We read with interest the study by Duerden et al. (
2012) about factors predicting the presence of self-injury in a well-characterized cohort of 241 autistic children and adolescents. This study used a hierarchical regression model to test whether seven considered factors (in order: atypical sensory processing, IQ, functional communication, social communication, age, need for sameness, rituals and compulsions) predicted self-injury. Two other influences on self-injury, gender and ADOS autism severity scores, were also explored. The authors concluded that “abnormal sensory processing was the strongest single predictor of self-injury followed by sameness, impaired cognitive ability and social functioning.” On this basis, they recommend treatments to reduce autistic self-injury which focus on the two strongest predictors. We agree with the authors about the importance of understanding self-injury in order to improve the situation of autistic individuals, and we believe their striking results deserve attention. However, we question how these results were reported, interpreted, and translated into proposed courses of action. …
