When Should Next Generation Sequencing be Used in Children at High-Risk for Autism Spectrum Disorder?

Thursday, May 12, 2016: 5:30 PM-7:00 PM
Hall A (Baltimore Convention Center)
T. Yuen1,2, W. Ungar3, P. Szatmari4 and M. T. Carter5, (1)Technology Assessment at Sick Kids, Hospital for Sick Chidlren, Toronto, ON, Canada, (2)Institute of Health Policy, Management and Evaluation, University of Toronto, Toronto, ON, Canada, (3)Hospital for Sick Chidlren, Toronto, ON, Canada, (4)Centre for Addiction and Mental Health, Toronto, ON, Canada, (5)Hospital for Sick Children, Toronto, ON, Canada
Background:   Next generation sequencing (NGS) is a novel but costly genetic testing technology that is currently reserved for syndromic children with autism spectrum disorder (ASD) who failed to receive a genetic diagnosis from traditional genetic testing. Using NGS earlier in the diagnostic pathway could generate information to aid diagnosis and lead to earlier treatment, but its optimal position in the diagnostic pathway remains unknown. 

Objectives:   The objective of this study was to construct a model of the diagnostic pathway from birth to age 6 in children at high risk for ASD to predict the health consequences of introducing NGS at different time points.

Methods:   The diagnostic pathway for children at high-risk for ASD was modeled using discrete event simulation to predict age at diagnosis, wait time for treatment, and total costs. Clinical progression was structured in consultation with clinicians and current Canadian guidelines. The model incorporates ASD risk factors, recognized phenotypes of genetic mutations and developmental trajectories based on two prospective cohort studies (the Infant Sibling Study and Pathways to ASD Study) and published literature.

Results:  The model accounts for heterogeneity in clinical pathway associated with individual characteristics such as gender, family history, congenital anomalies, ASD symptom severity and developmental milestones. Moreover, the model can simulate queues for health services and predict wait times for clincial assessment and treatment. In turn, the average age at diagnosis and age at treatment initiation can be estimated when NGS was introduced at different time points along the diagnostic pathway.

Conclusions:   The simulation model from this study can generate much needed information for clinicians and decision-makers on how to integrate NGS in clinical settings. This model will be used in economic evaluations to help decide whether NGS can replace cheaper but less effective genetic tests.  Reduction in unnecessary testing could also reduce delays in diagnosis and wait times for ASD services, which are critical issues in improving long term outcomes in ASD.