Risk markers for self-injurious behaviour in children with intellectual disability

Download Statistics

Downloads

Downloads per month over past year

Steenfeldt-Kristensen, Catherine (2019). Risk markers for self-injurious behaviour in children with intellectual disability. University of Birmingham. Clin.Psy.D.

Preview

SteenfeldtKristensen2019ClinPsyD_Vol_1.pdf
Text - Accepted Version
Available under License All rights reserved.
Download (3MB) | Preview

Abstract

Background: Previous research has identified a number of child and behavioural characteristics that are associated with self-injurious behaviour (SIB) in those with an intellectual disability and developmental delay. However, to date, few studies have explored the unique contribution of each risk marker to the presence of SIB, nor have any studies translated these risk markers into a clinical algorithm that can distinguish between the presence or absence of SIB. Therefore, the aim of the current study was to develop a model that classifies the presence or absence of SIB in children who have these known risk markers, as identified by the Self-injury, Aggression and Destruction-Screening Questionnaire (SAD-SQ).
Method: The study utilised existing data from previous studies that had recruited individuals with a confirmed or suspected intellectual disability or developmental delay and who had used the SAD-SQ as a measure in their studies. These data formed the training sample (N=1540) which was used to develop the risk model to predict presence or absence of SIB. Eight possible predictor variables were entered into the model. Binary logistic regression was used to identify the most predictive risk markers for SIB. These risk markers formed the final risk model and subsequent risk algorithm. The algorithm was then applied to a new test sample of children (N=320) and receiver operating characteristic (ROC) curve analysis used to assess the algorithm’s predictive accuracy.
Results: Diagnosis of autism, presence of health conditions, repetitive behaviour, impulsivity and age were predictive of SIB. Gender, diagnosis of a genetic syndrome and level of ability were the least predictive of SIB, and therefore, excluded from the final risk model. At the optimum cut-off of 0.28, the risk algorithm had a sensitivity of 77% and a specificity of 58%. When applied to the test sample at the same cut-off, the algorithm had a sensitivity of 87% and a specificity of 34% with positive and negative predictive values of 64.3% and 66.7%, respectively. ROC analysis provided an area under the receiver operating curve (AUC) value of .608 which is considered moderate. Analysis of the most severe cases of SIB did not alter the accuracy of the model.
Conclusion: The SAD-SQ is a sensitive screening tool that offers a simple and reliable way of screening those at risk of developing SIB in individuals with a suspected or confirmed intellectual disability or developmental delay. These results are discussed in relation to clinical and theoretical implications and areas for future research.

Type of Work:

Thesis (Doctorates > Clin.Psy.D.)

Award Type:

Doctorates > Clin.Psy.D.

Supervisor(s):