A Promis®-Based Approach to Item Development and Factor Analysis of the Emotion Dysregulation Inventory for Youth with ASD

Background: Poor emotion regulation maintains behavioral problems in autism spectrum disorder. Thus, emotional dysregulation is an ideal treatment target. However, measures of this construct are lacking for youth with ASD, and available measures that do exist are reliant on verbal ability. Little is known about the presentation of emotion dysregulation in youth with ASD – especially across the range from nonverbal to fully verbal and intellectual capacity.

Objectives: Using Patient-Reported Outcomes Measurement Information System (PROMIS) guidelines, this presentation will describe the creation of items tapping observable manifestations of poor emotion regulation in youth with ASD, and the dimensionality of emotion dysregulation in ASD. These represent the initial steps in the development of the Emotion Dysregulation Inventory (EDI).

Methods: A comprehensive literature search and the development of a conceptual model occurred to derive the item-bank. Candidate items underwent expert review, followed by interviews conducted with caregivers of 19 individuals with ASD who ranged in age, verbal ability, and cognitive functioning. The interviews combined a think-aloud and debriefing methodology; further detail will be provided. The 66 candidate items were evaluated via caregiver report in two samples: (1) Interactive Autism Network (IAN; n=1323, ages 6-17) representing the full range of ASD severity and adaptive functioning; and (2) Autism Inpatient Collection (AIC; n=432, age 4-20). The AIC is a six-site study of psychiatric inpatients with ADOS-confirmed ASD that includes participants with extreme forms of emotion dysregulation. Exploratory factor analysis (EFA) and confirmatory factor analysis (CFA) were conducted in non-overlapping subsamples (n=870 for EFA, 885 for CFA).

Results: Clinicians and researchers with expertise in ASD reviewed candidate items; recommended modifications were incorporated. Parent interviews indicated that items had sufficient coverage of emotional reactivity and regulation and were understandable to parents. Collectively, these steps provided evidence of face validity for the target construct. EFA identified a 2-factor model: (1) Reactivity, characterized by intense, rapidly escalating, sustained, and poorly regulated negative emotional reactions, and (2) Dysphoria, characterized by anhedonia, sadness, and nervousness. The factor structure did not differ for the IAN and AIC samples, despite lower verbal and cognitive ability in AIC compared to IAN. CFA in an independent sample demonstrated adequate to good fit for both factors. An overall total score to subsume the two factors was not supported.

Conclusions: The strategy used to develop the EDI was effective in creating an item bank with sufficient content coverage and ease of understanding. Factor analyses of EDI candidate items supported distinct dimensions of emotion dysregulation in ASD. Consistent with studies in non-ASD samples, the tendency to experience intense negative emotion and difficulty regulating emotion emerged as a single factor (i.e., Reactivity), which was separable from general unease (Dysphoria). Future directions include the downward age extension of the EDI to examine the presentation of emotion dysregulation in younger children.