In the United States, autism spectrum disorder (ASD) affects 1 of every 59 children (Centers for Disease Control and Prevention, 2019), contributing significantly to the caseloads of primary care providers. Even in 2004, when the prevalence of ASD was less than half the current rate, almost 50% of primary care providers had at least 10 patients with ASD (Dosreis, Weiner, Johnson, & Newschaffer, 2006). Nurse practitioners (NPs) are particularly affected by these developments: The American Academy of Pediatrics recommends children receive two screenings for ASD by 2 years of age (Johnson & Myers, 2007), and NPs may administer these screenings (Nadel & Poss, 2007). Thus, NPs are often among the first health professionals to have meaningful conversations with parents about ASD.
Parents may be distressed by an autism diagnosis (Crane, Chester, Goddard, Henry, & Hill, 2016) and need help regarding treatment selection. Unfortunately, NPs and other health care providers will be unable to facilitate treatment choice if they lack expertise regarding ASD management (Todorow, Connell, & Turchi, 2018; Will, Barnfather, & Lesley, 2013). This is evident to parents (Rhoades, Scarpa, & Salley, 2007), who are likely to turn to nonscientific sources, such as other parents, nonmedical professionals, or the Internet (Miller, Schreck, Mulick, & Butter, 2012; Schreck, 2014). Many of these sources are biased, pointing to nonscientific or harmful treatments (Hofer, Hoffmann, & Bachmann, 2017).
Many ASD treatments lack proven efficacy (Schreck, 2014), but even effective ones have reduced impact if they are delayed: Better outcomes are achieved with younger children (Johnson & Myers, 2007; Klintwall, Eldevik, & Eikeseth, 2013). Treatment delays also increase the public and private costs of supporting ASD individuals across their lifespans (Piccininni, Bisnaire, & Penner, 2017) because the cognitive, adaptive, and social skills children attain by school entry (6 years of age) predict functioning in adolescence and adulthood (Howlin, Savage, Moss, Tempier, & Rutter, 2014). Clearly, medical professionals should engage with parents of newly diagnosed children and recommend evidence-based approaches likely to yield positive outcomes.
Many parents opt for an “eclectic” approach that comprises a mixture of treatments delivered by professionals from such disciplines as special education, psychiatry, psychology, and speech, language, and occupational therapy (Dillenburger, 2011; Eldevik et al., 2009). These professionals individualize treatment by recommending specific treatment combinations (e.g., special education preschool and speech and occupational therapy) and amounts (e.g., five 30-minute sessions per week), based on each child's profile. Given the multiple domains affected by autism, eclectic treatments have substantial face validity.
Another option is applied behavior analysis (ABA), which—like eclectic treatment—is often a covered health benefit. Applied behavior analysis has an established history of application to young children with ASD. Treatment focuses on increasing behaviors critical to learning (e.g., communication, self-help, self-regulation, and social skills), while decreasing behaviors that are dangerous or interfere with development (Roane, Fisher, & Carr, 2016). Treatment protocols designed to improve weak or absent skills are developed and continually refined based on direct observation and measurement of behavior. New skills are taught by breaking learning into small steps, whereas problem behaviors are primarily addressed by teaching alternative, appropriate responses.
Applied behavior analysis is often included in eclectic treatment packages, but at subtherapeutic dosages averaging 10–15 hours per week (Howard, Stanislaw, Green, Sparkman, & Cohen, 2014). Research has shown that better outcomes occur when ABA is provided intensively (at least 25 hours per week) for several years (Eldevik et al., 2010; Roane et al., 2016). Accordingly, all references to ABA below refer to this higher dosage level.
The available evidence suggests ABA treatment is significantly more effective than eclectic treatment in improving cognitive, adaptive, and communication skills for young children (Virués-Ortega, 2010). However, previous studies focus on standard scores and other metrics that most parents find difficult to interpret. The literature also fails to directly address parental concerns that an ASD diagnosis portends a bleak future for the child and the family (Elder & D'Alessandro, 2009). Accordingly, this study focuses on the likelihood of young children with autism achieving normal levels of functioning. Treatment outcomes are presented for 82 children who were diagnosed with ASD at an early age and received either ABA or eclectic treatment for up to 7.5 years. Of particular interest is the likelihood of attaining normal cognitive and adaptive scores by school age, as these outcomes have predictive value for independent functioning in adolescence and adulthood (Howlin et al., 2014; Perry, Koudys, Prichard, & Ho, 2017).
Participants and treatment groups
Participants included 61 children diagnosed before the age of 4 years by experienced clinicians who were not associated with treatment delivery, using best practice, comprehensive, multimethod assessments. Children began receiving treatment between 1996 and 2002 and are fully described elsewhere (Howard, Sparkman, Cohen, Green, & Stanislaw, 2005). Another 21 children who were not available for the earlier study began treatment as late as 2006.
Children received either ABA (n = 50) or eclectic treatment (n = 32) for an average of 50.15 months (SD = 20.21, range = 11–90). Assignment was largely based on parental preference. Treatment costs were funded by state developmental disability agencies and public schools. Treatment protocols are described briefly here; complete details are available elsewhere (Howard et al., 2005, 2014).
Children in the ABA group received 25–40 hours of ABA treatment each week in settings that included a center, schools, and homes. Children in this group remained with ABA treatment throughout the study's duration. None received auxiliary treatments such as sensory integration or individual or group speech therapy. Clinical supervisors had specialized training in ABA; most held credentials issued by the Behavior Analyst Certification Board.
Eclectic treatments were delivered in school settings and consisted of two types. One was designed specifically for children with autism and averaged 10 hours of ABA therapy per week; the other was designed for children with any form of developmental delay. Half of the eclectic treatment children also received group or individual speech therapy several times each week; one quarter received occupational therapy. Some children switched back and forth between the two types of eclectic treatment, but previous analyses found no discernible differences in outcomes between the two types after up to 3 years of treatment (Howard et al., 2005, 2014). Accordingly, the analyses described below combined all eclectic children into a single group.
Children were diagnosed with ASD at a younger age, on average, in the ABA group (M = 30.10 months, SD = 5.14) than in the eclectic group (M = 37.13, SD = 5.72; t-test p < .001). However, the delay between diagnosis and treatment onset was shorter for eclectic children (M = 1.50 months, SD = 2.93) than that for ABA children (M = 3.94, SD = 4.38; p = .013). Treatment duration was comparable for both groups (M = 48.92 months, SD = 22.52 for ABA; M = 52.06, SD = 16.11 for eclectic; p = .464).
There were no significant differences between the two treatment groups in gender (82% of ABA and 91% of eclectic children were males; Fisher's exact p = .350). The two groups were also similar in ethnicity (both parents were non-Hispanic Caucasians in 64% of ABA and 55% of eclectic families; chi-square p = .621) and in the marital status of the children's parents when treatment began (both parents were married in 82% of ABA and 68% of eclectic families; Fisher's exact p = .180). However, parents of ABA children averaged one more year of formal education (M = 14.23, SD = 2.26) than parents of eclectic children (M = 13.02, SD = 1.58; t-test p = .005).
Cognition, motor skills, receptive and expressive language skills, and adaptive behavior (communication, self-help, social skills, and a composite measuring all three skills) were assessed at baseline and approximately annually thereafter for up to 7 more years. All assessments used established, age-appropriate tests detailed in Supplemental Digital Content 1 (http://links.lww.com/JAANP/A34). The examiners who conducted the follow-up assessments were not blind to treatment group assignment but were independent and not involved with treatment delivery.
Not all children were assessed every year, and in some years, only some domains were assessed for some children. The data reported here are the final assessment available for each domain for each child. These were made an average of 4 years after treatment began, although assessment of motor skills ended earlier because of the absence of age-appropriate standardized tests (Supplemental Digital Content 1, http://links.lww.com/JAANP/A34). Performance in each domain was quantified using standard scores when available; otherwise, developmental quotients were used (Howard et al., 2014).
Analyses were conducted to determine the likelihood of normal functioning in each domain and whether that likelihood differed with treatment type. This required dichotomizing each score as “normal” or “subnormal.” Normal functioning is often defined as the absence of intellectual disability (Flanagan et al., 2015). The Diagnostic and Statistical Manual (American Psychiatric Association, 2013) defines this as a standard score of at least 70. However, children with cognitive standard scores of 70–84 are routinely labeled “slow learners” and are at substantial risk of academic failure and of behavioral and emotional difficulties (Thompson, Lampron, Johnson, & Eckstein, 1990). Thus, “normal” was defined as the absence of delay, which dichotomized each child's standard score or developmental quotient as normal (≥85) or subnormal (<85) at baseline and at the final assessment. The Fisher exact test (Stata/MP version 13.1; StataCorp, 2013) was used to compare the proportion of children in each treatment group with normal final scores, excluding children whose scores were already normal at baseline.
Additional analyses of the dichotomized scores were also conducted. The percentage of children with normal final scores was higher in the ABA treatment group than in the eclectic treatment group for every domain. Accordingly, the number needed to treat (NNT) was calculated for ABA treatment, using eclectic treatment as the comparison. A separate NNT was calculated for each domain. This number estimated how many children in the eclectic group would need to receive ABA treatment, to expect one additional child to improve from subnormal at baseline to normal at the final assessment. These calculations included only those children who had subnormal baseline scores for the domain in question. The Bender method was used to determine 95% confidence intervals (CIs) for each NNT (Stata/MP version 13.1 procedure bcii; StataCorp, 2013).
Children were excluded from NNT calculations if they had normal baseline scores. However, the final scores for some of these excluded children were below normal, especially if they had received eclectic treatment. Accordingly, the number needed to harm (NNH) and the associated 95% CI were calculated for eclectic treatment, using ABA treatment as the comparison. For the motor domain, this calculation included all children with normal baseline motor scores. However, there were too few children with normal baseline scores in the other domains to calculate NNH for those domains individually. Instead, an overall NNH measure was calculated for the nonmotor domains by considering all scores that were normal at baseline. Children with several such scores contributed multiple data points to the NNH calculation; other children were excluded entirely if none of their baseline scores were normal.
At baseline, normal scores were common for motor functioning but rare in other domains (Figure 1). The percentage of children with normal baseline scores did not differ significantly between treatment groups for any domain. However, many children had normal final scores, particularly in the ABA group. In six domains—cognitive, expressive, receptive, adaptive composite, self-help, and communication—children were significantly more likely to improve from subnormal at baseline to normal at the final assessment if they had received ABA instead of eclectic treatment (Table 1). The percentage of children with normal final scores was also higher in the ABA than the eclectic group in the social and motor domains, but these differences were not significant.
Analyses of covariance described in Supplemental Digital Content 1 (http://links.lww.com/JAANP/A34) suggest that the superior outcomes for ABA treatment compared with eclectic treatment are not artifacts of differences in baselines scores for the two treatment groups, or the earlier age at diagnosis for ABA children, or the additional years of education acquired by the parents of ABA children. Even after controlling for these differences statistically, gains were significantly larger in the ABA treatment group than the eclectic treatment group.
The NNTs place these findings into perspective. For example, the NNT of 3.2 for the cognitive domain suggests that, had three children with subnormal baseline cognitive scores received ABA rather than eclectic treatment, one of those children would most likely have attained a normal final cognitive score, instead of remaining in the subnormal range. Similar expectations exist for the expressive, receptive, adaptive composite, self-help, and communication domains, which have NNTs ranging from 2.8 to 4.0. For every three or four children with a subnormal baseline score in these domains, providing ABA rather than eclectic treatment would be expected to result in one additional child attaining a normal final score.
Ideally, children with normal scores at baseline would remain in the normal range at the final assessment. However, this was not generally true for the eclectic treatment group: Fewer than half of the normal motor scores at baseline remained normal at the final assessment, and only one eighth of the normal baseline scores in the other domains remained normal (Table 2). By contrast, in the ABA group, 71% of the normal baseline motor scores remained normal at the final assessment, as did 75% of the scores in the other domains. These treatment group differences were not statistically significant for the motor domain (p = .102) but were for the other domains (p < .001). The NNH for the nonmotor domains implies that providing just two children with eclectic rather than ABA treatment would be expected to result in one additional child regressing from normal at baseline to subnormal at the final assessment.
A final set of analyses focused on the cognitive and adaptive composite standard scores because deficits in these two domains predict poorer functioning in adulthood (Howlin et al., 2014; Perry et al., 2017). Figure 2 illustrates baseline scores for both domains in the left panel and final scores in the right panel. The two treatment groups were comparable at baseline, with 82% of the eclectic children (gray circles) and 87% of the ABA children (black squares) scoring below normal in both the cognitive and adaptive composite domains (lower left quadrant in the left panel of Figure 2). After treatment, children in the eclectic group exhibited some improvements, but their gains were modest: only 3% attained normal functioning in both domains (upper right quadrant in the right panel), whereas 72% remained below normal in both domains. By contrast, 30% of those in the ABA group had normal final scores in both domains; only 28% remained below normal in both domains. These group differences are even more apparent in a dynamic version of Figure 2 available at https://vimeo.com/315600110.
Comparing the left and right panels of Figure 2 reveals that treatment affects cognition more than adaptive skills. Indeed, the mean adaptive composite score was 9.33 points higher than the mean cognitive score at baseline, but 11.15 points lower at the final assessment. Statistical analyses detailed in Supplementary Digital Content 1 (http://links.lww.com/JAANP/A34) confirm that cognitive scores improved more after treatment than did adaptive composite scores (p < .001). Applied behavior analysis treatment yielded greater gains across both domains than did eclectic treatment (p = .001); however, treatment had the same impact on the gap between the adaptive composite and cognitive scores in the eclectic group as it did in the ABA group (p = .67).
Parents should understand that the treatment decisions they make after an autism diagnosis affect the probability that their child will ultimately function within the normal range. Children respond differently to treatments, and a particular outcome cannot be guaranteed for any child who receives ABA or eclectic treatment. Even so, the data suggest that young children with autism who receive intensive ABA treatment are significantly more likely to test in the normal range of functioning on measures of cognitive, language, and adaptive skills than are children who receive an eclectic treatment based on a mixture of methods, including subtherapeutic dosages of ABA.
The findings for cognitive and adaptive functioning are particularly noteworthy because, at the final ages that were examined (7–8 years old), competency in these domains is predictive of adolescent and adult functioning (Howlin et al., 2014; Perry et al., 2017). Thus, parents who are unaware of, or unable to evaluate, the available evidence may choose an ineffective treatment in childhood whose consequences extend into adulthood.
An ASD diagnosis is hard for parents, but selecting the best treatment is also difficult (Association for Science in Autism Treatment, 2018). Decisions regarding treatment are complicated when parents look to health care providers for guidance and when those providers do not feel competent to guide treatment conversations. The power of personal testimonials and (mis)information on the Internet makes it imperative that NPs be informed about treatment options and their likely outcomes.
However, health care decisions are not always made rationally (Courtney, Spivey, & Daniel, 2014), and families need help realizing the impact of their treatment choices (Bonis, 2016). Decision aids can assist families in selecting treatments appropriately, as they have with diabetes and osteoarthritis (Stacey et al., 2017). Nurse practitioners may find it particularly useful to engage families with pictographs, which have been used by the Mayo Clinic (2019) and the UK's National Institute for Health and Care Excellence (2016) to create online tools that help patients decide whether to use statins.
Tools such as these can be used by NPs to guide discussions with families of children newly diagnosed with ASD. An example is a pictograph-based video at https://vimeo.com/315509078 that summarizes the cognitive outcomes for three studies of a total of 128 children who received 3 years of either ABA or eclectic treatment. Online tools, such as those provided by Visme (https://www.visme.co/make-infographics/), can easily be used by NPs to create similar videos or pictographs to illustrate the outcomes reported here or in other studies for specific domains of interest. Child- or family-specific decision aids can be developed in the future, after improved understanding of individual differences in treatment responses.
The children in this study received referrals, assessments, and treatment in community settings. Thus, the outcomes are likely to be similar to those that families will encounter under real-life conditions (Curran, Bauer, Mittman, Pyne, & Stetler, 2012). However, the findings of the present study are limited by the nonrandom assignment of children to treatment conditions and by irregular assessment intervals. Furthermore, parents were not blind to the intervention their children received.
These limitations are common in long-term studies of behavioral interventions and do not necessarily invalidate the results (Nathan & Gorman, 2015; Tiura, Kim, Detmers, & Baldi, 2017). As noted above, the differential outcomes that were observed in the two treatment groups cannot be attributed to baseline disparities, including differences in parental education and age at diagnosis. This suggests the superior outcomes associated with ABA treatment are not artifacts of the nonrandom assignment of children to treatments. There may yet be unidentified differences, with implications for outcomes, between families who choose ABA and families who choose eclectic treatments. However, in the absence of information regarding such possible confounds, NPs who facilitate discussions that culminate in parents choosing intensive ABA treatment should regard this as a logical, defensible choice given the available evidence.
Parents of children newly diagnosed with ASD need support when making treatment decisions. These decisions affect the likelihood their children will achieve normal functioning by school age and independent functioning in adolescence and adulthood. Nurse practitioners, together with other primary care providers, must be prepared to offer parents informed, evidence-based assistance in selecting treatment.
Eclectic treatment approaches for ASD are widely available and often appear compelling. However, the findings presented here suggest these approaches are less likely to help young children attain normal levels of functioning than intensive, comprehensive treatment based on ABA. Similar findings have been reported in other studies (Eldevik et al., 2009, 2010), providing guidance for NPs as they engage parents in discussions of treatment options for their children diagnosed with ASD.
American Psychiatric Association. (2013). Diagnostic and Statistical Manual of Mental Disorders (DSM-5®). Arlington, VA: Author.
Association for Science in Autism
Treatment. (2018). Making sense of autism
treatments: Weighing the evidence. Retrieved from https://www.asatonline.org/research-treatment/making-sense-of-autism-treatments-weighing-the-evidence
Bonis S. (2016). Stress and parents of children with autism
: A review of literature. Issues in Mental Health Nursing, 37, 153–163.
Centers for Disease Control and Prevention. (2019). Data & statistics on autism
spectrum disorder. Retrieved from https://www.cdc.gov/ncbddd/autism/data.html
Courtney M. R., Spivey C., Daniel K. M. (2014). Helping patients make better decisions: How to apply behavioral economics in clinical practice. Pharmaceutical Patent Analyst, 8, 1503–1512.
Crane L., Chester J. W., Goddard L., Henry L. A., Hill E. L. (2016). Experiences of autism
diagnosis: A survey of over 1000 parents in the United Kingdom. Autism
, 20, 153–162.
Curran G. M., Bauer M., Mittman B., Pyne J. M., Stetler C. (2012). Effectiveness-implementation hybrid designs: Combining elements of clinical effectiveness and implementation research to enhance public health impact. Medical Care, 50, 217–226.
Dillenburger K. (2011). The Emperor's new clothes: Eclecticism in autism
treatment. Research in Autism
Spectrum Disorders, 5, 1119–1128.
Dosreis S., Weiner C. L., Johnson L., Newschaffer C. J. (2006). Autism
spectrum disorder screening and management practices among general pediatric providers. Journal of Developmental & Behavioral Pediatrics, 27, S88–S94.
Elder J. H., D'Alessandro T. (2009). Supporting families of children with autism
spectrum disorders: Questions parents ask and what nurses need to know. Pediatric Nursing, 35, 243–253.
Eldevik S., Hastings R. P., Hughes J. C., Jahr E., Eikeseth S., Cross S. (2009). Meta-analysis of early intensive behavioral intervention for children with autism
. Journal of Clinical Child & Adolescent Psychology, 38, 439–450.
Eldevik S., Hastings R. P., Hughes J. C., Jahr E., Eikeseth S., Cross S. (2010). Using participant data to extend the evidence base for intensive behavioral intervention for children with autism
. American Journal on Intellectual and Developmental Disabilities, 115, 381–405.
Flanagan H. E., Smith I. M., Vaillancourt T., Duku E., Szatmari P., Bryson S., Georgiades S. (2015). Stability and change in the cognitive and adaptive behaviour scores of preschoolers with autism
spectrum disorder. Journal of Autism
and Developmental Disorders, 45, 2691–2703.
Hofer J., Hoffmann F., Bachmann C. (2017). Use of complementary and alternative medicine in children and adolescents with autism
spectrum disorder: A systematic review. Autism
, 21, 387–402.
Howard J. S., Sparkman C. R., Cohen H. G., Green G., Stanislaw H. (2005). A comparison of intensive behavior analytic and eclectic treatments for young children with autism
. Research in Developmental Disabilities, 26, 359–383.
Howard J. S., Stanislaw H., Green G., Sparkman C. R., Cohen H. G. (2014). Comparison of behavior analytic and eclectic early interventions for young children with autism
after three years. Research in Developmental Disabilities, 35, 3326–3344.
Howlin P., Savage S., Moss P., Tempier A., Rutter M. (2014). Cognitive and language skills in adults with autism
: A 40-year follow-up. Journal Child Psychology Psychiatry, and Allied Disciplines, 55, 49–58.
Johnson C. P., Myers S. M. (2007). Identification and evaluation of children with autism
spectrum disorders. Pediatrics, 120, 1183–1215.
Klintwall L., Eldevik S., Eikeseth S. (2013). Narrowing the gap: Effects of intervention on developmental trajectories in autism
, 19, 53–63.
Mayo Clinic. (2019). Statin choice decision aid. Retrieved from https://statindecisionaid.mayoclinic.org/
Miller V. A., Schreck K. A., Mulick J. A., Butter E. (2012). Factors related to parents' choices of treatments for their children with autism
spectrum disorders. Research in Autism
Spectrum Disorders, 6, 87–95.
Nadel S., Poss J. E. (2007). Early detection of autism
spectrum disorders: Screening between 12 and 24 months of age. Journal of the American Academy of Nurse Practitioners, 19, 408–417.
Nathan P. E., Gorman J. M. (2015). Preface. In Nathan P. E., Gorman J. M. (Eds.), A guide to treatments that work (4th ed.). (pp. ix–x). New York: Oxford University Press.
National Institute for Health and Care Excellence. (2016). Cardiovascular disease: risk assessment and reduction, including lipid modification. Retrieved from https://www.nice.org.uk/guidance/cg181/resources/patient-decision-aid-188102
Perry A., Koudys J., Prichard A., Ho H. (2017). Follow-up study of youth who received EIBI as young children. Behavior Modification, doi:10.1177/0145445517746916.
Piccininni C., Bisnaire L., Penner M. (2017). Cost-effectiveness of wait time reduction for intensive behavioral intervention services in Ontario, Canada. JAMA Pediatrics, 171, 23–30.
Rhoades R. A., Scarpa A., Salley B. (2007). The importance of physician knowledge of autism
spectrum disorder: Results of a parent survey. BMC Pediatrics, 7, 37.
Roane H. S., Fisher W. W., Carr J. E. (2016). Applied behavior analysis
as treatment for autism
spectrum disorder. Journal of Pediatrics, 175, 27–32.
Schreck K. A. (2014). Autism
, parents, and treatments for their children. In Patel V., Preedy V., Martin C. (Eds.), Comprehensive Guide to Autism
, (pp. 2283–2296). New York, NY: Springer.
Stacey D., Légaré F., Lewis K., Barry M. J., Bennett C. L., Eden K. B., Trevena L. (2017). Decision aids for people facing health treatment or screening decisions. The Cochrane Database of Systematic Reviews, 4, CD001431.
StataCorp. (2013). Stata Statistical Software: Release 13. College Station, TX: StataCorp LP.
Thompson R. J. Jr., Lampron L. B., Johnson D. F., Eckstein T. L. (1990). Behavior problems in children with the presenting problem of poor school performance. Journal of Pediatric Psychology, 15, 3–20.
Tiura M., Kim J., Detmers D., Baldi H. (2017). Predictors of longitudinal ABA treatment outcomes for children with autism
: A growth curve analysis. Research in Developmental Disabilities, 70, 185–197.
Todorow C., Connell J., Turchi R. M. (2018). The medical home for children with autism
spectrum disorder: An essential element whose time has come. Current Opinion in Pediatrics, 30, 311–317.
Virués-Ortega J. (2010). Applied behavior analytic intervention for autism
in early childhood: Meta-analysis, meta-regression and dose-response meta-analysis of multiple outcomes. Clinical Psychology Review, 30, 387–399.
Will D., Barnfather J., Lesley M. (2013). Self-perceived autism
competency of primary care nurse practitioners. Journal of Nurse Practitioners, 9, 350–355.