Early intervention programs for toddlers with Autism Spectrum Disorder: a systematic review

This review identified 8 randomized controlled trials (RCTs) that evaluated early intervention (EI) programs for infants in the age-group 18-48 months who either had been diagnosed with, or were considered at risk for, autism spectrum disorder (ASD). The studies were summarized in terms of participant characteristics, intervention characteristics, rigor of study/research and outcomes. Intervention characteristics included the provision of training to parents. All the studies used RCT design, with control subjects who were either toddlers of typical development (TD) or toddlers with ASD following “treatment as usual” (TAU) or another treatment, and all were rated as strong in terms of quality/rigor. Positive results were recorded for parental acceptability and satisfaction, and reduction of parenting stress. In most of the studies, the social communication and developmental skills of the toddlers were enhanced. We conclude that EI programs for ASD show promise, and may be beneficial for both the toddlers and the parents, but the limited number of RCTs and the wide variety in intervention programs and assessment instruments used indicates the need for additional research to evaluate the specific benefits.


Introduction
It is currently possible to diagnose autism spectrum disorder (ASD) at a very early stage, namely when the toddlers are aged around 2 years. This early identification of toddlers with ASD facilitates the implementation of early intervention (ΕΙ) for children even before they attend school (Mottron, 2017), which is considered to be "essential to achieving the best outcomes" (Pierce et al., 2016). Several studies have shown improved outcomes for toddlers and children with ASD after EI (Granpeesheh et al., 2009;Rogers et al., 2012;Zachor et al., 2007).
EI is addressed to toddlers and young children with disabilities and/or developmental delay, and their families, and can help them to cope with the difficulties that their condition causes in their everyday lives. Behavioral interventions are not aimed at "curing" ASD, which is a neurodevelopmental disorder already established in infancy (Landa et al., 2018), but one of the main goals of EI is to reduce the manifestation of ASD symptoms to a minimum. Other intervention goals include the development of social, language, cognitive, adaptive, and play skills (Green et al., 2017;Landa & Kalb, 2012).
The age of enrollment of children in EI programs is a factor that affects significantly their effectiveness and long-term outcomes, because the first two years of a child's life are characterized by rapid changes in many areas, especially in social, cognitive and language development. This means that the introduction of EI at around two years of age, when the developmental gaps between toddlers of typical development (TD) and those with ASD are still small, should bring the best results (Bradshaw et al., 2015). In their review, Granpeesheh and colleagues (2009) found that EI was more effective for younger participants (2.55.15 years) than for those who were older (5.157.14 years).
An EI program can be either evidence-based or associated with empirical data that validates its effectiveness (Stahmer et al. 2005). While many EI programs were based on applied behavior analysis (ABA) in the past, a method that was strongly supported by the research community (Lovaas 1987;Reichow 2012), contemporary EI tends to follow the principles of developmental psychology and other naturalistic methods. Thus, interventions have become directed more towards the child itself and are now conducted in more natural environments, such as the child's home (Schreibman 2014).
Several EI programs that are popular among researchers and clinicians have one thing in common, which is the integration of behavioral, naturalistic, and developmental strategies, and they are labeled "naturalistic developmental behavioral interventions" (NDBIs) (Bradshaw et al., 2015). Some of these are the Early Start Denver Model (ESDM) (Rogers and Dawson 2010), the Enhanced Milieu Teaching (Kaiser and Hester 1994), and the Pivotal Response Treatmen" (PRT) (Koegel and Koegel 2012). Systematic reviews of EI conducted to date include studies of a variety of EI programs (Bradshaw et al., 2015;Landa, 2018) or several studies implementing only one type of EI, e.g., the ESDM (Waddington et al., 2016).
The present review was focused on currently available studies of EI programs applied to children with ASD between the ages of 18 and 48 months. The effectiveness of these programs was evaluated through RCTs, where participants were assigned randomly to either to a treatment group or a control group. The results of the eligible studies were synthesized, and the most relevant findings are presented. Critical elements that were explored in this review were: a) the types of EI programs that were implemented, b) infant and parent outcomes, c) intensity and duration, and d) maintenance and generalization.

Methodology
The review focused on interventions for toddlers aged 18-48 months with ASD. The review methodology was based on the guidelines of Ahn & Kang (2018). The first step was the formulation of research questions. Next, the authors determined the inclusion and exclusion criteria for the studies that were to be analyzed, and conducted a rigorous literature search. The study selection was made with the application of the eligibility criteria, and the quality of the presented evidence was discussed. The final steps were data extraction and analysis, and presentation of results. Each study that met the predetermined criteria was analyzed and summarized in terms of a) participant characteristics, b) intervention approach, c) toddler and parent outcomes.

Research questions
1. What were the most popular early intervention programs for toddlers with ASD? 2. What were the intervention outcomes for toddlers with ASD after EI? Were they positive or negative?
3. Did the parents feel that they had benefited from the EI programs? What were the main effects on them?

Inclusion and exclusion criteria
Only RCTs were included in the review. To be included, a study had to meet the following criteria: a) empirical research evaluating the effects of an EI program, b) the participants of the EI were toddlers with an age of above 18 months and below 48 months at entry to the program, c) the toddlers had been diagnosed with ASD, d) the results of the study included at least one objective child measurement and one parent outcome measurement.
Articles were excluded from the review if they: a) were non-experimental (e.g., literature reviews, meta-analyses, case reports); b) did not include an EI program; c) did not include toddlers aged 18 to 48 months. Studies were included that primarily, but not exclusively, targeted children aged <48 months or whose mean age was <48 months at the start of the intervention. Grey literature (i.e., dissertations, chapters, etc.) was excluded.
To determine whether a study met the inclusion criteria, the first and the last authors independently completed the search and evaluated all the studies. The selected articles were then compared for reliability, which was calculated using percent agreement on the articles each author identified as meeting the inclusion criteria. Disagreement between the two authors was discussed until they came to an agreement.

Search procedure
The research papers were found by a search in the PubMed, Education Resources Information Centre (ERIC), Science Direct and Scopus databases for papers published in English, appearing in peer-reviewed journals since 2010. The keywords used were: ASD, autism, autistic, early intervention program, toddler. The initial search resulted in 4,348 studies after Research, Society andDevelopment, v. 10, n. 14, e103101421935, 2021 (CC BY 4.0) | ISSN 2525-3409 | DOI: http://dx.doi.org/10.33448/rsd-v10i14.21935 4 duplicates were removed. The researchers read the titles and the abstracts to exclude studies that did not incorporate experimental results of EI programs, and/or referred to disabilities not including ASD. The remaining articles were independently screened by the authors for the inclusion criteria.
An ancestral search was conducted using the reference lists of the studies that met the inclusion criteria and the "cited in" feature in Scholar Google, and a hand search was made in peer-reviewed articles. Finally, eight studies were identified that fulfilled the criteria (Figure 1). The overall interrater agreement (IRA) was 90% and consensus was reached to resolve the few disagreements.  Research, Society andDevelopment, v. 10, n. 14, e103101421935, 2021 (CC BY 4.0) | ISSN 2525-3409 | DOI: http://dx.doi.org/10.33448/rsd-v10i14.21935 5

Coding procedures
To map and synthesize the included studies, the following coding categories were used: a) child characteristics (number, age, and diagnosis), b) parent characteristics (number of parent participants), c) intervention approach (i.e., empirical and theoretical basis extracted from the description of the intervention; intensity and duration of the intervention, in terms of number of individual sessions over a set period of time, the length of each session), d) quality of the study/research rigor, e) child outcome measurements (e.g., scores on cognitive, language, and/or adaptive behavior assessment), and f) parental outcome measurements (e.g., changes in parenting stress, skills, responsivity, parental use of evidence-based strategies).
Finally, each study was coded to assess its quality based on the evaluative method for determining evidence-based practices in autism, which has been reported to have good to excellent reliability and validity (Reichow et al. 2008). To evaluate the rigor of the studies, two rubrics were developed; one for group research and one for single-subject research. These rubrics include two levels of methodological elements: primary quality indicators and secondary quality indicators. Three levels of rating were given to each study: strong, acceptable/adequate, weak; and demonstrating concrete evidence of quality, strong evidence in most, but not all areas, missing elements, and/or fatal flows. Primary quality indicators for group research include the quality of the description of participant characteristics, independent variable, comparison condition, dependent variable, the link between research question and data analysis, use of statistical tests. Secondary quality indicators were not deemed necessary for the establishment of the validity of the study and are related to random assignment, interobserver agreement, blind raters, fidelity, attrition, generalization and/or maintenance, effect size, social validity.
The second, the third and the fourth authors reviewed independently the included studies to determine whether each of them met the coding categories and the evaluative method for determining evidence-based practices in autism. They extracted data from each of the eight studies and created a summary, as shown in Tables 1 and 2. The authors compared the results for the coding, and any disagreement between the authors was discussed until they came to an agreement. Overall, the IRA for all the coding categories was 100%.

Results
The first search yielded in 4,348 papers, from which the final selection resulted in 8 studies that met the inclusion criteria and the coding procedures. Tables 1 and 2 provide a summary of the studies in this review in terms of a) participant characteristics, b) intervention characteristics, c) quality/rigor, and d) outcomes.

Child characteristics
The eight studies included a total of 485 participants aged between 17 and 48 months, of which 251 received EI and 234 were control subjects, who were either children of TD or children with ASD who received "treatment as usual" (TAU) or another treatment. All the studies required that the ASD participants either had been diagnosed with ASD or were considered to be at risk for ASD (i.e., they presented behavioral symptoms of ASD) prior to participating in the intervention. The study of Ibanez and colleagues (2018) did not have any children as direct participants in EI, as this study reported on training of the parents and its effect on their children.
The children had a diagnosis of ASD or high risk for ASD in all eight studies, although in one study, one child included in the experimental group had a diagnosis of pervasive developmental disorder, not otherwise specified (PDD-NOS) (Oosterling et al., 2010). The ASD diagnosis was based on the Autism Diagnostic Interview-Revised (ADI-R) and the Autism Diagnostic Observation Schedule (ADOS) . In one study the tool that was used was not defined, but a copy of the child's diagnostic report confirming an ASD diagnosis was one of the criteria for inclusion in the research (Ibanez et al., 2018). In three studies the participants were described as being at high risk of ASD, based on the Screening Tool for Autism in two-year-olds (STAT) (Carter et al., 2011), In 6/8 studies, exclusion of children was reported based on other medical, physical, genetic, or neurological conditions, specifically, a genetic disorder (Carter et al., 2011, Kasari et al., 2014, Turner-Brown et al., 2019, or severe auditory, visual, or motor impairments (Ibanez et al., 2018, Carter et al., 2011, Rogers et al., 2018, Turner Brown et al., 2019. In one study the researchers included only children with either a diagnosis of ASD in combination with a developmental age of at least 12 months or children with a diagnosis of PDD-NOS in combination with a developmental age of at least 12 months and a developmental quotient (DQ) below 80 (Oosterling et al., 2010). The gender of the participants is reported in all the studies; most of the participants were male (80%).

Parental characteristics
Almost all the interventions (6/8) used parent-mediated procedures; the parents were taught specific procedures, which they were expected to use with their children during the intervention sessions and in everyday life. The intervention strategies involved didactic sessions about treatment techniques, and a feedback session in which parents and their toddlers practiced the intervention while a therapist provided feedback about implementation. One study examined the effects of an interactive web-based tutorial for improving children's engagement in daily routines and social communication, and parenting efficacy and parental stress (Ibanez et al., 2018).

Intervention approach The empirical and theoretical basis
Most of the studies adapted intervention models that had been previously applied for toddlers. These included Hanen's More Than Words (HMTW) (Carter et al., 2011), psychoeducational intervention (PEI) , Joint Attention Symbolic Play, Engagement, and Regulation (JASPER) , focus parent training (Oosterling et al., 2010) and the ESDM (Rogers et al., 2018), which have all been used with toddlers and preschool-aged children. Two studies (Welterlin et al., 2012;Turner-Brown et al., 2019) applied similar intervention models: the Family Implemented TEACCH for toddlers (FITT) and the Home TEACCHing Program for toddlers and their families. These two early intervention programs follow the basic principles of the TEACCH program which was modified and altered to be implemented for toddlers and their families, mainly in at-home settings.
Two studies provided, respectively, an interactive, web-based parenting tutorial (Ibanez et al., 2018) and a selfdirected, web-based training course (online course/ tutorial; Kasari et al., 2014), which included 24hour accessibility, standardization of training, personalization/individualization (e.g., self-paced), risk-free environment, and the opportunity for interactive exercises and multimedia components. Self-directed, web-based parent training programs appear to be costeffective and easily available to the parents.

Intensity and duration
The duration of treatment ranged from 4 to 12 weeks in most of the interventions, and all were low-intensity, totaling no more than 2 hours of intervention per week. In one study (Turner-Brown et al., 2019), the sessions were carried out for 24 weeks, and in two studies the duration is not specified; in that of Carter and colleagues (2011), the intervention involved 8 group sessions with parents only, and 3 in-home individualized parent-child sessions, and in that of Ibanez and colleagues (2018), the entire tutorial was approximately 6 hours, with the parents reviewing the tutorial across at least 4 or 5 sessions.
All the studies reported data collection for evaluation at either two or three time-points, specifically at baseline and 1 year after the start of the intervention (Oosterling et al., 2018); baseline and approximately 7 months after (Turner-Brown et al., 2019), or time1: prior to randomization/baseline/pre-treatment time 2: 5/1/3/ months, post-treatment; time3: 9/2/12/6 months post-enrollment. In one study data were collected every 4 months on child and parent mastering of skills, and longterm observations of of child change were made (Rogers et al., 2018). One study included four data collection points, specifically prior to intervention, and at the fourth, eighth and twelfth weeks, the last being post-intervention (Welterlin et al.,

Research rigor
In terms of research rigor, all eight studies were rated as having a strong research design, according to the criteria developed by Reichow et al. (2008) and Reichow (2011) (Table 3). High quality was observed on all primary quality indicators (i.e., participant characteristics, independent variable, comparison condition, dependent variable, a link between research question and data analysis, use of statistical tests), and the studies showed evidence of four or more secondary quality indicators (i.e., random assignment, interobserver agreement, blind raters, fidelity, attrition, generalization and/or maintenance, effect size, social validity).

Child outcome measurements
The The child and parental outcomes of EI programs are presented in Table 2. Five studies reported at least one child outcome measurement of the child's social interaction and communication skills (Carter et al., 2011;Kasari et al., 2014;Rogers et al., 2018, Turner-Brown et al., 2019. One study focused on improving children's engagement in daily routines and social communication (Ibanez et al., 2018), and another on language development, engagement, and social communication (Oosterling et al., 2010). The outcome measurements included joint attention, initiating behavior requests, intentional and/or nonverbal communication, expressive and language skills, visual perception, functional and symbolic play, engagement during daily routines, and compliance and willingness to join in mutual activities. Three studies reported positive results for these outcome measurements, but Carter and colleagues (2011) found no major effects of treatment on child outcomes, either immediately after the parent-implemented treatment or at the follow-up assessment. Kasari and colleagues (2014) and Oosterling and colleagues (2010), also, found no significant differences between the two groups on joint attention and language skills. The study of Welterlin and colleagues (2012) focused on different types of skills, specifically on children's independent living skills, which they reported to be enhanced after the treatment, for most of the children that participated.

Study
Child outcomes Parental outcomes Carter et al. 14  Joint engagement more than doubled from entry to week 10 for the JASPER group, with a large effect size. The increase in the length of time spent jointly engaged was maintained at the 6-month follow-up and significant for the JASPER group compared with the PEI group.
The JASPER group increased more in types of functional play than the PEI group; however, these skills did not maintain at follow-up.
Children in the JASPER condition engaged with their teachers more in their early intervention classroom. These findings may be among the first indicating generalization of joint engagement skills from a parent-mediated intervention to new partners and contexts.
Parents coached in specific JASPER strategies were significantly more effective at engaging their children in play at post-treatment and follow-up than parents who received information about specific strategies through the PEI. Effect sizes were moderate to large.
Results indicated a reduction in parenting stress for families in the PEI condition While there were significant gains for both groups over time, there were no group differences in the degree of improvement in children's skills after 12 weeks of intervention.
There was a significant positive relationship between the degree of improvement in parental fidelity of implementation and increases in child social communication and decreases in autism symptoms on the proximal measure of change.
The rate of parental learning of the intervention was improved. Parents in the P-ESDM++ group demonstrated significantly increased sensitivity and skill in supporting child socialcommunicative development measured by increases in parent fidelity of implementation scores compared to the parents in the P-ESDM group.
Parents in both groups were extremely satisfied with the intervention that they received. This There were no significant differences for FITT and SAU groups at baseline. Children from the FITT group had higher PIA scores (a measure of autism symptom severity) and PIA imitation scores than children from the SAU group.
The evaluation form that was filled in by the parents revealed no regression in social interaction, cognitive skills, and communication of children from both groups. 43% of the children from the FITT group were reported to have made "a lot of progress" in social interaction skills (with a statistically significant difference from the SAU group).
Parents from the FITT group had lower levels of stress and parental distress. They also had better results regarding their quality of life (with the RAND-36 tool) and they reported high levels of satisfaction with the program. All parents exhibited high levels of engagement according to therapist evaluations.

Welterlin et
al. (2012) Independent functioning skills were enhanced for two of three pairs of participants. Some differences were noted for young participants from each pair, specifically from pair 2.
Subject HT-C2 had a decrease in their target skills, whereas WL-C2 had better results.
Regarding the first pair, both subjects (HT-C1 and WL-C1) had an increase in their outcomes, but variation was larger for WL-C1 after the treatment. Regarding the third group, subject HT-C3 showed a great response to the treatment, whereas subject WL-C3's response was smaller.
Children from the HTP group made progress in expressive language, as well as children from the WL group. Differences between children from the two groups were not statistically significant.
An increase in setup behavior was observed for all parents that participated in the program.
Also, there was an increase in effective prompts and a decrease in ineffective prompts for all parents, but variations were more significant for those that participated in the treatment group. There were no statistically significant differences between the HTP and the WL groups. Parent stress was had a slight decrease for HTP and a slight increase for WL participants, but again, group differences were not statistically significant.

Parent outcomes measurements
Positive results in parental responsivity were reported in 6/8 studies. The effect size was medium to large immediately after treatment, and moderate at the follow-up assessment (Carter et al., 2011). Nearly 80% of parents in the focused playtime intervention improved in their responsiveness Kasari et al., 2014), in parental use of evidence-based strategies, such as providing simple verbal instructions, using visual schedules, modifying routine steps (Ibanez et al., 2018), use of effective prompting during structured teaching implementation (Welterlin et al., 2012), and learning about the intervention by showing increasing sensitivity and skill in supporting child social-communicative development (Rogers et al., 2018). Parenting stress related either to challenges in the parent-child relationship or to the disorder itself declined significantly in the treatment group (Ibanez et al., 2018, Turner-Brown et al., 2019 and the parent education program , although the decrease in parental stress in the treatment group was not significant in the studies of Welterlin and colleagues, (2012), and Oosterling and colleagues (2010) reported that the training program did not significantly influence parental skills.

Social validity
A questionnaire about parent satisfaction was included in 4/8 studies. Positive results regarding feasibility, acceptability or satisfaction with the intervention were reported in three studies (Carter et al., 2011;Ibanez et al., 2018;Rogers et al., 2018). In the study of Ibanez and colleagues (2018), the parents in the tutorial group indicated high levels of satisfaction with the technical aspects and the clinical content. Rogers and colleagues (2018)

administered the Intervention Evaluation
Form for Parents, a Likert-type scale of 14 questions, at the end of the treatment period, and other researchers monitored treatment integrity by measuring fidelity of implementation (Kasari et al., 2014;. Oosterling and colleagues (2010) used professional observation, parent reports, and video recording of data collection. Turner-Brown and colleagues (2019) reported positive scores on parental satisfaction with the "Family Implemented TEACCH for Toddlers"; specifically, high ratings in the domains of satisfaction in general, and satisfaction with the goals set, with the intervention procedures, and with the outcomes.

Maintenance/Generalization
Maintenance and/or generalization probes were conducted in 6/8 studies. In three, only maintenance was reported on (Carter et al., 2011, Ibanez et al., 2018, Kasari et al., 2014, and in two studies both maintenance and generalization phases were included , Rogers et al., 2018. In the study of Carter and colleagues (2011) (2014) conducted a long-term follow-up, which showed lack of parental responsiveness; only those parents who showed responsiveness at baseline maintained their responsiveness to followup. Mixed results were reported by Kasari and colleagues (2015), since maintenance of joint engagement was limited, and the children's improvements in functional-play diversity and overall play level were not maintained at follow-up. The lack of follow-up data in the study of Rogers and colleagues (2018) prevents determination of the extent to which the treatment resulted in stable changes in parent delivery, or whether the results are generalizable to community settings. Kasari and colleagues (2015) explored the generalization of joint engagement in the classroom and reported that children in the JASPER program engaged more in their early intervention classroom.

Moderators of outcome
Two studies included moderator variables to uncover the effects of specific child and intervention characteristics on child and parent outcomes. Carter and colleagues (2011) identified limited object interest as a moderator for facilitating growth in communication for the HMTW group. Ibanez and colleagues (2018) identified the tutorial itself as leading to changes in the routine-specific strategies used by parents, and improvement in the behaviors exhibited by children. Kasari and colleagues (2014) revealed a possible relationship between the durability of the treatment and the long-term outcomes. Extending the duration of the intervention, or supplying "booster" sessions, may improve responsiveness and maintain positive changes in parental behavior. Kasari and colleagues (2015) indicated a reduction in parenting stress for families in the PEI program, who consulted with an expert about their children and gained greater knowledge about ASD. Oosterling and colleagues (2010) reported that the DQ may affect language improvement, engagement, and precursors of social communication. Welterlin et al.
(2012) noted that parents may need more time and practice opportunities to be more effective in implementing structured teaching, and that the implementation of only one baseline probe could not lead to potent conclusions. Lastly, Turner-Brown and colleagues (2019) discussed factors such as therapist consultation, in-home implementation of the early intervention program with particular emphasis given to understanding ASD, and implementation of parent groups, which may have a positive impact on parent outcomes.

Discussion
The purpose of this review was to evaluate EI programs for toddlers with, or at risk for, ASD. Using stringent criteria, eight relevant studies were identified, all of which were published since 2010. All of the studies included in the review were RCTs, and they examined EI for toddlers in the age range18-48 months at enrolment. The interventions varied in intensity and duration, ranging from 4 to 12 weeks, with no more than 2 hours per week . Mixed findings were reported regarding enhanced effectiveness over the TAU comparison groups in a range of outcome measures, including social skills in the children, and parenting stress.
A wide variety of EI programs was implemented in the eligible studies, including HMTW, JASPER, the Focus Parent Training, the Parent-implemented ESDM (P-ESDM, FITT, and others. No two (or more) studies implemented the same EI program, so each investigated the effectiveness of a different program, and therefore conclusions cannot be drawn on which program is more popular, simply based on the findings of this review.
Similarly, a various different instruments were used for outcome measurement in the children and the parents. Most of the studies reported some positive outcomes for the participating parents and/or toddlers, which is encouraging, and serves to demonstrate the need for further research. Positive parental outcomes were reported by 4/8 studies, which were maintained at follow-up, related to the use of evidence-based strategies, reduced parenting stress, increased parental sensitivity, and skills in supporting their children (Ibanez et al., 2010;Kasari et al., 2014;Rogers et al., 2018), although two studies detected no major effects on parental responsivity and skills (Carter et al.,2011;Oosterling et al., 2010). Most studies (Ibanez et al., 2018;Oosterling et al., 2010) also reported significant positive child outcomes, particularly in engagement during daily routines, communication skills, joint engagement, and language skills,, but others recorded no significant group differences in the degree of improvement in children's skills post-intervention (Carter et al., 2011;Kasari et al., 2014;Rogers et al., 2018).
Exploration of the moderator variables associated with outcome data may provide useful information about factors that can influence the effectiveness of an intervention. HMTW appears to be more effective with children who show less interest in objects, whereas children who had a high interest in objects exhibited growth attenuation (Carter et al., 2011). Other moderators appear to be the intensity of the intervention and the tutorial itself. Rogers and colleagues (2012) found that children who received more intervention hours appeared to benefit more.
It is common to find inconsistencies in the outcomes of studies of EI (Landa et al., 2018), which can be attributed, among other factors, to the individual differences and characteristics of the children (Howlin et al., 2009). Clinicians and therapists should therefore consider each child's strengths and weaknesses, and the family environment, very carefully before suggesting an EI program. Following enrolment, there should be constant contact with the family and meticulous gathering and evaluation of information.
The current review suggests that the various EI programs used in the reviewed studies, based on the high ratings for research rigor, offer promising treatment for toddlers with or at risk for ASD. All the studies included in the review were rated as being methodologically strong, which increases the certainty of the evidence. Two studies with strong ratings, however, did not report significant improvement for either the toddlers or the parents (Carter et al., 2011;Oosterling et al., 2010), but positive results from six of the eight studies support the EI programs that were used, as promising interventions for toddlers with ASD and their families.
It should be noted that EI is aimed at facilitating the participation of children with ASD in more inclusive settings, minimizing the developmental and behavioral obstacles that these children face (Landa et al., 2018). To this end, parents and clinicians should collaborate and decide on the best approach that fits their child's needs, as each program could have different effects on different children. The age of enrollment, the goals that are set before enrolling in an EI program, and the intensity and duration, must be tailored to the individual circumstances.

Final Considerations
This review has several limitations, and the conclusions presented are based upon a relatively small sample size. It is possible that some relevant studies were excluded based on the stringent criteria related to experimental design and/or publication in English-language peer-reviewed journals. Research groups investigating the effectiveness of EI must consider the impact of moderator variables and their effect on outcomes, so careful identification of the factors that might have an influence on the results is essential. Particular attention should be given to conducting maintenance and generalization probes, in order to examine the long-term benefits of an EI program. For future studies of EI programs for toddlers with ASD, larger sample sizes and application of various different intervention approaches would provide useful evidence.