Scientific Merit Paper

https://doi.org/10.1177/1362361316675120

Autism 2018, Vol. 22(3) 322 –334 © The Author(s) 2016 Reprints and permissions: sagepub.co.uk/journalsPermissions.nav DOI: 10.1177/1362361316675120 journals.sagepub.com/home/aut

Introduction

Autism spectrum disorder (ASD) is a neurodevelopmental disorder that is characterized by impairments in social communication and the presence of repetitive or restricted behavior (American Psychiatric Association, 2013). ASD frequently goes undiagnosed until 4 years of age or later (Daniels and Mandell, 2013); however, parents often report that they were concerned about their child’s devel- opment before their second birthday (Chakrabarti and Fombonne, 2005). Identifying early signs in children with ASD is crucial to ensure timely access to needed services that have the potential to improve functional outcomes (Dawson et al., 2010; Perry et al., 2008; Sallows and Graupner, 2005). One method for gaining information about early development is to collect information from parents using questionnaires or interviews. Such measures aid in obtaining information about behavioral features to assist with early identification of developmental disorders

(Eisert et al., 1991). Surveillance, as part of a more com- prehensive process in which referral decisions are based

The Autism Parent Screen for Infants: Predicting risk of autism spectrum disorder based on parent-reported behavior observed at 6–24 months of age

Lori-Ann R Sacrey1,2, Susan Bryson3, Lonnie Zwaigenbaum1,2, Jessica Brian4,5, Isabel M Smith3, Wendy Roberts5,6, Peter Szatmari5,6,7, Tracy Vaillancourt8, Caroline Roncadin5,9 and Nancy Garon10

Abstract This study examined whether a novel parent-report questionnaire, the Autism Parent Screen for Infants, could differentiate infants subsequently diagnosed with autism spectrum disorder from a high-risk cohort (siblings of children diagnosed with autism spectrum disorder (n = 66)) from high-risk and low-risk comparison infants (no family history of autism spectrum disorder) who did not develop autism spectrum disorder (n = 138 and 79, respectively). Participants were assessed prospectively at 6, 9, 12, 15, 18, and 24 months of age. At 36 months, a blind independent diagnostic assessment for autism spectrum disorder was completed. Parent report on the Autism Parent Screen for Infants was examined in relation to diagnostic outcome and risk status (i.e. high-risk sibling with autism spectrum disorder, high-risk sibling without autism spectrum disorder, and low-risk control). The results indicated that from 6 months of age, total score on the Autism Parent Screen for Infants differentiated between the siblings with autism spectrum disorder and the other two groups. The sensitivity, specificity, and positive and negative predictive validity of the Autism Parent Screen for Infants highlight its potential for the early screening of autism spectrum disorder in high-risk cohorts.

Keywords autism spectrum disorder, high-risk siblings, parent report, prospective, screening

1University of Alberta, Canada 2Glenrose Rehabilitation Hospital, Canada 3Dalhousie University, Canada 4Bloorview Research Institute, Canada 5University of Toronto, Canada 6The Hospital for Sick Children, Canada 7Centre for Addiction and Mental Health, Canada 8University of Ottawa, Canada 9Hamilton Health Sciences, Canada 10Mount Allison University, Canada

Corresponding author: Lori-Ann R Sacrey, Autism Research Centre—E209, Glenrose Rehabilitation Hospital, 10230-111 Avenue, Edmonton, AB T5G 0B7, Canada. Email: [email protected]

675120AUT0010.1177/1362361316675120AutismSacrey et al. research-article2016

Original Article

Sacrey et al. 323

on clinical judgment, or screening at check-ups, either uni- versal (level 1, regardless of pre-existing concerns) or tar- geted (level 2, based on concern or risk status), can assist with the diagnostic process and monitoring changes in behavior over time (Glascoe et al., 2007).

Parents provide uniquely valuable information from the perspective of their everyday observations of infants’ usual behavior across varied contexts. Siegel et al. (1986) reported high agreement between parent reports of child- hood behavior in the home and clinical impressions derived from diagnostic play sessions used to elicit spe- cific ASD-related behavior. Furthermore, Glascoe (2000) (Glascoe et al., 1997) reported that developmental diagno- ses generally correspond with parent reports of their con- cerns about children’s development. A recent examination of parent concerns in a longitudinal examination of chil- dren at risk of an ASD diagnosis indicated that parents rec- ognize atypical behavior in their children, in broad agreement with directly observed behavioral signs emerg- ing in the first 2 years (Sacrey et al., 2015).

The currently available parent-questionnaire screening tools for signs of ASD in children under 36 months of age with published evaluation data are presented in Table 1. The First Year Inventory (FYI; Ben-Sasson and Carter, 2012; Reznick et al., 2007; Watson et al., 2007) is a 63-item measurement tool for signs of ASD in 12-month olds. The current version of the Modified Checklist for Autism in Toddlers, the M-CHAT-Revised with Follow-up (M-CHAT-R/F; Robins et al., 2014), consists of a 20-item questionnaire administered between 18 and 30 months and a follow-up interview for positive screens. The Qualitative Checklist for Autism in Toddlers (Q-CHAT; Allison et al., 2008) covers a broad range of ASD symptoms, each rated on a 5-point scale, and is completed when target children between the ages 18 and 24 months. The Parent Observation of Early Markers Scale (POEMS; Feldman et al., 2012) is a parent-report measure designed to monitor the behavio- ral development of infants at risk for ASD (because they have older siblings diagnosed with ASD). The Early Screening of Autistic Traits Questionnaire (ESAT; Dietz et al., 2006) is a 14-item parent questionnaire for detecting ASD in young children in the general population (Swinkels et al., 2006). The Infant Toddler Checklist (ITC; Wetherby et al., 2008) is a broadband screener to identify children with ASD in the general population. The Brief Infant- Toddler Social and Emotional Assessment (BITSEA; Kruizinga et al., 2014) is a 43-item questionnaire that measures both developmental problems and acquisition of competencies in children aged 1–3 years and includes items designed to measure ASD symptoms (Briggs-Gowan et al., 2004). The Parent’s Observational Screen of Social Interactions (POSI; Smith et al., 2013) is a brief, 7-item screen of ASD-specific behaviors assessed for children between the ages of 16 and 48 months. There is a need for a reliable and valid screening instrument targeting a

high-risk sample that covers a range of behaviors emerg- ing during early development (6 months+).

The purpose of this study was to assess the potential for the Autism Parent Screen for Infants (APSI) to identify ASD in a high-risk cohort. Items were selected based on behaviors measured by the Autism Observation Scale for Infants (AOSI; Bryson et al., 2008), which has predictive validity as an observational tool by 12 months of age in infant siblings of children diagnosed with ASD (Bryson and Zwaigenbaum, 2014; Zwaigenbaum et al., 2005). Unique features of the AOSI include its use with infants as young as 6 months of age and being specifically designed for use in high-risk cohorts. Adapting the AOSI into a par- ent questionnaire (the APSI) may yield broader clinical applicability to community settings where trained observ- ers are not readily available. Thus, the APSI could comple- ment existing screening tools, particularly with respect to being potentially informative for children under 18 months of age and its application in high-risk infants (a need high- lighted by the American Academy of Pediatrics (Johnson and Myers, 2007)).

In this study, the APSI was completed by two groups of parents. The first group had at least one (older) child diagnosed with ASD (high-risk; HR) and the second had no family history of ASD (low-risk; LR). The primary caregiver completed the APSI at 6, 9, 12, 15, 18, and 24 months of age. Infants underwent a diagnostic assess- ment for ASD at 36 months of age. Our main objective was to examine whether scores on the APSI distinguished among HR infants who were diagnosed with ASD at 36 months, other HR infants, and comparison LR infants. We hypothesized that the total score on the APSI would differentiate HR infants subsequently diagnosed with ASD at 36 months from non-diagnosed HR infants and LR infants by 12 months of age. An additional objective was to assess the predictive ability of the APSI total score across the age range studied, that is, the sensitivity and specificity of APSI scores relative to 3-year ASD diagno- ses, assessed at an individual level.

Methods

Participants

Infant siblings of children with ASD were recruited between the ages of 6 and 12 months from families attend- ing one of five multidisciplinary ASD diagnostic and treat- ment centers in Canada: (locations blinded). The research ethics board at each institution approved this study and all families gave written informed consent during initial enrollment into the study.

For the HR group, diagnosis of ASD in the older sibling (i.e. proband) was confirmed by a clinical assessment or a review of diagnostic records, using Diagnostic and Statistical Manual of Mental Disorders (4th ed., text rev.;

324 Autism 22(3)

Table 1. Currently available ASD screens with evaluation data.

Author Assessment Ages assessed Outcome for ASD

Reznick et al. (2007)

First Year Inventory (FYI) 12 months FYI identified 44% of infants at 12 months (n = 699) who were diagnosed with ASD at 3 years (Turner-Brown et al., 2012) Children who met FYI 95th percentile cutoff had higher AOSI scores and lower MSEL scores (Ben-Sasson et al., 2012)

Robins et al. (2013)

Modified Checklist for Autism in Toddlers–Revised with follow-up (M-CHAT-R/F)

18–30 months PPV in community sample of 0.51 (i.e. 51% of children referred for ASD would receive a diagnosis of ASD; Robins et al., 2014)

Allison et al. (2008)

Qualitative Checklist for Autism in Toddlers (Q-CHAT)

18–24 months High sensitivity and specificity (91% and 89%, respectively) when scoring the 10 items that best discriminated ASD from other diagnostic groups, but further evaluation is warranted (Allison et al., 2012)

Feldman et al. (2012)

Parent Observation of Early Markers Scale (POEMS)

1–24 months Parents of HR infants (n = 108) completed POEMS at multiple times between 1 and 24 months; resulted in an overall sensitivity of 0.74 (range, 0.25–1.00) and overall specificity of 0.73 (range, 0.65–0.84) from outcomes at 3 years of age (Feldman et al., 2012)

Dietz et al. (2006)

Early Screening of Autistic Traits Questionnaire (ESAT)

14–15 months Only 25% of children with ASD screened positive on ESAT, although the remaining children did not show typical development (Dietz et al., 2006)

Wetherby et al. (2008)

Infant Toddler Checklist (ITC) 6–24 months Only 6.1% of children who screened positive for ASD were diagnosed with ASD at 3 years (Wetherby et al., 2008)

Kruizinga et al. (2014)

Brief Infant-Toddler Social Emotional Assessment (BITSEA)

12–36 months In a community sample of 2-year olds who screened positive showed good sensitivity and specificity for the problem scale (0.83 and 0.84, respectively) and competence scale (0.85 and 0.89, respectively; Briggs-Gowan et al., 2004)

Smith et al. (2013)

Parent’s Observational Screen of Social Interactions (POSI)

16–48 months Sensitivity and Specificity compared to M-CHAT in clinical sample and community referred sample showed higher sensitivity (but comparable specificity; 0.89 and 0.54) in clinical sample compared to M-CHAT (0.71 and 0.62, respectively) and higher specificity (but comparable sensitivity 0.74 and 0.83) in community sample compared to M-CHAT (0.84 and 0.50, respectively)

AOSI: Autism Observation Scale for Infants; ASD: autism spectrum disorder; HR: high risk; MSEL: Mullen Scales of Early Learning; PPV: positive predictive value.

DSM-IV-TR) criteria. Neither the high-risk (HR) infant siblings nor the probands had identifiable neurological or genetic conditions, or severe sensory or motor impair- ments. The LR controls, recruited from local communities, were included on the basis of having no first- or second- degree relatives with an ASD diagnosis. All participants were born at 36–42 weeks gestation and had a birth weight greater than 2500 g.

Children from our larger HR cohort were included in this study if (1) they had undergone a 3-year diagnostic assessment and (2) had at least one completed APSI. Of the 279 HR children with 3-year follow-up, 75 did not have any completed APSI forms and were excluded from further analyses. Of the 90 LR children who had completed 3-year follow-up, 10 did not have any com- pleted APSI forms (including one child who was diag- nosed with ASD at age 3 years). These children were recruited and completed their 3-year assessment prior to inclusion of the APSI in the larger longitudinal study and thus were excluded from further analyses. Therefore, a total of 204 HR infant siblings and 79 LR controls

participated in the study. Table 2 presents detailed par- ticipant characteristics.

Measures

Participants were assessed at 6, 9, 12, 15, 18, and 24 months of age, with the initial time point between 6 and 12 months, depending on the age of recruitment. Parent report of early signs was measured by the APSI, early behavior and devel- opment was measured using the AOSI, Mullen Scales of Early Learning (MSEL), and Vineland Adaptive Behavior Scale (VABS), and the diagnostic assessment was com- pleted using the Autism Diagnostic Observation Schedule (ADOS) and the Autism Diagnostic Interview–Revised (ADI-R).

The APSI is a 26-item forced-choice parent-report ques- tionnaire with content and format similar to the AOSI (Bryson et al., 2008; Bryson and Zwaigenbaum, 2014). It thus covers a wide range of pre-diagnostic behavioral symptoms, including impairments in eye contact, visual tracking, responding to name, imitation, language, social

Sacrey et al. 325

development, joint attention, gestures, play, visual exami- nation of objects, and emotional regulation (see Table 3 for items). For example, to the question, “Does your child use gestures, such as waving good-bye, nodding his or her head, or blowing a kiss?” response choices are “definitely,” “possibly,” or “no,” which are scored “0,” “1,” and “2,” respectively. The APSI was designed to monitor putative signs of ASD in infants aged 6–24 months and takes approximately 10–15 min to complete. The primary car- egiver completed the APSI. More items with scores indicat- ing the presence of ASD-like behavior resulted in a higher score. Table 4 displays reliability analyses for internal con- sistency and split-half reliability of the APSI in the sample of children who received an ASD diagnosis at 36 months.

AOSI. The AOSI (Bryson et al., 2008) is a semi-structured direct observational measure designed to identify early

behavioral markers of ASD in infants/toddlers (e.g. atypi- calities in social communication, engagement, affect shar- ing, attention, and behavioral regulation). The AOSI is designed for use with 6- to 18-month olds, and administra- tion takes approximately 15–20 min. Each behavior is rated on a scale from 0 to 2 or 3, where 0 implies typical function, and higher values indicate increasing atypicality. The AOSI has excellent inter-rater reliability (0.93 for total score at 12 months), fair-to-good test-retest reliability at 12 months (0.61 for total score: Bryson et al., 2008), and good predic- tive validity at 12 months for its original 16 items (Brian et al., 2008; Zwaigenbaum et al., 2015). We used a total score cutoff of ⩾7 to indicate risk, based on the evidence of good positive predictive value (0.75) and negative predictive value (0.98–0.99) in earlier work (Bryson and Zwaigenbaum, 2014). The AOSI was administered at 6 (where applicable), 9, 12, 15, and 18 months of age.

Table 2. Participant characteristics.

Characteristic LR HR-N HR-ASD

N 6 months 51 67 21 9 months 56 83 37 12 months 74 104 54 15 months 68 104 53 18 months 72 122 56 24 months 61 58 28 36 months 79 139 66 Sex 43M:36F 69M:69F 50M:16Fa,b

MSEL ELC M (SD) M (SD) M (SD) 6 months 92.3 (9.4) 91.3 (11.1) 92.2 (10.3) 12 months 98.2 (16.1) 94.0 (18.7) 89.1 (18.1)a

24 months 116.1 (20.0) 99.0 (17.1)a 88.7 (16.4)a

36 months 111.0 (14.0) 104.31 (21.1) 93.3 (15.4)a,b

VABS ABC M (SD) M (SD) M (SD) 12 months 104.8 (6.7) 103.8 (8.7) 99.1 (8.5)a

18 months 97.7 (7.1) 93.0 (7.2)a 85.2 (8.8)a,b

24 months 93.8 (8.7) 90.0 (9.3) 83.1 (10.6)a,b

36 months 99.0 (10.5) 95.1 (13.2) 81.3 (12.8)a,b

AOSI M (SD) M (SD) M (SD) 6 months 7.6 (3.8) 7.3 (3.2) 9.7 (3.7) 9 months 5.2 (3.6) 5.3 (3.0) 5.2 (3.9) 12 months 3.2 (3.2) 5.1 (3.4)a 7.3 (4.8)a,b

15 months 3.2 (2.8) 4.9 (3.8) 7.6 (4.8)a,b

18 months 3.3 (3.0) 5.1 (3.9)a 8.3 (4.2)a,b

ADOS Severity Score M (SD) M (SD) M (SD) 36 months 1.6 (1.1) 2.6 (1.7)a 6.2 (2.0)a,b

ADI-R M (SD) M (SD) M (SD) 36 months 4.4 (3.9) 5.9 (4.6) 22.4 (10.3)a,b

LR: low risk; HR-N: high-risk infants not diagnosed with autism spectrum disorder; HR-ASD: high-risk infants diagnosed with autism spectrum disorder; MSEL ELC: Mullen Scales of Early Learning Early Learning Composite; SD: standard deviation; VABS ABC: Vineland Adaptive Behavior Scales Adaptive Behavior Composite; AOSI: Autism Observation Scale for Infants; ADOS: Autism Diagnostic Observation Schedule; ADI-R: Autism Diagnostic Interview–Revised. aDifferent from LR. bDifferent from HR-N.

326 Autism 22(3)

The MSEL. The MSEL (Mullen, 1995) consists of five scales, four of which (Visual Reception, Receptive Lan- guage, Expressive Language, and Fine Motor) assess non- verbal, cognitive, and language ability, while the fifth scale measures gross motor development (from 0 to 29 months only). An Early Learning Composite is calcu- lated based on scores from the first four scales for chil- dren aged 0–69 months. Inter-rater and test–retest reliability are excellent (Mullen, 1995). The MSEL was administered at 6 (where applicable), 12, 24, and 36 months of age.

VABS. The VABS (Sparrow et al., 1984) is a semi-struc- tured parent interview designed to assess adaptive behav- ior across four subdomains—Communication, Daily Living, Socialization, and Motor skills (the last domain limited to children younger than 30 months), outlined by

typical developmental milestones that are anchored to spe- cific ages. The scale has excellent reliability and concur- rent validity and is sensitive to impairments experienced by children with ASD (Carter et al., 1998; Volkmar et al., 1993). The VABS was administered at 12, 18, 24, and 36 months of age.

ADOS. The ADOS (Lord et al., 2000) includes standard- ized activities and “presses,” which are used to elicit com- munication, social interaction, imaginative use of play materials, and repetitive behavior (Lord et al., 1989). Inter-rater reliability for the ADOS is excellent (Lord et al., 2000). The scoring algorithm was recently revised to optimize discrimination of ASD from other develop- mental disabilities and is organized into two domains, Social Affect (including Communication and Social items) and Restricted Repetitive Behaviors (Gotham et al., 2007). The ADOS consists of four modules, each of which is appropriate for individuals of differing language levels (Module 1 = minimal or no language, Module 2 = regular use of non-echoed 3-word phrases, Module 3 = child with fluent language, and Module 4 = adolescent or adult with fluent language), the first three of which were used to assess participants in this study; Module 1 alone was administered at the 18-month assessment. To optimize comparability across modules (and thus, across language levels), we used the 36-month ADOS severity metric (Gotham et al., 2009).

ADI-R. The ADI-R (Lord et al., 1994) is an investigator- directed interview that elicits information regarding social development, verbal and non-verbal communication skills, and the presence of repetitive, stereotyped interests and behavior required to make an International Classifica- tion of Diseases (10th ed.; ICD-10) or DSM-IV-TR diag- nosis of ASD. The questions are designed to distinguish qualitative impairments from developmental delays. The ADI-R discriminates well between ASD and other forms of developmental disability, and inter-rater reliability is excellent (Lord et al., 1994). The ADI-R was adminis- tered at 36 months of age.

Diagnostic procedure

At 3 years of age, each participant underwent an independ- ent diagnostic evaluation, conducted by an expert clinician blind to results from previous study visits. ASD diagnoses

Table 3. Items queried on the Autism Parent Screen for Infants.

1—Difficulty visually tracking a moving object 2—Visual fixation, or stare, at certain objects 3—Fail to respond to name 4—React to changes in facial expression 5—Anticipate the pleasure of social games 6—Imitation of sounds or actions of others 7—Vocalize back-and-forth with you 8—Difficulty in establishing eye contact 9—Smile in response to your smiles 10—Coordinate actions with eye gaze 11—Tend to be over-reactive or under-reactive 12—Cuddle into your body when holding them 13—Difficult to soothe 14—Show sustained interest and pleasure in interacting 15—Have difficulty with change 16—Difficulty using hands/holding objects 17—Unusual repetitive motor behaviors 18—Use another person’s hand as a tool 19—Have unusual sensory behaviors 20—Difficulty focusing attention on objects 21—Insist on holding/playing with certain toys 22—Resist others joining in play/have fixed play routines 23—Share interests in object/event with others 24—Point to objects/event at a distance 25—Use gestures 26—Loss of skill over past 2–3 months

Table 4. Reliability of the APSI.a

Age (months) 6 9 12 15 18 24

Internal consistency (α) 0.77 0.90 0.83 0.89 0.87 0.92 Split-half reliability (α) 0.35 0.89 0.82 0.91 0.91 0.94

APSI: Autism Parent Screen for Infants; HR-ASD: high-risk infants diagnosed with autism spectrum disorder. aReliability calculated on HR-ASD sample only.

Sacrey et al. 327

were assigned using DSM-IV-TR criteria, based on the best judgment of the clinician (developmental pediatri- cian, child psychiatrist, or clinical psychologist, all with at least 10 years of diagnostic experience), taking into account information from the ADI-R and ADOS, as well as concurrent developmental assessment using the MSEL and VABS.

Statistical analysis

Group membership was determined from the 3-year out- come assessments: high-risk infants diagnosed with ASD (HR-ASD), high-risk infants not diagnosed with ASD (HR-N), and low-risk infant not diagnosed with ASD (LR). Clinical characteristics of the groups were compared using linear mixed modeling with Group (LR, HR-N, HR-ASD) and Age (6, 9, 12, 15, 18, 24 months) as inde- pendent measures and scores on the various assessments as the dependent measures. Performance on the APSI was compared using linear mixed modeling with Group and Age as independent measures and scores on the question- naire as the dependent measure. Total scores were com- pared and group by age interactions were explored using Benjamini and Hochberg corrections (Benjamini and Hochberg, 1995). In this method, the p-values are ordered smallest to largest. The alpha level for each test is then set at (k m*α)/ with k corresponding to the p-value’s rank (e.g. lowest p-value = 1) and m corresponding to the num- ber of comparisons, which in this case was 18. The com- parisons stop once one of the t-tests is rejected. Thus, this method decreases the chance of false positives. Our main objective was to determine when differences appeared among the three groups of children, therefore planned comparisons were completed on all group by age interac- tions. Effect sizes were calculated for group differences between the two HR groups using Cohen’s d, with 0.2– 0.49 = small effect, 0.5–0.79 = medium effect, and 0.8 + = large effect.

To provide a preliminary assessment of the predictive utility of the APSI, we used receiver operator characteris- tic (ROC) analyses to assess the sensitivity and specificity of the APSI at each age with respect to ASD diagnosis at age 3. Analyses were limited to the HR cohort, to examine specifically the potential properties of the APSI within that context. To determine the optimal cut-point for the total score, Youden’s index was used, which is defined as the maximum vertical distance between the ROC curve and the diagonal or chance line (Youden’s index (J) = sensitiv- ity + specificity − 1 (Akobeng, 2007)). Other determinants of screening accuracy were (1) sensitivity, the proportion of children with ASD correctly classified by the APSI; (2) specificity, the proportion of children not diagnosed with ASD correctly classified by the APSI; (3) positive likeli- hood ratio (LHR+), the ratio of the probability of identify- ing a child as having ASD if the child does have ASD (true

positive) relative to the probability of identifying the child as having ASD if the child does not have ASD (false posi- tive) (sensitivity/(1 − specificity)); (4) negative likelihood ratio (LHR−), the ratio of the probability of not identifying the child as ASD if the child does have ASD (false nega- tive) to the probability of not identifying the child as ASD if the child does not have ASD (true negative) ((1 − sensi- tivity)/specificity); (5) odds ratio (OR), the ratio of the odds of a correctly identifying a child as ASD when having the “disorder” relative to the odds of a identifying a child as ASD when not having the “disorder” ((sensitivity*speci ficity)/((1 − sensitivity)*(1 − specificity)); (6) positive pre- dictive validity (PPV), the child with ASD is correctly identified as a child with ASD ((true positive/(true posi- tive + false positive))); and (7) negative predictive validity (NPV), the child without ASD is incorrectly identified as a child with ASD ((true negative/(true negative + false neg- ative))) (Fischer et al., 2003).

Results

Completers versus non-completers

A chi-square analysis was completed to determine if there were differences in the clinical characteristics of the chil- dren who have parents compete versus not complete APSI forms. There were no differences in the proportion of par- ents in each group who completed APSI forms (χ(2) = 3.56, p = 0.107). As well, there were no differences between the completers versus non-completers for the child’s severity score on the ADOS at 24 (t = .74, p = 0.46) and 36 months (t = 1.28, p = 0.20), nor the total score on the ADI at 36 months (t = 1.33, p = 0.18).

Respondent characteristics

The respondent who completed an APSI at each age assessed is presented in Table 5 as mother, father, both par- ents, or unidentified. Group differences for the respondent were assessed using chi-square analyses at each age. There were no group differences, except for age 18 months, where the difference was driven by the number of “uniden- tified” respondents in the HR-N group. The vast majority of respondents (78% or greater) were identified as “mother” across all groups and ages assessed.

The average age of the mother at study enrollment was significantly different between groups, with older mothers in the HR-ASD (34.35 ± 4.15 years) and HR-N (33.91 ± 4.25 years) groups compared to mothers in the LR group (31.72 ± 3.97 years; ps < .001). Similarly, the average age of the father at study enrollment was sig- nificantly different between groups, with older fathers in the HR-ASD (37.31 ± 5.83 years) and HR-N (36.35 ± 5.34 years) groups compared to the LR group (33.64 ± 4.09 years; ps < 0.001).

328 Autism 22(3)

Child characteristics

From the results of the 36-month diagnostic assessments, three groups were identified for comparison: (1) HR infant siblings who received a diagnosis of ASD (“HR-ASD”; n = 66; 50 boys and 16 girls); (2) HR infant siblings who did not receive a diagnosis of ASD (“HR-N”; n = 138; 69 boys and 69 girls); and (3) LR controls who did not receive a diagnosis of ASD (“LR”; n = 79; 43 boys and 36 girls). As shown in Table 2, a significant sex difference (χ = 11.26, p = .004) showed a higher boy-to-girl ratio in the HR-ASD group than in the LR and HR-N groups, who did not differ (ps < 0.01), as anticipated from previous analyses on this sample (citation blinded). The groups differed in age at ini- tial assessment (F(2,284) = 7.64, p < 0.01). The HR-ASD group was slightly older than the HR-N and LR groups, resulting from a lower proportion of first assessments at 6 months in the HR-ASD group, relative to the other two groups. The groups did not differ in age at the 36- month diagnostic assessment (F(2,283) = 0.20, p = 0.50). Descriptive data on developmental and behavioral features are also summarized for the three groups in Table 2. There were group differences for the MSEL, VABS, and AOSI at 12 months of age and older.

Total APSI scores

For each infant at each age, scores on each APSI item were computed and summed to yield a total score (n = 26; maxi- mum score = 52). To determine group differences for the total score, a mixed model analysis was performed. Overall, there were significant effects of Group (F(2, 1153) = 121.72, p < 0.001), Age (F(5, 1153) = 34.38, p < 0.001), and a Group × Age (F(10, 1153) = 1.81, p < 0.05) interaction. Post hoc analyses were run on the interaction effect using a Benjamini and Hochberg correc- tion (q < .036). As shown in Figure 1, the total score was higher for HR children with ASD compared to the other two groups beginning at 6 months and continuing to 24 months. Only 12-month APSI scores were higher for

the HR-N group compared to the LR group. Effect sizes were computed for comparisons between the two HR groups at 6–24 months, resulting in ds of 4.13, 7.25, 8.11, 10.64, 8.01, and 10.21, respectively.

Group comparisons were also completed on individual APSI questions for ages 6, 12, and 18 months. As shown in Table 6, item-level differences between the two HR groups at 6 months were detected on questions related to visual tracking and back-and-forth vocalizations. At 12 months, item-level group differences between the two HR groups were detected on 14 of 26 items, as detailed in Table 7. Finally, item-level group differences between the HR groups at 18 months were detected on 21 of 26 ques- tions, as presented in Table 8. Comparisons were cor- rected for multiple testing using the Benjamini and Hochberg (1995) correction.

ROC curve analyses

ROC curve analyses were completed to identify cutoff scores at each age that optimized the predictive utility of the APSI (total score), with respect to subsequent ASD diagnosis. The HR groups (HR-ASD vs HR-N) were com- pared using a series of ROC analyses at 6, 9, 12, 15, 18, and 24 months of age. The area under the curve (AUC) for the total score was significant at each age (i.e. the overall “area under the curve” was different from 0.5, the value expected by chance). Sensitivity, specificity, PPV, and NPV for the total scale score were 0.67, 0.86, 0.47, and 0.83, respectively (cutoff = 15), at 6 months, 0.59, 0.72, 0.65, and 0.76, respectively (cutoff = 10), at 12 months, and 0.65, 0.72, 0.68, and 0.77, respectively (cutoff = 9), at 18 months. Table 9 displays the results of the ROC analy- ses at each age. Youden indices indicated different optimal cutoffs for the total scale score at each age, with highest cutoffs at the youngest ages. When comparing the optimal parameters, discrimination between those HR children who would and would not be diagnosed with ASD was achieved at the youngest ages assessed, as determined by the highest sensitivity and specificity.

Table 5. Respondents for each group at each age assessed.

Age HR-ASD HR-N LR Significance

M F B U M F B U M F B U X p

6 months 84.85 6.06 6.06 3.03 88.17 1.08 5.38 5.38 93.55 0.00 4.84 1.61 7.01 0.32 9 months 89.47 2.63 7.89 0.00 82.29 1.04 6.25 10.42 94.12 0.00 4.41 1.47 11.29 0.08 12 months 86.57 1.49 5.97 5.97 80.85 2.13 5.67 11.35 93.10 0.00 5.75 1.15 10.79 0.09 15 months 86.00 2.00 4.00 8.00 81.03 1.72 4.31 12.93 85.71 2.86 7.14 4.29 4.89 0.56 18 months 85.07 4.48 7.46 2.99 78.21 3.21 4.49 14.10 90.59 1.18 5.88 2.35 15.34 0.02* 24 months 88.89 0.00 3.70 7.41 92.98 1.75 5.26 0.00 90.16 1.64 6.56 1.64 5.74 0.45

M: mother; F: father; B: both parents; U: unidentified; HR-N: high-risk infants not diagnosed with autism spectrum disorder. Values are percentages. *Difference is driven by the number of unidentified respondents in the HR-N group.

Sacrey et al. 329

Discussion

In this study, we described a parent-report questionnaire analogue of the AOSI, the APSI, which was completed by the primary caregivers of HR and LR infants at multiple time points from 6 to 24 months of age. Overall, the HR-ASD group had a higher mean APSI total score than the LR and HR-N groups at each time point between 6 and 24 months of age. Importantly, ROC analyses indicated that the APSI predicted group status beyond chance at the earliest age point assessed (6 months), differentiating between HR infants who would and would not receive an ASD diagnosis.

The earlier identification of high-risk siblings who would later be diagnosed with ASD reported here was based on differences in sensory behavior, imitation, responding to changes in facial emotion, back-and-forth communication, and sharing interests with others. This finding is consistent with those of the Avon Longitudinal Study of Parents and Children, a prospective investigation

Figure 1. Total score on the APSI (all questions combined) for ages 6–24 months. *Different from LR; ^different from HR-N.

Table 6. Group differences for individual questions at 6 months of age.

APSI at 6 months

HR-ASD (n = 21)

HR-non-ASD (n = 67)

LR-non-ASD (n = 52)

HR-ASD versus HR-N

HR-ASD versus LR

HR-N versus LR

Item % 1 % 2 % 1 % 2 % 1 % 2 p value p value p value

1—Visual tracking 19 5 – 2 – 2 0.001* 0.001* 0.65 2—Visual fixation 43 10 33 44 33 44 0.61 0.002* 0.001* 3—Respond to name 48 14 37 8 37 8 0.44 0.12 0.26 4—Response to facial emotion 35 45 48 19 48 19 0.02 0.05 0.75 5—Anticipatory social response 40 5 21 6 21 6 0.093 0.22 0.56 6—Imitation 24 62 53 13 54 14 0.02 0.001* 0.19 7—Vocalize with you 52 29 25 4 25 4 0.001* 0.001* 0.53 8—Eye contact 10 5 6 4 6 4 0.38 0.58 0.68 9—Reciprocal social smile 5 0 2 – 2 – 0.36 0.45 0.87 10—Coordinate actions with eye gaze 19 10 35 – 35 – 0.59 0.90 0.59 11—Reactivity 38 5 14 4 14 4 0.09 0.05 0.61 12—Cuddle with you 24 – 19 4 19 4 0.97 0.83 0.72 13—Difficult to soothe 24 5 10 4 10 4 0.66 0.23 0.28 14—Social interest and affect 24 – 4 – 4 – 0.66 0.07 0.05 15—Difficulty with change 5 – 10 – 10 – 0.48 0.42 0.87 16—Hand use/holding objects 19 – 12 – 12 – 0.43 0.50 0.05 17—Repetitive motor behaviors 5 10 10 6 10 6 0.16 0.86 0.09 18—Another person’s hand as tool 5 – 4 – 4 – 0.42 0.80 0.46 19—Unusual sensory behaviors 19 5 8 4 8 4 0.09 0.24 0.52 20—Focusing attention on objects 10 5 10 2 19 2 0.65 0.77 0.84 21—Insistence on particular object – – 4 – 4 – 0.73 0.18 0.16 22—Resist play/fixed play routines – 5 – – – – 0.08 0.05 0.67 23—Share interests with others 33 38 35 13 35 13 0.02 0.02 0.99 24—Distal point – 100 4 96 4 96 0.26 0.40 0.72 25—Use gestures 5 95 2 98 2 98 0.62 0.41 0.07 26—Loss of skill 5 – 2 – 2 – 0.37 0.46 0.86

APSI: Autism Parent Screen for Infants; HR-ASD: high-risk infants diagnosed with autism spectrum disorder; LR: low risk; HR-N: high-risk infants not diagnosed with autism spectrum disorder; “–”: no code at that score for item. *p < 0.006.

330 Autism 22(3)

of 14,541 families. The authors reported that differences in social skills and communication were evident by 6 months of age in those children who were later diagnosed with ASD (n = 68; Bolton et al., 2012).

The sensitivity and specificity of the APSI total score (derived from the Youden index) ranged from 0.58 to 0.67 for sensitivity and from 0.72 and 0.87 for specificity. Although moderate, the sensitivity and specificity esti- mates of the APSI are comparable to those of other ASD screening instruments. For example, recent evaluative data on the M-CHAT-R/F report sensitivity and specificity data (0.73 and 0.89, respectively) for a cutoff score of 3 (Robins et al., 2014). Similarly, the positive and negative predictive value of the APSI are similar to those of the ITC, which reports lower PPV at the lowest ages (0.43 at 6–8 months) and higher PPV at the highest ages examined (0.79 at 21–24 months) and NPV ranging between 0.87 and 0.99 across the ages examined (Wetherby et al., 2008). Although sensitivity and specificity are comparable, the

APSI has some unique strengths. For example, the M-CHAT/R (Robins et al., 2013) is administered between 18 and 30 months of age and requires a clinician-led fol- low-up interview in most cases to confirm positive screens. Although there are several screeners available, the APSI may fulfill a unique role, as it is informative as early as 6 months of age, and as a stand-alone instrument. Moreover, the APSI is particularly well suited for research with high-risk siblings because of its relationship and pos- sible convergence with the AOSI. Indeed, there is a need for early detection tools suited to infants who have an older child with autism, as this group has been identified as in need of intensified surveillance by Johnson and Myers (2007).

Current observational data suggest that social commu- nication differences in ASD may not be evident until around 12 months (Landa et al., 2012; Landa and Garrett- Mayer, 2006; Ozonoff et al., 2010; Sacrey et al., 2013; Zwaigenbaum et al., 2005). Prodromal features of ASD,

Table 7. Group differences for individual questions at 12 months of age.

APSI at 12 months

HR-ASD (n = 54)

HR-non-ASD (n = 104)

LR-non-ASD (n = 75)

HR-ASD versus HR-N

HR-ASD versus LR

HR-N versus LR

Item % 1 % 2 % 1 % 2 % 1 % 2 p value p value p value

1—Visual tracking 6 7 1 2 1 3 0.02* 0.04 0.74 2—Visual fixation 28 9 14 2 20 4 0.003* 0.10 0.19 3—Respond to name 20 19 14 2 11 3 0.001* 0.001* .73 4—Response to facial emotion 54 13 56 5 36 4 0.24 0.003* 0.02* 5—Anticipatory social response 13 9 8 2 4 – 0.005* 0.001* 0.19 6—Imitation 35 7 22 4 11 – 0.024* 0.001* 0.006* 7—Vocalize with you 33 11 22 3 8 – 0.001* 0.001* 0.009* 8—Eye contact 19 15 10 1 3 1 0.001* 0.001* 0.32 9—Reciprocal social smile 19 – 4 – – – 0.001* 0.001* 0.28 10—Coordinate actions with eye gaze 20 4 7 2 8 1 0.03* 0.05 0.88 11—Reactivity 22 17 11 4 5 1 0.001* 0.001* 0.10 12—Cuddle with you 31 13 18 3 19 – 0.001* 0.001* 0.57 13—Difficult to soothe 9 7 13 5 4 4 0.83 0.18 0.18 14—Social interest and affect 13 7 8 3 5 – 0.05 0.003* 0.17 15—Difficulty with change 20 7 16 2 16 3 0.10 0.22 0.73 16—Hand use/holding objects 9 9 9 3 4 – 0.07 0.002* 0.09 17—Repetitive motor behaviors 15 11 14 1 1 – 0.002* 0.001* 0.02* 18—Another person’s hand as tool 11 2 14 2 17 11 0.84 0.005* 0.002* 19—Unusual sensory behaviors 13 4 7 2 9 – 0.10 0.09 0.88 20—Focusing attention on objects 35 6 24 3 19 1 0.11 0.01* 0.21 21—Insistence on particular object 15 7 8 1 9 4 0.02* 0.13 0.46 22—Resist play/fixed play routines 9 – 2 1 – – 0.11 0.01* 0.24 23—Share interests with others 28 17 21 5 9 3 0.007* 0.001* 0.06 24—Distal point 19 54 25 41 28 36 0.35 0.10 0.38 25—Use gestures 19 24 28 11 4 5 0.06 0.001* 0.001* 26—Loss of skill 2 6 3 2 1 – 0.23 0.04 0.27

APSI: Autism Parent Screen for Infants; HR-ASD: high-risk infants diagnosed with autism spectrum disorder; LR: low risk; HR-N: high-risk infants not diagnosed with autism spectrum disorder; “–”: no code at that score for item. *p < 0.03.

Sacrey et al. 331

such as atypical visual orienting and face processing, have been detected by electroencephalogram (EEG) and eye tracking by 6–8 months, yet such technology-based methods are expensive and time-consuming (Chawarska et al., 2013; Cohen et al., 2013; Elison et al., 2013; Elsabbagh et al., 2012; Jones and Klin, 2013). Although eye tracking may ultimately be adapted for application in clinical settings, specialized equipment and technical skills will be needed. The APSI findings suggest that par- ents of high-risk siblings are detecting differences at around the same time as the technology measures and possibly earlier than laboratory- or clinic-based behavio- ral observations.

Our findings are promising, but not without limitations. We did not obtain inter-rater agreement, but rather requested responses from informants identified as the primary car- egiver. We did record who identified himself or herself as the primary caregiver (e.g. father, mother) and although the overwhelming majority were identified as “mother,” it is

possible that the different respondents may have added variability to the responses recorded. Additionally, children included in the analyses did not have had an APSI com- pleted at each age assessed. To be included, children only needed to have one APSI completed, reducing potential bias from parents who chose to participate versus non-par- ticipants. Finally, parents received feedback concerning their child’s performance at each visit. This in turn may have affected how the parent scored later APSIs. In line with this, parental ratings could also be influenced by their experience with the older sibling, affecting how the APSI was scored. As such, ratings reported in this research con- text, which provided the parents of high-risk children with ongoing feedback, might be fundamentally different from parental experiences when completing the APSI as part of a universal screen at general wellness visits. We also acknowledge that the APSI may perform differently in infants who are at increased risk of ASD for reasons other than family history. Previous research using the AOSI to

Table 8. Group differences for individual questions at 18 months of age.

APSI at 18 months

HR-ASD (n = 56)

HR-non-ASD (n = 122)

LR-non-ASD (n = 73)

HR-ASD versus HR-N

HR-ASD versus LR

HR-N versus LR

Item % 1 % 2 % 1 % 2 % 1 % 2 p value p value p value

1—Visual tracking 2 – – 2 – 1 0.46 0.84 0.57 2—Visual fixation 18 4 8 1 5 1 0.01* 0.02* 0.96 3—Respond to name 32 13 10 1 11 – 0.001* 0.001* 0.83 4—Response to facial emotion 41 7 39 6 27 – 0.46 0.008* 0.02* 5—Anticipatory social response 9 4 1 1 1 – 0.002* 0.002* 0.78 6—Imitation 21 5 14 – 8 – 0.002* 0.001* 0.23 7—Vocalize with you 20 7 14 1 3 – 0.002* 0.001* 0.03* 8—Eye contact 21 5 7 2 5 3 0.002* 0.003* 0.95 9—Reciprocal social smile 16 – 4 – 1 – 0.001* 0.001* 0.43 10—Coordinate actions with eye gaze 18 – 3 1 10 – 0.03* 0.33 0.25 11—Reactivity 23 11 8 3 5 – 0.001* 0.001* 0.14 12—Cuddle with you 16 5 7 1 7 – 0.002* 0.002* 0.80 13—Difficult to soothe 18 9 14 1 5 4 0.02* 0.009* 0.61 14—Social interest and affect 14 13 10 1 – 1 0.001* 0.001* 0.17 15—Difficulty with change 25 7 25 – 27 – 0.10 0.21 0.77 16—Hand use/holding objects 11 4 1 1 1 1 0.003* 0.05 0.37 17—Repetitive motor behaviors 16 9 15 2 3 – 0.02* 0.001* 0.01* 18—Another person’s hand as tool 16 18 18 11 22 21 0.39 0.33 0.03* 19—Unusual sensory behaviors 7 9 6 – 3 4 0.001* 0.03* 0.40 20—Focusing attention on objects 25 9 16 – 14 5 0.001* 0.02* 0.34 21—Insistence on particular object 18 13 14 3 16 10 0.04* 0.57 0.11 22—Resist play/fixed play routines 13 2 2 – – 1 0.003* 0.004* 0.89 23—Share interests with others 30 2 20 2 11 3 0.09 0.03* 0.45 24—Distal point 25 18 16 9 15 4 0.007* 0.001* 0.31 25—Use gestures 14 11 5 – – 1 0.001* 0.001* 0.50 26—Loss of skill 7 4 1 – – – 0.001* 0.001* 0.63

APSI: Autism Parent Screen for Infants; HR-ASD: high-risk infants diagnosed with autism spectrum disorder; LR: low risk; HR-N: high-risk infants not diagnosed with autism spectrum disorder; “–”: no code at that score for item. *p < 0.04 highlighted.

332 Autism 22(3)

Table 9. ROC characteristics for total score on the APSI for the high-risk sample only.

Total scale score

Age AUC (CI) Cutoff Sensitivity Specificity LHR+ LHR− OR J PPV NPV

6 0.78 (0.62–0.94) 14 0.67 0.77 2.93 0.34 6.80 0.44 0.45 0.82

15a 0.67 0.86 4.89 0.20 12.67 0.53 0.47 0.83

16 0.5 0.89 4.40 0.23 7.80 0.39 0.60 0.84

9 0.8 (0.69–0.92) 13 0.62 0.84 3.92 0.26 8.67 0.46 0.68 0.80

14a 0.62 0.87 4.70 0.21 10.73 0.49 0.70 0.79

15 0.52 0.9 4.98 0.20 9.35 0.42 0.74 0.78

12 0.66 (0.53–0.79) 9 0.59 0.67 1.76 0.57 2.86 0.26 0.63 0.79

10a 0.59 0.72 2.09 0.48 3.64 0.31 0.65 0.76

11 0.5 0.74 1.95 0.51 2.90 0.24 0.67 0.76

15 0.74 (0.61–0.87) 9 0.61 0.75 2.34 0.41 4.75 0.36 0.69 0.77

10a 0.58 0.83 3.48 0.29 6.92 0.41 0.73 0.76

11 0.45 0.88 3.61 0.28 5.76 0.33 0.80 0.75

18 0.68 (0.51–0.84) 8 0.74 0.6 1.85 0.54 4.25 0.34 0.63 0.78

9a 0.65 0.72 2.33 0.43 4.82 0.37 0.68 0.77

10 0.61 0.72 2.17 0.46 4.00 0.33 0.69 0.75

24 0.72 (0.55–0.90) 11 0.58 0.77 2.52 0.39 4.58 0.35 0.72 0.78

12a 0.58 0.85 3.76 0.27 7.56 0.43 0.79 0.77

13 0.53 0.85 3.42 0.29 6.11 0.37 0.85 0.77

AUC: area under the curve; CI: confidence interval (95%); LHR+: positive likelihood ratio; LHR−: negative likelihood ratio; OR: odds ratio; J: Youden’s Index; PPV: positive predictive value; NPV: negative predictive value. aShading indicates the optimal cutoff (via Youden’s J).

assess early signs of ASD in other at-risk groups indicates overlapping findings with early signs in younger siblings, but also important differences that may relate to underlying differences in biology and cognitive profiles (Jeste et al., 2016; Roberts et al., 2016; Yaari et al., 2016). Nevertheless, the strengths of our study include the large high-risk sibling sample, which included 66 children diagnosed with ASD, and availability of outcome data on the full sample, includ- ing “screen-negative” HR and LR children. Future work could assess the validity of the APSI in other high- and low- risk samples to assess potential generalizability beyond the high-risk context.

Based on the current findings, the APSI has potential to differentiate, starting at 6 months of age, high-risk sib- lings who will be diagnosed with ASD at 3 years of age from both high-risk siblings who do not have ASD and community controls. Although additional research is needed, the APSI shows promise as a simple, low-cost parent-report monitoring system. The use of the APSI may lead to earlier identification of children who could benefit from increased monitoring and/or early interven- tion to remediate problematic behavior in children at risk for an ASD diagnosis.

Acknowledgements

The authors would like to thank the research assistants at each site for their help with data collection and the parents and chil- dren who participated in our study.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was funded by Canadian Institutes of Health Research (CIHR) and NeuroDevNet.

References

Akobeng AK (2007) Understanding diagnostic tests 3: receiver operating characteristic curves. Acta Paediatrica 96: 644–647.

Allison C, Auyeung B and Baron-Cohen S (2012) Toward brief “red flags” for autism screening: the short autism spectrum quotient and the short quantitative checklist in 1,000 cases and 3,000 controls. Journal of the American Academy of Child and Adolescent Psychiatry 51(2): 202–212.e7.

Allison C, Baron-Cohen S, Wheelwright S, et al. (2008) The Q-CHAT (quantitative checklist for autism in toddlers): a normally distributed quantitative measure of autistic traits at 18–24 month of age: preliminary report. Journal of Autism and Developmental Disorders 38: 1414–1425.

Sacrey et al. 333

American Psychiatric Association (2013) Diagnostic and Statistical Manual of Mental Disorders. 5th ed. Arlington, VA: American Psychiatric Publishing.

Benjamini Y and Hochberg Y (1995) Controlling the false dis- covery rate: A practical and powerful approach to multiple testing. Journal of the Royal Statistical Society, Series B (Methodological) 57: 289–300.

Ben-Sasson A and Carter AS (2012) The application of the first year inventory for ASD screening in Israel. Journal of Autism and Developmental Disorders 42: 1906–1916.

Bolton PF, Golding J, Emond A, et al. (2012) Autism spec- trum disorder and autistic traits in the Avon Longitudinal Study of Parents and Children: precursors and early signs. Journal of the American Academy of Child and Adolescent Psychiatry 51(3): 249–260.

Brian J, Bryson SE, Garon N, et al. (2008) Clinical assessment of autism in high-risk 18-month-olds. Autism 12(5): 433–456.

Briggs-Gowan MJ, Carter AS, Irwin JR, et al. (2004) The brief infant-toddler social and emotional assessment: screening for social-emotional problems and delays in competence. Journal of Pediatric Psychology 29: 143–155.

Bryson SE and Zwaigenbaum L (2014) Autism Observation Scale for infants. In: Patel VB, Preedy VR and Martin CR (eds) Comprehensive Guide to Autism. New York: Springer, pp.299–310.

Bryson SE, Zwaigenbaum L, McDermott C, et al. (2008) The Autism Observation Scale for infants: scale development and reliability data. Journal of Autism and Developmental Disorders 38(4): 731–738.

Carter AS, Volkmar FR, Sparrow SS, et al. (1998) The Vineland Adaptive Behavior Scales: supplementary norms for indi- viduals with autism. Journal of Autism and Developmental Disorders 28(4): 287–302.

Chakrabarti S and Fombonne E (2005) Pervasive developmental disorders in preschool children: confirmation of high preva- lence. American Journal of Psychiatry 162: 1133–1141.

Chawarska K, Macari S and Shic F (2013) Decreased sponta- neous attention to social scenes in 6-month-old infants later diagnosed with autism spectrum disorders. Biological Psychiatry 74(3): 195–203.

Cohen IL, Gardner JM, Karmel BZ, et al. (2013) Neonatal func- tion and 4-month arousal-modulated attention are jointly associated with autism. Autism Research 6: 11–22.

Daniels AM and Mandell DS (2013) Explaining differences in age at autism spectrum disorder diagnosis: A critical review. Autism 18(5): 583–597.

Dawson G, Rogers S, Munson J, et al. (2010) Randomized, con- trolled trial of an intervention for toddlers with autism: the early start Denver model. Pediatrics 125(1): e17–e23.

Dietz C, Swinkels S, Van Daalen E, et al. (2006) Screening for autistic spectrum disorder in children aged 14–15 months. II: population screening with the early screening of autis- tic traits questionnaire (ESAT). Design and general find- ings. Journal of Autism and Developmental Disorders 36: 713–722.

Eisert DC, Sturner RA and Mabe PA (1991) Questionnaires in behavioural pediatrics: guidelines for selection and use. Journal of Developmental and Behavioral Pediatrics 12(1): 42–50.

Elison JT, Paterson SJ, Wolff JJ, et al. (2013) White matter microstructure and atypical visual orienting in 7-month-olds at risk for autism. American Journal of Psychiatry 170(8): 899–908.

Elsabbagh M, Mercure E, Hudry K, et al. (2012) Infant neural sensitivity to dynamic eye gaze is associated with later emerging autism. Current Biology 22(4): 338–342.

Feldman MA, Ward RA, Savona D, et al. (2012) Development and initial validation of a parent report measure of the behav- ioral development of infants at risk for autism spectrum dis- order. Journal of Autism and Developmental Disorders 42: 13–22.

Fischer JE, Bachmann LM and Jaeschke R (2003) A readers’ guide to the interpretation of diagnostic test properties: clinical example of sepsis. Intensive Care Medicine 29: 1043–1051.

Glascoe FP (2000) Evidence-based approach to develop- mental and behavioural surveillance using parents’ concerns. Child: Care, Health and Development 26: 137–149.

Glascoe FP, Foster M and Wolraich ML (1997) An economic analysis of developmental detection methods. Pediatrics 99(6): 830–837.

Glascoe FP, Macias MM, Wegner LM, et al. (2007) Can a broad- band developmental-behavioural screening test identify children likely to have autism spectrum disorder? Clinical Pediatrics 46(9): 801–805.

Gotham K, Pickles A and Lord C (2009) Standardizing ADOS severity scores for a measure of severity in Autism Spectrum Disorders. Journal of Autism and Developmental Disorders 39(5): 693–705.

Gotham K, Risi S, Pickles A, et al. (2007) The Autism Diagnostic Observation Schedule: Revised algorithms for improved diagnostic validity. Journal of Autism and Developmental Disorders 37(4): 613–627.

Jeste SS, Varcin KJ, Hellemann GS, et al. (2016) Symptom pro- files of autism spectrum disorder in tuberous sclerosis com- plex. Neurology 87(8): 766–772.

Johnson CP and Myers SM (2007) Identification and evalua- tion of children with autism spectrum disorders (American Academy of Pediatrics Council on Children With Disabilities). Pediatrics 120: 1183–1215.

Jones W and Klin A (2013) Attention to eyes is present but in decline in 2-6-month-old infants later diagnosed with autism. Nature 504(7480): 427–431.

Kruizinga I, Visser JC, Van Batenburg-Eddes T, et al. (2014) Screening for autism spectrum disorders with the brief infant-toddler social and emotional assessment. PLoS ONE 9(5): e97630.

Landa R and Garrett-Mayer E (2006) Development in infants with autism spectrum disorders: a prospective study. Journal of Child Psychology and Psychiatry 47(6): 629–638.

Landa RJ, Gross AL, Stuart EA, et al. (2012) Latent class anal- ysis of early developmental trajectory in baby siblings of children with autism. Journal of Child Psychology and Psychiatry 53(9): 986–896.

Lord C, Rutter M and LeCouteur A (1994) Autism diagnos- tic interview-revised: a revised version of the diagnos- tic interview for caregivers of individuals with possible

334 Autism 22(3)

pervasive developmental disorders. Journal of Autism and Developmental Disorders 24: 659–685.

Lord C, Risi S, Lambrecht L, et al. (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(3): 205–223.

Lord C, Rutter M, Goode S, et al. (1989) Autism diagnostic obser- vation schedule: a standardized observation of communicative and social behavior. Journal of Autism and Developmental Disorders 19(2): 185–212.

Mullen E (1995) Mullen Scales of Early Learning: American Guidance Services. Circle Pines, MN.

Ozonoff S, Iosif AM, Baguio F, et al. (2010) A prospective study of the emergence of early behavioral signs of autism. Journal of the American Academy of Child and Adolescent Psychiatry 49(3): 256–66.e1-2.

Perry A, Cummings A, Dunn Geier J, et al. (2008) Effectiveness of intensive behavioral intervention in a large, community- based program. Research in Autism Spectrum Disorders 2: 621–642.

Reznick JS, Baranek GT, Reavis S, et al. (2007) A parent-report instrument for identifying one-year-olds at risk for an even- tual diagnosis of autism: the first year inventory. Journal of Autism and Developmental Disorders 37: 1691–1710.

Roberts JE, Tonnsen BL, McCary LM, et al. (2016) Brief report: autism symptoms in infants with fragile X syndrome. Journal of Autism and Developmental Disorders. Epub ahead of print 14 September. DOI: 10.1007/s10803-016-2903-5.

Robins DL, Casagrande K, Barton M, et al. (2014) Validation of the modified checklist for autism in toddlers, revised with follow-up (M-CHAT-R/F). Pediatrics 133(1): 37–45.

Sacrey LA, Bryson SE and Zwaigenbaum L (2013) Prospective examination of visual attention during play in infants at high-risk for autism spectrum disorder: a longitudinal study from 6 to 36 months of age. Behavioural Brain Research 256: 441–450.

Sacrey LR, Zwaigenbaum L, Bryson S, et al. (2015) Can parents’ concerns predict autism spectrum disorder? A prospec- tive study of high-risk siblings from 6 to 36 months of age. Journal of the American Academy of Child and Adolescent Psychiatry 54(6): 470–478.

Sallows GO and Graupner TD (2005) Intensive behavioral treatment for children with autism: four-year outcome and predictors. American Journal on Mental Retardation 110(6): 417–438.

Siegel B, Anders TF, Ciaranello RD, et al. (1986) Empirically derived subclassification of the autistic syndrome. Journal of Autism and Developmental Disorders 16(3): 275–293.

Smith NJ, Sheldrick C and Perrin EC (2013) An abbreviated screening instrument for autism spectrum disorders. Infant Mental Health Journal 34(2): 149–155.

Sparrow SS, Balla D and Cicchetti D (1984) Vineland Adaptive Behavior Scales (Survey Form). Circle Pines, MN: American Guidance Service.

Swinkels SH, Dietz C, Van Daalen E, et al. (2006) Screening for autistic spectrum in children aged 14 to 15 months. I: the development of the early screening of autistic traits ques- tionnaire (ESAT). Journal of Autism and Developmental Disorders 36(6): 723–732.

Turner-Brown LM, Baranek GT, Reznick JS, et al. (2012) The first year inventory: a longitudinal follow-up of 12-month- old to 3-year-old children. Autism 17(5): 527–540.

Volkmar FR, Carter A, Sparrow SS, et al. (1993) Quantifying social development in autism. Journal of the American Academy of Child and Adolescent Psychiatry 32(3): 627–632.

Watson LR, Baranek GT, Crais ER, et al. (2007) The first year inventory: retrospective parent responses to a questionnaire designed to identify one-year-olds at risk for autism. Journal of Autism and Developmental Disorders 37(1): 49–61.

Wetherby AM, Brosnan-Maddox S, Peace V, et al. (2008) Validation of the infant-toddler checklist as a broadband screener for autism spectrum disorders from 9 to 24 months of age. Autism 12(5): 487–511.

Yaari M, Yitzhak N, Harel A, et al. (2016) Stability of early risk assessment for autism spectrum disorder in preterm infants. Autism 20(7): 856–867.

Zwaigenbaum L, Bryson SE, Brian J, et al. (2015) Stability of diagnostic assessment for autism spectrum disorder between 18 and 36 months in a high-risk cohort. Autism Research. Epub ahead of print 27 November. DOI: 10.1002/aur.1585.

Zwaigenbaum L, Bryson SE, Rogers T, et al. (2005) Behavioral manifestations of autism in the first year of life. International Journal of Developmental Neuroscience 23: 143–152.