32257
Practices Surrounding Autism over-Diagnosis and Under-Diagnosis: Results from a National Healthcare Provider Survey
Objectives: To measure rates of healthcare provider-reported over- and under-diagnosis of ASD in a national sample of U.S. ASD providers. To assess providers’ reasons for over- and/or under-diagnosis of ASD.
Methods: A mixed-mode (email/mail) survey was sent to all members of the Society of Developmental and Behavioral Pediatrics (SDBP) who indicated a clinical interest in ASD, performed at least 5 ASD evaluations in the prior year, and who were currently practicing in the US. Each respondent used a Likert-type scale to indicate: (1) whether he/she thought autism was over-diagnosed, under-diagnosed, or diagnosed at the appropriate rate; (2) whether he/she had personally diagnosed a child with ASD when the child did not meet DSM criteria (“over-diagnosis”); (3) whether other providers in his/her area over-diagnose children with ASD; (4) whether he/she had personally not diagnosed a child with ASD when the child did meet DSM criteria (“under-diagnosis”). Reasons for under- and over-diagnosis of ASD were assessed. Descriptive statistics were used to assess rates of self-reported over- and under- diagnosis, and multivariable logistic regression was used to assess variability in responses according to respondent gender, years in practice, and US region
Results: 217/347 (63%) of eligible providers completed the survey. Overall, 48% of providers thought that autism was diagnosed at an appropriate rate. More providers thought ASD was over diagnosed (32%) than underdiagnosed (19%; p = 0.008). 8.8% of providers self-reported over-diagnosing ASD at least sometimes; however, 58% reported that other providers over-diagnosed ASD at least sometimes. Commonly-cited reasons for ASD over-diagnosis included: assisting in access to school services (46.1%) and assisting in access to medical services through health insurance (47.6%). Male providers were significantly more likely to self-report over-diagnosing ASD compared to female providers (16.7% versus 5.7%; Adjusted Odds Ratio 3.24 [95% Confidence Interval 1.17-9.02]). 7.8% of providers self-reported under-diagnosing ASD. There were no significant provider demographic associations with under-diagnosis. Commonly-cited reasons for under-diagnosis included: parents really did not desire the diagnosis (35.4%) and parents weren’t ready to hear the diagnosis (43.4%).
Conclusions: Relatively few U.S. providers self-report over-diagnosis of ASD; however, they report that this practice pattern is widespread among local colleagues. Under-diagnosis was less common. Study results suggest that ASD over- and under-diagnosis may be occurring in a meaningful portion of reported ASD cases in the US. Over-diagnosis may present an effort on behalf of ASD providers to facilitate access to limited services.