Empirical Approaches to the Classification of Avoidant/Restrictive Food Intake Disorder.

Abstract

Avoidant/restrictive food intake disorder (ARFID) is a relatively new formal diagnosis for which empirical classification research (defined here as studies using latent class/latent profile analysis-type methods) is still emerging. Such research focused on ARFID is an important gap to fill given questions about (1) the boundaries between ARFID and phenotypically similar presentations (e.g., eating disorders [EDs] such as anorexia nervosa [AN], and pediatric feeding disorder [PFD]), and (2) within-ARFID heterogeneity. These questions have practical implications, including diagnostic reliability and treatment selection. This forum synthesizes the limited empirical classification literature seeking to quantitatively distinguish ARFID from non-ARFID EDs or from PFD, and/or characterize within-ARFID heterogeneity. To our knowledge, only five studies in clinical samples have used empirical classification methods to delineate ARFID from non-ARFID EDs and/or characterize within-ARFID heterogeneity; no studies have used such methods to delineate ARFID from PFD. Existing studies are mixed in determining how well ARFID can be distinguished from other EDs (particularly AN), but converge in identifying several potential ARFID subclasses (i.e., sensory sensitivity, low appetite, feared eating-related consequences, and subclass representing a combination of these) with some overlapping features. The existing ARFID empirical classification literature should guide future ARFID classification research priorities (e.g., incorporating mechanistic variables as classification indicators, incorporating longitudinal variables as classification validators) to inform differences between ARFID and other disorders and between ARFID presentations. Dimensional approaches to conceptualizing, studying, and modeling psychopathology (namely, the Hierarchical Taxonomy of Psychopathology [HiTOP] and the Research Domain Criteria [RDoC]) may offer useful insights.