OBJECTIVE This review summarises the recent trends in research in the field of self-injurious behaviour in people with intellectual disability and autism spectrum disorder. RESULTS New data on incidence, persistence and severity add to studies of prevalence to indicate the large scale of the clinical need. A number of person characteristics have been repeatedly identified in prevalence and cohort studies that: can be considered as risk markers (e.g. stereotyped behaviour, autism spectrum disorder) and indicate possible causal mechanisms (e.g. sleep disorder, anxiety). Studies have started to integrate traditional operant learning paradigms with known person characteristics and reviews and meta-analyses of applied behaviour analytic procedures can now inform practice. CONCLUSIONS Despite these positive developments interventions and appropriate support falls far short of the required need. Expansions in applied research are warranted to develop and evaluate innovative service delivery models that can translate knowledge of risk markers and operant learning paradigms into widespread, low cost routine clinical practice. Alongside this, further pure research is needed to elucidate the direction of causality of implicated risk factors, in order to understand and intervene more effectively in self-injury.