Abstract

In many lower-income countries, there is a paucity of assessment services for autism spectrum disorders (ASD)., Guidelines will be provided for conducting cross-cultural assessments in the context of limited validated resources in Tanzania. By examining behavioral, social, and adaptive differences we were able to provide differential diagnostic evaluations aligning with best practice standards for 41 children in Tanzania age 2–21 years. We describe the utility of a flexible, behavioral observation instrument, the Childhood Autism Rating Scales, Second Edition (CARS2), to gather diagnostic information in a culturally sensitive manner. We observed that the ASD group was characterized by significantly higher scores on the CARS2, F  =  21.09, p < .001, η2  =  .37, than the general delay comparison group. Additional recommendations are provided for making cultural adaptations to current assessment instruments for use in a country without normed instruments, such as Tanzania.

In developed nations, the assessment of autism spectrum disorders (ASD) has become highly advanced as a result of years of research and clinical practice. Procedures for ASD assessment, guided by research and outlined by groups such as the American Academy of Neurology (Filipek et al., 1999) and the American Academy of Child and Adolescent Psychiatry (Volkmar, Cook, Pomeroy, Realmuto, & Tanguay, 1999), help anchor practitioners to a set of clinical and research standards. Guidelines have been published establishing a minimal or sufficient ASD assessment battery (Ozonoff, Goodlin-Jones, & Solomon, 2005). Specifically, Ozonoff and colleagues outline that a sufficient, best-practice diagnostic assessment should include a thorough developmental history, an ASD-specific assessment (comprising both parent report and a minimally structured interaction and observation period with the child), intellectual and language assessments, and assessments of adaptive functioning. Despite this information, global disparities exist in the implementation of best-practice procedures for diagnosing ASD, rendering developing methods for disseminating these practices to low- and middle-income countries an important priority. This article reports on and tests the implementation of an observational ASD diagnostic assessment in Tanzania, a low-income nation with limited access to ASD-specific services.

Although there has been considerable research on the assessment and diagnosis of ASD in North America, Europe, and Asia, people in many low- and middle-income regions, such as many African communities, still do not have access to ASD assessment and diagnostic services (Elsabbagh et al., 2012). As a result, although the available research provides strong evidence that ASD are present in African nations (e.g., Bakare & Munir, 2011; Lotter, 1978; Mankoski et al., 2006), many people remain undiagnosed. This disparity exists in part because no guidelines or recommendations exist for assessing ASD in low-and middle-income countries where ASD diagnostic, cognitive, and adaptive measures are not widely available. Additional research is needed on the extension of ASD assessment and treatment to Africa; in turn, this article will focus on provision of ASD diagnostic services in the East African country, the United Republic of Tanzania.

Tanzania exemplifies a country that has many barriers to overcome before best-practice assessment (Ozonoff et al., 2005) and treatment (Rogers, 1998) services can be made readily available for children and families. Tanzania is among the small percentage of countries classified as a “heavily indebted poor country” in terms of world global economies (World Economic Outlook, 2012). In addition to economic limitations, a regional report provides evidence for a paucity of psychological services in Tanzania generally (Manji & Hogan, in press). Barriers to formal psychological assessments include low rates of mental health literacy among community members, restricted or no access to the nation's few psychologists and psychiatrists (e.g., less than 0.05 psychiatrists for every 10,000 people in Tanzania; World Health Organization, 2013), and a lack of assessment measures validated for use in Tanzania across cognitive, adaptive, and clinical diagnostic domains (Patel, 2007). At present there are no ASD diagnostic instruments validated for use in Swahili, a language that is spoken by millions of people in Kenya, Tanzania, and Uganda.

It is important that efforts be made to increase the validated translation and adaptation efforts in low and middle-income African countries using a systematic protocol (e.g., Bracken & Barona, 1991). Preliminary efforts have been made to validate ASD measures cross culturally in many developed nations (e.g., Bölte, Poustka, & Constantino, 2008; Fombonne, Marcin, Bruno, Tinoco, & Marquez, 2012; Pereira, Riesgo, & Wagner, 2008; Wong et al., 2004), yet more than 20 years after Bracken and Barona's suggestions, options remain limited in many African nations. This may be in part because validation efforts can be time intensive, involving multiple steps: (a) a multistep translation process involving source to target language translation, (b) blind back-translation, (c) back-translation repetition, (d) review by a bilingual committee, and (e) a range of additional validation procedures (e.g., pilot and field testing and regional norm development; Bracken & Barona, 1991).

Given that ASD measure translation is a costly and timely process, to provide services to families now in regions of need, clinicians in Tanzania, as described by K. Manji, a Professor in Pediatrics and Child Health at the Muhimbili University of Health and Allied Sciences in Dar Es Salaam, Tanzania (personal communication, January 13, 2013), often diagnose ASD by asking interview questions based on DSM-IV-TR criteria (American Psychiatric Association, 2000) or through reliance on measures that have not undergone formal translation and validation procedures (Mankoski et al., 2006). Given what we know about the importance of proper measure adaptation and translation, however, and research showing that increased diagnostic accuracy is achieved through the incorporation of both parent interview and diagnostic measures with a child observation component (Filipek et al, 1999; Volkmar et al., 1999), such as the Autism Diagnostic Observation Schedule (ADOS; Lord, Rutter, DiLavore, & Risi, 1999), the Autism Diagnostic Interview-Revised (ADI-R; Rutter, Le Couteur, & Lord, 2003) or the Childhood Autism Rating Scale (CARS; Schopler, Reichler, DeVellis, & Daly, 1980), these current practices are limiting.

One approach for managing the issue of insufficient measure translation in countries like Tanzania is to rely more heavily on judgments made by trained clinicians based on behavioral observation methods (Williams, Atkins, & Soles, 2009) and to use measures that do not require strict adherence to testing procedures (e.g., specific verbal prompts) and environmental settings to maintain psychometric validity. Measures with greater flexibility, which rely primarily on behavioral observation and are sufficiently flexible to allow for the use of materials with cultural relevance, may serve to minimize the impact of language and cultural differences thereby increasing diagnostic accuracy. This article suggests a method for the identification of ASD through the use of a combined parent interview and play-based interaction (see Klinger & Renner, 2010) designed to require limited language demands and with a mind for cultural sensitivity. Although measure translation and validation efforts are ongoing in Tanzania, without the existence of psychometrically supported measures adapted for use in Tanzania and available in Swahili at present, this article describes how best-practice ASD assessment methods can be implemented in low- and middle-income countries by trained clinicians.

The primary purpose of this article is to describe how, with minimal adaptations, existing measures with flexible administration guidelines can be utilized as part of an assessment battery for cross cultural use among non-English speaking families that adheres to best-practice standards. An example of the development and implementation of this assessment battery in Tanzania will help to highlight the feasibility and acceptability of such an approach in a country in Africa. With this approach, we can provide timely, rigorous, and targeted diagnostic assessment for families in need at present.

Method

Setting

Evaluations were conducted by an English-speaking clinician with autism assessment expertise across two clinical sites in Tanzania: the Gabriella Centre in Moshi, Tanzania, and a walk-in pediatric specialty clinic run by a consultant affiliated with Muhimbili University of Health and Allied Sciences (MUHAS) in Dar Es Salaam, Tanzania. The Gabriella Centre is a treatment facility designed to provide special education and intensive occupational therapy services for children with physical, behavioral, emotional, developmental, and intellectual disabilities. Educators and mental health workers in the Moshi community referred children to the clinic who exhibited symptoms consistent with ASD, such as language delays and repetitive motor behaviors. With regards to the MUHAS Medical Clinic, children considered to have ASD were referred to an assessment screening by the pediatric director. The directors of both clinics contacted and scheduled each participating family. Although the primary goal was clinical in nature, to allow for the clinical data to be reviewed for research purposes, informed consent was provided verbally to all families; and they were provided with a copy of the consent to take home.

Given the primary goal of the project was clinical service, and not research, this sample was obtained through a typical clinical referral process. Senior Tanzanian university students majoring affiliated with the two sites provided translation services for non-English-speaking families. The testing environment across sites consisted of a room with a mat on the floor, a small table, several chairs or benches, and a range of toys (see Figure 1 for an example).

Figure 1

Testing setting of the autism spectrum disorder (ASD) clinic that occurred at the Gabriella Centre in Moshi, Tanzania.

Figure 1

Testing setting of the autism spectrum disorder (ASD) clinic that occurred at the Gabriella Centre in Moshi, Tanzania.

Participants

Forty-one Tanzanian children completed a diagnostic evaluation to assess for the presence of ASD. Children were accompanied to the clinic by at least one family member or legal caregiver responsible for regular care: biological mother (63%), biological father (12%), sibling (10%), aunt or uncle (5%), grandparent (5%), or orphanage staff (5%). Families originated from a range of Tanzanian cities, including Mwanza, Nzega, Moshi, Arusha, and Dar Es Salaam. Participants ranged in age from 2 to 14 years with one 21-year-old (M  =  7.41, SD  =  3.67) and were primarily male (83%). All participating children were born in Tanzania except one child from Pakistan. Families from Tanzania represented 22 different tribal backgrounds. Patients had received a range of educational experiences (e.g., 12% no formal education, 5% home instruction only, 17% nonspecific short-term therapeutic instruction, 10% nonspecific long-term therapeutic instruction, 12% ongoing typical education, 5% discontinued typical education with no alternative schooling, and 39% special education).

Parent educational and occupational levels were used as an estimate of socioeconomic status. Parents varied greatly in their level of education. Specifically, the highest parental level of education in each family was as follows: no school (5%), primary school (27%), secondary school (17%), bachelor's degree (34%), and master's degree (12%). Parental level of education was not available for two children from orphanages (5%). The Hollingshead Scale was used to convert parental occupations to numerical data on a scale from zero to 9 with 9 reflecting the highest rating of occupational status (see Hollingshead, 1975 for further descriptions). Within the current study parents achieved the following ratings (in two-parent households, descriptives presented represent the highest rated profession in a household): 23% rating of 9, 13% rating of 8, 10% rating of 7, 8% rating of 6, 3% rating of 5, 18% rating of 4, 13% rating of 3, and 5% rating of 1. This information suggests that the parents who brought their children to the screening clinics were more highly educated than the majority of Tanzanians who complete primary school alone (National Bureau of Statistics, 2011).

Measures

Intake interview and developmental history

During the assessment, guardians first completed a brief intake interview designed to assess specific parental concerns, developmental milestones, medical and school history, and family structure (see  Appendix). The intake interview was used as a semi-structured method for obtaining information about the child's typical social, communication, and repetitive behaviors. The structure of this interview was derived from an early childhood intake questionnaire used at a university affiliated children's hospital in the United States, but was condensed and altered in an effort to minimize potentially culturally irrelevant elements. For example, given the diminished ASD services available in Tanzania (Manji & Hogan, in press), less time was devoted to questions regarding previous assessments, details of early interventions or other treatments, or information about individualized education programs as compared to North American assessments of children of similar age. Caregivers were asked generally about any medical concerns rather than inquiring about specific medical conditions, given concerns that many disease names may not be translatable or widely accepted in Tanzania (e.g., colic, asthma, lead poisoning, allergies). These kinds of modifications to typical intake assessments allowed for relevant and important information to be collected in a more efficient and culturally sensitive manner.

ASD diagnostic measure

Given the lack of ASD-specific diagnostic measures formally translated for use in Swahili, instrument selection was guided by identifying instruments that could be administered validly without requiring formal translation procedures. Research has shown that diagnostic errors are often made when using language based or interview measures outside of the cultural and language context for which they are normed (Magaña & Smith, 2012). As previously mentioned, relying on more observation-based methods may help to reduce some of the diagnostic error associated with using measures before undergoing formal translation procedures (Williams et al., 2009). Further, special attention was paid to the flexibility of the administration guidelines when selecting assessment measures to be used in an international context.

Based on these guiding principles, the Childhood Autism Rating Scale, Second Edition, Standard Version (CARS2; Schopler, Van Bourgondien, Wellman, & Love, 2010) was chosen to aid in clinician directed differential diagnosis of ASD. The CARS2 was developed so that the behavior ratings can be completed following observations of a range of informal play interactions. Thus, it is possible for this instrument to maintain its diagnostic integrity in a range of cultural and environmental contexts. Other excellent ASD diagnostic measures, such as the ADOS, require strict protocol adherence (both with regard to the exact verbal prompts and the specific nature of toy presentation). Though these types of measures have high functionality in the English-speaking cultures for which they were developed, rigorous adaption and translation procedures are necessary before using measures like the ADOS in different cultures and with non-English-speaking people (e.g., Tanzania). Given the flexibility of conducting behavioral observations necessary for completing the CARS2 (e.g., it is possible to use a play interaction with minimal language), this measure does not pose the same administration limitations as the ADOS for cross cultural use among non-English speaking families. The psychometric characteristics of the CARS2 have been established as having excellent validity and reliability for use in Western nations (Vaughn, 2011), and the original CARS provided good sensitivity and specificity as a diagnostic tool (Chlebowski, Green, Barton, & Fein, 2010). Further, the CARS2 has demonstrated good concurrent validity with other frequently used ASD diagnostic measures such as the ADOS (Schopler et al., 2010). The CARS2 form was used as anchor for assessment data collection. All of the information obtained through parent interview and child interaction was synthesized, and the CARS2 rating scale was completed to provide the clinician with diagnostic information across the fifteen symptom domains of the CARS2. Data were used to qualitatively inform diagnostic decisions made by a trained clinician, given a lack of Tanzanian norms on the CARS2. In addition to the CARS2 observational ratings, to gather additional information, we also adapted the CARS2 Questionnaire for Parents or Caregivers by administering it as a semi-structured interview to participants' guardians with the assistance of an interpreter (when required).

Play interaction

We used a semi-structured play observation, approximately 30 min long, which was designed to provide opportunities to observe behaviors relevant to the CARS2 behavior ratings with minimal language usage. This session included play tasks to specifically elicit deficits in social and communication domains, such as social orienting, joint attention, imitation, and functional play. Children were allowed access to a range of novel and potentially familiar toys from both the visiting clinician and the Tanzanian sites. We adapted activities from commonly used play assessments (i.e., ADOS; Lord et al., 2000), the Early Social Communication Scales (ESCS; Seibert, Hogan, & Mundy, 1982), the Social Communication Assessment for Toddlers with Autism (SCATA; Drew, Baird, Taylor, Milne, & Charman, 2007), and the Communication and Symbolic Behavior Scales (CSBS; Wetherby & Prizant, 1993). Adaptations were made in collaboration with local clinicians to create a culturally sensitive play interaction. For example, we included play themes that would be familiar to children in the region, such as food items and animals, but refrained from using potentially culturally unfamiliar toys. This play interaction consisted of the following activities: (a) Bubble and balloon play similar to the ADOS; (b) free play with culturally appropriate toys appropriate to different levels of play; (c) presentation of toys designed to elicit joint attention and social turn taking drawn from the ESCS; (d) response to name prompt; and (e) a toy imitation task similar to the ADOS. The CARS2 (Schopler et al., 2010) was completed following the play session.

Adaptive functioning

An interview version of the Malawi Development Assessment Tool (MDAT; Gladstone et al., 2008) was administered as a measure of adaptive functioning. The MDAT is an instrument created to assess adaptive behavior in rural Africa. The MDAT assesses domains of socialization, language, and motor skills and has established excellent reliability and validity (Gladstone et al., 2010). This measure was normed for use in Malawi and not Tanzania. Although there are vast cultural differences between the two nations, their regional proximity, similar historic colonization by the British, and subsequent parallel education and government structures result in many similarities (McGuigan, n.d.). As a result, a measure developed for Malawi, such as the MDAT, may be a more appropriate tool for use in Tanzania compared to measures developed for use in Western nations, which may contain more items or content unfamiliar to Tanzanian families (Gladstone et al., 2008). The MDAT has been used to assess adaptive functioning in many other African nations, including Burkina Faso and Gambia (personal correspondence with M. Gladstone, April 22, 2013). Of note, the preferred form of MDAT administration incorporates an interaction with the child, in addition to parent interview; however, in the interest of time constraints, the administration of the child interaction was impossible. Tanzanian caregivers were able to provide answers for their children on all of the items on the MDAT providing initial qualitative evidence that this measure is a feasible method for collecting data on adaptive functioning in Tanzania. As a result of the cultural differences between Malawi and Tanzania, Malawian norms were not used, and quantitative scores were not generated. Instead, this measure was used to qualitatively assess levels of adaptive functioning. Children were rated as having achieved or not achieved several developmental milestones across domains of Socialization, Language (receptive and expressive) and Daily Living Skills (see Table 1 for specific milestones rated).

Table 1

Malawi Developmental Assessment Tool: Milestones Examined

Malawi Developmental Assessment Tool: Milestones Examined
Malawi Developmental Assessment Tool: Milestones Examined

Cognitive assessment

To briefly assess nonverbal cognitive ability within this sample of Tanzanian children, three nonverbal subtests from the Kaufman Assessment Battery for Children (KABC; Kaufman & Kaufman, 1983) were selected, including Face Recognition, Hand Movements, and Triangles. The KABC has gained increasing use in African research due to the brevity of the nonverbal subtests, limited demands on verbalization, and its suitability for cultural modifications while preserving inherent psychometric properties (e.g., Bangirana et al., 2009; Holding et al., 2004; O'Donnell et al., 2012). Cognitive testing was attempted for children presenting at the 2013 clinics only (n  =  31).

DSM checklist

Differential diagnosis was also aided by the inclusion of a criterion checklist derived from DSM-IV-TR (American Psychiatric Association, 2000) and DSM 5 (American Psychiatric Association, 2013) and completed by the clinician. Final diagnostic decisions were established by integrating information from the parent interview and child observation and then completing the DSM checklist. Adherence to DSM criteria through gathering a clinical history is an important part of best practice diagnostic assessment (Lord et al., 2000; Perry, Condillac, & Freeman, 2002). Research has shown the structure and guidance of DSM criteria enhances diagnostic accuracy in some instances (i.e., raters who are less experienced diagnosing ASD), when compared to a diagnostic process not including DSM criteria (Klin, Lang, Cicchetti, & Volkmar, 2000).

Procedure

Each guardian/child dyad participated in one assessment session lasting approximately 2 hr. The assessment battery was comprised of four components: a brief intake, an adaptive screening interview, an ASD symptom questionnaire, and a semi-structured play-based interaction session. For the assessment the families met with one trained clinical psychologist with autism assessment expertise (first author) and one interpreter (when necessary). For families that did not speak fluent English, the interpreter translated from English to Swahili. The clinician minimized verbal communication with the child and instead primarily relied on social gestures and the introduction of novel toys during the play interaction so that interpretive services were not required during this component of the assessment.

Following the assessment, autism mental health literacy among caregivers was briefly assessed with a series of qualitative questions (e.g., “Have you ever heard of autism?” and “What do you know about autism?”). These questions helped the clinician to tailor the amount of psychoeducation incorporated into the family feedback session. Following these questions, families were provided with diagnostic impressions and a brief feedback form indicating symptoms their child exhibited consistent with an ASD. Further, to promote mental health literacy, caregivers were given a brochure in Swahili developed by staff at the Gabriella Centre and an affiliated organization, Autism Connects Tanzania (ACT). The clinician also reviewed general information about autism, included in the brochure, with families verbally (through an interpreter) to account for potential literacy barriers. During this feedback, parents were also provided with brief didactics in how to use basic behavioral strategies to improve behavior functioning (e.g., using positive reinforcement to encourage eye contact and requesting; how to use routines to increase success with adaptive skills such as toileting). The clinician described and modeled strategies to help families with the behaviors emerging as most impairing during the assessment.

Results

Diagnostic Findings

Given the variability in recruitment procedures across the two sites, (Moshi and Dar Es Salaam), between group analyses were used to compared children and parent characteristics to ensure that clinic site differences were not impacting diagnostic results. Children between the two sites did not differ on any examined variables: age, gender, language level, overall CARS score, symptom counts, diagnosis, or MDAT score (lowest p  =  0.24). Parents presenting at the Dar Es Salaam site, however, were significantly more educated (p  =  .002) and had jobs that received significantly higher scores on the Hollingshead Scale (p  =  0.01).

The 41 children assessed met criteria for one of two diagnostic groups: children meeting diagnostic criteria for ASD (n  =  30) and children with global delays (n  =  11). Diagnostic decisions were made by a clinical psychologist, who synthesized all information collected from the unique components of the assessment battery (parent report, observation during play, adaptive functioning) and used this information to complete either a DSM-IV or DSM 5 checklist depending on the year of the assessment (American Psychiatric Association, 2000; 2013). DSM-IV symptom counts were converted to DSM 5 criteria to enable a comparison of total number of ASD DSM symptoms for all participants. Codes for five of the seven DSM 5 criteria exactly included in the DSM-IV [e.g., (a) deficits in social emotional reciprocity, (b) deficits in nonverbal communication, (c) deficits developing and maintaining relationships, (d) adherence to routines, and (e) highly restricted interests] were directly transcribed. Children were coded as having the DSM 5 symptom of (6) stereotyped speech, motor movements, or use of an object if they had any of the following DSM-IV symptoms [(a) stereotyped or repetitive language, (b) repetitive or stereotyped language, or (c) preoccupation with parts of an object]. Two independent raters reviewed all aspects of the clinical data to determine if participants from 2012 met criteria for the newly added DSM 5 symptom of hyper- or hypo-reactivity to sensory input. Interrater reliability for this final item was good (κ  =  .92), and final decisions were reached through consensus coding. No group differences were observed in terms of gender, F  =  1.08, p  =  .31, or age, F  =  0.002, p  =  .97.

Consistent with group membership, a univariate analysis of variance (ANOVA) of between-group DSM 5 symptom counts revealed that, consistent with group membership, the ASD group had significantly more rated ASD symptoms than the group of children with general delays, F  =  49.27, p < .001, η2  =  .57. Although these effects were not significantly moderated by gender, F  =  0.04, p  =  .85, η2  =  .001, age emerged as a significant moderator in the model, F  =  4.35, p  =  .04. When follow-up analyses investigated the nature of the interaction, age was not significantly related to DSM 5 symptom counts in either diagnostic group. Figure 2 shows the distribution of rated DSM 5 symptoms.

Figure 2

Histogram depicting proportion of participants in both groups that met diagnostic criteria for each criterion A & B symptom in the Diagnostic and Statistical Manual (5th ed.).

Figure 2

Histogram depicting proportion of participants in both groups that met diagnostic criteria for each criterion A & B symptom in the Diagnostic and Statistical Manual (5th ed.).

Additional analyses were conducted to examine group differences on overall scores on the CARS2. A univariate ANOVA revealed that CARS2 scores were significantly higher for the ASD group than the general delay group, F  =  21.09, p < .001, η2  =  .37. This effect was not moderated by age, F  =  1.70, p  =  .20, η2  =  .05, or gender, F  =  0.01, p  =  .94, η2  =  .37. Specifically, the average CARS2 rating score for the children with ASD was 37.75 (SD  =  5.96), and the average CARS2 rating for the group with general delays was 27.15 (SD  =  6.79). Figure 3 shows the distribution of scores on the CARS2. Of note, children diagnosed with ASD in Tanzania scored in similar ranges to children from the United States diagnosed with ASD (Schopler et al., 2010). Further, analyses revealed that total score on the CARS2 was significantly correlated with the DSM5 symptom count in both the ASD group, r  =  .78, p < .001, and the general delay group, r  =  .66, p  =  .02. Although the CARS2 has not been validated for use among this population, this study provides preliminary evidence that group norms in Tanzania might align with group norms in the United States.

Figure 3

Histogram depicting distribution of total score on the CARS2 (range zero to 50) in each diagnostic group. CARS2  =  Childhood Autism Rating Scales (2nd ed.); ASD  =  autism spectrum disorder.

Figure 3

Histogram depicting distribution of total score on the CARS2 (range zero to 50) in each diagnostic group. CARS2  =  Childhood Autism Rating Scales (2nd ed.); ASD  =  autism spectrum disorder.

As discussed in detail here, though it was impossible to generate quantitative approximations of IQ or adaptive skills, based on parent report and behavioral observations, the other 11 children in the global delay group exhibited features that were likely consistent with a range of suspected syndromes (3 children with physical features consistent with fragile X syndrome, 1 child with Down syndrome, and 1 child with cerebral palsy) or intellectual disability (5 children). One additional child seemed to demonstrate symptoms more consistent with a language delay than a cognitive delay; however, this could not be formally assessed without appropriate cognitive and language assessments.

Results from cognitive testing on the KABC revealed that only a small proportion of children assessed accurately completed a sample item (13% Face Recognition, 17% Triangles, 20% Hand Movement). Testing was discontinued for these children that did not answer any of the sample items correctly. Of the 8 children that correctly completed a sample item correctly and were administered the tests, the mean number of items correct for each subtest was as follows: Face Recognition M  =  1.33, SD  =  1.03, range zero to 4; Hand Movements M  =  1.33, SD  =  1.03, range zero to 5; Triangles M  =  1.33, SD  =  1.21, range zero to 5. Of these 8 children, 7 met criteria for ASD, and one child did not (presented with a global delay only).

The Malawi Developmental Assessment Tool (MDAT) was used to qualitatively, and not quantitatively, provide adaptive estimates to inform diagnostic decisions. Standardized scores could not be generated because of a lack of Tanzanian MDAT norms and because out-of-level testing was required for many children given the pervasive developmental delays observed (the MDAT is designed for children birth to 6 years). Specific developmental milestones commonly assessed when taking a developmental history were examined in areas of socialization (4 items), communication (6 items), and daily living skills (6 items; see Table 1) and were rated as achieved (score of 1) or not achieved (score of zero). Within the ASD group, only two children achieved all communication milestones (M  =  3.21), 2 children achieved all daily living milestones (M  =  1.96), and 1 child achieved all socialization milestones (M  =  1.71). Similarly, within the general delay group, only 1 child achieved all language milestones (M  =  2.90), 1 child all daily living milestones (M  =  1.80), and no children achieved all socialization milestones (M  =  1.80). These data indicate that both groups demonstrated similar global impairment in adaptive domains.

Data collection on autism-specific mental health literacy began only in the second year of the screening clinic. As a result, data are only available among the subset of guardians participating in the 2013 clinics (n  =  29). Among the adults assessed, 38% had never heard of autism, 24% had heard of autism but knew little other information, and 38% had a working knowledge of autism.

The following case examples demonstrate how this multi-faceted approach enables the clinician to gather a range of diagnostic information to inform identification of DSM 5 symptoms. The selected diagnostic information is organized so that it is possible to determine from which component of the diagnostic assessment battery the information was derived. The cases presented are representative of the other children observed in both the ASD and globally delayed group and are meant to provide examples of how assessments were conducted and differential diagnoses were formed.

Representative ASD Case Example

Caregiver intake interview

Case 1 (“Joel”) is a 5-year-old male presenting for assessment with his father, who spoke fluent English during the assessment. Joel was born by Cesarean section following a typical pregnancy. He has no current medical problems. Joel experienced typical development until approximately 2½ years, when he started to regress in language ability. Before this regression, Joel was speaking in full sentences, but at present he primarily communicates in phrase or echolalic speech. Although he will point on occasion to have his needs met, he also frequently leads his parents by the hands when he needs something. At home, his father reports repetitive motoric flapping behavior.

Caregiver CARS2 questionnaire

According to parent responses on this questionnaire Joel very rarely uses toys in functional or make-believe play. Instead the majority of his play is repetitive, such as arranging shoes and pouring things from one cup to another.

Play observation

During the assessment session, Joel presented as hyperactive. He ran around the room without stopping and frequently eloped from the testing room. Joel spent much of his time spinning in circles with a train in his hand and making a repetitive screaming sound. Joel played with a ball by licking it and pressing it to his face. He responded to his name being called in session but made minimal eye contact.

Adaptive measure

The following adaptive information is based on parent report. According to Joel's father, he exhibits delays across all domains. In terms of expressive language ability, Joel communicates in single words but not yet sentences. He responds to his name and is able to follow two-part directions. Parent report revealed an even more pronounced delay in social skills. He smiles responsively and requests by pointing, but does not yet join in singing games or games requiring turn taking. In terms of daily living skills, Joel drinks by himself, eats using a spoon independently, and indicates the need to toilet; however, he is not yet using the toilet independently or helping to dress himself.

DSM 5 diagnostic checklist

Using a DSM 5 checklist, Joel met all three criterion A symptoms (Persistent deficits in social communication and social interaction) and 2 of the 4 criterion B symptoms (Restricted, repetitive patterns of behavior, interests, or activities). Overall, Joel met criteria for autism spectrum disorder. Although important to interpret with caution given the lack of psychometric data for the CARS2 in Tanzania, Joel's total score on the CARS2 was a 34.

Representative Non-ASD Case Example

The children not meeting criteria for an ASD seemed to display a commonality of language impairment. Although social skills were impaired among many of these children, these symptoms seemed most attributable to delays in cognition and language ability rather than a diagnosis of ASD. Moreover, none of these children exhibited significant repetitive or stereotyped behavior or interests.

Caregiver intake interview

Case 2 (“Oscar”) is an 11-year-old boy who presented for testing with his sister. There were no complications during pregnancy. Oscar has no known medical conditions. Oscar achieved his motor milestones on time but did not begin using words until 4 years. Oscar started speaking in phrases at 6 years and full sentences at 8 years. Although his language is progressing slowly, Oscar is not mastering any other academic domains in school. According to caregiver report, Oscar sits in school quietly but does not actively engage in learning.

Caregiver CARS2 questionnaire

Oscar's adult sister reported that he responsively smiles at others and shows a range of emotions at home. He often displays empathy by crying when he sees someone else is upset. Oscar does not respond negatively to changes in his environment and has no unusual interests or repetitive behaviors.

Play observation

Oscar exhibited adequate social pragmatics during the assessment session as evidence by socially directed smiles during turn taking, frequent initiation of joint attention through eye contact with the examiner, and appropriate response to his name. He appropriately played with all objects and imitated some creative play. Oscar did not exhibit any signs of repetitive or stereotyped behaviors.

Adaptive measure

Although Oscar is independent with regards to most adaptive daily living skills, he still requires some assistance using the toilet in unfamiliar locations. With regards to expressive language, Oscar has limited language flexibility and grammatical complexity.

DSM 5 diagnostic checklist

In sum, Oscar is exhibiting a language delay and probable cognitive impairment. He did not screen positive, however, for any of the DSM 5 criteria for ASD in either domain (social communication and interaction or Restricted, repetitive behaviors). Oscar did not meet criteria for an ASD. Although important to interpret with caution given the lack of psychometric data for the CARS2 in Tanzania, Oscar's total score on the CARS2 was a 24.

Discussion

Although global cross cultural measure development and validation procedures (Bracken & Barona, 1991) are essential, alternative efforts are needed at present to increase empirically-supported ASD diagnostic practices in underserved countries. This article aims to describe a process for trained clinicians to use preexisting instruments, incorporating clinical observation and culturally sensitive interviews, for differential ASD diagnoses in countries that do not have access to formally translated standardized measures such as Tanzania. By selecting pre-existing diagnostic tools with administration flexibility and reliance on observation methodology, it is possible to provide faster global access to ASD assessment services that align with best-practice standards of care (Ozonoff et al., 2005). The current study used an assessment battery comprised of parent report and observational measures to assess developmental (intake interview), cognitive (KABC), adaptive (MDAT), social (play interaction) and psychological (CARS2) domains to provide differential ASD diagnoses in Tanzania.

We provide information describing how to implement such an approach in low and middle-income countries that do not currently have psychometrically valid diagnostic instruments. We suggest minor modifications to make to existing measures to help to increase the feasibility and acceptability of using these instruments cross culturally. For example, instead of simply administering an intake designed for use in Western cultures, we removed items that would be considered irrelevant in Tanzania based on a working knowledge of cultural differences. Similarly, the play interaction with children included primarily toys and social routines that had a high likelihood of being familiar to children in the region. We also aimed to assess cognitive and adaptive functioning using measures specifically designed for use in sub-Saharan Africa (e.g., the MDAT; Gladstone et al., 2008) or that had been widely used in the region (e.g., the KABC). Finally, we carefully selected a diagnostic tool that had sufficient administration flexibility to be validly used in a culture and environment vastly different from where it originated, the CARS2 (Schopler et al., 2010) and paired this with a DSM checklist to increase the accuracy of our diagnostic decisions (Klin et al., 2000). These adaptations allowed for the collection of a range of diagnostic information from multiple sources (parent report and clinician observation) and from a range of important domains (e.g., cognitive, adaptive and behavioral) similar to diagnostic procedures used in Western nations.

Despite all the cultural differences that exist between Tanzania and Western countries, children that presented for assessment at the clinic and were diagnosed with ASD shared many similar features to children with ASD in the United States and as described by Kanner (1943). During the play-based observation, among the children diagnosed with ASD in Tanzania we observed social aloofness, (e.g., decreased initiations and engagement in social interactions and poor eye-contact), global impairments in verbal communication (e.g., non-verbal or primarily echolalic language) and nonverbal communication (e.g., diminished pointing, showing, and giving), and repetitive play themes and other repetitive or stereotyped behavior (e.g., unusual interests in a range of things including water faucets, lining up objects repetitively, filling containers, and manipulating small parts of a whole). Further, similar to data from the United States, individuals in the ASD group scored significantly higher than non-ASD individuals with other delays on the CARS2 and were rated as exhibiting significantly more DSM 5 symptoms. In addition, means on totals CARS2 scores for the two groups in Tanzania (ASD and general delay) align with published Western means (Schopler et al., 2010). Given the small size these results should be interpreted with caution; however, qualitative and quantitative data from the present study provides preliminary evidence to support that this method (including the use of the CARS2) might be useful for aiding in differential diagnoses in regions such as Tanzania where no diagnostic instruments have been translated for use in the local language.

It is notable that the children who presented to this clinic generally had significant developmental impairments, and those for whom a diagnosis of ASD was appropriate had prominent symptoms, similar to classic descriptions of ASD in developed countries. It is possible, if not likely, that there is a substantial group of children with ASD in Tanzania (and by extension other developing nations) who have forms of ASD with less severe developmental impairments or with mild symptom profiles. Identification of such children can be expected to be quite difficult in regions with fewer medical and educational resources. In addition to examining people with symptoms consistent with high functioning ASD, future work would also benefit from including people with delays but not suspected to have ASD, as was done in the present analysis. Greater variety in the functioning level and symptom presentation of the population examined would allow researchers to begin to identify Tanzanian behavioral norms on the CARS2 for groups of children with classic ASD and those with no ASD symptomatology.

Data ascertained from this ASD screening clinic helped to reveal that ASD psychoeducation among the community at large in Tanzania is greatly lacking. Further, formal diagnostic evaluations in this region are virtually nonexistent. Through the implementation of this small-scale clinic, families not only gained access to a formal diagnostic evaluation for their children, but were also provided with verbal and written information to promote ASD-specific mental health literacy in Tanzania. To help decrease the widely held stigmas in the region, families learned basic facts about ASD such as the role genetics play in the development of ASD, the effectiveness of behavioral treatments to improve functional outcomes, and Western prevalence rates. In addition to a psychoeducational component, through modeling and explanation, parents were also shown basic behavioral strategies for promoting verbal and non-verbal language development, increasing adaptive skills, and managing problematic behaviors specific to their child.

Despite the practical utility of this approach, there are still many limitations of note. Given this was a clinical rather than a research setting, no methodological controls were established; therefore, information gleaned should be considered preliminary information and should be used to inform the development of future clinical and research practice (Stake, 1978). To capture a full clinical presentation of the children, some parental report was required. This posed a potential problem given the limitations associated with parental rating scales (Johnston & Pennypacker, 1993) and the need to primarily rely on an interpreter to translate parent responses. Although the interpreters that aided in the clinical assessment were well versed in child behavior, they had not received training in performing formal interpretation services because such training was not available in the region. This was also a potential limitation in the assessment of adaptive functioning and language development through parent report and using a measure developed for use in another country. Global norm development may be one way to help provide a baseline of functioning in these domains for direct comparison. The approach outlined proved sufficient for identifying individuals with a more classic Kanner (1943) presentation, including global impairment (across all three ASD symptom domains) and with clearly observable unusual behaviors (e.g., little to no eye-contact or hand flapping) consistent with a diagnosis of ASD. However, this battery may not be sufficient for identifying less severe forms of ASD from global delays or people who present in a more atypical manner.

Obtaining a reliable assessment of cognitive ability was a limitation of the current study. There is burgeoning research in Africa that IQ can be validly assessed in many sub-Saharan African countries, using culturally relevant modifications to existing measures (e.g., Kitsao-Wekul, Holding, Taylor, Abubakar, & Connolly, 2012; Nampijja et al., 2010). There are several limitations, however, to this body of research. First, Tanzanian psychometric data for child intellectual assessment measures does not exist to our knowledge. Second, research demonstrating the successful adaptation of Western cognitive measures, such as the KABC and KABC-II, for use in Africa has focused mostly on typically developing children or children with medical problems (i.e., malaria, HIV), with less focus on children with developmental disabilities. Given the limited imitation ability of many of the children, 80%–87% of the children assessed did not complete any of the KABC sample or test items correctly across the three subtests, so group comparisons in cognitive ability were not indicated. To more accurately assess cognitive ability in Tanzanian children with developmental disabilities (i.e., to provide data about intellectual disability), future research would benefit from identifying culturally appropriate, nonverbal subtests designed for toddlers and younger children (i.e., items from the Bayley Scales of Infant Development; Bayley, 1993).

Conclusion

In sum, the current study addresses how existing tools and clinical resources can be used to enhance ASD assessment initiatives globally and, in particular, in regions of limited socioeconomic and psychological resources. Although formal translation and adaptation procedures comprise an essential step in promoting ASD assessment in low- and middle-income countries, there are children and families in need of services that cannot wait for these developments. By collaborating with people indigenous to a region and making as many cultural alterations as possible to existing measures, particularly observational measures, we can achieve a compromise between the best practice outlined for cross cultural assessment and helping people with ASD in need. We believe that a similar approach could be used to enhance ASD assessment efforts in other low- and middle-income countries for which standardized diagnostic tools are not yet available; however, we also must determine how to more widely disseminate these types of approaches within low- and middle-income countries. Although this approach can be easily modeled and taught to local clinicians (and was in Tanzania), the shortage of clinicians presents as a primary hurdle to overcome. As a result, dissemination of best-practice assessment standards may need to start with the few existing centers serving these regions. Finally, research indicates that in areas with limited clinical services (e.g., remote areas of the United States) parents may be a viable option for implementing basic interventions (Koegel, Symon, & Koegel, 2002); thus, our future work will be geared toward both parent and teacher trainings in Tanzania so that our efforts might be even further reaching.

Acknowledgments

This work was supported by a NIMH T32 fellowship to Ashley J. Harrison (5T32MH019927_20). We are grateful to all of the families that participated. We are also appreciative of the collaborators in Tanzania that aided in conducting this work: Anthony Ephraim and Brenda Shuma at the Gabriella Centre for Occupational Therapy and Special Education Services and Kilimanjaro Christian Medical University College, Moshi, Tanzania. We appreciate the continued support of organizations like EdPowerment and Autism Connects Tanzania for facilitating these collaborations. Additional support was provided to Eric M. Morrow through the Career Award in Medical Science from the Burroughs Wellcome Fund.

References

References
American Psychiatric Association (Ed.)
. (
2000
).
Diagnostic and statistical manual of mental disorders: DSM-IV-TR
.
Arlington, VA
:
Author
.
American Psychiatric Association (Ed.)
. (
2013
).
Diagnostic and statistical manual of mental disorders: DSM-5
.
Arlington, VA
:
Author
.
Bakare
,
M. O
.,
&
Munir
,
K. M
.
(
2011
).
Autism spectrum disorders (ASD) in Africa: A perspective
.
African Journal of Psychiatry
,
14
,
208
210
. .
Bangirana
,
P
.,
John
,
C. C
.,
Idro
,
R
.,
Opoka
,
R. O
.,
Byarugaba
,
J
.,
Jurek
,
A. M
.,
&
Boivin
,
M. J
.
(
2009
).
Socioeconomic predictors of cognition in Ugandan children: implications for community interventions
.
PLoS One
,
4
,
e7898
. .
Bayley
,
N
.
(
1993
).
The Bayley Scales of Infant Development-II
.
San Antonio, TX
:
The Psychological Corporation
.
Bölte
,
S
.,
Poustka
,
F
.,
&
Constantino
,
J. N
.
(
2008
).
Assessing autistic traits: cross-cultural validation of the social responsiveness scale (SRS)
.
Autism Research
,
1
,
354
363
. .
Bracken
,
B. A
.,
&
Barona
,
A
.
(
1991
).
State of the art procedures for translating, validating and using psychoeducational tests in cross-cultural assessment
.
School Psychology International
,
12
(
1–2
),
119
132
. .
Chlebowski
,
C
.,
Green
,
J. A
.,
Barton
,
M. L
.,
&
Fein
,
D
.
(
2010
).
Using the childhood autism rating scale to diagnose autism spectrum disorders
.
Journal of Autism and Developmental Disorders
,
40
,
787
799
. .
Drew
,
A
.,
Baird
,
G
.,
Taylor
,
E
.,
Milne
,
E
.,
&
Charman
,
T
.
(
2007
).
The Social Communication Assessment for Toddlers With Autism (SCATA): An instrument to measure the frequency, form and function of communication in toddlers with autism spectrum disorder
.
Journal of Autism and Developmental Disorders
,
37
,
648
666
. .
Elsabbagh
,
M
.,
Divan
,
G
.,
Koh
,
Y. J
.,
Kim
,
Y. S
.,
Kauchali
,
S
.,
Marcín
,
C
.,
&
Fombonne
,
E
.
(
2012
).
Global prevalence of autism and other pervasive developmental disorders
.
Autism Research
,
5
,
160
179
. .
Filipek
,
P. A
.,
Accardo
,
P. J
.,
Baranek
,
G. T
.,
Cook
,
E. H
., Jr,
Dawson
,
G
.,
Gordon
,
B
.,
&
Volkmar
,
F. R
.
(
1999
).
The screening and diagnosis of autistic spectrum disorders
.
Journal of Autism and Developmental Disorders
,
29
,
439
484
.
Fombonne
,
E
.,
Marcin
,
C
.,
Bruno
,
R
.,
Tinoco
,
C. M
.,
&
Marquez
,
C. D
.
(
2012
).
Screening for autism in Mexico
.
Autism Research
,
5
,
180
189
. .
Gladstone
,
M. J
.,
Lancaster
,
G. A
.,
Jones
,
A. P
.,
Maleta
,
K
.,
Mtitimila
,
E
.,
Ashorn
,
P
.,
&
Smyth
,
R. L
.
(
2008
).
Can Western developmental screening tools be modified for use in a rural Malawian setting
?
Archives of Disease in Childhood
,
93
,
23
29
. .
Gladstone
,
M
.,
Lancaster
,
G. A
.,
Umar
,
E
.,
Nyirenda
,
M
.,
Kayira
,
E
.,
van den Broek
,
N. R
.,
&
Smyth
,
R. L
.
(
2010
).
The Malawi Developmental Assessment Tool (MDAT): The creation, validation, and reliability of a tool to assess child development in rural African settings
.
PLoS medicine
,
7
(
5
),
e1000273
. .
Holding
,
P. A
.,
Taylor
,
H. G
.,
Kazungu
,
S. D
.,
Mkala
,
T
.,
Gona
,
J
.,
Mwamuye
,
B
.,
&
Stevenson
,
J
.
(
2004
).
Assessing cognitive outcomes in a rural African population: Development of a neuropsychological battery in Kilifi District, Kenya
.
Journal of the International Neuropsychological Society
,
10
,
246
260
. .
Hollingshead
,
W
.
(
1975
).
The Hollingshead four factor index of socioeconomic status
.
Unpublished manuscript, Yale University, New Haven, CT
.
Johnston
,
J. M
.,
&
Pennypacker
,
H. S
.
(
1993
).
Strategies and tactics of human behavioral research (2nd ed.)
.
Hillsdale, NJ
:
Psychology Press
.
Kanner
,
L
.
(
1943
).
Autistic disturbances of affective contact
.
Nervous Child
,
2
,
217
250
.
Kaufman
,
A. S
.,
&
Kaufman
,
N. L
.
(
1983
).
K-ABC–Kaufman Assessment Battery for Children
.
Circle Pines, MN
:
American Guidance Service
.
Kitsao-Wekulo
,
P. K
.,
Holding
,
P. A
.,
Taylor
,
H. G
.,
Abubakar
,
A
.,
&
Connolly
,
K
.
(
2012
).
Neuropsychological testing in a rural African school-age population: Evaluating contributions to variability in test performance
.
Assessment
. .
Klin
,
A
.,
Lang
,
J
.,
Cicchetti
,
D. V
.,
&
Volkmar
,
F. R
.
(
2000
).
Brief report: Interrater reliability of clinical diagnosis and DSM-IV criteria for autistic disorder: Results of the DSM-IV autism field trial
.
Journal of Autism and Developmental Disorders
,
30
,
163
167
.
Klinger
,
L. G
.,
&
Renner
,
P
.
(
2000
).
Performance-based measures in autism: implications for diagnosis, early detection, and identification of cognitive profiles
.
Journal of Clinical Child Psychology
,
29
,
479
492
. .
Koegel
,
R. L
.,
Symon
,
J. B
.,
&
Koegel
,
L. K
.
(
2002
).
Parent education for families of children with autism living in geographically distant areas
.
Journal of Positive Behavior Interventions
,
4
,
88
103
. .
Le Couteur
,
A
.,
Lord
,
C
.,
&
Rutter
,
M
.
(
2003
).
The Autism Diagnostic Interview–Revised (ADI-R)
.
Los Angeles, CA
:
Western Psychological Services
.
Lord
,
C
.,
Risi
,
S
.,
Lambrecht
,
L
.,
Cook
,
E. H
., Jr,
Leventhal
,
B. L
.,
DiLavore
,
P. C
.,
&
Rutter
,
M
.
(
2000
).
The Autism Diagnostic Observation Schedule—Generic: A standard measure of social and communication deficits associated with the spectrum of autism
.
Journal of Autism and Developmental Disorders
,
30
,
205
223
.
Lord
,
C
.,
Rutter
,
M
.,
DiLavore
,
P
.,
&
Risi
,
S
.
(
1999
).
Autism Diagnostic Observation Schedule: Manual
.
Los Angeles, CA
:
Western Psychological Services
.
Lotter
,
V
.
(
1978
).
Childhood autism in Africa
.
Journal of Child Psychology and Psychiatry
,
19
,
231
244
. .
Magaña
,
S
.,
&
Smith
,
L. E
.
(
2012
).
The use of the Autism Diagnostic Interview-Revised with a Latino population of adolescents and adults with autism
.
Journal of Autism and Developmental Disorders
,
43
,
1
8
. .
Manji
,
K. P
.,
&
Hogan
,
M. N
.
(
in press
).
Identifying gaps in knowledge, prevalence, and care of children with autism spectrum disorders in Tanzania—A qualitative review
.
Tanzanian Medical Journal
.
Mankoski
,
R. E
.,
Collins
,
M
.,
Ndosi
,
N. K
.,
Mgalla
,
E. H
.,
Sarwatt
,
V. V
.,
&
Folstein
,
S. E
.
(
2006
).
Etiologies of autism in a case-series from Tanzania
.
Journal of Autism and Developmental Disorders
,
36
,
1039
1051
. .
McGuigan
,
B
.
(
n.d.
).
Culture of East Africa
.
In
D
.,
Media
,
(
Ed.
),
USA Today
. .
Nampijja
,
M
.,
Apule
,
B
.,
Lule
,
S
.,
Akurut
,
H
.,
Muhangi
,
L
.,
Elliott
,
A. M
.,
&
Alcock
,
K. J
.
(
2010
).
Adaptation of Western measures of cognition for assessing 5-year-old semi-urban Ugandan children
.
British Journal of Educational Psychology
,
80
,
15
30
.
National Bureau of Statistics
. (
2011
).
Tanzania demographic and health survey
.
Calberton, MD, USA
:
ICF Macro
.
O'Donnell
,
K
.,
Murphy
,
R
.,
Ostermann
,
J
.,
Masnick
,
M
.,
Whetten
,
R. A
.,
Madden
,
E
.,
&
Whetten
,
K
.
(
2012
).
A Brief Assessment of Learning for Orphaned and Abandoned Children in Low and Middle Income Countries
.
AIDS and Behavior
,
16
,
480
490
.
Ozonoff
,
S
.,
Goodlin-Jones
,
B. L
.,
&
Solomon
,
M
.
(
2005
).
Evidence-based assessment of autism spectrum disorders in children and adolescents
.
Journal of Clinical Child and Adolescent Psychology
,
34
,
523
540
. .
Patel
,
V
.,
Araya
,
R
.,
Chatterjee
,
S
.,
Chisholm
,
D
.,
Cohen
,
A
.,
De Silva
,
M
.,
&
van Ommeren
,
M
.
(
2007
).
Treatment and prevention of mental disorders in low-income and middle-income countries
.
The Lancet
,
370
(
9591
),
991
1005
.
Pereira
,
A
.,
Riesgo
,
R. S
.,
&
Wagner
,
M. B
.
(
2008
).
Childhood autism: Translation and validation of the Childhood Autism Rating Scale for use in Brazil
.
Jornal de Pediatria
,
84
,
487
494
. .
Perry
,
A
.,
Condillac
,
R. A
.,
&
Freeman
,
N. L
.
(
2002
).
Best practices and practical strategies in the assessment/diagnosis of autism
.
Journal on Developmental Disabilities
,
9
,
61
75
.
Rogers
,
S. J
.
(
1998
).
Empirically supported comprehensive treatments for young children with autism
.
Journal of Clinical Child Psychology
,
27
,
168
179
. .
Rutter
,
M
.,
Le Couteur
,
A
.,
&
Lord
,
C
.
(
2003
).
Autism Diagnostic Interview, Revised
.
Los Angeles, CA
:
Western Psychological Services
.
Schopler
,
E
.,
Reichler
,
R. J
.,
DeVellis
,
R. F
.,
&
Daly
,
K
.
(
1980
).
Toward objective classification of childhood autism: Childhood Autism Rating Scale (CARS)
.
Journal of Autism and Developmental Disorders
,
10
,
91
103
. .
Schopler
,
E
.,
Van Bourgondien
,
M. E
.,
Wellman
,
G. J
.,
&
Love
,
S. R
.
(
2010
).
Childhood Autism Rating Scale, Second Edition (CARS2)
.
San Antonio, TX
:
Pearson Inc
.
Seibert
,
J. M
.,
Hogan
,
A. E
.,
&
Mundy
,
P. C
.
(
1982
).
Assessing interactional competencies: early social-communication scales
.
Infant Mental Health Journal
,
3
,
244
258
. .
Stake
,
R. E
.
(
1978
).
The case study method in social inquiry
.
Educational Researcher
,
7
(
2
),
5
8
. .
Vaughan
,
C. A
.
(
2011
).
Test Review. (From the book Childhood Autism Rating Scale by E. Schopler, M. E. Van Bourgondien, G. J. Wellman, & S. R. Love [Eds.], 2010
.
Journal of Psychoeducational Assessment
,
29
,
489
493
.
Volkmar
,
F
.,
Cook
,
E
., Jr,
Pomeroy
,
J
.,
Realmuto
,
G
.,
&
Tanguay
,
P
.
(
1999
).
Summary of the Practice Parameters for the Assessment and Treatment of Children, Adolescents, and Adults with Autism and Other Pervasive Developmental Disorders (American Academy of Child and Adolescent Psychiatry)
.
Journal of the American Academy of Child and Adolescent Psychiatry
,
38
,
1611
1616
.
Wetherby
,
A. M
.,
&
Prizant
,
B. M
.
(
1993
).
Profiling communication and symbolic abilities in young children
.
Communication Disorders Quarterly
,
15
,
23
32
. .
Williams
,
M. E
.,
Atkins
,
M
.,
&
Soles
,
T
.
(
2009
).
Assessment of autism in community settings: Discrepancies in classification
.
Journal of Autism and Developmental Disorders
,
39
,
660
669
. .
World Economic Outlook
. (
2012
).
World economic and financial surveys
. .
World Health Organization
. (
2008
).
mhGAP: Mental Health Gap Action Programme: Scaling up care for mental, neurological and substance use disorders
.
Geneva, Switzerland
:
World Health Organization
.
World Health Organization
. (
2013
).
World health statistics
.
Geneva, Switzerland
:
World Health Organization Press
.
Wong
,
V
.,
Hui
,
L. H
.,
Lee
,
W. C
.,
Leung
,
L. S
.,
Ho
,
P. K
.,
Lau
,
W. L
.,
&
Chung
,
B
.
(
2004
).
A modified screening tool for autism (Checklist for Autism in Toddlers [CHAT-23]) for Chinese children
.
Pediatrics
,
114
,
e166
e176
. .

Appendix

Author notes

Ashley J. Harrison, Postdoctoral Fellow, Department of Psychiatry and Human Behavior, Alpert Medical School of Brown University; Eric H. Zimak, Postdoctoral Fellow, Department of Psychiatry and Human Behavior, Alpert Medical School of Brown University; Stephen J. Sheinkopf, Assistant Professor, Department of Psychiatry and Human Behavior, Alpert Medical School of Brown University; Karim K. Manji, Professor, Department of Pediatrics, Muhimbili University of Health and Allied Sciences, and Eric M. Morrow, Assistant Professor, Department of Psychiatry and Human Behavior, Alpert Medical School of Brown University.

Address correspondence concerning this article to Ashley Johnson Harrison, Bradley Hospital, Developmental Disorders Research Program, 1011 Veterans Memorial Parkway, Providence, RI 02906 (e-mail: Ashley_johnson@brown.edu).