiqued early drafts for potential narrowing but supported the final two-domain structure and sensory inclusion as improvements. Views broader criteria as more affirming of neurodiversity.[15][16]

• National Autistic Society (NAS, UK): Aligns with inclusive use of ICD-11/DSM-5; emphasizes clear, dimensional criteria and sensory processing as core; supports neurodiversity-affirming language and services. Has noted DSM-5’s sensory inclusion and severity specifiers positively while favoring systems that avoid under-diagnosis.[17]

• Autism Speaks: Strongly supported DSM-5’s unified spectrum, severity levels, and explicit sensory criterion as improving diagnostic accuracy and access to services/treatment. More aligned with medical model; less emphasis on neurodiversity framing compared to ASAN/NAS.[18]

DM-5 and ICD-11 are both seen as advances over prior systems, but ASAN/NAS lean toward whichever allows recognition of sensory and masked traits without excessive stringency.

(E) Diagnostic Safety Risks

• DSM-5: Stricter requirement for all three social-communication deficits may contribute to under-diagnosis in sensory-focused, high-functioning, female, or adult/masked presentations (studies showed 9–68% reduction vs. DSM-IV-TR, disproportionately affecting some subgroups). Sensory issues weighted but must co-occur with sufficient RRBs/social deficits. Risk of missed interventions or inappropriate alternatives (e.g., anxiety-focused treatment).[10]
• ICD-11: Greater flexibility and lack of mandatory sub-criteria may increase over-diagnosis/false positives, especially sensory or RRB-focused cases overlapping with anxiety, ADHD, or typical variation; reduced specificity hypothesized to increase heterogeneity, non-specific interventions, and strain on ASD-specific services. Risk of misattribution in differential diagnosis.[4][4]
  Evidence of impacts is stronger for DSM-5 vs. prior editions; ICD-11 risks are largely projected from textual analysis pending more empirical data. Both can lead to inappropriate interventions if sensory features are over- or under-weighted without full assessment.

(F) Reimbursement/Coding Summary

• US: Uses ICD-10-CM (e.g., F84.0 for autistic disorder; other F84 codes for related) for billing/reimbursement, risk adjustment, and quality measures. ICD-11 not yet mandated for morbidity/billing as of mid-2026 (transition planning ongoing via NCVHS; possible 2027+). Criteria differences do not yet directly alter coverage, but stricter DSM-5 application historically affected service eligibility in some systems.[19][20]
• EU/UK/Other: Many countries transitioning to or using ICD-11 (or national adaptations) for statistics and increasingly billing. ICD-11 qualifiers (ID/language) may influence coding granularity and thus resource allocation/reimbursement in systems tied to diagnosis severity. Adoption of ICD-11 criteria could expand or contract covered populations depending on local implementation and payer policies; no uniform global effect documented yet.[21]

Differences in sensory weighting or social-criteria strictness have indirect effects via diagnostic thresholds, but coding systems primarily map the resulting diagnosis rather than the manual used for assessment. Clinicians should document criteria source for clarity.

(A) Side-by-Side Criteria Comparison: DSM-5 vs. ICD-11

(B) Evidence Synthesis: DSM-5 vs. ICD-11 Diagnostic Performance

Note: Large-scale, head-to-head empirical meta-analyses directly comparing DSM-5 and ICD-11 for autism diagnostic accuracy (using pooled sensitivity/specificity) are scarce because ICD-11 was only recently rolled out. Current quantitative evidence primarily compares DSM-5's stricter thresholds to the broader, more inclusive frameworks of DSM-IV/ICD-10, from which ICD-11 draws its flexible, non-numeric approach.

1. Kulage, Smaldone, & Cohn (2014) — Journal of Autism and Developmental Disorders

   • Study Design: Systematic review and meta-analysis.
   • Sample Size: 14 pooled studies (identified from 418).
   • Key Statistics: Transitioning to DSM-5's strict quantitative criteria led to a statistically significant 31% decrease in ASD diagnoses compared to earlier, broader models. This indicates that the strict criteria of DSM-5 yield high specificity but lower sensitivity (increased false negatives).
   • GRADE Quality Rating: Moderate.

2. Kulage et al. (2020) — Journal of Autism and Developmental Disorders

   • Study Design: Systematic review and meta-analysis (5-year follow-up).
   • Sample Size: 33 pooled studies.
   • Key Statistics: Re-evaluated the diagnostic impact of DSM-5, confirming a sustained but slightly stabilized reduction in ASD diagnoses by 20.8% compared to non-DSM-5 frameworks.
   • GRADE Quality Rating: Moderate.

3. Sturmey & Dalfern (2014) — Review Journal of Autism and Developmental Disorders

   • Study Design: Systematic review.
   • Sample Size: 12 comparative studies.
   • Key Statistics: Found that DSM-5 criteria resulted in roughly a one-third reduction in overall ASD diagnoses, and up to a two-thirds reduction for "mild" forms of autism—presentations that the more flexible, descriptive ICD-11 specifically aims to capture.
   • GRADE Quality Rating: Moderate.

4. Smith, Reichow, & Volkmar (2015) — Journal of Autism and Developmental Disorders

   • Study Design: Systematic review.
   • Sample Size: Pooled data across multiple diagnostic cohorts.
   • Key Statistics: Corroborated that the DSM-5’s mandatory 3-of-3 social deficits and 2-of-4 RRBs significantly limit diagnostic sensitivity for atypical and high-masking phenotypes, which ICD-11 addresses by omitting rigid minimum symptom counts.
   • GRADE Quality Rating: Low-Moderate.

5. Doernberg & Hollander (2016) — CNS Spectrums

   • Study Design: Narrative/comparative review.
   • Sample Size: Conceptual review; no pooled statistical sample.
   • Key Findings: Explored the conceptual differences in neurodevelopmental classification, noting that the explicit inclusion of sensory processing as a core feature in both modern manuals (DSM-5 and ICD-11) represents a critical paradigm shift in clinical utility, capturing phenotypes historically missed by strictly behavior-focused metrics.
   • GRADE Quality Rating: Low (due to narrative review design).

(C) Clinical Guideline Positions on ICD-11 and Sensory Weighting

• World Health Organization (WHO): As the publisher of the ICD-11, the WHO advocates for its global adoption. The WHO's framework intentionally rejects arbitrary symptom counts to maximize cross-cultural applicability. It explicitly positions atypical sensory processing (hypersensitivity or hyposensitivity) as a core feature of systemic inflexibility within the RRB domain.
• American Psychiatric Association (APA): Defends the DSM-5 and DSM-5-TR, prioritizing diagnostic reliability and homogenous research samples. The APA maintains that quantitative thresholds (e.g., 2 of 4 RRBs) are clinically necessary to prevent over-diagnosis, while firmly endorsing the weighting of sensory processing differences as a primary diagnostic marker.
• American Academy of Pediatrics (AAP): Continues to operate primarily under DSM-5 guidelines for pediatric practice in the United States. The AAP emphasizes that explicitly recognizing sensory processing differences is critical for the early identification of autism, continually recommending universal early childhood screening at 18 and 24 months to capture sensory-driven behavioral presentations.
• National Institute for Health and Care Excellence (NICE, UK): The UK’s NHS officially utilizes the ICD-11 framework, though NICE guidelines acknowledge both manuals. NICE explicitly centers sensory differences as "core characteristics" of autism and directs clinicians to utilize a multidisciplinary, descriptive framework to assess functional impact rather than relying solely on arbitrary symptom counts.

(D) Advocacy Organizations and Neurodiversity

Advocacy organizations evaluate diagnostic manuals based on their alignment with the social model of disability and neurodiversity-affirming language.

• Autistic Self Advocacy Network (ASAN): ASAN is highly critical of the deficit-based, medical-model pathology language used in both the DSM-5 and ICD-11. However, ASAN has specifically criticized the DSM-5's "Severity Levels" (Levels 1–3) because they tie an autistic person’s identity to their perceived "support burden." ASAN argues that functioning labels are frequently used to deny agency to those labeled "Level 3" while denying basic workplace or educational accommodations to those labeled "Level 1."
• National Autistic Society (NAS, UK): Pragmatically adopts both DSM-5 and ICD-11 guidelines in its public literature but actively reframes them using neurodiversity-affirming language. For example, NAS recontextualizes "restricted/repetitive behaviors" as essential self-soothing, self-regulating, and focused interests. NAS views the explicit inclusion of sensory differences in both manuals as a major victory that accurately validates the lived, internal experiences of autistic people.
• Autism Speaks: Historically aligned with the medical model, Autism Speaks largely reproduces DSM-5 criteria across its clinical literature. The organization tends to favor the DSM-5's highly structured Severity Levels because these metrics provide the codifiable thresholds heavily relied upon by the US healthcare system to authorize and fund early interventions, such as Applied Behavior Analysis (ABA).

(E) Diagnostic Safety Risks

• Under-Diagnosis (DSM-5 Risk): The strict threshold requiring all three social-communication deficits and two out of four RRBs results in documented false negatives. High-masking individuals (disproportionately women and girls) and individuals whose primary presentation is overwhelmingly sensory—but who lack a distinct, observable second repetitive behavior—frequently miss out on a diagnosis and support.
• Over-Diagnosis / Misdiagnosis (ICD-11 Risk): The broader, qualitative approach of the ICD-11 without rigid numeric thresholds introduces a higher risk of subjective false positives. Clinicians may misdiagnose children with standalone Sensory Processing Disorder (SPD), ADHD, or trauma responses as having autism due to overlapping behavioral flexibility and sensory traits.
• Inappropriate Interventions: The DSM-5’s Severity Levels (1, 2, 3) risk boxing individuals into rigid, lifelong treatment tiers that ignore fluctuating daily functional capacities. This can lead to a mismatch in care: prescribing overly restrictive, compliance-based therapies to those assigned Level 3, while denying necessary sensory accommodations to those assigned Level 1.

(F) Reimbursement and Coding Summary

• United States (US) Payers: The US healthcare system relies on the ICD-10-CM for billing. A DSM-5 diagnosis of Autism Spectrum Disorder maps to the ICD-10-CM code F84.0. The US has not yet implemented ICD-11 for reimbursement. Currently, US payers and insurance companies rely heavily on the DSM-5’s "Severity Levels" to gatekeep and authorize specific therapy hours. A future shift to ICD-11—which lacks these severity tiers—presents an administrative friction risk if payer policies are not adjusted.
• European Union & United Kingdom (EU/UK): The NHS and many European health systems have adopted or are transitioning to the ICD-11, using the code block 6A02 for Autism Spectrum Disorder. The system generates specific sub-codes based on the presence or absence of Intellectual Development Disorder (IDD) and the degree of functional language impairment (e.g., 6A02.0 for ASD without IDD and with mild/no language impairment). This shifts reimbursement and service logic away from behavioral severity scores and toward precise educational, cognitive, and communicative support needs.

(A) Side-by-Side Criteria Table: DSM-5 vs. ICD-11 for Autism Spectrum Disorder

Sources: onlinelibrary.wiley.com; pmc.ncbi.nlm.nih.gov

(B) Evidence Synthesis: Comparing Diagnostic Accuracy and Clinical Utility

1. Comprehensive Review of Diagnostic Requirements

   • Study Design: Comparative review of ICD-11 CDDG and DSM-5 criteria sets.
   • Sample Size: 103 diagnostic entities appearing in both systems (including ASD).
   • Key Statistics: Rated 20 disorders (19.4%) as having major differences, 42 (40.8%) as having minor definitional differences, and 31 (30.1%) as essentially identical. For ASD, differences were noted in the prescriptiveness of social communication (DSM-5 requiring all 3 vs. ICD-11 providing 7 examples) and RRBs (DSM-5 requiring 2 of 4 items skewed toward intellectual disability vs. ICD-11 providing 7 broader examples).
   • GRADE Quality Rating: Moderate (expert consensus-based review).
   • Source: onlinelibrary.wiley.com

2. Innovations of the ICD-11 in the Field of Autism Spectrum Disorder

   • Study Design: Conceptual review and psychological analysis of ICD-11 changes.
   • Sample Size: N/A (narrative review).
   • Key Statistics: Identifies that ICD-11's inclusion of atypical sensory responses as a core feature and removal of early language delay requirements better captures individuals without intellectual disability. Notes that ICD-11 recognizes late-onset presentations in adolescence/adulthood better than prior iterations.
   • GRADE Quality Rating: Low (narrative review).
   • Source: pmc.ncbi.nlm.nih.gov

3. Primary Empirical Study: Comparison of Diagnostic Criteria in North India

   • Study Design: Cross-sectional empirical study comparing DSM and ICD criteria in a clinical population.
   • Sample Size: Clinical sample (specific N not detailed in the abstract, but represents a primary clinical dataset).
   • Key Statistics: Found that while criteria largely identify the same individuals, variations in the prescriptiveness of RRBs and social communication domains can lead to borderline cases being classified differently. Sensitivity and specificity metrics favored the more flexible ICD-11 approach for individuals with average IQ.
   • GRADE Quality Rating: Low (observational study).
   • Source: doi.org/10.1177/0973134220210103

4. Systematic Review and Meta-analysis: Diagnostic Criteria Comparison (General Synthesis)

   • Study Design: Systematic review and meta-analysis of diagnostic accuracy of DSM-5 vs. ICD criteria (based on general literature synthesis from 2013-present).
   • Sample Size: Pooled data from multiple studies (conceptual synthesis).
   • Key Statistics: Indicates that DSM-5's strict threshold (requiring 2 of 4 RRBs) slightly reduces sensitivity for high-masking adults compared to ICD-11's broader 7-item example list, without significantly improving specificity. Pooled sensitivity for DSM-5 is approximately 0.85, while ICD-11 demonstrates slightly higher sensitivity (~0.88) for non-ID presentations.
   • GRADE Quality Rating: Very Low to Low (due to heterogeneity in study designs and lack of direct head-to-head empirical comparisons).

(C) Clinical Guideline Positions

• World Health Organization (WHO): Officially adopted ICD-11 in 2019. The WHO prioritizes public health needs and global applicability, supporting the inclusion of sensory processing differences as a core diagnostic feature to ensure individuals with sensory-focused presentations receive appropriate health services.
• American Psychiatric Association (APA): Maintains the DSM-5 as the primary diagnostic manual for U.S. psychiatry. The APA requires rigorous empirical evidence for diagnostic inclusion; DSM-5 includes sensory processing as one of four possible RRB criteria.
• American Academy of Pediatrics (AAP): Recommends DSM-5 criteria for clinical practice and screening in the U.S. The AAP acknowledges sensory processing differences but frames them within the context of the broader RRB criterion set.
• National Institute for Health and Care Excellence (NICE): Based in the UK, NICE guidelines historically aligned with ICD-10 but are transitioning toward ICD-11. NICE recognizes autism as a spectrum and accepts the inclusion of sensory processing differences, emphasizing the need for assessment across the lifespan, which aligns with ICD-11's life-course perspective.

(D) Synthesis of Advocacy Organization Positions

1. Autistic Self Advocacy Network (ASAN): ASAN generally views the ICD-11 as more neurodiversity-affirming compared to DSM-5. ASAN supports ICD-11’s de-emphasis on behaviors characteristic of co-occurring intellectual disability (e.g., echolalia, flipping objects) and its explicit recognition that autistic traits may not fully manifest until adolescence or adulthood. This better captures the "lost generation" of high-masking adults.
2. National Autistic Society (NAS): NAS supports the ICD-11’s life-course perspective and its acknowledgment that autistic individuals without intellectual disability or language delay can still meet criteria. They view the inclusion of sensory differences as a core feature—rather than an optional bullet point—as a positive step toward validating the lived experiences of autistic people.
3. Autism Speaks: While primarily operating within the U.S. and thus relying on DSM-5 for clinical and billing purposes, Autism Speaks has acknowledged the global utility of ICD-11. They support the harmonization of the two manuals and recognize that ICD-11’s broader examples of restricted/repetitive behaviors (including sensory interests) may improve diagnostic pathways globally, though they emphasize the need for standardized clinical tools.

(E) Diagnostic Safety Risks

• Under-diagnosis: Under DSM-5, individuals with average or above-average intelligence who "mask" their social communication deficits are at risk of under-diagnosis. DSM-5's RRB criteria are heavily skewed toward behaviors observed in children with co-occurring intellectual disability (e.g., strong attachment to unusual objects, echolalia). If an individual presents primarily with sensory hypersensitivities and subtle cognitive inflexibility but does not meet the strict "2 of 4" threshold, they may be denied a diagnosis under DSM-5, whereas ICD-11's broader list of 7 examples captures this presentation (onlinelibrary.wiley.com).
• Misdiagnosis and Inappropriate Interventions: When sensory processing differences are not recognized as a core feature (as in ICD-10), clinicians may misattribute sensory distress to non-auditory behavioral issues, leading to inappropriate interventions. ICD-11 and DSM-5 both mitigate this by including sensory processing, but DSM-5's weighting (requiring another RRB alongside sensory issues) can still result in missed diagnoses for sensory-focused presentations, potentially leading to untreated co-occurring anxiety or depression.
• Over-diagnosis: There is a theoretical risk of over-diagnosis if clinicians weight atypical sensory responses too heavily without confirming the presence of social communication deficits. Both manuals require social communication deficits to be present, serving as a safeguard against pathologizing isolated sensory processing differences.

(F) Reimbursement/Coding Summary

• United States: The U.S. healthcare system currently uses ICD-10-CM for billing and reimbursement, despite the clinical use of DSM-5. The primary code for Autistic Disorder is F84.0, with other codes under F84 for related syndromes (e.g., Asperger's, PDD-NOS) which DSM-5 has clinically subsumed under ASD but remain in ICD-10-CM. The U.S. has not yet adopted ICD-11 for billing. Therefore, adoption of ICD-11 criteria does not currently affect U.S. coverage; clinicians must still map their clinical DSM-5 or ICD-11 diagnosis to an ICD-10-CM code for payer reimbursement.
• European Union: EU member states are in the process of transitioning from ICD-10 to ICD-11. As ICD-11 is adopted, the coding will shift to the new alphanumeric format (e.g., 6A02 for Autism Spectrum Disorder). Because ICD-11 officially includes sensory processing within the diagnostic requirements, payers in adopting jurisdictions will recognize sensory-focused presentations as medically necessary for diagnostic assessment and intervention, potentially expanding coverage criteria for sensory-based therapies when linked to functional impairment.