Accessibility of Health Services and Age at Autism Diagnosis

Background:   Early diagnosis of autism is important to the success of early intervention.  Yet there is a well-documented delay ranging 2-4 years between a parent’s concern and eventual autism diagnosis; some children are missed entirely.  The relation between geographic access to health services and the timing of autism diagnosis has not been studied.
Objectives:   We evaluated the impact of several health services accessibility factors at birth on the age at autism diagnosis among children with autism.
Methods:   We included 8-year-old children who met the standardized criteria for autism used by the Autism and Developmental Disabilities Monitoring Network in North Carolina and had a matching NC birth certificate.  This multiple-source, active, records-based surveillance network relies on documented behavioral symptoms in developmental records, from which we obtained the presence of and age at a prior autism diagnosis.  Health accessibility exposures were measured as provider density, distance to the nearest provider or facility, and residence within a Health Professional Shortage Area (HPSA).  Addresses for primary care physicians serving children, specialist physicians likely to diagnose autism, and psychologists were obtained from a medical and psychologist board licensure database.  Other diagnostic facilities included the Division of Treatment and Education of Autistic and related Communication handicapped Children, the Child Development Service Agencies of the Early Intervention Program (CDSA), and NC teaching hospitals.  Child and provider addresses were geocoded.  Provider density and distances were calculated using geographic information systems.  A multivariable Cox survival model was used to estimate hazard ratios (HR) of the age at previous autism diagnosis,  adjusting for social and demographic factors from the surveillance record (race and child’s cognitive status), birth certificate (maternal education, maternal age, marital status), and census (block group median household income). 
Results:  Of 220 children with confirmed autism and a matching birth certificate, 51 (23%) had no previous autism diagnosis.  The remainder had a mean age of diagnosis of 57 months (standard deviation: 22 months, median: 52 months).  Several accessibility factors were associated with earlier autism diagnosis, as indicated by elevated adjusted HRs, including:  higher count of specialist physicians within a 20-mile road network radius (HR and 95% confidence interval of 1.6 [1.0, 2.4] comparing 250+ providers to < 75), shorter distance to a CDSA (1.4 [0.8, 2.2] comparing < 5 miles to > 20 miles), shorter distance to a teaching hospital (1.5 [1.0, 2.2] comparing proximity < 5 miles to > 20 miles), and not residing within a HPSA (1.3 [0.7, 2.1]).  HRs for accessibility measures and early diagnosis were elevated to a greater degree among children with developmental loss or plateau, with married parents, or living in urban areas. 
Conclusions:   Better geographic accessibility of health services appears to be associated with an increased likelihood of an earlier age at autism diagnosis.  The impact of accessibility was found using several different geographic accessibility measures.  Many determinants of accessibility were not measured in this case-only study.  Considering the location of health services may be useful in targeting efforts to improve early identification of autism.