The Stepped Care and Personalised Health Model in Autism

Based on: The Lancet Commission on the Future of Care and Clinical Research in Autism (Lord, Charman et al., The Lancet, 2022; 399: 271–334)

1. Overview and Rationale

• Fill the gap between existing clinical knowledge and real-world practice
• Allocate limited resources equitably and efficiently
• Respect individual and family preferences, cultures, and contexts
• Function across both high-income countries (HICs) and low/middle-income countries (LMICs)

2. What Is Stepped Care? Origins and Principles

"A system of treatment delivery and monitoring in which the least resource-intensive service is offered first, and then gradually stepped up to more intensive or specialist-delivered treatments if necessary."

• Least restrictive first: Begin with the lowest cost, most accessible intervention
• Monitor and escalate: Systematically monitor outcomes and escalate intensity only when needed
• Task sharing: Services are provided by the least expensive, most accessible provider where possible, supervised by more highly trained professionals
• Cost containment: Reduces financial burden on health systems and families

• Personal and family costs (time, effort, stress, disruption)
• Individual and family preferences
• The life costs of inappropriate or insufficient treatment
• Participation and autonomy of autistic people and their families

3. The Novel Stepped Care and Personalised Health Model - Core Architecture

1. Personalised medicine approaches - tailored to each individual's profile of strengths, difficulties, co-occurring conditions, preferences, and circumstances
2. Measurement-based care - systematic monitoring of progress using validated tools at each step
3. Shared decision making - involving both autistic individuals and families at every stage
4. Transdiagnostic thinking - addressing autism alongside other neurodevelopmental conditions rather than in diagnostic isolation

3.1 Starting Point: Identification of Needs

1. Safety issues (e.g. a child who wanders, engages in repetitive self-injury, or is at risk in the community) - these always take first priority
2. Individual preferences regarding treatment type (medication vs behavioural, individual vs group)
3. Family circumstances including life events, resources, and capacity to participate
4. Individual characteristics - cognitive and language skills, autism severity, strengths, interests, mental health

4. Figure 5 - The Central Diagram: Stepped Care and Personalised Health Interventions

Figure 5 is the most important diagram in the paper and represents the heart of the model.

Column 1: Assessment of the Individual

• Select one or more priority needs to start, in collaboration with family and patient
• Define the goal of treatment (e.g. improvement or remission)
• Consider individual factors:
  • Age or developmental status
  • Preference for medical vs behavioural or individual vs group strategies
  • Severity of symptoms and adaptive functioning
  • Cognitive and language skills
  • Location of difficulties (at school, at home, with peers)
  • Strengths and interests

Column 2: Family and Contextual Factors

• Preference for medical vs behavioural or individual vs group strategies
• Motivation and ability to participate
• Acceptance of the diagnosis and of specific interventions
• Life events and risks that may affect treatment uptake or effectiveness
• Factors affecting families more broadly

Column 3: Accessibility and Cost - The Steps

"Begin with the least costly approach. Costs include not just economic impact, but burden on family and the person with autism in terms of time, effort, financial cost, and stress."

• Home-based treatment (if easier for family, e.g. Social ABCs, Early Social Interaction)
• School or preschool-based treatment (e.g. JASPER, TEACCH, LEAP, ESDM classrooms)
• Based on personal schedule
• Telehealth (where internet access and comfort permit)

• Treatments at schools or preschools (e.g. TEACCH, JASPER)
• Home-based interventions (e.g. Social ABCs)
• Supported employment programmes in the workplace (e.g. Project SEARCH, Ready, Willing & Able)

"Some travel in the local community required, requires some caregiver effort."

• Treatment in groups (nearby clinic)
• Medication management with regular local physician visits
• Moderate parent-mediated treatment (e.g. PACT, Early Social Interaction - more intensive version)
• Clinic-based cognitive behaviour therapy groups

"Substantial travel required, high family investment of time, restrictedness (inpatient service), intensive hours."

• Highly specialised care requiring considerable travel (tertiary care hospital or clinic)
• Intensive hours at home or in clinic
• Inpatient treatment
• Some naturalistic developmental behavioural interventions
• Parent-Child Interaction Therapy (requires clinic visits by multiple family members)
• Intensive early behavioural programmes (e.g. 20+ hours/week applied behaviour analysis or ESDM)

Key Caveat on Stepping

5. Figure 6 - Sources of Support and Locations of Treatment

Developmental Periods:

• The family role is dominant and central
• Child receives specific short-term therapies
• Community support in childcare/preschool

• Family psychoeducation
• Family coaching around core features
• Parent-mediated treatments (JASPER, Early Social Interaction, PACT)
• Behaviour management (as advocates)

• Specific short-term therapies (JASPER)
• General approaches (PRT, Project ImPACT, DTT)
• Comprehensive curricula (ESDM, Lovaas approach)
• Specific disciplines (occupational therapy, speech therapy)

• Support in childcare and preschool
• ESDM, TEACCH, LEAP classrooms

• Both family and school/community are central; child becomes a more active participant
• School provides the greatest number of intervention hours

• Behaviour management (RUBI - Research Units in Behavioural Intervention)

• General approaches (PRT, DTT)
• Academic skills
• Social skills (PEERS)
• Specific disciplines (psychopharmacology)
• CBT (Coping Cat, Facing your Fears)

• School inclusion
• Classes, special education
• Sports and community programmes

• The adolescent begins to take a more prominent role
• Family transitions to a support/advocacy role
• Community involvement through school inclusion continues

• CBT (Coping Cat, BIACA - Behavioural Interventions for Anxiety in Children with Autism)
• Social skills (PEERS)
• Psychopharmacology
• Academic skills

• School inclusion
• Special education
• Sports and community

• The adult is the primary participant; family serves as advocate
• Community support becomes the main vehicle

• CBT
• Social skills (PEERS)
• Psychopharmacology
• Pre-employment training

• Support in education
• Support in employment
• Support in housing
• Sports and activities
• Project SEARCH, Project ImPACT job training

6. Figure 3 - Influences on the Path of Typical Development

7. Figure 4 - Societal Response and Services Can Optimise Outcomes

• Profound autism likely; long-term support needed (bottom left cluster)
• Mixed outcomes with potential abilities but substantial specific needs (middle)
• Post-secondary education, employment and independence severely limited
• There is an "effective ceiling on opportunities" shown graphically

• The green curve rises moderately
• Mixed outcomes improve
• Some individuals can access post-secondary opportunities

• The green curve reaches its highest level
• Post-secondary education, employment, and independence become possible
• The effective ceiling disappears

8. Figure 8 - Assessment Flow and Stepped Assessment Model

Assessment Levels:

• Conducted at every health visit (immunisation, routine checkups)
• Observe communication, interaction, and behaviour
• Ask if there are any concerns
• Monitor development over time
• Instruments: CREDI, GMCD, ASQ, PEDS, MDAT, TQSI, ITC; Emotional screeners: SDQ, ASEBA; ASD screeners: M-CHAT, PAAS, TIDOS, SCQ, SRS, AQ

• Ask the family open questions about support needs and resources
• Brief assessment of the individual's strengths, challenges, and needs
• Re-evaluate as needed
• Instruments: SDQ with Impact Supplement, WHODAS, ASEBA; More specific: VABS, ABAS, CARS

1. Estimate level of verbal and non-verbal development
   • Brief: WASI, SB5 Routing subtests, KBIT, BINS, INTER-NDA
   • Comprehensive: WPPSI, WISC, WAIS, DAS, RPM, MSEL, Bayley, M-P-R, PEP, RNDA
2. Estimate level of language functioning
   • Brief: CELF screening test, PLS screening, CDI
   • Comprehensive: CELF, PLS, OSEL
3. Assess ASD signs by history and in current daily life
   • Gather information from parents/caregivers, from multiple settings
   • Brief: SRS, SCQ, M-CHAT, AQ, CCC, PAAS, CAST, ASRS, ASSQ, SCDC
   • Comprehensive: ADI-R, DISCO, 3-Di
4. Assess ASD signs by observational assessment
   • Directly observe and interact with the individual
   • Brief: STAT, SORF, AOSI, CARS, BOSCC, AMSE, TIDOS
   • Comprehensive: ADOS-2
5. Estimate level of adaptive functioning
   • Brief: SDQ Impact Supplement, WHODAS
   • Comprehensive: VABS, ABAS
6. Screen for emotional and behavioural problems and stressful life events
   • Query: anxiousness, mood, concentration, hyperactivity, disruptive behaviour, thought problems, eating, sleeping, adverse life events
   • Brief: SDQ, ASEBA, Inter-NDA, ABC, Conners, ECI, CSI, MINI, ACE-Q
   • Comprehensive: PAPA, CAPA, K-SADS, SCID
7. Screen for medical problems
   • Medical history and physical examination as minimum

• Monitoring progress and changes in needs
• Early identification of risk factors and emerging co-occurring disorders
• Timed at points of transition and by indication in between
• Use same brief instruments over time for monitoring
• Stepped assessment as needed

The Stepped Assessment Principle

9. Figure 9 - Probability-Based Approach to Assessment

How It Works:

• Score ≤6 (low): probability increases to 63% alone, or 88% if combined with a positive ADOS-2
• Score >6 (average): probability decreases to 29% alone, or 7% if combined with a negative ADOS-2

• >99% probability → High probability (rule in): Integrate evidence with clinical judgment; assess severity; stepped assessment for co-occurring treatment needs
• Moderate probability → Step up assessment (e.g. more specific, comprehensive instruments)
• <1% probability → Low probability (rule out): Assess alternative treatment needs; re-assess if new concerns arise

10. How the Intervention and Assessment Models Integrate

1. Assessment informs intervention: A brief needs assessment is sufficient to initiate broad, low-cost interventions (Step 1). More detailed diagnostic assessment informs targeted, higher-intensity interventions (Steps 2 and 3).
2. Intervention informs ongoing assessment: Response to treatment determines whether to step up, maintain, or step down the intensity of both assessment and intervention. Measurement-based care - using validated brief tools at each follow-up - is essential to this iterative cycle.

"The result of the absence of this information is a dependence on the clinician, the autistic individual (if possible), and their family to provide the first impetus for a treatment plan... Thus, stepped care and personalised health begins with the identification of family and individual concerns."

11. Figure 2 - Neurobiological and Experiential Influences Across Development

• Prenatal: Autism-specific neurobiological differences
• Early childhood: Anxiety and phobias; ADHD and oppositional/conduct issues
• Adolescence/adulthood: Anxiety and social phobia; depression; obsessive-compulsive disorder

• Atypical social and communicative understanding and behaviour (infancy/early childhood)
• Problems in social understanding and social participation (adolescence)
• Restricted economic integration and achievement (adulthood)
• Pregnancy complications and prematurity
• Limited educational achievement
• Limited vocational or professional training

12. Figure 1 - Developmental Milestones Predictive of Adult Functioning

• Age 2: Repetitive behaviours; delays in expressive language, gross motor skills
• Age 3: Higher ADOS CSS (Calibrated Severity Score)
• Age 4: Delays in fine motor skills
• Age 5: Delayed adaptive skills
• Age 9: High verbal intelligence quotient and overall adaptive skills; strengths in academic skills; limited social skills; scarce peer connections
• Age 18+: Adult outcomes across all domains

13. Applying the Model in Practice: Key Worked Examples

Example 1: Minimally verbal 10-year-old with eating problems

• Priority need (safety/health): Substantial feeding problem
• Step 1 approach: A behaviour programme developed by an expert in feeding difficulties in similar children, who then demonstrates techniques and coaches the parents (home-based, parent-mediated)
• General educational approach: An inclusive school programme with support to foster peer interactions
• If insufficient: Step up to a social skills group

Example 2: 10-year-old with near-typical language and strong academic skills

• Priority need: Anxiety and outbursts related to unpredictability
• Stepped approach: Cognitive behaviour therapy to alleviate anxiety (clinic-based, moderate cost - Step 2)
• This child does not need the intensive feeding intervention example above

Example 3: Adolescent with anxiety

• Clinician assesses: is the goal adapting CBT for autism, or is there a need for autism-specific diagnostic documentation?
• For a known autistic adolescent needing CBT adaptation: a moderate probability threshold for autism diagnosis documentation is sufficient - avoid unnecessary full assessment burden
• For long-term treatment planning and intensive autism-focused behavioural treatment: a high probability threshold is warranted

14. What the Model Requires: System-Level Conditions

14.1 Measurement-Based Care at Every Step

• Points of developmental transition (school entry, adolescence, adulthood)
• Specified intervals between transitions
• Any time concerns arise or needs change

14.2 Shared Decision Making

"Participatory decision making should be incorporated into each step of clinical practice and systems."

• Autistic individuals (where possible)
• Families (especially during childhood)
• Attention to cultural and linguistic preferences at each step

14.3 Psychoeducation at Every Stage

14.4 Task Sharing and Workforce Capacity

• Non-specialist providers (teachers, community health workers, parents) trained and supervised to deliver Step 1 interventions
• Specialists retained for more complex steps and for supervision
• Avoidance of the assumption that more is always better - matching intensity to need

14.5 Integration Across Sectors

15. Summary of Key Messages Regarding the Stepped Care Model

1. Start with needs, not diagnosis. Intervention begins at the point of concern, not after a formal diagnosis is received. A diagnosis clarifies and refines intervention planning but should not be a prerequisite for action.
2. Begin with the least costly approach. Cost means more than money - it includes time, effort, travel, disruption, and family burden. The first step must be the most accessible.
3. Always monitor progress. If progress is not happening, ask why. Step up, step down, or shift approach based on data - not assumptions.
4. Personalise within each step. Two children at the same step may need entirely different interventions. The stepped model is a framework for resource allocation, not a prescription.
5. Co-occurring conditions must be treated. The same stepped and personalised approach applies to anxiety, ADHD, depression, sleep problems, and other conditions - not just autism core features. Treating co-occurring conditions often enables better engagement with autism-specific interventions.
6. Families are partners, not passive recipients. Their preferences, resources, values, and cultural context determine feasibility and effectiveness at every step.
7. The model is transdiagnostic. It applies to autism and other neurodevelopmental disorders. Not having a diagnosis of autism should not be an exclusion criterion for access to an effective intervention.
8. Systems must support the model. Individual clinicians cannot implement stepped care without supportive health-care systems that allow navigation, coordination, continuity, and monitoring. The triple aim of improvement science - better patient-perceived quality, better population health, lower cost per capita - applies here.
9. The model applies globally, with adaptation. In LMICs, the same stepped logic applies, but with adjusted calibration: community-based delivery, task sharing, culturally adapted tools, and recognition that most families receive no formal support at all.
10. Profound autism is not forgotten. The Commission proposes the administrative term "profound autism" for individuals with substantial intellectual disability and minimal functional language, to ensure that stepped care models do not inadvertently serve only more able autistic people - those with the highest support needs must have pathways explicitly designed for them within each step.

Autism Spectrum Disorder (ASD) - A Detailed Overview

1. Definition and Historical Background

"DSM-5 collapses the core diagnostic impairments into two domains: deficits in social communication, and restricted and repetitive behaviors." - Kaplan & Sadock's Synopsis of Psychiatry

2. Epidemiology

• Prevalence: At least 1-2% of the population. Rigorous epidemiological studies estimate prevalence as high as 2.6% for the full spectrum of social disability - far higher than estimated decades ago.
• Global burden: Approximately 78 million people worldwide have autism (Lancet Commission, 2022).
• Sex ratio: ASD occurs predominantly in males, with a 3-4:1 male-to-female ratio in the general population, though the ratio is lower (closer to 2:1) among those with severe intellectual disability. Under-recognition in females may account for a proportion of this difference.
• Intelligence: 31% of children with ASD have intellectual disability (IQ <70), 25% are in the borderline range (IQ 71-85), and 44% have IQ scores in the average to above-average range (IQ >85).
• Age of onset: Must be detectable before 3 years of age by definition, though some children with mild presentations are not identified until adolescence or adulthood.
• Age of diagnosis: Average age 3.1 years (autistic disorder), 3.9 years (PDD-NOS), 7.2 years (former Asperger disorder). Severe language deficits and motor repetitive behaviors lead to earlier identification.
• Universality: ASD occurs in all countries, cultures, and every socioeconomic group.

3. DSM-5 Diagnostic Criteria

Criterion A: Persistent Deficits in Social Communication and Social Interaction

• Abnormal social approach and failure of normal back-and-forth conversation
• Reduced sharing of interests, emotions, or affect
• Failure to initiate or respond to social interactions

• Poorly integrated verbal and nonverbal communication
• Abnormalities in eye contact and body language
• Deficits in understanding and use of gestures
• Total lack of facial expressions and nonverbal communication

• Difficulties adjusting behavior to suit various social contexts
• Difficulties in sharing imaginative play or in making friends
• Absence of interest in peers

Criterion B: Restricted, Repetitive Patterns of Behavior, Interests, or Activities

• Simple motor stereotypies, lining up toys or flipping objects
• Echolalia (immediate or delayed)
• Idiosyncratic phrases

• Extreme distress at small changes
• Difficulties with transitions
• Rigid thinking patterns
• Greeting rituals
• Need to take same route or eat same food every day

• Strong attachment to or preoccupation with unusual objects
• Excessively circumscribed or perseverative interests

• Apparent indifference to pain/temperature
• Adverse responses to specific sounds or textures
• Excessive smelling or touching of objects
• Visual fascination with lights or movement

Criterion C: Symptoms must be present in the early developmental period

Criterion D: Symptoms cause clinically significant impairment in social, occupational, or other important areas of current functioning.

Criterion E: These disturbances are not better explained by intellectual disability or global developmental delay.

DSM-5 Severity Levels (Table 90.1, Bradley & Daroff's Neurology)

DSM-5 Specifiers

• With or without accompanying intellectual impairment
• With or without accompanying language impairment
• Associated with a known medical/genetic condition or environmental factor
• Associated with another neurodevelopmental, mental, or behavioral disorder
• With catatonia

4. Clinical Features

4.1 Social Communication and Interaction

• Infants: May not develop a social smile; may lack anticipatory posture for being picked up; decreased response to name; declining eye gaze (evidenced to begin declining as early as 2-6 months of age in prospective eye-tracking studies).
• Attachment: Atypical but not absent. Children with ASD may not differentiate parents from strangers as expected. May not react strongly to being left with a stranger.
• School age: Social skills may increase, but deficits remain in spontaneous peer play, back-and-forth conversation, shared interests, and use of body/facial gestures.
• Adolescents/adults: Many desire friendships and romantic relationships. Higher-functioning individuals may be aware of their social deficits as a barrier to forming relationships.

"The inability to mentalize, or 'mind blindness,' means individuals with ASD have difficulty inferring the feelings or emotional state of others around them... The lack of a 'theory of mind' produces difficulties interpreting the social behavior of others and leads to a lack of social reciprocation." - Kaplan & Sadock's Synopsis; Eric Kandel, Principles of Neural Science

4.2 Restricted, Repetitive Behaviors

• Play: Restricted exploratory play from the first years of life. Toys used ritualistically rather than symbolically. Less imitative play and abstract pantomime.
• Routines: Insistence on sameness. Changes in routine (furniture rearranged, meal order changed, new route taken) can evoke panic, fear, or temper tantrums.
• Motor stereotypies: Spinning, banging, hand-flapping, toe-walking, rocking - more common with increasing intellectual disability.
• Self-injurious behavior: Higher rates in those with severe intellectual disability.
• Restricted interests: Intense, idiosyncratic preoccupation with narrow range of topics or objects.

4.3 Language

• Up to one-third of people with autism are nonverbal.
• Echolalia (immediate and delayed) - repeating words or phrases heard previously.
• Abnormal prosody - unusual rhythm, stress, or intonation in speech.
• Language regression - in up to 25% of cases, language develops and is subsequently lost (most commonly around 15-18 months).
• Pragmatic language deficits - difficulty with the social use of language even when structural language (grammar, vocabulary) is intact.
• Individuals with conversational language by age 5-6 do significantly better than those without.

4.4 Cognitive Profile

• Children with ASD are frequently more skilled in visual-spatial tasks than in verbal reasoning tasks.
• Some demonstrate "islets of ability" - areas of remarkable talent (savant skills) in domains such as calendar calculation, music, drawing, or mathematical computation.
• Weak central coherence: Tendency to focus on detail rather than integrating information into a global whole (can be a relative strength in some tasks, e.g. embedded figures tests).
• Executive dysfunction: Difficulties with planning, cognitive flexibility, inhibition, and working memory.
• Normal range IQ is a positive prognostic sign.

4.5 Associated Physical Features

• No pathognomonic physical features.
• Higher rates of minor physical anomalies (ear malformations) reflecting abnormalities in fetal development.
• Delayed or absent handedness and lateralization.
• Abnormal dermatoglyphics (fingerprints), suggesting disturbance in neuroectodermal development.

5. Developmental Course and Onset Patterns

1. Early-onset: Atypical development apparent from the first year of life (most common).
2. Regression (setback): Normal or near-normal development for 12-24 months, followed by loss of skills. Developmental regression affects approximately one-third of children with ASD, at an average age of 1.78 years. Language regression alone: ~25%; language + social regression: ~38%. A period of relatively typical development followed by decline in social behaviors beginning around 9 months is actually more common than previously recognized (prospective sibling studies).
3. Late recognition: Core features present from early life but not clinically apparent until social demands increase (e.g. school entry, higher education).

• 12 months: Early social behaviors, use of gestures, response to name, eye contact
• 18 months: Language milestones, joint attention
• 2 years: First words; higher ADOS calibrated severity scores
• 3 years: Phrase speech; non-verbal IQ
• 5 years: Conversational language - a critical prognostic marker
• School age: Executive function, adaptive skills, peer connections

6. Etiology and Pathophysiology

6.1 Genetic Factors

• Twin studies: Concordance ~60-90% in monozygotic twins vs ~0-30% in dizygotic twins.
• The genetic architecture is polygenic (many common variants of small effect), with a smaller contribution from rare variants of large effect.

• Chromosome 16p11.2 deletions/duplications: >10-fold increased ASD risk; 0.5-1% of cases
• Other risk CNVs: 1q21, 15q11-13, 3q29, 22q13 (SHANK3 deletion), 2p16 (NRXN1 deletion), 7q11.23 duplication

• Neuroligins (NLGN3, NLGN4X): Synaptic adhesion molecules; mutations found in ASD
• Neurexins (NRXN1): Pre-synaptic cell adhesion molecules
• SHANK3: Scaffolding protein at postsynaptic density
• SYNGAP1: Ras GTPase-activating protein at synapses

6.2 Neural Circuits Implicated in ASD

• Orbitofrontal cortex (OFC) - social reward processing
• Anterior cingulate cortex (ACC) - social monitoring and motivation
• Amygdala - emotional processing, threat detection, social significance attribution

• Superior temporal sulcus (STS) - biological motion perception, gaze direction, voice processing; impaired in ASD
• Fusiform gyrus (FG) - face processing; hypoactivation in ASD
• Inferior frontal gyrus (IFG) / Broca's area - language production; also associated with mirror neuron system

• Striatum - habit and routine behavior; dysfunction may underlie insistence on sameness and repetitive motor behaviors

• Mirror neuron system dysfunction is hypothesized - these neurons fire both during action execution and observation of the same action, forming a neural basis for imitation and empathy
• Reduced long-range connectivity between brain regions (underconnectivity theory)
• Amygdala hyperactivation during face processing contributes to gaze aversion
• Cerebellar abnormalities (reduced Purkinje cell numbers in some postmortem studies)

6.3 Neuroimaging Findings

• Early brain overgrowth: Accelerated brain volume increase in the first 2 years of life in high-risk infants who later develop ASD.
• Reduced gray matter in the superior temporal sulcus and related social brain regions.
• Functional MRI: Reduced activation of fusiform face area during face processing; reduced STS activation during biological motion and gaze perception; reduced frontal-posterior connectivity.
• No findings are specific or sensitive enough to be used as clinical biomarkers.

6.4 Environmental Factors

• Advanced parental age (both maternal and paternal)
• Preterm birth and perinatal complications
• Prenatal valproate exposure (increases ASD risk ~6-fold)
• In utero immune activation (thalidomide, viral infections during pregnancy)
• Genetic syndromes listed above

7. Co-occurring Conditions

8. Assessment and Diagnosis

8.1 Developmental Surveillance

• No babbling by 12 months
• No gesturing (pointing, waving) by 12 months
• No single words by 16 months
• No two-word spontaneous phrases by 24 months
• Any loss of any language or social skills at any age

8.2 Screening Instruments

• M-CHAT-R/F (Modified Checklist for Autism in Toddlers, Revised with Follow-up) - most widely researched; sensitivity 0.83, specificity 0.51 in meta-analysis
• SCQ (Social Communication Questionnaire)
• SRS (Social Responsiveness Scale)

8.3 Gold-Standard Diagnostic Instruments

• Direct, semi-structured observation of social communication and play
• 4 modules based on developmental/language level
• Calibrated Severity Scores allow comparison across developmental levels
• AUC 0.93 (best single instrument in the MoBa study)

• Structured parent interview covering developmental history
• Assesses social communication, language/communication, and repetitive behaviors
• AUC 0.90

8.4 Comprehensive Assessment Components

1. Detailed developmental history (especially regression, language milestones)
2. Assessment of ASD signs by history (ADI-R, SCQ) and observation (ADOS-2)
3. Cognitive testing (verbal and non-verbal IQ - WISC-5, DAS-II, MSEL for younger children)
4. Language assessment (CELF, PLS)
5. Adaptive behavior assessment (Vineland Adaptive Behavior Scales - VABS)
6. Screening for co-occurring emotional and behavioral problems (SDQ, ASEBA)
7. Medical evaluation (see below)

8.5 Medical Evaluation

9. Management

9.1 Behavioural and Developmental Interventions

• Based on principles of operant conditioning
• Older form: Discrete Trial Training (DTT) - highly structured, instructor-led, intensive (20-40 h/week)
• Modern form: Naturalistic developmental behavioural interventions (NDBIs) - more child-led, embedded in natural routines

• JASPER (Joint Attention, Symbolic Play, Engagement, and Regulation) - targets joint attention and symbolic play
• ESDM (Early Start Denver Model) - comprehensive developmental curriculum, can be parent-delivered or therapist-delivered
• PACT (Preschool Autism Communication Trial) - parent-mediated, focuses on social communication
• Pivotal Response Treatment (PRT) - naturalistic ABA targeting pivotal areas
• Social ABCs - caregiver-mediated, designed for toddlers

• TEACCH - structured teaching using visual supports and physical organization
• PEERS (Program for the Education and Enrichment of Relational Skills) - social skills
• RUBI (Research Units in Behavioural Intervention) - parent training for behavioural problems
• CBT (Coping Cat, Facing Your Fears) - anxiety treatment adapted for ASD

• PEERS - social skills (evidence extends to adults)
• CBT (modified for ASD - accounting for cognitive style, communication, and sensory differences) - for anxiety and depression
• Project SEARCH, Ready Willing & Able - supported employment

9.2 Psychopharmacological Interventions

"Psychopharmacological interventions in autism spectrum disorder help ameliorate behavioral symptoms rather than core features of autism spectrum disorder." - Kaplan & Sadock's Synopsis

9.3 Educational Interventions

• Developed at University of North Carolina, 1970s
• Structured physical environment with visual supports and picture schedules
• Targets academic subjects and socially appropriate responses
• Promotes autonomy through predictability

9.4 Co-occurring Conditions

• Sleep: Parent-mediated behavioural intervention (sleep hygiene, graduated extinction) as first line; melatonin when behavioural approaches insufficient
• Epilepsy: Standard antiepileptic treatment per epilepsy guidelines (no ASD-specific trials)
• ADHD: Psychostimulants (methylphenidate preferred), then atomoxetine, then α-agonists
• Anxiety: Modified CBT (e.g. Coping Cat, BIACA); SSRIs as adjunct
• Gastrointestinal: Dietary assessment, nutritional support; specialist referral for persistent symptoms

10. Prognosis

1. IQ - most important predictor; normal range IQ is the strongest positive prognostic factor
2. Language acquisition by age 5-6 - those with functional conversational language by age 5 have significantly better adult outcomes
3. Early intensive intervention - before age 3 with active family involvement
4. Joint attention skills at age 2-3
5. Adaptive behavior skills at school age

• Poor outcomes in ~78% of cases in older cohort studies
• Only 4 of 120 subjects living independently in one prospective study
• Only 5-15% competitively employed, leading independent lives, in higher-functioning groups
• Psychiatric problems common across the ability spectrum
• In more recent cohort studies with early identification and intervention, outcomes have improved substantially

• Autism is not progressive. Special educational programmes and professional support often lead to improvements in behaviour and adaptive functioning with age.
• Approximately 10% of clearly affected children show sufficient improvement with little or no evidence of social disability later in life ("optimal outcome").
• Adults and the general community increasingly recognize that autistic individuals contribute significantly to society across many domains.

11. Neurodiversity Perspective and Social Context

• Advocates for accommodations rather than "cures"
• Values autistic identity and self-advocacy
• Recognizes that many autistic individuals do not want to be "fixed" but want to be understood and supported
• Calls for greater societal acceptance, inclusion, and structural accommodations

12. Summary of Key Points

Naturalistic Developmental Behavioural Interventions (NDBIs) in ASD

1. What Are NDBIs? Definition and Origins

1. Applied Behaviour Analysis (ABA) - the science of understanding and changing behaviour through systematic manipulation of environmental contingencies (reinforcement, prompting, shaping)
2. Developmental science - child-directed learning, social-emotional development, relationship building, and the recognition that children learn best through intrinsically motivated, meaning-laden interactions

"Other interventions embed ABA principles within naturalistic developmental behavioral approaches. These emphasize the importance of responding to the child's cues, building learning episodes on the child's initiative and choices, establishing emotional attunement and positive relationship between child and adult, and creating learning environments that resemble the 'real world' where targeted skills would naturally be utilized."

• Kaplan & Sadock's Comprehensive Textbook of Psychiatry

2. The Intervention Continuum

"Although these intervention approaches are sometimes depicted as mutually exclusive, they are better characterized as falling along a multidimensional continuum from more structured to more naturalistic, from more adult directed to more child directed." - Kaplan & Sadock's Comprehensive Textbook

3. Defining Characteristics of All NDBIs

3.1 Child-Initiated Learning Opportunities

3.2 Natural Environments and Routines

3.3 Positive Affect and Warm Relationships

3.4 Natural/Intrinsic Reinforcement

3.5 Targeting Developmental Pivotal Skills

• Joint attention (coordinating attention with another person toward a shared external object)
• Symbolic play and imitation
• Social motivation and responsiveness
• Language initiation (spontaneous communication, not just prompted)
• Self-regulation
• Emotional reciprocity

3.6 Manualised and Operationalised Strategies

3.7 Adult as Responsive Partner

4. The Theoretical Foundation: Why NDBIs Target What They Target

4.1 The Transactional Model

4.2 Joint Attention as a Pivotal Target

• JA is the foundation of early word learning (mapping words to referents requires coordinated attention)
• JA is the substrate for social referencing (looking to another to interpret an ambiguous situation)
• Longitudinal data show that JA skills at ages 2-3 predict language and social outcomes years later
• JA has cascading effects: teaching JA leads not just to more JA, but to improvements in language and cognition

"Thus, an intervention can teach particular skills that have immediate value (eg, how to ask for help or how to say no) and support pivotal skills (eg, joint attention and imitation) that contribute to change in processes that have cascading developmental effects on language and cognition." - Lancet Commission on Autism

4.3 Brain Plasticity and the Timing Argument

5. Major NDBI Programmes: Detailed Descriptions

5.1 Early Start Denver Model (ESDM)

• Comprehensive developmental curriculum covering all domains (communication, social, cognitive, motor, adaptive, play)
• Integrates ABA procedures (discrete trial teaching, reinforcement systems) within a relationship-focused, play-based framework
• Targets joint attention, imitation, nonverbal communication, and early language as priority domains
• Uses a joint activity routine structure - adult and child engage in predictable, enjoyable interactions with clear roles, shared attention, and turn-taking
• Parents are trained as co-therapists, implementing ESDM techniques throughout daily routines
• Intensive delivery: typically 20-25 hours/week of direct intervention + parent-mediated hours throughout the day

5.2 JASPER (Joint Attention, Symbolic Play, Engagement, and Regulation)

• Focuses specifically on joint engagement (the state of coordinated attention between child and adult around a shared object or activity), symbolic play, and self-regulation
• Targets the quality of engagement not just quantity of interaction
• Play routines are systematically varied to move through developmental levels: unengaged → onlooker → object play → supported joint engagement → coordinated joint engagement
• Incorporates explicit regulation strategies for children who are dysregulated
• Can be delivered in brief sessions (e.g. 30 min, 2-3x/week) by trained parents or teachers

• Multiple RCTs demonstrating improvements in joint attention, play, language, and generalised social interaction
• Pivotal RCT (Kasari et al., 2014 - Sequential Multiple Assignment Randomised Trial/SMART): Studied optimal sequencing of interventions for minimally verbal children with ASD. Children receiving JASPER + speech-generating device showed superior language outcomes to speech-generating device alone or JASPER alone
• Longitudinal follow-up data show that improvements in joint engagement mediate later improvements in social communication and language (Kasari et al., 2012)
• Community implementation studies show JASPER can be effectively delivered by teachers in school settings with adequate training

5.3 Pivotal Response Treatment (PRT)

• Based on the concept of pivotal behaviours - behaviours that, when changed, produce broad improvements across many untrained behaviours and settings
• Four core pivotal areas targeted:
  1. Motivation (teaching the child to be motivated to communicate and interact)
  2. Responsivity to multiple cues (preventing tunnel focus on one feature of a stimulus)
  3. Self-management (teaching children to monitor and regulate their own behaviour)
  4. Self-initiation (encouraging the child to spontaneously seek social and informational interactions)
• Uses natural reinforcement: the reinforcer is directly related to the child's attempt (e.g. the child says "ball" and receives the ball)
• Interspersed maintenance tasks: mixing new (acquisition) targets with already mastered tasks to maintain motivation and reduce frustration
• Child choice: allowing the child to select materials and activities wherever possible

• Extensive evidence base from single-case designs and RCTs demonstrating improvements in language, social communication, and play
• Randomised controlled trial (Gengoux et al., 2019) in school-aged children showed significant improvements in social communication behaviours
• Neuroimaging prediction: Kaplan & Sadock's Comprehensive Textbook describes fMRI studies (Yang et al.) using machine learning (multivariate pattern analysis) of pretreatment brain activity in social processing circuits (posterior STS, fusiform gyrus, amygdala, orbitofrontal cortex, striatum) to predict individual PRT response - a step toward precision medicine in autism intervention

"Pivotal Response treatment (PRT) is an evidence-based treatment for young children with ASD. Extensive research has shown its efficacy in promoting language acquisition and social communication development in young children." - Kaplan & Sadock's Comprehensive Textbook of Psychiatry

5.4 PACT (Preschool Autism Communication Trial / Parent-mediated Intervention)

• Focus on parent responsiveness and synchrony as the mechanism of change
• Therapist coaches parent to improve the quality of their communication with their child, not to deliver a curriculum
• Uses video feedback: sessions are recorded, and therapist reviews short video clips with the parent to identify moments of synchrony and communication success
• The parent learns to: follow the child's focus of attention, respond to the child's communicative attempts, add language that matches the child's communicative level, and avoid over-directing
• Embeds strategies across all daily routines - meals, dressing, play, bath time
• Manual available; has been adapted and tested in South Asia (PASS Plus) for non-specialist delivery

• Green et al. (2010, Lancet): Significant improvements in parent-child social communication in the PACT group at 13 months
• Pickles et al. (2016, Lancet long-term follow-up): At 6 years post-randomisation, PACT group showed significantly reduced autism severity (ADOS scores) compared to control group - one of the first demonstrations that a psychosocial intervention for autism can show sustained long-term reduction in core symptoms
• Findings on parental synchrony: Changes in parental synchrony during intervention predict child language outcomes, providing mechanistic evidence

5.5 Social ABCs (Caregiver-Mediated Intervention for Toddlers)

• Specifically designed for very young toddlers at the earliest stage of intervention
• Teaches caregivers three core NDBI strategies: animation, following the child's lead, and rewarding with natural reinforcement
• Brief (30 min/week coaching) yet embedded throughout the day
• Cross-site RCT demonstrated improvements in vocalisations, caregiver responsiveness, and child social communication

5.6 Hanen More Than Words Programme

• 8-session group programme teaching parents to become facilitators of their child's communication
• Teaches parents: observe-wait-listen before responding; follow the child's lead; communicate at the child's level; set up communication opportunities
• Video feedback component
• Evidence: Variable results across studies; Hanen MTW showed improvements in parental responsiveness but not always in child language outcomes in RCTs (Carter et al., 2011)

6. Common Active Ingredients Across All NDBIs

7. Evidence Summary: What the Research Shows

7.1 Project AIM - Updated Meta-Analysis (Sandbank et al., BMJ, 2023) - PMID 37963634

• NDBIs showed significant effects on:
  • Adaptive behaviour (g = 0.23, p = 0.03)
  • Language (g = 0.16, p = 0.04)
  • Play (g = 0.19, p = 0.03)
  • Social communication (g = 0.35, p < 0.001) ← strongest and most consistent effect
  • Measures of diagnostic characteristics/autism severity (g = 0.38, p = 0.002)

• NDBIs retained significant effects on social communication (g = 0.36, p < 0.001) and autism diagnostic characteristics (g = 0.44, p = 0.002)

• Only NDBIs on autism diagnostic characteristics remained significant (g = 0.30, p = 0.03)

7.2 Umbrella Review (Gosling et al., Molecular Psychiatry, 2022) - PMID 35790873

• Evidence for NDBIs was rated as "suggestive" (the second highest credibility tier) for improving social communication deficits, global cognitive abilities, and adaptive behaviours in preschool children
• NDBIs were classified alongside early intensive behavioural interventions, developmental interventions, and parent-mediated interventions as having the strongest evidence bases

7.3 NDBIs + AAC for Minimally Verbal Children (Pope et al., J Autism Dev Disord, 2025) - PMID 38848009

• NDBIs had a strong positive impact on language for minimally verbal children with ASD
• Combining AAC (augmentative and alternative communication) within NDBI procedures produced notably larger effect sizes than NDBIs without AAC
• Conclusion: For minimally verbal children, integrating AAC (speech-generating devices, picture exchange systems) with NDBI procedures should be standard practice

7.4 NDBIs and Family Quality of Life (Duncan et al., Autism, 2024) - PMID 38318790

• NDBIs neither improved nor worsened family quality of life
• Despite being described as "family-friendly" and integrated into daily routines, improvements in the child's development did not translate into improvements in overall family QoL
• Clinical implication: Practitioners must explicitly address family wellbeing, parental stress, and carer burden as separate targets, not assume these will improve as a byproduct of child improvement

7.5 Forest Plot Data from the Lancet Commission (Sandbank et al., 2020 - original Project AIM)

• Social communication: g = 0.42 (95% CI 0.23-0.62) - significant
• RCTs only for social communication: g = 0.17 (not significant without quality restriction)
• This gap highlights the risk of bias from unblinded outcome assessment when parents who know their child received intervention report on outcomes

8. Who Delivers NDBIs? Task Sharing and Scalability

8.1 Parent/Caregiver Delivery

• Massively increases the number of intervention hours the child receives (a therapist seeing a child 3x/week provides far fewer opportunities than a parent using NDBI strategies across 10+ hours of daily interaction)
• Promotes generalisation - skills learned in the presence of parents generalise more readily than those learned only with a therapist
• Is more culturally embedded - parents adapt strategies to their family's routines and values
• Is cost-effective and scalable

8.2 Teacher Delivery

8.3 Non-Specialist Community Delivery (LMICs)

• Translation into local languages
• Manualised clinical decision algorithms
• Video feedback using local families' recordings
• Training and supervision cascade with fidelity monitoring

8.4 WHO Caregiver Skills Training

• Non-specialist deliverable
• Evaluated in multiple international sites
• Designed as a first-step (Step 1 in the stepped care model) accessible intervention worldwide

9. NDBIs in the Stepped Care Model

10. Limitations and Ongoing Challenges

10.1 Effect Sizes Are Modest

10.2 Not All Children Respond

10.3 Short-Term Outcome Focus

10.4 Generalisation Remains a Challenge

10.5 Fidelity in Community Implementation

10.6 Family Wellbeing Is a Separate Target

10.7 Adverse Events Undermonitored

11. Summary Table: Core NDBI Programmes

The Stepped Care and Personalised Health Model for Autism

Lord C et al., The Lancet Commission on the Future of Care and Clinical Research in Autism (Lancet 2022; 399: 271-334)

1. Background and Rationale: Why a New Model Was Needed

• At least 78 million people worldwide live with autism, yet formal documentation exists for only a subset of countries
• Most autistic people globally do not receive adequate support or care, particularly in low- and middle-income countries (LMICs)
• Existing frameworks had concentrated on categorical diagnosis and university-based clinic trials rather than on scalable, community-deliverable care
• Interventions existed but lacked guidance on "what works for whom, when, at what intensity, and at what cost" - the fundamental clinical questions remained unanswered
• A single assessment or single treatment is never sufficient given autism's lifelong and heterogeneous nature

"Although autism affects at least 78 million people worldwide, formal documentation of their existence is limited to a subset of countries. Formal documentation through governmental health-care, education, and social care systems for people with autism would be a first step in determining the needs and addressing the potential inequalities faced by these individuals." - Lancet Commission Key Messages

2. Origins of Stepped Care as a Framework

1. Offer the least resource-intensive service first
2. Monitor outcomes against predefined goals
3. Step up to more intensive/specialist treatment only if needed
4. The decision to step up is data-driven, not arbitrary

3. The Commission's Novel Adaptation: Stepped Care + Personalised Health

"We outline a novel, adapted, precision health-integrated, stepped care model for intervention, which includes aspects of personalised medicine approaches and recognises the wide range of strengths, needs, preferences, and circumstances of autistic people and their families worldwide."

• Time and effort demands on the family
• Stress and burden on caregivers
• Cultural acceptability
• Impact on siblings, grandparents, wider household
• Sensory challenges (e.g. transportation costs for children challenged by sensory aspects of travel)

4. Figure 5: The Architecture of the Stepped Care and Personalised Health Interventions

4.1 Personalisation Axis (Left Side of Figure 5)

• Age and developmental status
• Preference for medical vs behavioural vs individual vs group strategies
• Severity of symptoms and adaptive functioning
• Cognitive and language skills
• Location of difficulties (school, home, with peers)
• Strengths and interests

• Safety issues
• Motivation and ability to participate
• Acceptance
• Life events and risks

4.2 Accessibility and Cost Axis (Right Side of Figure 5)

• Home-based delivery (if easier for family)
• Based on personal schedule
• At school
• Via telehealth

• Some travel in local community required
• Requires some caregiver effort
• In groups
• With medication

• Substantial travel required
• High family investment of time
• Restrictedness (inpatient service)
• Intensive hours

5. The Three Categories of Intervention Within the Model

5.1 Building Skills

• Social interaction (shared enjoyment, turn-taking, joint attention)
• Communication (spoken language, comprehension, symbol use, AAC)
• Reading and academic skills
• Adaptive skills (toileting, dressing, community navigation)
• Emotional regulation
• Vocational skills

5.2 Minimising Barriers

• Development of alternative strategies for aggression, tantrums, severe distress
• Treatment of depression and irritability (CBT, medication)
• Management of hyperactivity and inattention (stimulants, behavioural approaches)
• Anxiety treatment (CBT adaptations for ASD)

5.3 Optimising Person-Environment Fit

• Environmental modifications (TEACCH-style visual supports, sensory accommodations)
• Inclusive education programmes with peer support
• Supported employment in the workplace (Project SEARCH, Ready Willing & Able)
• Augmentative and alternative communication systems
• Physical activity and nutritional support

6. Figure 6: Sources of Support by Developmental Period - Shifting Roles Over Time

6.1 Preschool Age (Before 6 Years)

• Family psychoeducation
• Family coaching around core features
• Parent-mediated treatments: JASPER, Early Social Interaction, PACT
• Behaviour management (as advocates)

• Specific short-term therapies (JASPER)
• General approaches: PRT, Project ImPACT, DTT
• Comprehensive curricula: ESDM, Lovaas approach
• Specific disciplines: occupational therapy, speech and language therapy

• Support in childcare and preschool
• ESDM classrooms, TEACCH classrooms, LEAP classrooms

6.2 School Age (6-11 Years)

• Behaviour management (RUBI - Research Units in Behavioural Intervention)

• General approaches: PRT, DTT
• Academic skills
• Social skills: PEERS (Program for the Education and Enrichment of Relational Skills)
• Specific disciplines: psychopharmacology
• CBT: Coping Cat, Facing Your Fears

• School inclusion
• Classes (special education)
• Sports and community programmes

6.3 Adolescence (12-17 Years)

• As supports in group interventions

• CBT (Coping Cat, BIACA - Behavioural Interventions for Anxiety in Children with Autism)
• Social skills: PEERS
• Psychopharmacology
• Academic skills

• School inclusion
• Special education
• Sports and community
• Project SEARCH (supported work experience)

6.4 Adulthood (18 Years and Older)

• As advocates

• CBT
• Social skills: PEERS (adult version)
• Psychopharmacology
• Pre-employment support

• Support in education (higher education)
• Support in employment (supported employment, job coaching)
• Support in housing
• Sports and activities
• Project ImPACT job training
• Project SEARCH

7. The Evidence Base: Figure 7 Forest Plot Data

Behavioural Interventions (primarily EIBI/DTT)

Developmental Interventions

NDBIs

8. The Stepped Assessment Model (Figure 8)

Level 1: Developmental Surveillance

• Developmental screeners: CREDI, GMCD, ASQ, PEDS, MDAT, TQSI, ITC
• Emotional/behavioural screeners: SDQ, ASEBA
• ASD screeners: M-CHAT, PAAS, TIDOS, SCQ, SRS, AQ

Level 2: Brief Needs Assessment

• SDQ with Impact Supplement, WHODAS, ASEBA
• VABS, ABAS, CARS

Level 3: In-Depth (Diagnostic) Assessment

1. Estimate verbal and non-verbal developmental level - at least one verbal + one non-verbal problem-solving test
   • Tools: WPPSI, WISC, WAIS, DAS, RPM, MSEL, Bayley, Merrill-Palmer-Revised, PEP, RNDA; brief: WASI, SB5, KBIT, BINS
2. Estimate language functioning - observe + caregiver report on complexity of speech
   • Tools: CELF, PLS, CDI; full: CELF, CIAT, PRISM, BESA
3. Estimate adaptive skills - daily living, socialisation, communication, motor
   • Tools: VABS, ABAS, Scales of Independent Behavior
4. ASD-specific assessment - at least one standardised instrument from caregiver report AND one from direct observation
   • Gold standard: ADI-R (Autism Diagnostic Interview-Revised) + ADOS-2 (Autism Diagnostic Observation Schedule-2nd edition)
5. Emotional and behavioural screening - co-occurring conditions
   • Tools: ASEBA, SDQ, Conners, SCARED, RCADS
6. Medical evaluation - physical exam, growth, dysmorphic features, hearing, vision; EEG or MRI only if indicated; genetic testing if indicated

The Probabilistic Decision Rule (Figure 9)

• 6% had sufficiently high ADOS-2 scores - indicating confirmed autism without further testing needed
• 80% had very low ADOS-2 scores - potentially reducing the need for further autism-specific testing
• 14% remained in the uncertain middle range - requiring further evaluation

9. Co-occurring Conditions: Integrated Into Every Step

"Integration of co-occurring conditions and core features is essential for the stepped care approach."

• A common first presentation (anxiety, aggression, sleep problems often precede autism identification)
• A major moderator of intervention response
• A source of ongoing need even when autism core features are stable

• Intellectual disability (co-occurring in ~30%)
• ADHD - stimulants, behavioural approaches
• Anxiety and depression - CBT adaptations for ASD (visual/multimodal strategies, explicit problem-solving, incorporating special interests)
• Sleep problems - behavioural sleep programmes, low-dose melatonin
• Eating disorders (especially adolescent girls)
• Suicidal ideation and self-harm (actively screen; risk is higher than general population)
• Obesity and metabolic conditions (nearly 3x increased risk of premature mortality)
• Gender dysphoria (increased prevalence in autism; must not be dismissed as autistic trait)
• Psychotic symptoms in adolescence and adulthood
• Epilepsy

10. Key Implementation Principles

10.1 Task Sharing (Essential for Scalability)

1. Expanding scope: One professional does tasks traditionally allocated to others
2. Delegating downward: Tasks requiring less training are delegated so specialists do the most complex work
3. Professional-to-non-professional task sharing: Parents and community members deliver interventions while experts retain specific responsibilities
4. Broadening skills across the workforce: General practitioners, teachers, and community workers gain enough autism knowledge to deliver first-step interventions

10.2 Measurement-Based Care and Progress Monitoring

• Treatment goals are set at each step
• Progress is measured with validated instruments at specified intervals
• If goals are not being met, the reason is actively sought (is the problem the intervention, the dose, the fidelity, the family capacity, or the environment?)
• Brief caregiver-reported validated measures can supplement clinical assessment between formal reviews

10.3 Psychoeducation at Every Step

• Understanding of autism in general
• Understanding of the specific individual's profile
• Understanding of the evidence base for and against available interventions
• Understanding of potential harms (adverse effects, opportunity costs)
• Support in navigating and accessing local resources
• Help reducing or eliminating ineffective or no-longer-needed services

10.4 Cultural Adaptation Is Not Optional

• Transported models (UK/USA programmes used in other countries) can be effective with proper adaptation processes, ongoing supervision, and fidelity monitoring
• Homegrown models may be needed where contextual differences are profound
• PASS Plus (PACT adapted for South Asia) is the Commission's primary exemplar: a specialist-designed intervention repackaged for non-specialist delivery in a low-resource, culturally distinct setting, maintaining fidelity while adapting language, materials, and clinical decision pathways
• WHO Caregiver Skills Training (CST) is presented as a transdiagnostic first-step tool designed for non-specialist delivery globally, covering children with any developmental delay, not just autism - making it the most scalable possible first step

10.5 Vulnerable Groups Requiring Specific Attention

• Minimally verbal individuals: require AAC alongside NDBIs; specialist communication systems expertise at Step 3
• Females and girls: under-diagnosed due to different phenotypic presentation (less severe RRBs, higher social motivation, "camouflaging"); clinicians must actively compensate
• Minority ethnic groups: less likely to receive timely diagnosis; cultural competence in assessment tools is limited
• Low-SES families: financial and time costs of stepped interventions fall disproportionately on them
• Adults with profound autism: most neglected group; inpatient and supported-living services are almost entirely evidence-free
• Adolescents and adults broadly: majority of the autistic population; fewest services; least studied

11. Economic Dimension

• Lifetime cost of supporting an autistic person with intellectual disability: US$2.4 million (USA) / £2.2 million (UK)
• Lifetime cost of supporting an autistic person without intellectual disability: US$1.4 million (both countries)
• In childhood: special education services and parental productivity loss are the dominant costs
• In adulthood: residential care and lost individual earnings are dominant

12. Research Design Priorities for the Stepped Care Model

12.1 SMART Designs (Sequential Multiple Assignment Randomised Trials)

• Participants are randomised at Step 1
• Those who respond adequately complete the trial
• Those who do not respond are re-randomised at Step 2 to different options
• This generates evidence about optimal sequences of interventions, not just single-point comparisons
• Example from JASPER research: children non-responsive to JASPER alone were randomised to JASPER + speech-generating device vs device alone

12.2 Multiphase Optimisation Strategy (MOST)

• Tests active components or combinations of components
• Can identify the minimal effective package (most scalable) vs full programme
• Useful for hybrid efficacy/implementation designs

12.3 Hybrid Effectiveness-Implementation Designs

• Test both clinical effectiveness AND implementation strategies simultaneously
• Particularly important for bridging the gap between university-based trial results and community real-world outcomes

12.4 Prediction of Treatment Response

• Which baseline features (age, language, cognition, social communication profile, co-occurring conditions) predict response to which intervention, at which step?
• Current evidence: children with better baseline joint attention, imitation, and cognitive skills tend to respond better to NDBIs; children with more severe intellectual disability may respond better to structured behavioural approaches
• Future evidence needed: neuroimaging biomarkers (e.g. fMRI prediction of PRT response via STS/fusiform/amygdala/striatal activity patterns) may eventually enable truly personalised step selection

13. Key Messages (Panel - 10 Actionable Recommendations)

1. Formal documentation of autistic individuals in health, education, and social care systems is the first step toward addressing inequalities
2. Autism requires personalised assessments and intervention strategies. A stepped care and personalised health model to assess and direct interventions can increase effectiveness. Governments and health-care systems must recognise the need for integration across systems across development.
3. A single assessment or single treatment is never sufficient. Follow-up assessments and personalised treatment plans that focus on individual strengths, difficulties, and context changes across the life span are needed.
4. Interventions should begin as soon as signs are noticed, then monitored with comprehensive assessment once begun. No one should wait for months or years to start treatment because they are unable to find an appropriate assessment. However, within a reasonable period of time, assessments do need to be supported and undertaken to identify personalised needs.
5. Focused research strategies should be prioritised with emphasis on clinical practice
6. Co-occurring conditions are integral, not additive - the same standards of diagnosis and treatment apply to people with autism as to others
7. Systems integration is essential - health, education, social care, employment, and housing must be coordinated around the individual
8. Autistic adults are the largest underserved group - adult services need urgent expansion and rigorous evaluation
9. Cultural diversity and LMICs require genuine adaptation, not simple translation; homegrown and hybrid models are legitimate
10. Families and autistic adults should be involved at each step, but should never be expected to assume societal and community responsibilities for individuals who need support

14. Critical Reception and Limitations

• The call for personalised, stepped models of care
• The emphasis on "what works, for whom, when, and at what intensity"
• Recognition of health inequalities and premature mortality
• The integration of LMICs and cultural diversity

• The introduction of "profound autism" as a term was controversial - neurodiversity advocates argued it risks reifying a hierarchy that may entrench different standards of care
• The model's reliance on existing RCT evidence means it inherits the demographic biases of that literature (predominantly white, male, preschool-age, higher-IQ subjects in HICs)
• The personal and societal costs section focuses heavily on economic burden rather than quality of life as defined by autistic people themselves
• Implementation evidence for stepped care in autism specifically remains thin - the model is a logical framework more than an empirically validated care pathway