Journal of Primeasia

1. Introduction

Autism Spectrum Disorder (ASD) has increasingly been recognized not as a single, uniform condition, but rather as a broad and highly heterogeneous neurodevelopmental spectrum characterized by persistent difficulties in social communication, reciprocal interaction, and behavioral flexibility (American Psychiatric Association [APA], 2013; Yu et al., 2020). Although these defining features typically emerge during the early developmental years, their expression varies substantially across individuals, ranging from mild social-communication impairments to profound functional limitations requiring lifelong support. In many cases, the condition extends far beyond diagnostic criteria alone, influencing educational attainment, emotional regulation, family dynamics, adaptive independence, and long-term quality of life (Eapen et al., 2013; Volkmar et al., 2005). Over the last two decades, ASD has gradually shifted from being viewed primarily as a rare childhood disorder toward recognition as a major global public health and developmental concern.

The rising prevalence of ASD has intensified both scientific and clinical attention. Epidemiological estimates now suggest that ASD affects approximately 1 in 59 to 1 in 100 individuals worldwide, although prevalence rates differ across regions depending on diagnostic practices, healthcare access, cultural awareness, and surveillance methodologies (Brugha et al., 2015; Eapen et al., 2013). Some scholars argue that increased diagnostic sensitivity and broader conceptualization of the spectrum partially explain this growth, whereas others suggest that environmental, genetic, and epigenetic contributors may also be involved. Regardless of the exact cause, the increasing number of diagnosed individuals has created significant pressure on healthcare systems, educational infrastructures, and family support networks (Ganz, 2006). The economic burden associated with ASD is particularly substantial, encompassing not only direct medical and therapeutic costs but also indirect consequences such as caregiver stress, reduced workforce participation, specialized education requirements, and long-term social care demands.

At the clinical level, ASD rarely exists in isolation. Beyond the core impairments in communication and social reciprocity, many individuals experience co-occurring behavioral and emotional difficulties that can, at times, become equally disabling. Anxiety disorders, irritability, aggression, hyperactivity, sensory dysregulation, sleep disturbances, and repetitive self-stimulatory behaviors are frequently reported across different age groups (Tarver et al., 2019). These associated conditions often complicate treatment planning and may interfere with educational participation, peer relationships, and adaptive functioning. For families, the unpredictability and chronic nature of these behavioral challenges can contribute to substantial emotional exhaustion and long-term psychological strain. As LeBlanc and Gillis (2012) observed, behavioral difficulties frequently become one of the primary reasons families seek clinical intervention, sometimes even more urgently than deficits in communication itself.

In response to these challenges, behavioral intervention approaches have emerged as the most widely implemented and empirically supported treatment strategies for ASD. Much of this work has been grounded in the principles of Applied Behavior Analysis (ABA), a framework that systematically applies learning theory to improve socially meaningful behaviors through reinforcement, prompting, shaping, and environmental modification (Yu et al., 2020). Historically, ABA-based interventions were developed from the assumption that structured teaching and repeated behavioral practice could enhance communication, reduce maladaptive behaviors, and promote adaptive independence. Over time, these interventions became central to autism treatment programs worldwide, particularly because pharmacological therapies have shown limited effectiveness for addressing the core social and communicative symptoms of ASD (Ospina et al., 2008).

Among behavioral approaches, Early Intensive Behavioral Intervention (EIBI) has perhaps received the greatest research attention. EIBI generally involves highly structured one-to-one therapy delivered intensively—often between 20 and 40 hours per week—during early childhood, particularly before the age of five (Reichow et al., 2012; Smith, 2010). The theoretical rationale behind this model is closely linked to developmental neuroplasticity, which suggests that early intervention may influence the formation and reorganization of neural pathways during critical developmental windows (Eapen et al., 2013). Several systematic reviews and meta-analyses have reported improvements in cognitive performance, language acquisition, adaptive behavior, and social engagement following intensive early intervention programs (Charman et al., 2009; Reichow et al., 2018). Yet, despite these encouraging findings, outcomes remain remarkably variable. Some children demonstrate dramatic developmental gains, while others experience only modest improvements despite receiving similar treatment intensity (Ospina et al., 2008). This inconsistency continues to raise important questions regarding individual responsiveness, intervention timing, treatment fidelity, and the role of family or environmental factors in moderating therapeutic outcomes.

As the field evolved, concerns gradually emerged regarding the rigidity and ecological validity of highly structured intervention models. Critics argued that although intensive behavioral programs could produce measurable improvements in controlled settings, some learned behaviors did not always generalize effectively into natural social environments. Consequently, more naturalistic and developmentally informed models began to gain prominence. Naturalistic Developmental Behavioral Interventions (NDBIs), for instance, attempt to integrate the empirical rigor of behavioral science with developmental theories emphasizing play, social reciprocity, and child-led interaction (Schreibman et al., 2015). Rather than relying exclusively on repetitive structured drills, these interventions embed learning opportunities within everyday social exchanges and daily routines.

One of the most influential examples of this newer generation of interventions is the Early Start Denver Model (ESDM). Developed specifically for toddlers and preschool-aged children, ESDM combines developmental relationship-based strategies with behavioral teaching techniques in a more flexible and socially engaging framework (Dawson et al., 2010). Randomized controlled trials have suggested that ESDM may improve cognitive abilities, receptive language skills, and social engagement among young children with ASD, particularly when implemented early and consistently (Shi et al., 2021). Importantly, these approaches also appear to place greater emphasis on emotional reciprocity and family involvement, recognizing that developmental progress often occurs most meaningfully within natural caregiving environments rather than isolated clinical contexts.

At the same time, intervention priorities shift considerably as individuals with ASD move into adolescence and adulthood. While early childhood interventions primarily target language acquisition and foundational social communication, older individuals often experience heightened emotional distress related to social anxiety, peer rejection, academic pressure, and increasing awareness of interpersonal difficulties. In this context, Cognitive Behavioral Therapy (CBT) has emerged as a promising intervention for managing co-occurring anxiety, depression, and emotional dysregulation among individuals with high-functioning ASD (Sofronoff et al., 2005; Ung et al., 2014). Modified CBT protocols frequently incorporate visual supports, structured routines, and concrete cognitive strategies tailored to the learning profiles associated with ASD. Although evidence supporting CBT continues to grow, questions remain regarding long-term effectiveness, accessibility, and the extent to which conventional CBT models require adaptation for autistic populations.

Despite the substantial expansion of behavioral intervention research, the evidence base remains characterized by considerable methodological and conceptual heterogeneity. Intervention studies vary widely in treatment duration, intensity, outcome measures, participant characteristics, and implementation settings, making direct comparison difficult (Howlin et al., 2009; Ospina et al., 2008). Some studies prioritize cognitive outcomes, whereas others focus on adaptive functioning, emotional regulation, or parental stress reduction. Furthermore, methodological limitations—including small sample sizes, inconsistent diagnostic criteria, lack of blinded assessment, and high attrition rates—continue to affect the reliability and generalizability of findings (Reichow et al., 2012). There is also an ongoing debate regarding what constitutes meaningful clinical improvement in ASD, particularly given the spectrum’s intrinsic variability and the growing neurodiversity perspective, which challenges deficit-oriented interpretations of autism.

Another important gap in the literature concerns lifespan outcomes. The majority of behavioral intervention studies focus heavily on preschool-aged children, often overlooking adolescents, adults, and aging individuals with ASD (Brugha et al., 2015; Wright et al., 2013). As a result, relatively little is known about the sustainability of treatment effects across developmental transitions or about the behavioral support needs of autistic adults navigating employment, higher education, relationships, and independent living. This imbalance has important implications for healthcare planning and service development because ASD is fundamentally a lifelong condition rather than a disorder confined to early childhood.

Given these complexities, there remains a clear need for updated and critically synthesized evidence regarding the effectiveness of behavioral interventions across different age groups and functional domains. The present systematic review therefore aims to evaluate the current evidence surrounding behavioral intervention models for ASD, including structured ABA-based therapies, naturalistic developmental approaches, parent-mediated interventions, and modified CBT strategies. Beyond simply determining whether interventions are effective, this review seeks to examine how intervention characteristics—such as intensity, timing, delivery model, and parental involvement—influence clinical outcomes. Particular attention is also directed toward methodological rigor, risk of bias, and the broader question of how intervention success should be conceptualized within increasingly diverse autistic populations.

Ultimately, understanding “what works, for whom, and under what circumstances” remains one of the most pressing challenges in autism intervention research. By synthesizing existing findings and identifying unresolved gaps, this review aims to contribute to a more balanced and clinically meaningful understanding of behavioral intervention effectiveness in autism spectrum disorder.

2. Methodology

2.1 Study Design and Review Framework

This study was conducted as a systematic review with quantitative narrative synthesis to evaluate the effectiveness of behavioral interventions for autism spectrum disorder (ASD) across cognitive, adaptive, emotional, and family-centered outcome domains. The review methodology followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA 2020) guidelines to ensure transparency, reproducibility, and methodological rigor throughout the review process (Page et al., 2021) as represented in Figure 1. The overall review structure, study selection strategy, evidence synthesis, and reporting procedures were additionally informed by recommendations outlined in the Cochrane Handbook for Systematic Reviews of Interventions (Cumpston et al., 2019).

The review aimed to synthesize evidence regarding major behavioral intervention models for ASD, including Early Intensive Behavioral Intervention (EIBI), Early Start Denver Model (ESDM), Cognitive Behavioral Therapy (CBT), parent-mediated interventions, and other Naturalistic Developmental Behavioral Interventions (NDBIs). Particular attention was directed toward intervention intensity, participant age, therapeutic framework, clinical outcomes, and methodological quality.

2.2 Literature Search Strategy

A comprehensive literature search was conducted across multiple electronic databases, including PubMed, Scopus, Web of Science, PsycINFO, Embase, and Google Scholar. The search strategy combined controlled vocabulary terms and free-text keywords related to autism and behavioral intervention research. Search terms included combinations of: “autism spectrum disorder,” “ASD,” “behavioral intervention,” “Applied Behavior Analysis,” “ABA,” “Early Intensive Behavioral Intervention,” “EIBI,” “Early Start Denver Model,” “ESDM,” “Naturalistic Developmental Behavioral Intervention,” “NDBI,” “Cognitive Behavioral Therapy,” “CBT,” “parent-mediated intervention,” “systematic review,” and “meta-analysis.”

Reference lists from eligible articles and previous reviews were also manually screened to identify additional relevant studies. The literature search focused primarily on peer-reviewed English-language publications involving children, adolescents, or adults diagnosed with ASD.

2.3 Eligibility Criteria

Studies were included if they met the following criteria: (1) involved participants formally diagnosed with Autism Spectrum Disorder; (2) evaluated behavioral, developmental, psychosocial, or parent-mediated interventions; (3) reported measurable clinical, cognitive, adaptive, emotional, or behavioral outcomes; (4) employed randomized controlled trial (RCT), controlled clinical trial (CCT), pre-post intervention, systematic review, or meta-analytic designs; and (5) provided sufficient methodological or outcome-related information for evidence synthesis.

Studies were excluded if they: (1) focused exclusively on pharmacological interventions; (2) lacked primary outcome data; (3) were conference abstracts, editorials, or opinion papers; or (4) were unavailable in English. Duplicate publications and studies with insufficient methodological detail were also excluded from the final synthesis.

2.4 Study Selection and Data Extraction

Study selection was performed in a stepwise manner following PRISMA recommendations (Page et al., 2021). Titles and abstracts were initially screened for relevance, followed by full-text assessment of potentially eligible studies. Disagreements regarding study eligibility were resolved through discussion and consensus-based evaluation.

A standardized data extraction framework was developed to ensure consistency across included studies. Extracted variables included author name, publication year, study design, intervention category, sample size, participant age, intervention duration, treatment intensity, outcome measures, effect sizes, confidence intervals, and methodological quality indicators. Clinical outcomes extracted from eligible studies included intelligence quotient (IQ), adaptive behavior, communication, socialization, anxiety symptoms, disruptive behaviors, and parenting stress.

2.5 Quantitative Synthesis and Effect Size Analysis

Where quantitative synthesis was appropriate, pooled effect estimates were interpreted using standardized mean differences, including Hedges’ g and Cohen’s d. Meta-analytic principles followed the statistical framework outlined by Borenstein et al. (2009), particularly regarding effect size interpretation and pooled estimate synthesis. Random-effects modeling was conceptually adopted because substantial clinical and methodological heterogeneity was anticipated across intervention studies, populations, and outcome measures. The random-effects approach was based on the Pathak et al., (2020) method, which accounts for both within-study and between-study variability in pooled analyses (Pathak et al., 2020).

Inter-study heterogeneity was evaluated using the I² statistic proposed by Higgins et al. (2003). Values exceeding 50% were interpreted as indicating moderate-to-substantial heterogeneity. Forest plot interpretation focused on pooled effect size magnitude, confidence interval width, and consistency across intervention categories. Funnel plots were additionally examined to explore potential publication bias and small-study effects.

2.6 Risk of Bias and Methodological Quality Assessment

Methodological quality and risk of bias were evaluated using information reported within the included systematic reviews and meta-analyses. Particular attention was directed toward randomization procedures, blinding, selection bias, attrition bias, and measurement bias. Reviews reporting predominantly randomized controlled trials with low attrition and standardized outcome assessment were interpreted as methodologically stronger than reviews relying heavily on uncontrolled or observational evidence.

Publication bias was conceptually assessed through visual funnel plot inspection and guided by the principles described by van Enst et al., (2014). However, because several intervention categories included relatively few pooled studies, publication bias findings were interpreted cautiously (van Enst et al., 2014). Overall methodological interpretation was additionally informed by the Cochrane Handbook recommendations regarding evidence certainty, consistency, and risk-of-bias evaluation  (Barcot et al., 2019).

3. Evolution of Behavioral Approaches in Autism Intervention

Autism Spectrum Disorder (ASD) has traditionally been described as a neurodevelopmental condition characterized by persistent challenges in social communication, restricted interests, repetitive behaviors, and difficulties in adaptive functioning. Yet, over the past decade or so, the scientific understanding of autism has become noticeably more nuanced. Researchers increasingly argue that ASD cannot be fully interpreted through diagnostic criteria alone because autistic experiences vary enormously across individuals, developmental stages, family environments, and sociocultural contexts. Autism, in many respects, reflects not simply a collection of behavioral symptoms, but a highly heterogeneous developmental trajectory shaped by biological predisposition, environmental interaction, educational opportunity, and access to supportive care systems (American Psychiatric Association [APA], 2013; Chung et al., 2021).

This broader conceptual shift has substantially influenced the evolution of behavioral interventions. Earlier therapeutic frameworks were largely grounded in deficit-oriented assumptions that emphasized behavioral normalization and symptom reduction. Contemporary approaches, however, increasingly attempt to balance measurable clinical outcomes with emotional well-being, developmental individuality, family quality of life, and meaningful social participation. Rather than focusing

Figure 1: PRISMA 2020 Flow Diagram of Literature Search and Study Selection Process for Behavioral Interventions in Autism Spectrum Disorder. The diagram illustrates the identification, screening, eligibility assessment, and inclusion of studies following PRISMA 2020 guidelines, resulting in nine studies included in the quantitative synthesis

exclusively on eliminating autistic behaviors, many intervention models now seek to promote communication, autonomy, emotional regulation, and adaptive functioning in ways that are developmentally supportive and contextually realistic.

3.1 Early Intensive Behavioral Intervention and the Foundations of ABA

The early history of autism intervention research was dominated by highly structured behavioral therapies derived from Applied Behavior Analysis (ABA). Among these approaches, Early Intensive Behavioral Intervention (EIBI) emerged as one of the most influential and extensively researched treatment models for young autistic children (Howlin et al., 2009; Reichow et al., 2012). EIBI generally involves intensive one-to-one therapeutic instruction delivered for approximately 20–40 hours per week during early childhood, typically before five years of age. The theoretical foundation of this approach is closely tied to the principle of neuroplasticity, which suggests that early developmental periods provide unique opportunities for shaping cognitive, behavioral, and communicative functioning through targeted environmental input (Eapen et al., 2013).

A considerable body of evidence has supported the effectiveness of EIBI, particularly regarding improvements in language development, adaptive behavior, and cognitive performance. Reichow et al. (2012), for example, reported moderate-to-large improvements in intelligence quotient (IQ) and adaptive functioning among children receiving intensive behavioral intervention. Similarly, Howlin et al. (2009) observed that children who began treatment earlier and received greater intervention intensity often demonstrated stronger developmental gains over time. These findings contributed to the widespread perception of EIBI as a “gold standard” intervention within early autism care.

Nevertheless, enthusiasm surrounding EIBI has gradually become more tempered as researchers and clinicians have begun acknowledging its practical and conceptual limitations. Although developmental gains are frequently reported, the intervention itself often demands extraordinary levels of financial investment, caregiver participation, and emotional endurance. For many families—particularly those in under-resourced healthcare systems or low-income settings—the logistical burden associated with intensive therapy can become overwhelming (LeBlanc & Gillis, 2012). Moreover, intervention outcomes remain highly variable. Some autistic children demonstrate substantial improvements in communication and independence, whereas others show only partial or modest progress despite years of intensive support. This inconsistency has prompted growing debate regarding the factors that predict treatment responsiveness and whether current measures of “success” adequately capture meaningful quality-of-life outcomes.

3.2 The Emergence of Naturalistic Developmental Behavioral Interventions

As criticism surrounding rigid behavioral models intensified, intervention research gradually shifted toward approaches that integrate behavioral science with developmental psychology and relationship-based learning. These models, commonly referred to as Naturalistic Developmental Behavioral Interventions (NDBIs), attempt to combine the empirical structure of ABA with child-centered developmental principles emphasizing emotional reciprocity, shared attention, and social engagement (Shi et al., 2021).

Unlike traditional therapist-directed interventions, NDBIs embed learning opportunities within play, daily routines, and naturally occurring social interactions. This transition reflects a broader recognition that children often learn more effectively within emotionally meaningful and socially responsive environments than within highly controlled clinical settings. The emphasis on flexibility, responsiveness, and child-led interaction represents not only a methodological shift, but also a philosophical one. Increasingly, autism intervention research appears less concerned with behavioral conformity and more focused on supporting functional communication, emotional connection, and authentic developmental participation.

Among the most widely recognized NDBIs is the Early Start Denver Model (ESDM), which integrates behavioral teaching strategies with developmental and relational frameworks (Eapen et al., 2013; Shi et al., 2021). ESDM emphasizes imitation, joint attention, communication, and shared engagement within playful, interactive contexts. Emerging evidence suggests that ESDM can produce meaningful gains in cognitive and verbal functioning, particularly when intervention begins early in development. Shi et al. (2021) reported especially strong improvements in cognitive performance among children receiving comprehensive naturalistic interventions.

Importantly, the appeal of ESDM extends beyond clinical outcomes alone. Because intervention can be incorporated into family routines, childcare settings, and community-based environments, the model is often perceived as more accessible and sustainable than highly intensive clinic-centered programs (Eapen et al., 2013). In many ways, the increasing popularity of naturalistic approaches reflects the evolving priorities of autism care itself—priorities that increasingly value emotional well-being, relational development, and practical family integration alongside measurable behavioral improvement.

3.3 Cognitive Behavioral Therapy and Emotional Well-Being in Adolescence

While early childhood interventions remain central within ASD research, adolescence introduces a different constellation of developmental and psychological challenges. Anxiety disorders are particularly common among autistic adolescents and young adults and often contribute to social withdrawal, emotional dysregulation, academic stress, and reduced quality of life (Kreslins et al., 2015). As awareness of these co-occurring mental health concerns has grown, Cognitive Behavioral Therapy (CBT) has emerged as one of the most extensively studied psychosocial interventions for autistic individuals experiencing anxiety-related difficulties.

Modified CBT protocols for ASD typically incorporate structured routines, visual supports, concrete language, and explicit emotional recognition exercises tailored to the cognitive and communicative profiles commonly associated with autism (Weston et al., 2016). Meta-analytic findings generally indicate that CBT can significantly reduce anxiety symptoms among autistic youth, especially when interventions are individualized and developmentally adapted (Kreslins et al., 2015; Weston et al., 2016).

However, the literature also highlights an intriguing discrepancy between external evaluations of improvement and the self-reported experiences of autistic individuals themselves. Parents and clinicians often describe noticeable reductions in anxiety-related behaviors following intervention, whereas self-reported emotional improvements are sometimes smaller or statistically insignificant. This divergence raises important methodological and conceptual questions regarding emotional self-awareness, alexithymia, communication barriers, and the adequacy of existing outcome measures for capturing internal emotional experiences in autism.

3.4 Parent-Mediated Interventions and Family-Centered Care

Another increasingly significant development within behavioral intervention research involves the rise of parent-mediated interventions (PMIs). Earlier therapeutic models often positioned caregivers as passive observers while professionals delivered treatment. More recent evidence, however, increasingly recognizes parents as essential therapeutic partners capable of influencing communication, emotional regulation, and behavioral consistency within the child’s everyday environment (Tarver et al., 2019).

PMIs generally involve training caregivers in reinforcement techniques, communication facilitation strategies, emotional support methods, and behavioral management approaches that can be integrated into ordinary daily interactions. Research suggests that these interventions may generate benefits extending beyond child behavior alone. Tarver et al. (2019) reported moderate reductions in disruptive behaviors alongside measurable improvements in caregiver stress and family functioning. Similarly, Twum-Antwi et al. (2020) emphasized that empowering parents often strengthens parent–child interaction, improves intervention consistency, and enhances long-term family resilience (Twum-Antwi et al., 2020).

This family-centered perspective is particularly important because autism-related challenges rarely affect only the diagnosed individual. Emotional strain, financial burden, and caregiving demand often influence the psychological stability and daily functioning of the entire household. Consequently, interventions that support family adaptation may ultimately produce broader and more sustainable developmental benefits than child-focused approaches alone.

3.5 Persistent Gaps and the Need for Lifespan-Oriented Research

Despite substantial advances in behavioral intervention science, several important gaps remain evident throughout the literature. Perhaps most notably, autism intervention research continues to focus disproportionately on early childhood populations, even though ASD is a lifelong developmental condition (Benevides et al., 2020). Evidence regarding behavioral support for autistic adults—including vocational independence, social relationships, community participation, mental health, and aging—remains comparatively limited.

Benevides et al. (2020) argued that autistic adults remain significantly underserved in both research funding and clinical service development despite experiencing substantial mental health and quality-of-life challenges. This imbalance raises important ethical and practical concerns regarding continuity of care across the lifespan. As increasing numbers of autistic individuals transition into adulthood, there is growing recognition that intervention research must extend beyond early developmental gains and address long-term participation, autonomy, emotional well-being, and societal inclusion.

Overall, the behavioral intervention landscape in ASD appears to be undergoing a gradual yet meaningful transformation. Contemporary approaches increasingly emphasize individualized support, emotional well-being, developmental diversity, and family participation rather than focusing exclusively on symptom reduction or behavioral compliance. Still, continued methodological rigor, culturally adaptable intervention models, and lifespan-oriented research remain critically necessary if behavioral therapies are to adequately address the diverse and evolving needs of autistic individuals and their families.

4. Results

4.1 Overview of Included Studies and Intervention Distribution

A total of nine primary studies satisfied the eligibility criteria for inclusion in this systematic review, collectively representing 539 participants across multiple behavioral intervention frameworks for autism spectrum disorder (ASD). The included studies encompassed four major intervention categories: Early Intensive Behavioral Intervention (EIBI), Early Start Denver Model (ESDM), Cognitive Behavioral Therapy (CBT), and parent-mediated interventions. The distribution of these intervention models across the included literature is illustrated in Figure 2, where ESDM and CBT appeared most frequently (k = 3 each), followed by EIBI (k = 2) and parent-mediated intervention approaches (k = 1).

Although all interventions fell broadly within the behavioral or psychosocial treatment spectrum, the conceptual orientation of the studies varied considerably. Some investigations emphasized intensive therapist-led behavioral conditioning grounded in Applied Behavior Analysis (ABA), whereas others adopted more naturalistic developmental frameworks centered on social engagement, emotional reciprocity, or caregiver participation. This diversity is particularly evident in Table 5, which compares the theoretical foundations, delivery modes, parental roles, and target outcomes associated with each intervention model. EIBI, for example, remained heavily structured and therapist-directed, while ESDM and JASPER incorporated more developmental and play-based interactional components. CBT interventions primarily targeted anxiety regulation among older autistic youth, whereas parent training programs focused more directly on behavioral management within family settings.

The methodological composition of the included evidence bases also reflected notable variation. Most studies employed randomized controlled trial (RCT) designs, including Dawson et al. (2010), Smith et al. (2000), Chalfant et al. (2007), Murphy et al. (2017), and Bearss et al. (2015). However, several studies utilized controlled clinical trial (CCT) designs or pre-post intervention frameworks, including Howard et al. (2005) and Eapen et al. (2013), introducing additional heterogeneity into the pooled evidence structure. These methodological differences become especially relevant when interpreting comparative intervention effectiveness because study rigor, allocation concealment, and blinding procedures varied substantially across the literature.

4. 2 Characteristics of Study Populations

Participant characteristics differed substantially across intervention categories, particularly regarding age range, developmental stage, and intervention intensity. As shown in Figure 3 and detailed in Tables 1 and 3, most early intervention studies focused on preschool-aged children younger than five years, with mean participant ages typically ranging between 21 and 50 months. Dawson et al. (2010), and Eapen et al. (2013) all concentrated on toddlers or preschool-aged children receiving ESDM-based interventions, while EIBI investigations such as Smith et al. (2000) and Howard et al. (2005) targeted similarly young populations undergoing intensive ABA-oriented treatment programs.

In contrast, CBT-focused studies generally recruited older children, adolescents, or young individuals with high-functioning ASD and co-occurring anxiety symptoms.

Table 1. Characteristics, Outcome Measures, and Methodological Quality of Primary Behavioral Intervention Studies Included in the Systematic Review. This table summarizes the design features, participant characteristics, intervention models, primary clinical outcome measures, and quality assessments of the principal behavioral intervention studies included in the review.

Table 2. Meta-Analytic Effect Size Summary Across Behavioral Intervention Categories. This table presents pooled effect sizes for major intervention domains identified in the systematic reviews and meta-analyses. Positive effect sizes indicate improvements favoring active intervention conditions.

Table 3. Characteristics of Primary Behavioral Intervention Studies. This table summarizes the methodological and therapeutic characteristics of influential EIBI and NDBI studies frequently cited in behavioral autism intervention research.

Figure 2: Intervention Categories Represented Across Included Studies. Bar graph showing the number of studies for ESDM, EIBI, CBT, and parent-mediated intervention.

Sofronoff et al. (2005), for instance, reported a mean participant age exceeding 10 years, while Murphy et al. (2017) specifically targeted adolescents with ASD-related anxiety presentations. This developmental distinction is clinically important because intervention goals differed markedly between younger and older populations. Early intervention models primarily emphasized communication, cognitive development, and adaptive functioning, whereas CBT studies concentrated more heavily on emotional regulation, anxiety reduction, and coping strategies.

Sample sizes also varied considerably across the included evidence base. Bearss et al. (2015) represented the largest individual trial, enrolling 180 participants within a parent-training randomized controlled framework, while Eapen et al. (2013) involved only 26 participants within a pre-post ESDM intervention design. Several EIBI studies similarly relied on relatively modest participant numbers, often fewer than 30 individuals per group. This imbalance between large-scale and small-scale studies may partially explain the variability observed in pooled effect estimates and confidence intervals across intervention categories.

Intervention duration and treatment intensity likewise demonstrated striking heterogeneity. EIBI studies generally delivered the highest treatment intensity, often ranging from 25 to 40 hours per week over periods extending from 24 to 48 months. Howard et al. (2005), Cohen et al. (2006), and Sallows et al. (2005) all implemented highly intensive behavioral schedules exceeding 35 hours weekly. By comparison, ESDM programs tended to involve somewhat lower weekly intensity but incorporated more naturalistic and socially embedded therapeutic interactions. CBT and parent-mediated interventions generally relied on shorter outpatient or community-based treatment schedules, emphasizing flexibility and family integration over therapeutic intensity alone.

4.3 Outcomes Associated with Early Intensive Behavioral Intervention

Across both Tables 3 and 4, EIBI consistently demonstrated moderate-to-large improvements in cognitive and adaptive functioning outcomes among young autistic children. Meta-analytic pooling revealed a significant positive effect on intelligence quotient (IQ), with a pooled effect size of g = 0.76 (95% CI [0.40, 1.11], p < 0.0001), based on four studies comprising 172 participants (Reichow et al., 2012). Similarly, adaptive behavior outcomes demonstrated a pooled effect size of g = 0.69 (95% CI [0.38, 1.01]), also reaching strong statistical significance.

These findings broadly align with the individual primary studies summarized in Table 3. Smith et al. (2000), Sallows et al. (2005), and Howard et al. (2005) all reported meaningful gains in developmental functioning following intensive ABA-oriented interventions delivered during early childhood. Notably, these interventions often involved exceptionally high treatment exposure, with some programs exceeding 35 hours of therapy weekly over multiple years. Such intensity likely contributed to the substantial developmental gains observed, although it simultaneously raises questions regarding scalability and accessibility in routine clinical practice.

Still, despite generally encouraging outcomes, the EIBI evidence base appeared methodologically fragile in several respects. Table 6 indicates that many EIBI reviews contained elevated risks of selection bias, limited blinding, and substantial heterogeneity across intervention protocols. Reichow et al. (2012), for example, concluded that EIBI appeared effective, yet simultaneously emphasized the low overall quality of the available evidence due to methodological weaknesses within the included studies. Similarly, Warren et al. (2011) argued that evidence supporting early intervention remained insufficiently robust because many trials relied on small sample sizes, non-randomized designs, or inconsistent outcome measurement approaches.

Consequently, although EIBI continues to demonstrate promising developmental benefits, the interpretation of these findings requires a degree of caution. The intervention appears capable of producing measurable cognitive and adaptive improvements, yet the consistency, generalizability, and long-term sustainability of these gains remain somewhat uncertain across broader clinical populations.

4.4 Cognitive and Developmental Outcomes Associated with ESDM and NDBIs

Among all intervention categories examined, ESDM demonstrated the largest pooled effect size for cognitive outcomes. As presented in Table 4 and Figure 4, the pooled cognitive effect estimate reached g = 1.37 (95% CI [0.95, 1.80]), based on two studies involving 45 participants (Shi et al., 2021). These findings suggest that developmental-behavioral hybrid models may produce

Figure 3: Sample Size Distribution Across Primary Studies. Horizontal bar graph comparing total sample size for each included study.

Figure 4: Pooled Effect Sizes Across Intervention Models and Outcome Domains. Forest plot showing Hedges’ g values and 95% confidence intervals.

Figure 5: Funnel Plot for Publication Bias Assessment. Scatter plot of effect size versus standard error for outcomes with complete CI data.

Table 4. Pooled Effect Sizes Across Major Clinical Outcome Domains. This table presents pooled meta-analytic estimates for behavioral intervention effectiveness across cognitive, adaptive, social, emotional, and family-centered outcome domains in ASD.

especially strong benefits in early cognitive functioning when implemented during critical developmental periods.

The strongest individual evidence emerged from Dawson et al. (2010), whose randomized controlled trial demonstrated meaningful improvements in cognitive performance, adaptive behavior, and autism symptom severity following intensive ESDM delivery among toddlers with ASD. Eapen et al. (2013) similarly reported favorable developmental outcomes within a community-based group ESDM program, despite employing a less rigorous pre-post design. Collectively, these findings appear to support the growing theoretical argument that interventions combining behavioral structure with social-developmental engagement may be particularly beneficial during early childhood neurodevelopment.

Interestingly, ESDM interventions generally relied on lower treatment intensity than classical EIBI models. While EIBI frequently exceeded 30–40 hours per week, Dawson et al. (2010) reported an average weekly intensity of approximately 15 hours. Yet despite lower therapeutic intensity, developmental gains remained substantial. This pattern may indicate that intervention quality, relational engagement, and contextual integration are potentially as important as treatment duration alone.

Nevertheless, inferential uncertainty surrounding ESDM findings remains difficult to ignore. Table 6 demonstrates that reviews evaluating ESDM and related NDBI frameworks frequently reported moderate methodological quality rather than strong certainty. Shi et al. (2021) acknowledged moderate risks of bias and considerable between-study heterogeneity, while (Doty et al., 2017) classified as high risk due to allocation and blinding limitations. Thus, although ESDM currently appears highly promising, additional large-scale randomized trials remain necessary before definitive conclusions can be established regarding comparative superiority over more traditional ABA-based approaches.

4.5 CBT Outcomes for Anxiety and Emotional Regulation

CBT interventions demonstrated consistently favorable outcomes for anxiety reduction among autistic children and adolescents. Figure 4 and Table 4 indicate a pooled clinician-rated effect size of g = 1.05 (95% CI [0.45, 1.65], p = 0.0006) across six studies involving 208 participants (Kreslins et al., 2015). Parent-reported anxiety outcomes also showed strong improvement, with a pooled effect estimate of g = 1.00 (95% CI [0.21, 1.80]).

Studies by Chalfant et al. (2007), Sofronoff et al. (2005), and Murphy et al. (2017) collectively suggest that adapted CBT protocols can substantially reduce anxiety symptoms when interventions are tailored to the communication and cognitive profiles associated with ASD. Modifications frequently included visual supports, structured emotional identification exercises, and highly concrete therapeutic language. These adaptations likely contributed to the favorable outcomes observed across both clinician- and parent-reported domains.

However, the CBT literature also revealed greater variability than some other intervention categories. Confidence intervals surrounding pooled anxiety effects remained relatively wide, suggesting notable between-study heterogeneity. Moreover, Weston et al. (2016) reported that informant-reported improvements in ASD symptoms were more modest overall (g = 0.48), even though statistically significant. This discrepancy may indicate that anxiety reduction does not necessarily translate into broad behavioral or social transformation across all domains of autistic functioning.

Methodological concerns were also apparent. Murphy et al. (2017) and Chalfant et al. (2007) were both classified as high risk due to attrition, blinding limitations, and allocation concerns. Table 6 similarly highlights elevated risks of measurement bias within CBT reviews because many outcomes relied heavily on informant reports or self-report instruments. Despite these limitations, the cumulative evidence still suggests that CBT remains one of the most consistently supported psychosocial interventions for anxiety management in autistic youth populations.

4.6 Parent-Mediated Interventions and Family Outcomes

Parent-mediated interventions demonstrated meaningful effects not only on child behavioral outcomes but also on family functioning and caregiver stress. As summarized in Table 4, pooled analyses identified a moderate effect size for disruptive behavior reduction (g = 0.67, 95% CI [0.49, 0.85]) alongside a smaller yet statistically significant reduction in parenting stress (g = 0.37) (Tarver et al., 2019).

Bearss et al. (2015), which represented the largest trial included in this review, demonstrated that structured parent training significantly improved behavioral outcomes compared with parent education alone. These findings support the growing recognition that caregiver’s function not merely as observers of therapy but as active therapeutic agents capable of shaping emotional regulation, behavioral consistency, and communication opportunities within everyday environments.

Importantly, the methodological quality of parent-mediated intervention research appeared comparatively stronger than several other intervention categories. Table 6 classified Tarver et al. (2019) as low risk across most major bias domains, with strong overall methodological quality. This relative robustness strengthens confidence in the reliability of parent-mediated intervention findings and suggests that family-centered models may represent one of the more evidence-supported directions within contemporary ASD intervention research.

4.7 Publication Bias and Overall Methodological Quality

The funnel plot presented in Figure 5 demonstrated a broadly symmetrical distribution of intervention effect estimates around the pooled mean effect size (g ≈ 0.91), with no obvious visual evidence of severe publication bias. Nevertheless, interpretation remained limited because only five eligible intervention-outcome estimates could be included in the analysis. Under such conditions, formal statistical asymmetry testing would likely remain underpowered and potentially unreliable.

More broadly, the overall methodological landscape across behavioral intervention research remained mixed. Table 6 reveals persistent concerns regarding blinding, selection bias, measurement heterogeneity, and non-standardized outcome reporting across many intervention categories. Reviews focusing on adult ASD interventions, particularly Benevides et al. (2020) and Brugha et al. (2015), emphasized the substantial lack of rigorous adult-focused evidence and the continued absence of standardized outcome frameworks for adulthood functioning.

Taken together, the findings suggest that behavioral interventions for ASD generally demonstrate favorable outcomes across cognitive, adaptive, emotional, and behavioral domains. However, the strength of evidence varies meaningfully between intervention categories, and substantial methodological uncertainty persists throughout the literature. While ESDM and CBT produced some of the largest pooled effects, parent-mediated interventions arguably displayed the strongest balance between effectiveness and methodological reliability. EIBI, although historically influential, continued to exhibit variability in evidence quality despite favorable developmental outcomes. Overall, the results indicate a field that is progressively evolving toward more individualized, developmentally integrated, and family-centered intervention paradigms while still confronting important methodological and translational challenges.

5. Discussion

5.1 Comparative Effectiveness, Methodological Challenges, and Emerging Trends in Behavioral Interventions for Autism Spectrum Disorder

The present systematic review synthesized evidence across multiple behavioral intervention frameworks for autism spectrum disorder (ASD), including Early Intensive Behavioral Intervention (EIBI), Early Start Denver Model (ESDM), Cognitive Behavioral Therapy (CBT), and parent-mediated interventions. Collectively, the findings suggest that behavioral interventions can produce meaningful improvements across cognitive, adaptive, emotional, and behavioral domains, although the magnitude and consistency of these effects varied considerably according to intervention type, participant age, methodological quality, and outcome domain. Importantly, the evidence also reflects a broader conceptual evolution within autism intervention science—from highly structured therapist-driven behavioral conditioning toward more individualized, naturalistic, and family-centered approaches.

One of the clearest patterns emerging from this review involves the sustained effectiveness of early intervention approaches, particularly EIBI and ESDM, during critical developmental periods. Across the included studies, early intervention programs targeting toddlers and preschool-aged children consistently demonstrated moderate-to-large improvements in cognitive performance and adaptive functioning (Tables 2–4). EIBI interventions showed pooled effect sizes of g = 0.76 for intelligence outcomes and g = 0.69 for adaptive behavior, while ESDM demonstrated an even larger pooled cognitive effect size of g = 1.37 (Reichow et al., 2012; Shi et al., 2021). These findings, visualized in Figure 4, reinforce the longstanding hypothesis that early neurodevelopmental plasticity may increase responsiveness to intensive behavioral learning during the first years of life.

Table 5. Comparative Frameworks of Behavioral Intervention Models in ASD. This table compares the theoretical orientation, therapeutic goals, delivery settings, parental involvement, and evidence strength associated with major behavioral intervention approaches for ASD.

Table 6. Methodological Quality and Risk of Bias Across Systematic Reviews. This table summarizes methodological quality assessments and major conclusions reported across systematic reviews evaluating behavioral interventions for ASD.

Yet the interpretation of these findings requires some caution. Although EIBI has historically been considered one of the most empirically supported autism interventions, the methodological quality of supporting evidence remains inconsistent. Table 6 illustrates that many EIBI-focused systematic reviews reported elevated risks of selection bias, non-blinded outcome assessment, and substantial heterogeneity across intervention protocols. Reichow et al. (2012) themselves acknowledged that while EIBI appeared effective, the overall certainty of evidence remained low due to methodological limitations. This tension between promising developmental outcomes and imperfect evidence quality has become a recurring theme throughout autism intervention research. In practice, the field still struggles to determine whether positive outcomes reflect the intervention model itself, exceptionally intensive treatment exposure, family motivation, or broader environmental influences.

Another particularly important observation concerns the apparent shift away from strictly therapist-directed intervention toward more naturalistic developmental approaches. The increasing prominence of ESDM and other Naturalistic Developmental Behavioral Interventions (NDBIs) reflects growing recognition that autistic children may benefit more effectively from socially embedded and emotionally responsive learning environments than from highly repetitive behavioral drills alone (Schreibman et al., 2015). As summarized in Table 5, ESDM differs from classical ABA-oriented approaches by emphasizing play-based interaction, joint attention, imitation, and emotional reciprocity within natural caregiving settings. Interestingly, despite requiring lower weekly intervention intensity than traditional EIBI programs, ESDM still produced some of the strongest pooled cognitive outcomes in the current review (Figure 4). This pattern raises an important possibility that intervention quality, relational engagement, and developmental appropriateness may sometimes matter as much as sheer therapeutic intensity.

The findings related to CBT also deserve careful consideration. Anxiety disorders remain highly prevalent among autistic adolescents and young adults, often contributing substantially to emotional distress, social withdrawal, and reduced quality of life. In the present review, CBT interventions produced large pooled effects for clinician-reported anxiety reduction (g = 1.05) and parent-reported anxiety improvement (g = 1.00) (Table 4; Figure 4). These findings broadly align with previous systematic reviews suggesting that modified CBT protocols can effectively address anxiety symptoms in autistic youth when interventions incorporate structured routines, visual supports, and concrete communication strategies (Kreslins et al., 2015; Weston et al., 2016).

However, the CBT literature also revealed important complexities. Improvements in anxiety symptoms did not always translate into equally strong improvements in broader ASD-related functioning. Weston et al. (2016), for example, reported more modest pooled effects for ASD symptom reduction itself (g = 0.48), despite stronger anxiety outcomes. This discrepancy may suggest that emotional regulation and core social-communication difficulties represent partially distinct therapeutic targets requiring different intervention strategies. It also highlights a broader challenge within ASD research: outcome measures often capture only selected dimensions of autistic functioning while potentially overlooking quality-of-life changes, emotional resilience, or subjective well-being.

Parent-mediated interventions emerged as another especially promising area within the current review. Unlike earlier intervention models that positioned parents primarily as observers or treatment facilitators, contemporary parent-training approaches increasingly recognize caregivers as active therapeutic agents capable of shaping communication, behavioral regulation, and emotional consistency within everyday settings. As summarized in Tables 2 and 4, parent-mediated interventions demonstrated moderate effects for disruptive behavior reduction (g = 0.67) and smaller but significant improvements in parenting stress (Tarver et al., 2019). Notably, these interventions also demonstrated comparatively stronger methodological quality than several other intervention categories, with lower overall risk of bias reported in Table 6.

This family-centered perspective may represent one of the most meaningful conceptual developments within autism intervention research. ASD rarely affects only the diagnosed individual; rather, it often reshapes family dynamics, caregiver mental health, financial stability, and household functioning. Consequently, interventions that strengthen parent confidence, improve behavioral consistency within natural environments, and reduce caregiver burden may ultimately generate broader and more sustainable developmental gains than exclusively clinic-based therapies. The findings of Bearss et al. (2015) strongly support this interpretation, particularly given the relatively large sample size represented in Figure 3.

Despite these encouraging findings, several important limitations continue to affect the broader behavioral intervention literature. Perhaps the most persistent issue involves methodological heterogeneity. As shown throughout Tables 1–6, studies varied considerably regarding participant age, intervention intensity, duration, outcome measures, randomization procedures, and comparison conditions. Such variability complicates direct comparison across studies and likely contributes to the wide confidence intervals observed in several pooled analyses (Figure 4). Furthermore, many studies relied heavily on parent-report or clinician-report measures, increasing vulnerability to measurement and expectancy bias. Funnel plot analysis (Figure 5) suggested no obvious severe publication bias, although the relatively small number of pooled estimates substantially limits interpretive certainty.

Another major concern involves the disproportionate emphasis on early childhood populations. Most intervention studies included in this review focused almost exclusively on children younger than five years of age (Figure 3), whereas evidence for adolescents, adults, and aging autistic populations remained remarkably limited. This imbalance is concerning because ASD is fundamentally a lifelong neurodevelopmental condition rather than a childhood-limited disorder. Benevides et al. (2020) and Brugha et al. (2015) both emphasized the substantial lack of rigorous adult-focused intervention research, particularly regarding employment, social relationships, emotional well-being, and independent living outcomes. Consequently, the field currently possesses far stronger evidence for improving early developmental skills than for supporting autistic individuals across adulthood transitions and long-term quality-of-life challenges.

Taken together, the evidence synthesized in this review suggests that behavioral interventions for ASD are generally beneficial across multiple developmental and psychosocial domains. ESDM and CBT demonstrated some of the largest pooled effects, EIBI continued to show meaningful developmental benefits despite methodological concerns, and parent-mediated interventions emerged as particularly promising family-centered approaches with comparatively stronger evidence quality. At the same time, the field appears to be gradually shifting toward more individualized, relational, and ecologically grounded models of support. Rather than focusing exclusively on symptom reduction or behavioral normalization, contemporary intervention science increasingly recognizes the importance of emotional well-being, developmental diversity, caregiver participation, and long-term social functioning. Future research will likely need to prioritize larger multicenter trials, standardized outcome frameworks, culturally adaptable interventions, and lifespan-oriented support models capable of addressing the evolving needs of autistic individuals across childhood, adolescence, and adulthood.

5.3 Limitations

Several limitations should be considered when interpreting the findings of this review. First, substantial methodological heterogeneity existed across the included studies regarding intervention intensity, duration, outcome measures, participant age, and study design, limiting direct comparability between intervention categories. Many studies also relied heavily on parent-report or clinician-report measures, increasing susceptibility to expectancy and measurement bias. Second, several intervention domains included relatively small sample sizes, reducing statistical precision and potentially inflating pooled effect estimates. Third, risk-of-bias assessments revealed persistent concerns involving non-blinded outcome evaluation, allocation limitations, and inconsistent randomization procedures, particularly within older EIBI studies. Additionally, publication bias could not be fully excluded because the number of eligible pooled studies remained limited for several analyses. Finally, most available evidence focused primarily on preschool-aged children, while rigorous intervention data for adolescents, adults, and aging autistic populations remained comparatively sparse, restricting broader lifespan generalizability.

6. Conclusion

Behavioral interventions for ASD generally demonstrate meaningful benefits across cognitive, adaptive, emotional, and behavioral domains, although intervention effectiveness varies considerably according to developmental stage, therapeutic model, and methodological quality. ESDM and CBT produced some of the strongest pooled outcomes, while parent-mediated interventions emerged as particularly promising family-centered approaches. Contemporary intervention research increasingly emphasizes emotional well-being, developmental individuality, and relational engagement rather than behavioral normalization alone. Future progress will likely depend on larger multicenter trials, lifespan-oriented intervention frameworks, culturally adaptable treatment models, and more standardized outcome measures capable of capturing meaningful quality-of-life changes across diverse autistic populations.

Author Contributions

Md S. Islam: Conceptualization, Methodology, Formal Analysis, Investigation, Writing – Original Draft, Project Administration. H. Akter: Data Curation, Validation, Visualization, Writing – Review & Editing.

Acknowledgements

The authors sincerely thank the researchers, clinicians, and families whose published work formed the evidence base of this systematic review. The authors also acknowledge the broader autism research community for their continued efforts to advance understanding of behavioral intervention science across diverse autistic populations.

Conflict of Interest

The authors declare no conflict of interest.

Financial Disclosure

Md S. Islam and H. Akter confirm that this research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors. No financial relationship, sponsorship, or competing interest of any kind influenced the study design, literature selection, data synthesis, interpretation of findings, or decision to publish.

References