Article Summary

Contents lists available at ScienceDirect

Research in Autism Spectrum Disorders

journal homepage: www.elsevier.com/locate/rasd

A systematic review of factors related to parents’ treatment decisions for their children with autism spectrum disorders

Meghan Wilson⁎, David Hamilton, Thomas Whelan, Pamela Pilkington School of Psychology, Faculty of Health Sciences, Australian Catholic University, 115 Victoria Parade, Fitzroy VIC 3065, Australia

A R T I C L E I N F O

Number of reviews completed is 2

Keywords: Autism spectrum disorder ASD Treatment decisions Parents Systematic review

A B S T R A C T

Background: There are many treatment options for children with Autism Spectrum Disorder (ASD). Misinformation and easy access to ineffective treatments complicates the decision-making process for parents. Research on implicit factors (e.g., parent or child characteristics) and de- clared factors (e.g., parent-reported reasons) contributes to an understanding of what influences these decisions. Method: The aim of this systematic review was to examine the significance of factors associated with treatment selection. The review was conducted in accordance with the PRISMA protocol. Results: The search revealed 51 studies which contained data on implicit and/or declared factors associated with treatment selection. The data were tabulated by factor and synthesised. The severity of a child’s behavioural problems, parental stress, and parent beliefs about ASD were consistently identified as implicit factors associated with the use of particular treatments. A wide range of reasons for treatment choices were declared by parent respondents, including; the in- dividual needs of their child, recommendations from others, practical reasons (e.g., cost), child age, hope for recovery, hope for improvement, and concerns about side-effects. Conclusion: A better understanding of these factors will inform targeted educational approaches which encourage evidence-based practice and a more informed view of treatments not yet sup- ported by research.

1. Introduction

Following a diagnosis of Autism Spectrum Disorder (ASD), parents are encouraged to access an intervention for their child. This can be challenging given that there are many options. Green et al. (2006) identified 111 different treatments for ASD. The list included a wide range of options such as dietary interventions (e.g., restricted diets or vitamin supplements), other alternative therapies (e.g., detoxification treatments), educational or clinical approaches (e.g., Applied Behaviour Analysis programs or speech therapy), and combined programs (e.g., Floor Time). The commitment of resources (e.g., time or cost) and ease of implementation can vary greatly between approaches (Green, 2007). The selection of interventions is further complicated in that it is common for professionals to recommend treatments that are not evidence-based (Miller, Schreck, Mulick, & Butter, 2012) and the internet provides a forum for misinformation (Matson, Adams, Williams, & Rieske, 2013). Not surprisingly, choosing treatments can be overwhelming for parents. Exploring the reasons treatments are chosen is a worthwhile step in understanding the scope of this problem and developing meaningful strategies to assist with choice making. Therefore, the present review aimed to identify and understand the significance of factors associated with the selection of ASD treatments.

https://doi.org/10.1016/j.rasd.2018.01.004 Received 28 July 2017; Received in revised form 18 December 2017; Accepted 9 January 2018

⁎ Corresponding author. E-mail address: [email protected] (M. Wilson).

Research in Autism Spectrum Disorders 48 (2018) 17–35

Available online 03 February 2018 1750-9467/ © 2018 Elsevier Ltd. All rights reserved.

T

Intervention research has largely focussed on programs based on behavioural principles (e.g., ABA programs) or educational approaches (e.g., Treatment and Education of Autistic and Related Communication Handicapped Children) (Myers & Johnson, 2007). Such programs are implemented to teach new skills and address maladaptive behaviours. Behavioural interventions are supported by the best available evidence (Anagnostou et al., 2014; Myers & Johnson, 2007). Along with traditional intensive behavioural inter- ventions, there is emerging evidence for variations to these approaches, for example, developmental, play-based, or social skills interventions (Weitlauf et al., 2014). Yet, evidence-based treatments do not result in equal gains for every child, progress can be slow, and there is no expectation of a cure (Myers & Johnson, 2007).

The high prevalence of comorbidity in children with ASD (e.g., ADHD or intellectual disability) adds to the difficulty of choosing an appropriate intervention (Matson & Williams, 2015). Some common approaches used for children with ASD (e.g., restricted diets or drug treatments), may be warranted for comorbid problems, but are not currently recommended to treat the core features of ASD (National Institute for Health and Care Excellence, 2013).

Treatments outside of the realm of conventional practice (known as complementary and alternative medicine, CAM) continue to be used (Matson et al., 2013; Whitehouse, 2013). In addition, parents often access multiple treatments simultaneously. For example, Smith and Antolovich (2000) found that, of 121 children engaged in ABA therapy, parents reported accessing an average of seven additional treatments. Commonly used CAM treatments in the paediatric ASD population are the use of vitamins (e.g., vitamin B6/ Magnesium) and restrictive diets (e.g., a gluten-free/casein-free diet) (Levy & Hyman, 2008; Whitehouse, 2013). Other examples are detoxification treatments, mind-body practices, hyperbaric oxygen therapy and sensory integration therapies (Levy & Hyman, 2008; Whitehouse, 2013). CAM practices may be ineffective or pose unnecessary risks (e.g., nutritional imbalances) (Levy & Hyman, 2008; Whitehouse, 2013). Other concerns about using CAM include high financial costs and missing out on treatments supported by research (Matson et al., 2013).

It appears that the research evidence guiding professional practice is often not the driving force behind parent decisions (Matson & Williams, 2015). Indeed, many factors have been hypothesised to influence parents’ decisions about treatments. Implicit factors are those characteristics associated with the use of treatments, but not necessarily cited by parents as a reason for choosing a treatment. Parent demographics (e.g., education or age), child characteristics (e.g., age, gender or ASD severity), and family demographics (e.g., income or ethnicity) are examples of implicit factors that have been explored (Matson & Williams, 2015). Declared factors are reasons or influences that parents cite regarding their intervention choices. A systematic review of 16 studies (Carlon, Carter, & Stephenson, 2013) examined factors parents declared to have influenced treatment choices for their child with ASD. Recommendations (by health professionals or others) was the most cited reason for choosing a treatment. Other frequently declared factors included practical reasons (e.g., availability, accessibility, cost, time constraints, funding, and availability of other interventions), perception of pro- gress, use and perceived effectiveness of other interventions, needs of the child, research evidence, child’s resistance, side effects, and compatibility with other interventions (Carlon et al., 2013).

In a recent discussion paper, Matson and Williams (2015) identified concerns about the process of ASD treatment selection and highlighted the importance of researching parent decision-making. Both implicit and declared factors contribute to a complete understanding of why treatments are selected (Carlon et al., 2013). To date, there has been no systematic review incorporating both implicit and declared findings.

Knowledge of the relationship between implicit factors and treatment use may be useful in understanding the context in which parents choose treatments. If groups with specific characteristics are likely to choose particular treatments, this information could inform the development of targeted educational strategies. In some instances, factors that influence decision-making (e.g., beliefs about ASD) may be modifiable. Equally, the explanations provided by parents are key to understanding what is important or not important to their decision-making. The present systematic review of the literature was not limited to specific study designs. It aimed to synthesise (a) the implicit factors (e.g., child or family characteristics) significantly associated with the use of any treatment reported by parents for their children with ASD and (b) the reasons reported by parents of children with ASD to influence or explain their decision to use any treatment.

2. Method

A systematic search of the literature was conducted in accordance with the PRISMA guidelines (Moher, Liberati, Tetzlaff, & Altman, 2009). The review protocol was registered on the PROSPERO International prospective register of systematic reviews (Regis- tration number: CRD42016033955).

2.1. Inclusion and exclusion criteria

Included studies reported on factors associated with the use of treatments or declared reasons for selecting treatments for children with ASD. Included studies met the following criteria.

(a) Studies were published after 1993. This timeframe was selected to target studies where children were more likely to have been diagnosed under recent criteria and a similar range of treatments would have been available.

(b) Respondents were mothers, fathers, or the child's primary caregivers. (c) Children reported on in the studies had a primary diagnosis of ASD (as indicated by the mother, father, or primary caregivers or

independently confirmed). A study was excluded if it was specified that criteria prior to the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV) were used (i.e., DSM-III) or if it was not clear that a sample or sub-sample of the children

M. Wilson et al. Research in Autism Spectrum Disorders 48 (2018) 17–35

18

had a diagnosis of ASD. Comorbid conditions (e.g., intellectual disability or ADHD) often occur with ASD, thus it was expected that some children would have secondary diagnoses.

(d) Studies on services that are not intervention or treatment types (e.g., respite or recreational activities) or not chosen by parents (e.g., exclusively school-based interventions) were excluded.

(e) Review or discussion papers, meta-analyses, conference papers, case studies, and dissertations were excluded. (f) Studies on declared factors included in a review by Carlon et al. (2013) were excluded. Given that the review had similar

inclusion criteria to the current review, these studies were excluded to avoid replication.

2.2. Search strategy

A systematic search of the databases Medline, CINAHL, PsychINFO, ERIC, Scopus and Web of Science was first conducted in May of 2016 and repeated in December 2016. The search terms used were; (autis* or ASD or asperger*) AND (mother* or father* or parent* or family or families) AND (treat* or intervention* or therap*) AND (decision* or selection or choice or choose). The same strategy was used in each database. Relevant subject headings (MESH terms) were used in Medline, CINAHL and PsycINFO databases. Additional studies were identified through hand-searching the references and a forward citation search. The search strategy for Medline is included as Appendix A.

2.3. Quality assessment

All included studies were assessed for quality using the Standard Quality Assessment Criteria for Evaluating Primary Research Papers (Kmet, Lee, & Cook, 2004). A quality checklist for each included study was completed by the first author (MW). Checklists for 35% (18 studies; 14 quantitative and 4 qualitative) of the studies were completed by the last author (PP) to ensure accuracy. The studies for double rating were randomly selected using the random number generator function in Microsoft Excel. The initial inter-rater agreement was calculated by dividing agreed item scores by the total scores and multiplying by 100. Agreement was 85% for quantitative studies and 75% for qualitative studies. Any discrepancies in ratings were resolved through discussion and re-checking the papers in question.

2.4. Data extraction and synthesis

Data were extracted on study characteristics: publication year, design, data source, methodology, treatment type investigated, sample size, age of the children, and key findings. Data extraction was completed by the first author (MW), and 50% (26 papers) of studies were coded by the fourth author (PP) to ensure accuracy. The initial inter-rater agreement was calculated by dividing the number of agreed studies by the total number of studies checked and multiplying by 100. The agreement was 80.8%. Discrepancies in data extraction for five papers were resolved through discussion and re-checking the papers.

2.4.1. Implicit factors Studies were examined to identify all factors (e.g., child, parent or family characteristics) that were investigated for associations

with treatment use. Factors were tabulated according to frequency (i.e., number of papers in which they appeared). Any factors which appeared in fewer than three studies were listed, but the results were not extracted. Across studies, 20 implicit factors were identified.

Statistics with p < .05 were considered significant. For studies which presented more than one statistical analysis, the main analysis relevant to the factor was selected for the synthesis. If parents of children with ASD were a subsample, only data relevant to the subsample were extracted. When findings were included on services that are not clearly interventions or treatments (e.g., respite services, recreational activities, or title of professional) data were not extracted on these services. Data on school-based services or specific classes of medication were not extracted, since these treatments are not clearly chosen by parents. Data on use of any medication (in general) were extracted. The synthesis involved observing trends and providing a narrative overview of the sig- nificance of the associations between each factor and treatment use. Appendix B presents an overview by study of the p-values and odds ratios (where applicable) for each implicit factor.

2.4.2. Declared factors Declared factors were identified by tabulating reasons or influences for treatment choice cited by parents. For qualitative studies,

this was achieved by listing the key themes identified by authors. For survey studies, key themes or percentages relating to declared reasons were extracted. The most common reasons declared by parents across studies were presented in a narrative synthesis.

3. Results

3.1. Search results

The database search produced 1167 records. A further 475 records were identified through forward citation searching and hand- searching of references. With duplicates removed, 1034 records remained. Titles and abstracts were screened for eligibility by the first author (MW) and 25% of abstracts were screened by the third author (TW). The agreement between screeners was 92.4%. Discrepancies were resolved by checking the papers in question. The full texts of 147 studies were checked for eligibility by the first

M. Wilson et al. Research in Autism Spectrum Disorders 48 (2018) 17–35

19

author (MW) and 96 were excluded. Therefore, a total of 51 studies were included in the review. Fig. 1 presents a PRISMA flow chart summarising the phases of the search. Included studies are denoted in the reference list by an asterisk after the title.

3.2. Quality assessment

The Standard Quality Assessment Criteria for Evaluating Primary Research Papers consists of items designed to measure research quality. The scorer assigns a 2 (yes), 1 (partial) or 0 (no) for each item. A summary score for each study is calculated by totalling the item scores and dividing by the total possible score. Possible scores range from 0 to 1, with higher scores indicating higher meth- odological quality. Due to the exploratory nature of review, the assessment was completed to provide an overall indication of strengths and weaknesses in the literature and to identify quality issues that could be considered in future investigations. Exceeding a quality threshold or cut-off score was not a requirement for inclusion in the current review.

Records identified through database searching

(n = 1167)

S cr ee ni ng

In cl ud ed

E lig ib ili ty

Id en tif ic at io n

Additional records identified through other sources

(n = 475)

Records after duplicates removed (n = 1034)

Records screened (n = 1034)

Records excluded (n = 887)

Full-text articles assessed for eligibility

(n = 147)

Full-text articles excluded, with reasons

(n = 96)

No relevant factor or treatment (n = 39) Review or dissertation (n = 26) No parent respondents (n = 12) ASD not primary diagnosis (n = 10) Studies included in prior review (n = 7) Not in English (n = 2)

Studies included in qualitative synthesis

(n = 51)

Fig. 1. PRISMA flow-chart summary of search strategy and results.

Table 1 Characteristics of included studies (n = 51).

Study characteristics Number of studies (%)

Mean age of children with ASD Under 5 years 11 (21.6) 5–12 years 22 (43.1) 13–15 years 2 (3.9) Sample 18 or younger* 13 (25.5) Not specified 3 (5.9)

Sample size (N) < 50 8 (15.7) 50–249 20 (39.2) 250–499 10 (19.6) > 500 13 (25.5)

Treatment type investigated Conventional**/CAM*** 24 (47.1) CAM 16 (31.4) Conventional 5 (9.8) Medications 4 (7.8) Communication interventions 2 (3.9)

* Mean age not reported. ** Educational or behavioural therapies (including speech therapy and occupational therapy). *** Treatment approaches other than educational or behavioural therapies.

M. Wilson et al. Research in Autism Spectrum Disorders 48 (2018) 17–35

20

For the present review summary scores for quantitative studies (n = 45) ranged between 0.67 and 1.00. Most studies adequately specified an objective and design. In 67.7% of studies, sampling procedures were not well defined or were likely to have introduced bias (e.g., convenience sampling). Participant characteristics were well described in most studies. In 35.6% of studies the outcome measures were not well described (e.g., the categorisation of treatments was unclear). The majority of studies adequately reported the results and conclusions. Summary scores for qualitative papers (n = 6) ranged between 0.60 and 0.85. In most of these studies the research question and design were well described. Data collection procedures, analysis and use of verification strategies were suf- ficient for most studies. The majority of studies were rated less than adequate for items regarding sampling procedures, reflexivity of the account, and clarity of the conclusions. Appendix C provides the obtained quality summary scores for each study.

3.3. Description of included studies

All studies used a survey or interview to obtain information on treatment use from parents. Nine studies (18%) were based on retrospective survey or interview data. The search was limited to studies published after 1993, however, all of those included were published after 1999. A summary of included studies is provided in Table 1.

3.4. Implicit factors

3.4.1. Child factors 3.4.1.1. Age. Of the 25 studies which included child age as a variable, 13 (Alnemary, Aldhalaan, Simon-Cereijido, & Alnemary, 2017; Bowker, D’Angelo, Hicks, & Wells, 2011; Goin-Kochel, Myers, & Mackintosh, 2007; Memari, Ziaee, Beygi, Moshayedi, & Mirfazeli, 2012; Mire, Gealy, Kubiszyn, Burridge, & Goin-Kochel, 2015 ; Mire, Nowell, Kubiszyn, & Goin-Kochel, 2014; Mire, Raff, Brewton, & Goin-Kochel, 2015; Owen-Smith et al., 2015; Pringle, Colpe, Blumberg, Avila, & Kogan, 2012; Rosenberg et al., 2010; Salomone et al.,

Table 2 Summary of findings on the relationship between child characteristics and treatment use.

Child characteristic No. of studies Findings

Age 25 Mixed results Gender 17 NS* related to treatment use**

Diagnostic subtypes 11 Mixed results ASD severity 11 Mixed results Comorbidity 10 Mixed results Cognitive/adaptive behaviour 8 Mixed results Child medication use 5 Mixed results Time since diagnosis 5 Mixed results Age at diagnosis 4 Mixed results Challenging behaviour 3 Scores indicating challenging behaviour were associated with the use of CAM treatments

* NS = not significant. ** One study reported that girls were more likely to use mind-body treatments.

Table 3 Summary of findings on the relationship between parent characteristics and treatment use.

Parent characteristic No. of studies Findings

Education level 23 Mixed results Age 7 NS* associated with treatment use ASD beliefs 5 Associated with treatment use Marital status 4 Mixed results Stress 3 Associated with treatment use

* NS = not significant.

Table 4 Summary of findings on the relationship between family characteristics and treatment use.

Family characteristic No. of studies Findings

Ethnicity 14 Mixed results Income 11 Mixed results Location 4 Mixed results Family size 3 NS associated with treatment use*

Family member with ASD 3 NS associated with treatment use

* In one study, family size was associated with CAM when “spiritual healing” was later excluded from the analysis.

M. Wilson et al. Research in Autism Spectrum Disorders 48 (2018) 17–35

21

2016; Thomas, Ellis, McLaurin, Daniels, Morrissey, 2007; Witwer & Lecavalier, 2005) reported at least one significant association between age and treatment use. There were two trends which emerged across a number of these studies; older children were more likely to use drug treatments (Goin-Kochel et al., 2007; Memari et al., 2012; Mire et al., 2014; Mire, Raff et al., 2015; Rosenberg et al., 2010; Thomas, Ellis et al., 2007; Witwer & Lecavalier, 2005), and younger children were more likely to use behavioural or conventional interventions (Bowker et al., 2011; Goin-Kochel et al., 2007; Mire, Raff et al., 2015; Pringle et al., 2012; Salomone et al., 2016).

Another nine studies which focussed specifically on CAM (Bilgiç et al., 2013; Granich, Hunt, Ravine, Wray, & Whitehouse, 2014; Hanson et al., 2007; Levy, Mandell, Merhar, Ittenbach, & Pinto-Martin, 2003; McIntyre & Barton, 2010; Salomone, Charman, McConachie, & Warreyn, 2015; Winburn et al., 2014; Wong & Smith, 2006; Wong, 2009) reported no significant associations with child age. A further three studies (Dardennes et al., 2011; Irvin, McBee, Boyd, Hume, & Odom, 2012; Miller et al., 2012) reported no association between child age and any type of treatment, including both conventional and CAM interventions.

3.4.1.2. Gender. Gender was not associated with treatment use across 16 studies (Alnemary et al., 2017; Bilgiç et al., 2013; Granich et al., 2014; Hanson et al., 2007; Irvin et al., 2012; Levy et al., 2003; Memari et al., 2012; Owen-Smith et al., 2015; Patten, Baranek, Watson, & Schultz, 2013; Perrin et al., 2012; Rosenberg et al., 2010; Salomone et al., 2016; Valicenti-McDermott et al., 2014; Witwer & Lecavalier, 2005; Wong & Smith, 2006; Wong, 2009) investigating CAM, conventional or both. As an exception, Salomone et al. (2015) found that girls were more likely than boys to use mind-body practices (e.g., sensory integration therapy, auditory integration training, or massage). The authors noted that this finding should be interpreted with caution given that girls constituted a minority of the sample.

3.4.1.3. Diagnostic subtypes. The DSM-IV conceptualised different subtypes of ASD (i.e., autism, Aspergers and PDD-NOS). These subtypes are sometimes used as a proxy for the severity of the ASD traits. In eight studies (Bowker et al., 2011; Christon, Mackintosh, & Myers, 2010; Goin-Kochel et al., 2007; Green et al., 2006; Hanson et al., 2007; Perrin et al., 2012; Rosenberg et al., 2010; Thomas, Ellis et al., 2007) the use of particular treatments was associated with diagnostic category. A pattern emerged in four studies (Goin- Kochel et al., 2007; Green et al., 2006; Perrin et al., 2012; Thomas, Ellis et al., 2007) which all found that children with Asperger’s were less likely to have tried special diets, relative to children with autism. Another three studies (Bilgiç et al., 2013; Granich et al., 2014; Owen-Smith et al., 2015), which examined CAM use, found no association between diagnostic subtype and CAM.

3.4.1.4. ASD severity. In one study (Horovitz, Matson, & Barker, 2012) it was reported that a group of children using psychotropic medications had higher scores on a measure of ASD severity (i.e., the Baby and Infant Screen for Children with Autism Traits – BISCUIT, Part 1). In two studies (Christon et al., 2010; Hall & Riccio, 2012) it was found that use of CAM treatments was more frequent among children with higher ASD severity, measured by parents’ report of severity. The remaining eight studies which reported on this factor (Alnemary et al., 2017; Dardennes et al., 2011; Granich et al., 2014; Irvin et al., 2012; McIntyre & Barton, 2010; Memari et al., 2012; Patten et al., 2013; Pickard & Ingersoll, 2015) reported no association between ASD severity and treatment use (CAM or conventional).

3.4.1.5. Comorbidity. The presence of comorbid conditions, such as intellectual disability, ADHD, anxiety, depression, allergies or epilepsy, were examined for a relationship with treatment use in ten studies. Some studies reported significant associations between comorbidities and psychotropic medication use (Rosenberg et al., 2010; Zablotsky et al., 2015) or other treatments including CAM (Levy et al., 2003; Perrin et al., 2012; Thomas, Ellis et al., 2007; Valicenti-McDermott et al., 2014; Zablotsky et al., 2015). Both studies which examined medications reported that use was more likely when comorbidities were present. In other studies no association was found between comorbidities and CAM (Harrington, Rosen, Garnecho, & Patrick, 2006; Memari et al., 2012; Wong, 2009), or treatments in general (Alnemary et al., 2017).

3.4.1.6. Cognitive and adaptive behaviour. Scores on cognitive measures (e.g., Mullen Scales of Early Learning) or adaptive behaviour measures (e.g., Vineland Adaptive Behaviour Scales) were explored for associations with treatment use in eight studies. Three (Mire, Gealy et al., 2015; Mire et al., 2014; Witwer & Lecavalier, 2005) reported that treatment use was associated with cognitive or adaptive behaviour scores. One study found that lower scores on a cognitive scale was associated with the use of medication (Mire et al., 2014). Another reported that children with higher adaptive behaviour scores were less likely to use modified diets (Witwer & Lecavalier, 2005). Higher scores on a verbal cognitive scale were associated with the use of intensive behavioural therapy (Mire, Gealy et al., 2015). Other studies found that scores on cognitive or adaptive behaviour scales were not related to CAM use (Akins, Krakowiak, Angkustsiri, Hertz-Picciotto, & Hansen, 2014; McIntyre & Barton, 2010), private speech or occupational therapy (Irvin et al., 2012), or treatments in general (Carter et al., 2011; Patten et al., 2013).

3.4.1.7. Child medication use. Child medication use was associated with CAM use in four studies (Granich et al., 2014; Owen-Smith et al., 2015; Perrin et al., 2012; Salomone et al., 2015). In three of these investigations those taking prescription medications were more likely to use other CAM treatments, alternatively Perrin et al. (2012) reported that children taking prescription medications had a lower use of special diets. Another study, Valicenti-McDermott et al. (2014), reported that CAM use was not related to medication use.

3.4.1.8. Time since diagnosis. In two studies (Hanson et al., 2007; Salomone et al., 2016) an association between time since diagnosis and treatment use was reported. Hanson et al. (2007) found the likelihood of CAM use increasing with time since diagnosis. Salomone

M. Wilson et al. Research in Autism Spectrum Disorders 48 (2018) 17–35

22

et al. (2016) found that time since diagnosis predicted the use of behavioural, developmental, relationship-based and speech intervention. Two studies (Bilgiç et al., 2013; Valicenti-McDermott et al., 2014) reported no association between time since diagnosis and the use of CAM. Another investigation (Miller et al., 2012) reported no association between time since diagnosis and empirically supported treatments.

3.4.1.9. Age at diagnosis. In one study (Zuckerman, Lindly, & Chavez, 2017) it was reported that the use of a behavioural intervention was less likely and psychotropic medication was more likely amongst children diagnosed later in childhood (relative to those diagnosed early in childhood). Across three other studies, child’s age at diagnosis was not found to be associated with CAM use (Granich et al., 2014; Valicenti-McDermott et al., 2014) or treatment type in general (Alnemary et al., 2017).

3.4.1.10. Challenging behaviour. Scores on children’s behaviour scales were reported to be associated with a higher use of CAM treatments across three studies. Witwer and Lecavalier (2005) adopted the Nisonger Child Behaviour Rating Form, NCBRF (Aman, Tassé, Rojahn, & Hammer, 1996) and found that lower scores on the compliant/calm subscale and higher scores on the hyperactivity subscale were predictive of the use of psychotropic medication. No association was found between NCBRF scores and vitamins or supplement use. Perrin et al. (2012) found that higher total scores on the Child Behaviour Checklist (Achenbach & Rescorla, 2000) were associated with the use of CAM treatments. Valicenti-McDermott et al. (2014) reported that higher scores on the Aberrant Behaviour Checklist (Aman, Singh, Stewart, & Field, 1985) were associated with the use of CAM treatments. Table 2 summarises findings on child characteristics and treatment use.

3.4.2. Parent factors 3.4.2.1. Education level. In eight studies which focused on CAM, it was reported that children’s use was higher when parents had a higher level of education (Akins et al., 2014; Bilgiç et al., 2013; Hall & Riccio, 2012; Hanson et al., 2007; Owen-Smith et al., 2015; Patten et al., 2013; Salomone et al., 2015; Wong & Smith, 2006). In another three studies (Alnemary et al., 2017; Salomone et al., 2016; Thomas, Ellis et al., 2007) other associations were found between years of education and the use of specific treatments (e.g., one investigation reported that the use of a picture exchange system and hippotherapy was more likely when parents had a college education). In twelve studies (Al Anbar, Dardennes, Prado-Netto, Kaye, & Contejean, 2010; Dardennes et al., 2011; Granich et al., 2014; Harrington et al., 2006; McIntyre & Barton, 2010; Memari et al., 2012; Miller et al., 2012; Pickard & Ingersoll, 2015; Rosenberg et al., 2010; Valicenti-McDermott et al., 2014; Wong, 2009; Zuckerman, Lindly, Sinche, & Nicolaidis, 2015) there was no association between treatment use (CAM or conventional) and parent education level.

3.4.2.2. Age. Parent age was not associated with the use of conventional or CAM treatments across seven studies (Al Anbar et al., 2010; Alnemary et al., 2017; Dardennes et al., 2011; Miller et al., 2012; Valicenti-McDermott et al., 2014; Wong & Smith, 2006; Wong, 2009).

3.4.2.3. ASD beliefs. The Revised Illness Perception Questionnaire – Modified for Autism (IPQ-RA) was used to measure health beliefs about ASD in three studies (Al Anbar et al., 2010; Mire, Gealy et al., 2015; Zuckerman et al., 2015) and another two investigations (Bilgiç et al., 2013; Dardennes et al., 2011) enquired about parents’ beliefs regarding ASD aetiology. Three of these studies (Al Anbar et al., 2010; Bilgiç et al., 2013; Dardennes et al., 2011) found that some specific causal beliefs were related to the treatments that parents chose. For example, Bilgiç et al. (2013) found that genetic or congenital causal beliefs were related to a lower rate of CAM use and immunisation causal beliefs were related to more frequent CAM use. Three of the studies (Al Anbar et al., 2010; Mire, Gealy et al., 2015; Zuckerman et al., 2015) reported significant associations between other beliefs about ASD and treatment use. For example, Zuckerman et al. (2015) indicated that parents who considered ASD to be a lifelong condition were more likely to use psychotropic medications, while Mire, Gealy et al. (2015) found that parents who considered ASD to be a lifelong condition were less likely to use speech therapy as an intervention.

3.4.2.4. Marital status. In one study which focused on CAM, it was reported that parents who were married were more likely to access CAM for their children with ASD (Hall & Riccio, 2012). Another study (Owen-Smith et al., 2015) found a bivariate association between married parents and CAM use. Other studies found that parental marital status was not related to psychotropic medication use (Memari et al., 2012) or the uptake of the EarlyBird intervention program (Birkin, Anderson, Seymour, & Moore, 2008).

3.4.2.5. Stress. Parental stress has been measured with the Parenting Stress Index (Abidin, 1995) or the Questionnaire for Resources and Stress (Friedrich et al., 1983). Valicenti-McDermott et al. (2014) reported that higher levels parent stress were associated with a greater use of CAM. Similarly, Thomas, Ellis et al. (2007) found that higher parent stress was associated with the use of medication and supplements, the Picture Exchange Communication System (PECS) and hippotherapy. Irvin et al. (2012) found that parents with higher stress were more likely to utilise private occupational therapy for their child. Table 3 summarises the findings on the relationship between parent characteristics and treatment use.

3.4.3. Family factors 3.4.3.1. Ethnic background. Analyses regarding ethnicity typically investigated differences in treatment use between those of Caucasian, Hispanic, and African American family backgrounds. Studies reported associations between ethnicity and CAM (Akins et al., 2014; Levy et al., 2003; Valicenti-McDermott et al., 2014), psychotropic medication (Rosenberg et al., 2010; Zuckerman et al.,

M. Wilson et al. Research in Autism Spectrum Disorders 48 (2018) 17–35

23

2015) and other interventions (Birkin et al., 2008; Thomas, Ellis et al., 2007). In most investigations, children from minority groups were less likely to use CAM and other treatments. As an exception, Levy et al. (2003) indicated that children with a Latino background were more likely to use CAM treatments. In another five studies which focused on CAM treatments, ethnicity was not associated with CAM use (Granich et al., 2014; Hall & Riccio, 2012; Hanson et al., 2007; Harrington et al., 2006; Owen-Smith et al., 2015). A final two investigations (Irvin et al., 2012; Patten et al., 2013) found no association between ethnicity and any treatment (CAM or conventional).

3.4.3.2. Income. In three studies (Alnemary et al., 2017; Pickard & Ingersoll, 2015; Thomas, Ellis et al., 2007) income was related to treatment choice. Alnemary et al. (2017) reported that lower income was associated with using fewer non-medical treatments (e.g., ABA therapy or sensory integration therapy) and Thomas, Ellis et al. (2007) found that higher income was related to increased chances of accessing speech/language therapy. Pickard and Ingersoll (2015) reported that level of income predicted the use of evidence-based practices. Some studies reported that income was not associated with CAM (Granich et al., 2014; Harrington et al., 2006; McIntyre & Barton, 2010; Owen-Smith et al., 2015), psychotropic medication (Memari et al., 2012) or treatment in general (Miller et al., 2012; Patten et al., 2013; Zuckerman et al., 2015).

3.4.3.3. Location (urban/rural). Alnemary et al. (2017) found that those living in a major city used more non-medical treatments. Rosenberg et al. (2010) found that those living in larger metropolitan areas used less psychotropic medications (not significant in multivariate analysis). Another two studies (Birkin et al., 2008; Thomas, Ellis et al., 2007) found that urban or rural living was not associated with the use of treatment.

3.4.3.4. Family size. Family size (Bilgiç et al., 2013; Birkin et al., 2008; Wong, 2009) was not related to the use of any treatment across studies.

3.4.3.5. Family member with ASD. Having a sibling or other family member with ASD or DD (Levy et al., 2003; Valicenti-McDermott et al., 2014; Wong & Smith, 2006) was not associated with treatment use. Table 4 summarises findings on family characteristics and treatment use.

3.4.4. Factors not frequently examined across studies Factors which appeared in two or fewer studies were: vaccination status of the child, parent gender, location of treatments, ASD

knowledge, socio-economic status, knowledge of treatments, empirical support, immediacy of outcome, cost, availability, parent age at child's birth, CAM characteristics, US born parents or other, seeing another provider prior to intake, appointment wait time, number of services received, ABA hours, service hours, school hours, atypical behaviours, parent college major or occupation, in- surance type, ASD core features, age of problem onset, classroom type, progression of ASD, number of ER visits, sensory processing difficulties, social networks, country median income, identifying with a major treatment approach (e.g., ABA), and religion. Two studies which met inclusion criteria (Call, Delfs, Reavis, & Mevers, 2015; Thomas, Morrissey, & McLaurin, 2007) did not include any of the common factors included in the synthesis.

3.5. Declared factors

There were 11 studies which reported on factors declared by parents to influence treatment decisions for their children. In six of these investigations, a qualitative interview approach was used. The other five investigations surveyed parents as part of a larger interview or questionnaire. In total, there were seven factors that were reported on by three or more studies. Of these, four factors (recommendations, child’s individual needs, practicalities and side effects) were also identified as main findings in the recent review on parent-declared factors (Carlon et al., 2013). In addition, three new factors (hope for cure or recovery; child’s age; and hope for improvement) which were identified by only one or two studies in the previous review, emerged more prominently in the current review.

3.5.1. Child’s individual needs Individual child’s needs were identified by parents in four studies as an influential factor. Carlon, Carter, and Stephenson (2015)

asked parents to rate how important a variety of factors were in their early intervention decision-making. The particular needs of a child was rated as the most important in a list of provided factors. Two qualitative investigations (Finke, Drager, & Serpentine, 2015; Serpentine, Tarnai, Drager, & Finke, 2011) found that a child’s need was important to choosing communication interventions. An- other qualitative study (Hebert, 2014) found that the individual needs of a child influenced decisions made for treatments in general.

3.5.2. Recommendations Recommendations from others was reported to be important to parents’ treatment choices in four studies. Carlon et al. (2015)

reported that advice from therapists, service providers, teachers, doctors, other parents and friends and relatives were all rated important by parent participants. According to Wong (2009), 42.5% of parents took into account advice from family members and 32.5% of parents considered the advice of medical professionals. In two qualitative studies (Finke et al., 2015; Grant, Rodger, & Hoffmann, 2016), advice was revealed as a key theme.

M. Wilson et al. Research in Autism Spectrum Disorders 48 (2018) 17–35

24

3.5.3. Practicalities (affordability, availability and accessibility) In four studies, parents recognised the importance of the practicalities of treatment (e.g., affordability, availability and accessi-

bility) when making treatment decisions. Carlon et al. (2015) found that parents rated availability, funding, cost, and accessibility as important factors in their early intervention decision-making. In two qualitative studies (Hebert, 2014; Serpentine et al., 2011), cost was identified as a key theme. In another study (Tzanakaki et al., 2012), 20% of parents reported that the availability of the treatment was a part of their decision to pursue an intensive behaviour intervention for their child.

3.5.4. Cure or recovery Parents indicated that the hope for a cure was influential in their treatment decisions in four studies. According to Provenzi,

Saettini, Barello, and Borgatti (2016), 58.1% parents chose CAM treatments hoping that they would bring about a cure for ASD. Similarly, Carlon et al. (2015) reported that parents rated hope for a cure as an important factor in their early intervention decision- making. Two qualitative studies (Finke et al., 2015; Hebert, 2014) identified hope for a cure as a key theme.

3.5.5. Child age In three studies child age was identified as a factor relevant to choosing treatments. Parents in one investigation (Carlon et al.,

2015) rated child age as important to their early intervention decision-making. Two qualitative studies (Hebert, 2014; Serpentine et al., 2011) identified child’s age as a key theme.

3.5.6. Hope for improvement In three studies hope for improvement was identified as an important factor. In one study (Carlon et al., 2015), parents rated hope

that the intervention will work as important in their early intervention decisions. Finke et al. (2015) identified hope for improvement as a key theme for choosing communication interventions in a qualitative investigation. Tzanakaki et al. (2012) reported that 16.7% of parents in their sample identified hope for their child as part of their reason for pursuing an early intensive intervention program.

3.5.7. Concerns about side effects Concerns about the side effects of other treatments appeared in three studies. Carlon et al. (2015) reported that parents rated

consideration of side effects as an important factor in their early intervention decision-making. In contrast, two studies found only a relatively small number of parents concerned about this factor. Wong (2009) reported that 12.5% of parents hoped that CAM would lower the toxicity of conventional medicine. Bilgiç et al. (2013) indicated that only 6% of parents chose CAM treatments to avoid the side effects of pharmacotherapy.

3.5.8. Factors not frequently examined across studies A number of declared factors were cited by parents in two or fewer studies. These factors were: empowerment, confidence, self-

reliance, resourcefulness, wanting to do anything that might help, parenting style, parents’ intuition, parents’ personal experiences, preference for natural therapies, perceptions of ASD, child enjoyment, ideas about how children learn, better outcomes, improving general health, relaxation, to address particular symptoms, comorbidities, to integrate, enhancing conventional treatments, quality of life, choosing a familiar intervention, trial and error, staff attributes, causal beliefs, lack of improvement with conventional treat- ments, program philosophy, service characteristics, ASD specific programs, program intensity, commitment required, specific in- formation sources, perceived effectiveness, and compatibility with other treatments. There were two studies (Edwards, Brebner, McCormack, & MacDougall, 2016; Granich et al., 2014) that met inclusion criteria, but did not examine any of the synthesised common factors.

4. Discussion

The aim of this systematic review was to synthesise factors associated with parents’ selected treatments for their children with ASD. A search of the literature identified 51 studies which examined implicit or declared factors related to treatment choice.

4.1. Implicit factors

There are three factors, child challenging behaviour, parental stress, and parents’ beliefs about ASD, that were consistently associated with treatment use. Mixed findings emerged for most other implicit factors, making it difficult to draw conclusions about their role in treatment decisions.

Challenging behaviour was related to psychotropic medication use (Witwer & Lecavalier, 2005) and the use of CAM in general (Perrin et al., 2012; Valicenti-McDermott et al., 2014). Interestingly, conceptually similar factors (ASD severity and diagnostic subtype) were not consistently associated with any treatments. It may be that it is not the severity of ASD specific traits that lead parents to select alternative treatments, but instead, challenging behaviours in general. Parents may not necessarily be targeting core ASD features (e.g., social-communication impairments and repetitive behaviours) through intervention. This notion is supported by Granich et al. (2014) who reported that parents most often chose CAM to treat non-core ASD symptoms (e.g., hyperactivity or aggression) rather core ASD features. Similarly, (Green, 2007) asked 14 parents about their child’s experience of using a combination of vitamin B6 and magnesium and noted that four of the parents were mainly using the treatment for health reasons and did not necessarily consider it a treatment for ASD.

M. Wilson et al. Research in Autism Spectrum Disorders 48 (2018) 17–35

25

Parental stress is another factor found to be associated with a higher likelihood of using specific treatments, including conven- tional and CAM (Irvin et al., 2012; Thomas, Ellis et al., 2007; Valicenti-McDermott et al., 2014). Matson and Williams (2015) identified that parents may feel anxiety about choosing treatments and an urge to try anything that might be helpful. This approach can lead to accessing a number of treatment options simultaneously.

Causal beliefs about ASD (Al Anbar et al., 2010; Bilgiç et al., 2013; Dardennes et al., 2011) were found to be related to treatment use. Some beliefs were related to the likelihood of choosing conventional treatments. For example, Dardennes et al. (2011) found that parents who endorsed early trauma as a causal factor were less likely to use behaviour therapy and PECS. Other beliefs were associated with CAM use, such as Bilgiç et al., (2013) who found that the rate of CAM was lower in parents who suspected the causal role of genetic factors and higher for those who held immunisation casual beliefs. Additionally, beliefs about the course of ASD (e.g., belief that ASD is chronic) were found to be associated with the choice of specific treatments (Al Anbar et al., 2010; Mire, Gealy et al., (2015); Zuckerman et al., 2015). There were different associations presented in each study and no clear pattern emerged. Further research is warranted to explore the influence of specific beliefs to understand the overall impact of beliefs on decision-making.

Of note, these three factors relate to the experience of parents (i.e., parental stress, beliefs and perceptions of their child’s behaviour). Since parents are the primary decision-makers in their child’s treatment, it makes sense that their experience would be related to their chosen treatments. In addition, these three factors are modifiable. The potential to make positive change in these areas has implications for guiding parents with decision-making. Recognising when parents are under stress and providing appro- priate supports might help parents to receive accurate information about treatment options. Identifying and discussing misconcep- tions about ASD could lead to more informed treatment choices. Further, discussing parents’ concerns about challenging aspects of their child’s behaviour may lead to a better understanding of parents’ priorities when selecting treatments.

The findings related to child challenging behaviour, parent stress and parents’ beliefs about ASD should be considered pre- liminary, since these factors were only investigated in a small number of studies (n = 3–5). It is also important to consider that the direction of the relationship is not established by these findings (e.g., it could be that accessing a particular intervention results in higher parental stress). Nevertheless, the pattern of findings suggests that parent perceptions are associated with treatment choice and play an important role in decisions.

For the majority of implicit factors (i.e., child age, diagnostic subtype, ASD severity, comorbidities, cognitive/adaptive behaviour, child medication use, time since diagnosis, age at diagnosis, parent education level, marital status, ethnicity, income and location) the findings were mixed. Even so, there were some factors (i.e., parent age, child gender, family size, and having a family member with ASD) that were not associated with treatment selection across studies. Overall, these findings suggest that it is almost impossible to predict which families are more likely to choose CAM treatments.

4.2. Declared factors

Across studies, seven main factors were declared by parents as instrumental in their treatment choice. Four of the most commonly cited factors (i.e., recommendations, practicalities, needs of the child, and side effects) were also identified as important in a previous review (Carlon et al., 2013). This indicates that these are relatively stable factors in parent decision-making.

Child age emerged as a declared factor in the current review. In contrast, as an implicit factor, the findings on the relationship between child age and treatment use were mixed. This finding suggests that parents consider their child’s age when selecting an intervention, but whether this consideration leads to differences in treatment use is less clear. A trend noted among some studies was that families with younger children were more likely to use conventional treatments and families with older children tended to favour drug treatments. It could be that parent decision-making changes as children grow. Parents of older children may have exhausted certain treatment options, noticed a change in their child’s needs, or discovered a new treatment type that seems promising. Understanding the relationship between child age and treatment choices warrants further investigation since it is important to ensure that evidence-based practices remain a priority as children grow into adolescents and adults.

Hope for improvement and hope for a cure were cited as common reasons for choosing treatments in the present review. In a previous review (Carlon et al., 2013) these factors were only identified in one unpublished study. These factors may indicate that parents focus on anticipated outcomes when they choose treatments. It appears that it would be helpful for clinicians to explore parent hopes during times of intervention decision-making. Green (2007) investigated parents experience of using treatments with varying levels of empirical support (i.e., ABA, sensory integration and vitamin B6-Mg), and found that expectations varied between treatments. For example, parents using sensory integration with their child had hopes specifically related to improving their child’s sensory experience. Across all types of treatments, some parents had specific hopes (e.g., “I wanted my child to learn to hold a conversation”) whereas others had very general hopes (e.g., “I wanted improvement”). When clinicians understand what parents hope to achieve from an intervention, they might be better able to communicate the way that the intervention works, set goals for desired outcomes, and manage expectations.

Overall, the findings on declared factors in the present review revealed that parents cited diverse reasons for choosing treatments and many reasons were cited in two or fewer papers. There is scope for future research to explore what parents prioritise when making treatment decisions. Given the wide range of factors considered by parents, it would be beneficial for clinicians to discuss treatment choice in the context of each family’s individual situation (e.g., their resources, perceived needs, hopes and expectations of outcome).

M. Wilson et al. Research in Autism Spectrum Disorders 48 (2018) 17–35

26

4.3. Limitations & strengths

Methodological limitations within studies were revealed by the quality assessment and should be considered when interpreting these results. In many studies, convenience samples were used. Partly due to the use of internet survey methods in many studies, the diagnosis of children with ASD was often based on parent report and not independently confirmed by researchers.

Many studies analysed broad categories of treatment type (e.g., CAM), rather than specific treatment modalities. The measure- ment of outcome variables was not clear in some cases and it may be that some measures resulted in under or over-reporting of treatments used. In future investigations, it would be helpful to ask parents about the treatments their child has used and additionally present a list of options to review. In some instances, parents may not recall all of the approaches that have been tried or they may not consider a non-clinical approach (e.g., taking vitamins) as a “treatment”. Given that there can be overlap and confusion pertaining to names of ASD treatments, it is also worth ensuring that parents have an accurate understanding of the treatment type, perhaps by providing a description.

There are limitations which apply to the synthesis of the current review. Methodologies varied substantially among the included studies. First, the definition and categorisation of treatments varied across studies. For example, two studies (Hanson et al., 2007; Valicenti-McDermott et al., 2014) categorised sensory integration therapy as a conventional treatment, due to its general acceptance and wide use. A second limitation is that studies varied in the way that information about treatment use was obtained and the way that “treatment use” was operationalized (e.g., current use verses ever used). Although treatments not clearly chosen by parents (e.g., school-based treatments) were not included in this review, in some studies the location of delivery was not specified. It was also not possible to ascertain the degree of choice parents had when selecting treatments. Treatments may have been selected because they were the only ones available. As a consequence, the synthesis was only able to explore broad trends in the available literature and a quantitative or meta-analysis was not possible.

Many implicit factors (e.g., child and family characteristics) have been explored in the existing literature, however, they have not previously been investigated in the context of a systematic review. The methodology used for this paper has provided an important contribution by ensuring that the available data on both implicit and declared decision-making factors was located, evaluated and synthesised. The strength of this approach is that it has resulted in a comprehensive examination of all factors that have been found to be associated with the use of a diverse range of treatments. The breadth of information resulting from this work will be helpful both to support parent decision-making and to extend the related research.

4.4. Future research

In order to understand the impact of factors with mixed findings in relation to treatment use (e.g., parent education or child age) it would be useful for future systematic reviews to adopt a narrower focus (e.g., an investigation of ASD symptom severity and treatment use).Further exploration of the findings of this review could be achieved by examining the role of child challenging behaviour, parent ASD beliefs, and stress in treatment selection. This could involve investigation of the relative and combined impact of these factors on decision-making. Mediating and moderating effects between factors could be explored to obtain more specific information on how these relationships function. For example, perhaps child age is only associated with treatment choice within a diagnostic subtype, or the combined impact of co-morbidities and low cognitive scores could lead to particular choices. Models aimed to explain the choice of particular treatments could be hypothesised and tested. In particular, the relative impact of the child’s presentation (e.g., age, level of functioning) and the attributes of the parent as the decision-maker (e.g. parent cognitions, beliefs and stress) could be explored.

There are many factors (both implicit and declared) that were identified in very few studies (two or fewer) and were not included in the synthesis. In terms of child and family factors, two areas that seem to be prevalent are the presentation of the ASD (e.g., age of onset, observed features) and parents’ approaches to decision-making (e.g., problem solving approach, resilience). In terms of de- clared factors, future investigations could identify the attributes that parents are looking for in a service (e.g., number of hours, staff attributes and physical environment).

Given the lack of research evidence and possible risk, it is unsurprising that many of the studies on parent decisions have focussed on CAM treatment. There are far fewer studies that examine parent decision-making regarding conventional treatments. A better understanding of how parents come to choose conventional, evidence-based interventions will be an important future direction. This knowledge can inform ways to encourage use of evidence-based approaches and thus increase the numbers of children receiving these interventions.

4.5. Conclusion

A systematic review of the literature identified that a number of implicit factors have been associated with parents’ treatment choices for their children with ASD. Factors relating to the experience of parents (i.e., child challenging behaviour, parental stress and beliefs about ASD) were associated with the use of particular treatments. Mixed findings were revealed for most implicit factors. Many reasons were identified by parents for their treatment choices including, child’s individual needs, recommendations, practi- calities of accessing treatment, child age, hope for cure, hope for improvement, and concerns about sideeffects. Knowledge of both implicit and declared factors is important to understanding treatment choice and has implications for educational approaches to support parents with this complex decision-making process.

M. Wilson et al. Research in Autism Spectrum Disorders 48 (2018) 17–35

27

Conflict of interest

None declared.

Acknowledgments

The first author (M. Wilson) received an Australian Government Research Training Program Scholarship. The funding source had no role in the study design, analysis or interpretation of data.

Appendix A

Medline search strategy

# Query

S16 S11 AND S14 (limit results 1994–2016) S15 S11 AND S14 S14 S12 OR S13 S13 TI (decision* OR selection OR choice OR choose) OR AB (decision* OR selection OR choice OR choose) S12 (MH “Decision Making”) OR (MH “Choice Behavior”) S11 S7 AND S10 S10 S8 OR S9 S9 TI (treat* OR intervention* OR therap*) OR AB (treat* OR intervention* OR therap*) S8 (MH “Early Intervention (Education)") S7 S3 AND S6 S6 S4 OR S5 S5 TI (mother* OR father* OR parent* OR family OR families) OR AB (mother* OR father* OR parent* OR family OR families) S4 (MH “Parents”) OR (MH “Single Parent”) OR (MH “Single-Parent Family”) OR (MH “Family”) OR (MH “Mothers”) OR (MH

“Fathers”) S3 S1 OR S2 S2 TI (autis* OR ASD OR asperger*) OR AB (autis* OR ASD OR asperger*) S1 (MH “Autism Spectrum Disorder”) OR (MH “Autistic Disorder”) OR (MH “Asperger Syndrome”)

Appendix B

See Table B1

Table B1 Key findings of included studies which reported on implicit factors (n = 41).

Study N Age in years, mean (SD) Key findings by study

Akins et al. (2014) 453^ 3.8 (0.82) Parent education: College degree – increased CAM, relative to parents without a degree (indicated in text only; statistic for total ASD/DD sample). Ethnicity: Hispanic ethnicity – lower CAM use, relative to those not of Hispanic ethnicity (indicated in text only; statistic for ASD/DD sample). NS: Cognitive/adaptive behaviour.

Al Anbar et al. (2010) 89 13.11 (IC 95% = 11.04–15.19) ASD beliefs: Higher beliefs in the seriousness of the disorder – increased odds of educative treatments (OR = 1.28**); higher beliefs in cyclic timeline – increased odds of drug treatments (OR = 1.27*); higher beliefs in personal control – lower odds of metabolic treatments (OR = 0.72**), special diets (OR = 0.83*), vitamins (OR = 0.77*), & drug treatments (OR = 0.81*); higher negative perceptions – lower odds of using PECs (OR = 0.84*) & educative treatments (OR = 0.84*); environmental attributions – lower odds of educative treatments (OR = 0.83**) & increased odds of metabolic treatments (OR = 1.38***), vitamins (OR = 1.33**), & special diets (OR = 1.33**); hereditary attributions – increased odds of metabolic treatments (OR = 1.50*) & vitamin supplements (OR = 1.62**). NS: parent education, parent age.

(continued on next page)

M. Wilson et al. Research in Autism Spectrum Disorders 48 (2018) 17–35

28

Table B1 (continued)

Study N Age in years, mean (SD) Key findings by study

Alnemary et al. (2017) 205 8.0 (3.5) Child age: An increase in child age – increased use of non-medical interventions (NMD)* & biomedical interventions (BMD)**. Parent education: Fathers with ≤ high school diploma – decreased BMD, relative to those with higher education**; mothers without college degree – increased cultural or religious treatments (CR), relative to higher education*. Income: Income below the sufficiency line – decreased use of NMD treatments*. Location: Residents of major cities – increased use of NMD, relative to residents of other cities*. NS: child gender, ASD severity, comorbidity, age at diagnosis, parent age.

Bilgiç et al. (2013) 172 8.8 (3.7) Parent education: Higher maternal & paternal education – increased CAM (p = 001 & p = .002) when ‘spiritual healing’ excluded from analysis. ASD beliefs: Genetic/congenital causal beliefs – lower CAM (p = .008); Immunization causal beliefs – higher CAM (p = .030). Family size: More children in the family – decreased CAM*** (when ‘spiritual healing’ was excluded from analysis). NS: child age, child gender, diagnostic subtype, time since diagnosis.

Birkin et al. (2008) 77 5.5 (3.2) Ethnicity: Ethnic minorities less likely to participate in the EarlyBird program (p = .0001). NS: marital status (family structure), location, family size.

Bowker et al. (2011) 970 0–5 (41%), 6–12 (46%), 13–18 (9.6%), > 18 (3.4%)

Child age: Early childhood – higher rate of standard therapies, skills training, ABA, physiological, alternative, & relationship-based treatments, relative to middle childhood, adolescents, & adults. Middle childhood – higher rate of skill-based treatments & medications, relative to early childhood, adolescents, & adults (indicated in text). Diagnostic subtype: AS group – lower rate of ABA***, vitamins, & detoxification treatments*, & higher rate of relationship-based treatments*** (relative to expected counts). Autistic group – higher rate of ABA***, & fewer relationship-based treatments***, (relative to expected counts). PDD-NOS group – higher rate of diets, relationship- based treatments, & detoxification* (relative to expected counts).

Carter et al. (2011) 84 3.5 (0.61) NS: Cognitive/adaptive behaviour (measure: Griffiths Mental Developmental Scales-Extended Revised).

Christon et al. (2010) 248 8.6 (4.4) Diagnostic subtype: Autism or PDD–NOS – tried more CAM, relative to AS (p = .004). ASD severity: Parent reported severe or moderate ASD – tried more CAM, relative to mild ASD***.

Dardennes et al. (2011) 78 13.5 (range: 2.3–44.5) ASD beliefs: Beliefs in chemical imbalance – increased odds of special diets (OR = 2.36*) & vitamins (OR = 2.48**); beliefs in illness during pregnancy – increased odds of using medications (OR = 2.76***); beliefs in brain abnormalities – lower odds of vitamins (OR = 0.45*); beliefs in early trauma – lower odds of using behaviour therapy (OR = 0.69*) & PECs (OR = 0.59**); genetic beliefs – increased odds of TEACCH (OR = 1.76*); food allergy beliefs – increased odds of chelation (OR = 4.27**), special diets (OR = 2.38**) & vitamins (OR = 2.29**) & lower odds of drug treatments (OR = 0.50**). NS: child age, ASD severity, parent education, parent age.

Goin-Kochel et al. (2007) 479 8.3 (4.3) Child age: Early childhood & middle childhood – more behavioural/ educational/alternative treatments, relative to adolescents***. Adolescents tried & used more drug treatments relative to middle childhood & early childhood***. Early childhood had tried more diets than older children***. Diagnostic subtype: Autism or PDD-NOS – had tried*** or were using more special diets, relative to AS (p = .029). AS had tried more drug treatments relative to autism (p < .02). AS/PDDNOS were using more drug treatments relative to autism. Autism had tried more diets than those with AS (p = .027). Autism & PDD-NOS had tried & were using more behavioural/educational/alternative treatments relative to AS***. Statistics for age & subtype group differences for specific treatments are also reported in paper.

Granich et al. (2014) 169 8.57 (4.8) (continued on next page)

M. Wilson et al. Research in Autism Spectrum Disorders 48 (2018) 17–35

29

Table B1 (continued)

Study N Age in years, mean (SD) Key findings by study

Child medication use: Psychotropic medication more among CAM users, relative to non-CAM users (p = .036). NS: child age, child gender, diagnostic subtype, ASD severity, age at diagnosis, parent education, ethnicity, income.

Green et al. (2006) 552 0–5 (34%), 6–10 (36%), 11–14 (18%), ≥15 (12%)

Diagnostic subtype: AS – lower use of standard therapies***, skills based therapies***, ABA therapies***, medications*, physiological therapies***, alternative diets***, relationship-based treatments*** & combined programs***, relative to autism.

Hall and Riccio (2012) 452 Child age not reported. ASD severity: Severity (parent reported) – predictive of total CAM used (p = .006) as well as the use of specific CAM (reported in paper)**. Parent education: Parents with a graduate degree – more likely to use CAM than those with technical school/some college (p = .02). Marital status: Married parents – more likely to use CAM, relative to divorced parents (p = .02). NS: ethnicity.

Hanson et al. (2007) 112 < 5 (17%), 5–10 (49%), > 10 (34%)

Diagnostic subtype: Children with GDD/MD & autism – higher CAM use relative to those with PDD-NOS or other***. Time since diagnosis: More years since diagnosis – increased chances of CAM use (p = .02; sig. in multivariate analysis only). Parent education: Higher maternal education – increased use of CAM (p = .04; sig. in univariate analysis only). NS: child age, child gender, ethnicity.

Harrington et al. (2006) 77 7.2 (range: 2–19) NS: comorbidity, parent education, ethnicity & income.

Horovitz et al. (2012) 78^ 2.3 (0.39) ASD severity: Those using psychotropic medication – higher severity, relative to no medication ASD group**.

Irvin et al. (2012) 137 3.97 (0.61) Parent stress: Higher level of stress – more likely to use private OT services (p = .031). NS: child age, child gender, ASD severity, cognitive/adaptive behaviour, ethnicity. Data on school-based services and dosage of therapy – not extracted.

Levy et al. (2003) 284 4.6 (2.6) Comorbidity: Children with comorbidities – lower odds (aOR = 0.3*) of CAM use, relative to those without. Ethnicity: Latino background – increased odds (aOR = 6.5*) of CAM use, relative to Caucasian reference group. NS: child age, child gender, family member with ASD.

McIntyre and Barton (2010)

73 4.6 (1.0) NS: Child age, ASD severity, adaptive behaviour, parent education, income (data on CAM use extracted).

Memari et al. (2012) 345 7–8 (39.8%), 9–10 (31.9%), 11–12 (20.4%), 13–14 (8.0%)

Child age: Increased odds (OR = 6.42*) of using 3 or more psychotropic medications concurrently in 11–12 years group, relative to 7–8 years. NS: child gender, ASD severity, comorbidity, parent education, marital status, income.

Miller et al. (2012) 400 9.0 (6.0) NS: child age, time since diagnosis, parent education, parent age, income.

Mire et al. (2014) 1605 8.7 (3.3) Child age: Child age – increased use of psychotropic medication***. Cognitive: Higher FSIQ – lower use of psychotropic medication***.

Mire, Gealy et al. (2015) 68 8.74 (3.7) Child age: As age increased – lower odds of biomedical treatments (OR = 0.789, p = .037). Cognitive: Higher verbal cognitive scores – lower odds of using intensive behavioural interventions (OR = 0.997, p = .013). ASD Beliefs: Attributing child symptoms to ASD – increased odds of behavioural interventions (OR = 1.321, p = .027) & lower odds of psychotropic medication (OR = 0.820, p = .037). Perceptions of control over treatment – increased odds of OT (OR = 1.328, p = .008), intensive interventions (OR = 1.609, p = .042), & psychotropic medications (OR = 1.494, p = .001). Believing ASD to be chronic – lower odds of speech therapy (OR = 0.792, p = .008). Only data on current study sample/main analysis extracted.

Mire, Raff et al. (2015) 2758 8.6 (3.6) Child age: 6 year olds – more likely to use private speech therapy***, private OT** & intensive behavioural treatment**, relative to older children (11 & 16 years). 11 year old & 16 year olds – more likely to use psychotropic medication** relative to 6 year olds.

(continued on next page)

M. Wilson et al. Research in Autism Spectrum Disorders 48 (2018) 17–35

30

Table B1 (continued)

Study N Age in years, mean (SD) Key findings by study

Owen-Smith et al. (2015) 1084 0–4 (9.2), 5–9 (34.2), 10–14 (37.9), 15–18 (18.6)

Child age: ≤4 years of age (aOR 3.20***) & 5–9 years (aOR = 1.97***) – increased CAM, relative to 15–18 year old group. Younger children – increased odds of using CAM products (0–4 years, aOR = 3.97***; 5–9 years aOR = 1.93**) relative to 15–18 group. Child medication use: Children using prescription medications – increased odds of CAM (aOR = 2.16***) & CAM products (aOR = 2.08***) relative to those not taking medications. Parent education: Graduate college – higher odds of CAM (aOR = 2.27*) & CAM products (aOR = 2.19**) relative to ≤ high school. NS: child gender, diagnostic subtype, marital status (sig. in bivariate analysis only), ethnicity, income.

Patten et al. (2013) 70 4.2 (1.4) Parent education: Higher education – increased use of gluten/casein free diets & vitamin therapy (maternal, p = .014 & paternal p = .042). NS: child gender, ASD severity, cognitive/adaptive behaviour, ethnicity, income.

Perrin et al. (2012) 3173 2–5 (56.4%), 6–11 (33.5%), 12–18 (10.2%)

Diagnostic subtype: AS or PDD-NOS – lower odds of CAM, relative to autism (ORs = 0.62* & 0.66*). PDD-NOS or AS – lower odds of special diets, relative to autism (ORs = 0.44* & 0.65*). PDD-NOS or AS – lower odds of other CAM, relative to autism (OR = 0.67* & 0.72*). Comorbidity: GI problems – increased CAM use (OR = 1.88*), special diets (OR = 2.38*), & other CAM (OR = 1.82*). Seizures – increased odds of CAM (OR = 1.58*), special diets (OR = 1.97*) & other CAM (OR = 1.66*). Child medication use: Reported psychotropic medication – lower odds of special diets (OR = 0.69*). Challenging behaviour: Higher challenging behaviour (CBCL score) – increased CAM (OR = 1.29*) & special diets (OR = 1.34*). NS: gender.

Pickard and Ingersoll (2015)

244 6.41 (2.57) Income: Income – predictor of evidence-based practices used**. NS: ASD severity, parent education.

Pringle et al. (2012) 1420 Range: 6–17 years Child age: Children 6–11 years – more likely to use speech therapy or OT, relative to those 12–17 years*.

Rosenberg et al. (2010) 5181 0–2 (.9%), 3–5 (27.3%), 6–11 (51.6%), 12–17 (20.1%)

Child age: 6–11 years & 12−17 years increased use psychotropic medications, relative to 3–5 years, (ORs = 2.4 & 4.4, respectively***). Diagnostic subtype: AS – more likely to use psychotropic medication**

(sig. in bivariate analysis only). Comorbidity: ID – increased odds of psychotropic medications (OR = 1.3, p = .012), relative to no ID. No comorbidity – lower odds of psychotropic medication use (OR = 0.3***), relative to any comorbidity. Ethnicity: Hispanic families – less likely to use psychotropic medication, relative to non-Hispanic families** (sig. in bivariate analysis only). Location: Residents of large metropolitan areas – less likely to be using psychotropic medication** (sig. in bivariate analysis only). NS: child gender, parent education.

Salomone et al. (2015) 1680 4.8 (1.2) Child gender: Male – lower odds of mind-body practices (OR = 0.68, p = 0.010). Child medication use: Increased odds of diets & supplements (OR = 1.62***). Parent education: Higher education – increased odds of diets & supplements (OR = 1.35, p = 0.013) & mind-body practices (OR = 1.64***). NS: child age.

Salomone et al. (2016) 1680 4.8 (1.2) Child age: Older children – decreased odds of behavioural, developmental & relationship-based interventions (OR = 0.98***). Time since diagnosis: ≥1 since diagnosis – increased odds of behavioural, developmental & relationship interventions (OR = 1.92***) & speech intervention (OR = 2.06***). Parent education: Higher education – increased odds of behavioural, developmental & relationship-based interventions (OR = 1.54***). NS: child gender (statistics for specific regions in Europe are also provided in paper).

Thomas, Ellis et al. (2007)

383 6.0 (1.8) Child age: ≤4 years – increased odds of supplements (OR = 2.24*), PECs (OR = 2.09*) & speech therapy (OR = 2.49*) & lower odds of

(continued on next page)

M. Wilson et al. Research in Autism Spectrum Disorders 48 (2018) 17–35

31

Table B1 (continued)

Study N Age in years, mean (SD) Key findings by study

medication (OR = 0.53*) & social skills training (OR = 0.38*), relative to 5–8 year olds. Children 9–11 years – lower odds of PECs (OR = 0.24*) & sensory integration therapy (OR = 0.38*). Diagnostic subtype: AS – increased medication (OR = 2.11*), lower odds of PECS (OR = 0.32*) & special diets (OR = 0.26*), relative to autism. Comorbidity: ID – increased odds (OR = 2.09*) of sensory integration therapy, relative to those with no ID. Parent education: College degree – increased odds of PECs (OR = 2.19*) & hippotherapy (OR = 3.93*), relative to high school. Parent stress: Stress – increased odds of medications (OR = 1.08*), supplements (OR = 1.07*), PECS (OR = 1.07*) & hippotherapy (OR = 1.10*). Income: Higher income – increased odds (OR = 2.49*) of speech therapy, relative to lower income. Ethnicity: Minority groups – lower odds of sensory integration therapy (OR = 0.25*), relative to Caucasian reference group. NS: location.

Valicenti-McDermott et al. (2014)

50* 8.8 (3.0) Challenging behaviour: Correlations between total CAM & child irritability***. Children who used ≥2 types of CAM were more likely to have Aberrant Behaviour Checklist irritability scores above the 85th percentile (p = .03) & hyperactivity scores above the 85th percentile**. Those who used CAM were more likely to have an irritability score > 85th percentile, relative to those who do not use CAM (p = .04). Comorbidity: Children with food allergies were more likely to use CAM, relative to those without food allergies**

Parent stress: Correlation between total CAM used and Parenting Stress Index score***. Ethnicity: Hispanic mothers reported using fewer types of CAM (p = .03) & non-Hispanic families – more likely to use ≥2 CAM types*. NS: child gender, time since diagnosis, age at diagnosis, child medication use, parent education, parent age, family member with ASD.

Winburn et al. (2014) 258 < 2.11 (2%), 3–5.11 (31%), 6–11 (67%)

NS: child age (indicated in text).

Witwer and Lecavalier (2005)

353 9.5 (3.9) Child age: Older age – increased odds of psychotropic medication (OR = 1.19***), younger age – increased odds of modified diet (OR = 0.78***). Adaptive behaviour: Higher scores on Scales of Independent Behaviour – lower odds of modified diet (OR = 0.48*). Challenging behaviour: Lower calm/compliant scores – increased odds of psychotropic medication (OR = −0.89*) & higher hyperactivity scores – increased odds of psychotropic medication (OR = 1.08***). Modified diet – lower insecure/anxious scores* (preliminary analysis). NS: child gender, (data on specific medication classes included in paper).

Wong (2009) 98^ 0– < 3 (3.1%), 3– < 5 (27.6%), 5– < 10 (48.0%), 10– < 15 (17.3%), 15– < 18 (4.1%)

NS: child age, child gender, comorbidity, parent age, parent education, family size.

Wong and Smith (2006) 50* 9 (range 14–17) Parent education: University degree, college or diploma – higher CAM use, relative to those with high school or less (indicated in text only; statistic reported for combined ASD & control group). NS: child age, child gender, parent age, family member with DD.

Zablotsky et al. (2015) 1420* 6–11 (54.8%), 12–17 (45.2%) Comorbidity: ASD & ID – increased use of medication**, sensory integration*, CBT***, physical therapy*, speech therapy*, relative to ASD only. Children with co-occurring psychiatric diagnoses in the ASD group – more likely to be using medications*.

Zuckerman et al. (2015) 1420 6–8 (20.9%), 9–11 (33.7%), 12–14 (25.6%), 15–17 (19.7%)

ASD beliefs: Beliefs that ASD is a lifelong condition – increased odds of using psychotropic medications (aOR = 1.89, p = .003) & beliefs that ASD is a mystery – lower odds of behaviour intervention (aOR = 0.66, p = .026). Ethnicity: Black (non-Hispanic) background – lower odds of using psychotropic medication (aOR = 0.41) & non-Hispanic background – lower odds of behavioural intervention (aOR = 0.37), indicated in text. NS: parent education, income.

(continued on next page)

M. Wilson et al. Research in Autism Spectrum Disorders 48 (2018) 17–35

32

Appendix C

See Table C1

Table B1 (continued)

Study N Age in years, mean (SD) Key findings by study

Zuckerman et al. (2017) 722 8.9 (1.5) Age at diagnosis: 4 years or older – higher use of psychotropic medication (aOR = 3.09***) & lower odds of behavioural intervention (aOR = 0.55, p = .039) relative to those diagnosed before 4 years. Older age at diagnosis (continuous variable) – increased use of psychotropic medication***.

AS = Asperger’s syndrome, PDD-NOS = Pervasive Developmental Disorder, Not Otherwise Specified, OR = odds ratio, aOR = adjusted odds ratio, NS = not sig- nificant. Note: only data on synthesised factors included in table.

^ ASD subsample. * p < .05. ** p < .01. *** p < .001.

Table C1 Quality assessment summary scores (n = 51).

Score range Author (year) Summary score Score range Author (year) Summary score

≥0.60 Finke et al. (2015) 0.60 ≥ 0.90 Owen-Smith et al. (2015) 0.90 Pringle et al. (2012) 0.90

≥0.65 Wong and Smith (2006) 0.67 Salomone et al. (2015) 0.90 Salomone et al. (2016) 0.90

≥0.70 Grant et al. (2016) 0.70 Al Anbar et al. (2010) 0.94 Hall and Riccio (2012) 0.72 Christon et al. (2010) 0.94 Winburn et al. (2014) 0.72 Dardennes et al. (2011) 0.94

McIntyre and Barton (2010) 0.94 ≥0.75 Edwards et al. (2016) 0.75 Mire, Gealy et al. (2015) 0.94

Tzanakaki et al. (2012) 0.75 Mire et al. (2014) 0.94 Birkin et al. (2008) 0.78 Patten et al. (2013) 0.94 Miller et al. (2012) 0.78 Provenzi et al. (2016) 0.94

Wong (2009) 0.94 ≥0.80 Serpentine et al. (2011) 0.80

Granich et al. (2014) 0.83 ≥ 0.95 Perrin et al. (2012) 0.95 Harrington et al. (2006) 0.83 Rosenberg et al. (2010) 0.95 Memari et al. (2012) 0.83 Zuckerman et al. (2015) 0.95 Pickard and Ingersoll (2015) 0.83 Carlon et al. (2015) 1.00 Valicenti-McDermott et al. (2014) 0.83 Green et al. (2006) 1.00 Witwer and Lecavalier (2005) 0.83 Zablotsky et al. (2015) 1.00

Zuckerman et al. (2017) 1.00 ≥0.85 Bowker et al. (2011) 0.85

Hebert (2014) 0.85 Mire, Raff et al. (2015) 0.85 Akins et al. (2014) 0.86 Call et al. (2015) 0.86 Alnemary et al. (2017) 0.89 Bilgiç et al., (2013) 0.89 Carter et al. (2011) 0.89 Goin-Kochel et al. (2007) 0.89 Hanson et al. (2007) 0.89 Horovitz et al. (2012) 0.89 Irvin et al. (2012) 0.89 Levy et al. (2003) 0.89 Thomas, Ellis et al. (2007) 0.89 Thomas, Morrissey et al. (2007) 0.89

M. Wilson et al. Research in Autism Spectrum Disorders 48 (2018) 17–35

33

References

Abidin, R. R. (1995). Parenting stress index: professional manual (3rd ed.). Odessa, FL: Psychological Assessment Resources, Inc. Achenbach, T., & Rescorla, L. (2000). Child behavior checklist. Burlington, VT: ASEBA. Akins, R. S., Krakowiak, P., Angkustsiri, K., Hertz-Picciotto, I., & Hansen, R. L. (2014). Utilization patterns of conventional and complementary/alternative treatments

in children with autism spectrum disorders and developmental disabilities in a population-based study*. Journal of Developmental and Behavioral Pediatrics, 35(1), 1–10. http://dx.doi.org/10.1097/DBP.0000000000000013.

Al Anbar, N. N., Dardennes, R. M., Prado-Netto, A., Kaye, K., & Contejean, Y. (2010). Treatment choices in autism spectrum disorder: The role of parental illness perceptions*. Research in Developmental Disabilities, 31(3), 817–828. http://dx.doi.org/10.1016/j.ridd.2010.02.007.

Alnemary, F. M., Aldhalaan, H. M., Simon-Cereijido, G., & Alnemary, F. M. (2017). Services for children with autism in the Kingdom of Saudi Arabia*. Autism, 21(5), 592–602. http://dx.doi.org/10.1177/1362361316664868.

Aman, M. G., Singh, N. N., Stewart, A. W., & Field, C. J. (1985). The Aberrant Behavior Checklist: A behavior rating scale for the assessment of treatment effects. American Journal of Mental Deficiency, 89, 485–491.

Aman, M. G., Tassé, M. J., Rojahn, J., & Hammer, D. (1996). The Nisonger CBRF: A child behavior rating form for children with developmental disabilities. Research in Developmental Disabilities, 17(1), 41–57.

Anagnostou, E., Zwaigenbaum, L., Szatmari, P., Fombonne, E., Fernandez, B. A., Woodbury-Smith, M., et al. (2014). Autism spectrum disorder: Advances in evidence- based practice. Canadian Medical Association Journal, 186(7), 509–519. http://dx.doi.org/10.1503/cmaj.121756.

Bilgiç, A., Cöngöloğlu, A., Hergüner, S., Türkoğlu, S., Bahali, K., Gürkan, K., et al. (2013). Use of complementary and alternative medicine in children with autism spectrum disorders: A multicenter study*. Archives of Neuropsychiatry, 50(3), 237–243. http://dx.doi.org/10.4274/Npa.y6389.

Birkin, C., Anderson, A., Seymour, F., & Moore, D. W. (2008). A parent-focused early intervention program for autism: Who gets access?*. Journal of Intellectual & Developmental Disability, 33(2), 108–116. http://dx.doi.org/10.1080/13668250802036746.

Bowker, A., D’Angelo, N. M., Hicks, R., & Wells, K. (2011). Treatments for autism: Parental choices and perceptions of change*. Journal of Autism and Developmental Disorders, 41(10), 1373–1382. http://dx.doi.org/10.1007/s10803-010-1164-y.

Call, N. A., Delfs, C. H., Reavis, A. R., & Mevers, J. L. (2015). Factors influencing treatment decisions by parents for their children with autism spectrum disorder*. Research in Autism Spectrum Disorders, 15-16, 10–20. http://dx.doi.org/10.1016/j.rasd.2015.04.001.

Carlon, S., Carter, M., & Stephenson, J. (2013). A review of declared factors identified by parents of children with autism spectrum disorders (ASD) in making intervention decisions. Research in Autism Spectrum Disorders, 7(2), 369–381. http://dx.doi.org/10.1016/j.rasd.2012.10.009.

Carlon, S., Carter, M., & Stephenson, J. (2015). Decision-making regarding early intervention by parents of children with autism spectrum disorder*. Journal of Developmental & Physical Disabilities, 27(3), 285–305. http://dx.doi.org/10.1007/s10882-014-9415-z.

Carter, M., Roberts, J., Williams, K., Evans, D., Parmenter, T., Silove, N., et al. (2011). Interventions used with an Australian sample of preschool children with autism spectrum disorders*. Research in Autism Spectrum Disorders, 5(3), 1033–1041. http://dx.doi.org/10.1016/j.rasd.2010.11.009.

Christon, L. M., Mackintosh, V. H., & Myers, B. J. (2010). Use of complementary and alternative medicine (CAM) treatments by parents of children with autism spectrum disorders*. Research in Autism Spectrum Disorders, 4(2), 249–259. http://dx.doi.org/10.1016/j.rasd.2009.09.013.

Dardennes, R. M., Al Anbar, N. N., Prado-Netto, A., Kaye, K., Contejean, Y., & Al Anbar, N. N. (2011). Treating the cause of illness rather than the symptoms: Parental causal beliefs and treatment choices in autism spectrum disorder*. Research in Developmental Disabilities, 32(3), 1137–1146. http://dx.doi.org/10.1016/j.ridd. 2011.01.010.

Edwards, A., Brebner, C., McCormack, P., & MacDougall, C. (2016). The early intervention message: Perspectives of parents of children with autism spectrum disorder*. Child: Care Health and Development, 43(2), 202–210. http://dx.doi.org/10.1111/cch.12428.

Finke, E., Drager, K., & Serpentine, E. C. (2015). ‘It’s not humanly possible to do everything’: Perspectives on intervention decision-making processes of parents of children with autism spectrum disorders*. Perspectives on Language Learning and Education, 22(1), 13–21. http://dx.doi.org/10.1044/lle22.1.13.

Friedrich, W., Greenberg, M., & Crnic, K. (1983). A short-form of the questionnaire on resources and stress. American Journal of Mental Deficiency, 88(1), 41–48. Goin-Kochel, R. P., Myers, B. J., & Mackintosh, V. H. (2007). Parental reports on the use of treatments and therapies for children with autism spectrum disorders*.

Research in Autism Spectrum Disorders, 1(3), 195–209. http://dx.doi.org/10.1016/j.rasd.2006.08.006. Granich, J., Hunt, A., Ravine, D., Wray, J., & Whitehouse, A. J. O. (2014). High use of complementary and alternative medication among children with autism is not

associated with the severity of core symptoms*. Journal of Autism, 1(1), http://dx.doi.org/10.7243/2054-992x-1-4 4. Grant, N., Rodger, S., & Hoffmann, T. (2016). Intervention decision-making processes and information preferences of parents of children with autism spectrum

disorders*. Child: Care, Health and Development, 42(1), 125–134. http://dx.doi.org/10.1111/cch.12296. Green, V. A., Pituch, K. A., Itchon, J., Choi, A., O'Reilly, M., & Sigafoos, J. (2006). Internet survey of treatments used by parents of children with autism. Research in

Developmental Disabilities, 27(1), 70–84. http://dx.doi.org/10.1016/j.ridd.2004.12.002. Green, V. A. (2007). Parental experience with treatments for autism*. Journal of Developmental and Physical Disabilities, 19(2), 91–101. http://dx.doi.org/10.1007/

s10882-007-9035-y. Hall, S. E., & Riccio, C. A. (2012). Complementary and alternative treatment use for autism spectrum disorders*. Complementary Therapies in Clinical Practice, 18(3),

159–163. http://dx.doi.org/10.1016/j.ctcp.2012.03.004. Hanson, E., Kalish, L. A., Bunce, E., Curtis, C., McDaniel, S., Ware, J., et al. (2007). Use of complementary and alternative medicine among children diagnosed with

autism spectrum disorder*. Journal of Autism and Developmental Disorders, 37(4), 628–636. http://dx.doi.org/10.1007/s10803-006-0192-0. Harrington, J. W., Rosen, L., Garnecho, A., & Patrick, P. A. (2006). Parental perceptions and use of complementary and alternative medicine practices for children with

autistic spectrum disorders in private practice*. Developmental and Behavioral Pediatrics, 27(2), 156–161 0196-206X/06/2702-0156. Hebert, E. B. (2014). Factors affecting parental decision-making regarding interventions for their child with autism*. Focus on Autism and Other Developmental

Disabilities, 29(2), 111–124. http://dx.doi.org/10.1177/1088357614522291. Horovitz, M., Matson, J. L., & Barker, A. (2012). The relationship between symptoms of autism spectrum disorders and psychotropic medication use in infants and

toddlers*. Research in Autism Spectrum Disorders, 6(4), 1406–1411. http://dx.doi.org/10.1016/j.rasd.2011.05.013. Irvin, D. W., McBee, M., Boyd, B. A., Hume, K., & Odom, S. L. (2012). Child and family factors associated with the use of services for preschoolers with autism spectrum

disorder*. Research in Autism Spectrum Disorders, 6(1), 565–572. http://dx.doi.org/10.1016/j.rasd.2011.07.018. Kmet, L. M., Lee, R. C., & Cook, L. S. (2004). Standard quality assessment criteria for evaluating primary research papers from a variety of fields. HTA Initiative #13. Alberta

Heritage Foundation for Medical Research1–21. Levy, S. E., & Hyman, S. L. (2008). Complementary and alternative medicine treatments for children with autism spectrum disorders. Child and Adolescent Psychiatric

Clinics of North America, 17(4), 803–820. http://dx.doi.org/10.1016/j.chc.2008.06.004 [ix]. Levy, S. E., Mandell, D. S., Merhar, S., Ittenbach, R. F., & Pinto-Martin, J. A. (2003). Use of complementary and alternative medicine among children recently

diagnosed with autistic spectrum disorder*. Journal of Developmental and Behavioral Pediatrics, 24(6), 418–423. http://dx.doi.org/10.1097/00004703-200312000- 00003.

Matson, J. L., & Williams, L. W. (2015). The curious selection process of treatments for autism spectrum disorders. Research in Autism Spectrum Disorders, 9, 21–25. http://dx.doi.org/10.1016/j.rasd.2014.10.004.

Matson, J. L., Adams, H. L., Williams, L. W., & Rieske, R. D. (2013). Why are there so many unsubstantiated treatments in autism? Research in Autism Spectrum Disorders, 7(3), 466–474. http://dx.doi.org/10.1016/j.rasd.2012.11.006.

McIntyre, L. L., & Barton, E. E. (2010). Early childhood autism services: How wide is the research to practice divide?*. Behavioral Development Bulletin, 10, 34–43. http://dx.doi.org/10.1037/h0100518.

Memari, A. H., Ziaee, V., Beygi, S., Moshayedi, P., & Mirfazeli, F. S. (2012). Overuse of psychotropic medications among children and adolescents with autism spectrum disorders: Perspective from a developing country*. Research in Developmental Disabilities, 33(2), 563–569. http://dx.doi.org/10.1016/j.ridd.2011.10.001.

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*.

M. Wilson et al. Research in Autism Spectrum Disorders 48 (2018) 17–35

34

Research in Autism Spectrum Disorders, 6(1), 87–95. http://dx.doi.org/10.1016/j.rasd.2011.03.008. Mire, S. S., Nowell, K. P., Kubiszyn, T., & Goin-Kochel, R. P. (2014). Psychotropic medication use among children with autism spectrum disorders within the simons

simplex collection: Are core features of autism spectrum disorder related?*. Autism, 18(8), 933–942. http://dx.doi.org/10.1177/1362361313498518. Mire, S. S., Gealy, W., Kubiszyn, T., Burridge, A. B., & Goin-Kochel, R. P. (2015). Parent perceptions about autism spectrum disorder influence treatment choices*. Focus

on Autism and Other Developmental Disabilities, 1–14. http://dx.doi.org/10.1177/1088357615610547. Mire, S. S., Raff, N. S., Brewton, C. M., & Goin-Kochel, R. P. (2015). Age-related trends in treatment use for children with autism spectrum disorder. Research in Autism

Spectrum Disorders, 15-16, 29–41. http://dx.doi.org/10.1016/j.rasd.2015.03.001. Moher, D., Liberati, A., Tetzlaff, J., & Altman, D. G. (2009). Preferred reporting items for systematic reviews and meta-analyses: The PRISMA statement. PLoS Medicine,

6(7), 1–6. http://dx.doi.org/10.1371/journal.pmed.1000097. Myers, S. M., & Johnson, C. P. (2007). Management of children with autism spectrum disorders. American Academy of Pediatrics, 120(5), 1162–1182. http://dx.doi.org/

10.1542/peds.2007-2362. National Institute for Health and Care Excellence (2013). Autism spectrum disorder in under 19s: Support and management – Clinical guideline [CG170]. Owen-Smith, A. A., Bent, S., Lynch, F. L., Coleman, K. J., Yau, V. M., Pearson, K. A., et al. (2015). Prevalence and predictors of complementary and alternative

medicine use in a large insured sample of children with autism spectrum disorders*. Research in Autism Spectrum Disorders, 17, 40–51. http://dx.doi.org/10.1016/j. rasd.2015.05.002.

Patten, E., Baranek, G. T., Watson, L. R., & Schultz, B. (2013). Child and family characteristics influencing intervention choices in autism spectrum disorders*. Focus on Autism and Other Developmental Disabilities, 28(3), 138–146. http://dx.doi.org/10.1177/1088357612468028.

Perrin, J. M., Coury, D. L., Hyman, S. L., Cole, L., Reynolds, A. M., & Clemons, T. (2012). Complementary and alternative medicine use in a large pediatric autism sample*. Pediatrics, 130(Suppl. 2), S77–S82. http://dx.doi.org/10.1542/peds.2012-0900E.

Pickard, K. E., & Ingersoll, B. R. (2015). From research settings to parents: The role of parent social networks in the choices parents make about services for their child with autism spectrum disorder*. Clinical Psychological Science, 3(2), 256–269. http://dx.doi.org/10.1177/2167702614534240.

Pringle, B. A., Colpe, L. J., Blumberg, S. J., Avila, R. M., & Kogan, M. D. (2012). Diagnostic history and treatment of school-aged children with autism spectrum disorder and special health care needs*. Hyattsville, MD: U.S. Department of Health and Human Services.

Provenzi, L., Saettini, F., Barello, S., & Borgatti, R. (2016). Complementary and alternative medicine (CAM) for children with special health care needs: A comparative usage study in Italy*. Journal of Integrative Medicine, 8(2), 115–121. http://dx.doi.org/10.1016/j.eujim.2015.06.005.

Rosenberg, R. E., Mandell, D. S., Farmer, J. E., Law, J. K., Marvin, A. R., & Law, P. A. (2010). Psychotropic medication use among children with autism spectrum disorders enrolled in a national registry, 2007–2008*. Journal of Autism and Developmental Disorders, 40(3), 342–351. http://dx.doi.org/10.1007/s10803-009- 0878-1.

Salomone, E., Charman, T., McConachie, H., & Warreyn, P. (2015). Prevalence and correlates of use of complementary and alternative medicine in children with autism spectrum disorder in Europe*. European Journal of Pediatrics, 174(10), 1277–1285. http://dx.doi.org/10.1007/s00431-015-2531-7.

Salomone, E., Beranova, S., Bonnet-Brilhault, F., Briciet Lauritsen, M., Budisteanu, M., Buitelaar, J., et al. (2016). Use of early intervention for young children with autism spectrum disorder across Europe*. Autism, 20(2), 233–249. http://dx.doi.org/10.1177/1362361315577218.

Serpentine, E. C., Tarnai, B., Drager, K. D., & Finke, E. H. (2011). Decision making of parents of children with autism spectrum disorder concerning augmentative and alternative communication in Hungary*. Communication Disorders Quarterly, 32(4), 221–231. http://dx.doi.org/10.1177/1525740109353938.

Smith, T., & Antolovich, M. (2000). Parental perceptions of supplemental interventions received by young children with autism in intensive behavior analytic treatment. Behavioral Interventions, 15, 83–97.

Thomas, K. C., Ellis, A. R., McLaurin, C., Daniels, J., & Morrissey, J. P. (2007). Access to care for autism-related services*. Journal of Autism and Developmental Disorders, 37(10), 1902–1912. http://dx.doi.org/10.1007/s10803-006-0323-7.

Thomas, K. C., Morrissey, J. P., & McLaurin, C. (2007). Use of autism-related services by families and children*. Journal of Autism and Developmental Disorders, 37(5), 818–829. http://dx.doi.org/10.1007/s10803-006-0208-9.

Tzanakaki, P., Grindle, C., Hastings, R. P., Hughes, J. C., Kovshoff, H., & Remington, B. (2012). How and why do parents choose early intensive behavioral intervention for their young child with autism*. Education and Training in Autism and Developmental Disabilities, 47(1), 58–71.

Valicenti-McDermott, M., Burrows, B., Bernstein, L., Hottinger, K., Lawson, K., Seijo, R., et al. (2014). Use of complementary and alternative medicine in children with autism and other developmental disabilities: Associations with ethnicity, child comorbid symptoms, and parental stress*. Journal of Child Neurology, 29(3), 360–367. http://dx.doi.org/10.1177/0883073812474489.

Weitlauf, A. S., McPheeters, M. L., Peters, B., Sathe, N., Travis, R., Aiello, R., et al. (2014). Therapies for children with autism spectrum disorder: Behavioral interventions update. comparative effectiveness review no. 137. (Prepared by the vanderbilt evidence-based practice center under contract No. 290–2012-00009-I.). Rockville, MD: Agency for Healthcare Research and Quality.

Whitehouse, A. J. (2013). Complementary and alternative medicine for autism spectrum disorders: Rationale, safety and efficacy. Journal of Paediatrics and Child Health, 49(9), 438–442. http://dx.doi.org/10.1111/jpc.12242.

Winburn, E., Charlton, J., McConachie, H., McColl, E., Parr, J., O’Hare, A., et al. (2014). Parents’ and child health professionals’ attitudes towards dietary interventions for children with autism spectrum disorders*. Journal of Autism and Developmental Disorders, 44(4), 747–757. http://dx.doi.org/10.1007/s10803-013-1922-8.

Witwer, A., & Lecavalier, L. (2005). Treatment incidence and patterns in children and adolescents with autism spectrum disorders*. Journal of Child and Adolescent Psychopharmacology, 15(4), 671–681. http://dx.doi.org/10.1089/cap.2005.15.671.

Wong, H. H., & Smith, R. G. (2006). Patterns of complementary and alternative medical therapy use in children diagnosed with autism spectrum disorders*. Journal of Autism and Developmental Disorders, 36(7), 901–909. http://dx.doi.org/10.1007/s10803-006-0131-0.

Wong, V. C. (2009). Use of complementary and alternative medicine (CAM) in autism spectrum disorder (ASD): Comparison of Chinese and western culture (Part A)*. Journal of Autism and Developmental Disorders, 39(3), 454–463. http://dx.doi.org/10.1007/s10803-008-0644-9.

Zablotsky, B., Pringle, B. A., Colpe, L. J., Kogan, M. D., Rice, C., & Blumberg, S. J. (2015). Service and treatment use among children diagnosed with autism spectrum disorders*. Journal of Developmental and Behavioral Pediatrics, 36(2), 98–105. http://dx.doi.org/10.1097/DBP.0000000000000127.

Zuckerman, K. E., Lindly, O. J., Sinche, B. K., & Nicolaidis, C. (2015). Parent health beliefs, social determinants of health, and child health services utilization among US school-age children with autism*. Journal of Developmental and Behavioral Pediatrics, 36(3), 146–157. http://dx.doi.org/10.1097/DBP.0000000000000136.

Zuckerman, K., Lindly, O. J., & Chavez, A. E. (2017). Timeliness of autism spectrum disorder diagnosis and use of services among U.S. elementary school-aged children*. Psychiatric Services, 68(1), 33–40. http://dx.doi.org/10.1176/appi.ps.201500549.

M. Wilson et al. Research in Autism Spectrum Disorders 48 (2018) 17–35

35