Developing ASD Screening Criteria for the Brief Infant Toddler Social Emotional Assessment (BITSEA)

Background: Retrospective and prospective studies have confirmed that symptoms of ASD emerge during the first two years of life (Ozonoff et al., 2011). Yet, the national average age of diagnosis is above 36 months (CDC, 2012), with a greater delay for children from low socioeconomic statuses, minority groups, and non-native English speaking families (Valicenti-McDermott et al., 2012). With increasing evidence that early intervention significantly improves outcomes for children with ASD, early detection and diagnosis is critically needed (Seida et al., 2009). The AAP and CDC recommend that pediatricians administer developmental screeners at routine visits; universal screening can potentially ameliorate health disparities regarding age of diagnosis. As public awareness of the disorder grows, the number of screening tools available to identify children at risk for developing ASD has also increased. The Brief Infant Toddler Social Emotional Assessment (BITSEA) (Briggs-Gowan & Carter, 2006) is a screening tool designed to identify toddlers with social-emotional/behavioral problems and/or delays/deficits in social-emotional competencies. Of the 42 BITSEA items, 17 items reflect behaviors consistent with ASD symptoms. 

Objectives: The goal of this preliminary study was to determine the feasibility of developing cut-points on the new BITSEA ASD-Problem, ASD-Competency, and ASD-Total scales that reliably discriminate children with ASD from those without. The new scales are derived from existing BITSEA items. 

Methods: Data are presented on 436 toddlers (71.3% male) between the ages of 12 and 48 months. 51% of toddlers were diagnosed with ASD using gold-standard tools and clinical judgment, while the remaining children were typically developing or diagnosed with intellectual/developmental disabilities or non-ASD psychopathology. Mothers of participants completed the BITSEA. Analyses examined the following subscales: ASD-Problem (including nine items focused on ASD negative symptoms), ASD-Competency (including eight items focused on ASD positive symptoms), and ASD-Total (sum of the ASD-Problem and reverse-scored ASD-Competency subscales).

Results:  ROC plots were developed for the ASD subscales (see Figure 1). The following cut-points were determined as most effective for the ASD-Problem, ASD-Competency, and ASD-Total subscales respectively: 4.5, 10.5, and 14.5. The cut-points on all subscales evidenced moderate-to-high discriminative power and moderate-to-high sensitivity, specificity, and PPV.  Because the analyses yielded fractionated cut-points, optimal whole number cut-points were determined to increase clinical utility. The following whole number cut-points were calculated as the most effective for the ASD-Problem, ASD-Competency, and ASD-Total subscales respectively: 5, 10, and 14 (see Table 1). Of the three subscales, the ASD-Total was determined as the most effective. The clinical cut-point of 14 yielded 82.43% sensitivity, 84.21% specificity, and 80.26% PPV. False positives had significantly lower cognitive and language abilities than true negatives. False negatives showed higher language abilities, than true positives. 

Conclusions:  The BITSEA ASD-Total scale was shown to be a highly effective screening measure for ASD in toddlers. These findings contribute to the overarching goal of increased early detection and diagnosis of ASD and give pediatric healthcare providers an efficient and effective ASD screening tool within a broadband behavioral health screener. This preliminary work sets the stage for testing cut-points prospectively in early intervention and pediatric settings.