3.30 Predicting Service Urgency in Children and Youth with ASD: The Development of an Algorithm

Title3.30 Predicting Service Urgency in Children and Youth with ASD: The Development of an Algorithm
Publication TypeConference Paper
Year of Publication2022
AuthorsKing G.KC, Van Dyke J.N, Poss J.W, Stewart S.L
Conference NameJournal of the American Academy of Child and Adolescent Psychiatry
Issue10 Supplement
Keywords*algorithm, *juvenile, Child, clinical decision making, community mental health, conference abstract, controlled study, Decision Making, decision tree, diagnostic test accuracy study, Female, health care quality, Human, human experiment, major clinical study, Male, mental health, Ontario, patient triage

Objectives: ASD is characterized by an array of core features, but there is a broad range of impairment associated with this disorder. As a result, there is a need to improve triaging and prioritization of services in children and youth with ASD. Currently, there is no empirically supported system to assess service urgency needs among children and youth with ASD seeking support. To address this gap in the literature, the current study describes the clinical decision-making and empirical methods used to validate and develop an algorithm for service urgency for children and youth with ASD. Method(s): The interRAI Child and Youth Mental Health (ChYMH) instruments were used to assess trends in service urgency for those with probable ASD. Data from 53,387 ChYMH-Screener (ChYMH-S) assessments were completed as part of standard of care in community mental health agencies across the Province of Ontario. Since service urgency, but not diagnostic information, was available for the ChYMH-S, specific behaviors that were endorsed on ChYMH assessments with a confirmed diagnosis of ASD (n = 2584) were used as indicators of a probable diagnosis within the ChYMH-S sample. This method identified 1598 children and youth with probable ASD, totaling 3% of the ChYMH-S sample. Using these individuals, a decision-tree tool was used to identify predictors of service urgency. Informed by a k-means cluster algorithm, branches of the decision tree were grouped together, each representing distinct levels of likelihood of service urgency. Logistic regression was used to assess the fit of the algorithm. Result(s): The final algorithm identified 12 different items, which are also aligned with previous research, to predict service urgency in the sample. The algorithm produced 5 groups, ranging from lowest urgency to highest. For the highest-urgency group, 36.1% required services within 7 days, 11.6 times more likely than the lowest-urgency group. Conclusion(s): This algorithm is the first empirically and clinically supported decision-making tool that can be used to inform the choices related to triaging services for children and youth with ASD. By prioritizing mental health resources for these individuals, efficient and appropriate access to community-based mental health resources can be provided. ASD, TREAT, SACCopyright © 2022

Reseach Notes