Is tele-health fit for purpose in autism assessments?
I came across a recent study that I found absolutely fascinating. It looked at the history of the revisions of the diagnostic criteria for autism in the DSM, and then pivoted to the changes made in the diagnostic process because of the COVID-19 pandemic. Researchers wondered at the validity of the new methods.
You see, the pandemic necessitated adaptations to standard autism diagnostic assessment practices in order to accommodate physical distancing requirements. This included increased use of tele-health platforms. Tools like the Brief Observation of Symptoms of Autism (BOSA) were developed specifically for remote evaluation in response to these needs. They note that the BOSA demonstrates good discriminant validity and inter-rater reliability. However, its accuracy compared to in-person assessments remains to be fully examined, especially for more subtle cases. Other remote assessment tools, like the TELE-ASD-PEDS (TAP) and Telehealth Evaluation of Development for Infants (TEDI), were already in development prior to the pandemic to improve accessibility of evaluations. The researchers note that these show positive preliminary findings related to feasibility and provider/caregiver acceptance. However, further research is still needed to firmly establish the reliability and validity of these tele-health assessment instruments. They further note that ongoing growth and developments is expected in the use of tele-health, mobile applications, and artificial intelligence approaches for improving early screening and detection of autism.
Additionally, the researchers note that there is a great need to address disparities in the autism assessment process for underserved groups. This includes focused efforts on developing more culturally and linguistically appropriate tools, with greater representation in assessment standardisation samples and validation in diverse populations.
This study jumped out at me because properly validating newly developed and adapted autism assessment instruments is crucial before they are widely implemented in research or clinical practice. Without establishing strong psychometric properties through validation research, the accuracy of instruments in identifying and diagnosing autism cannot be relied upon. This could lead to serious negative consequences like under- or over-diagnosis of individuals being assessed. Appropriate validation typically involves assessments of reliability, such as test-retest reliability, inter-rater reliability, and internal consistency to demonstrate the tool’s stability and precision across administrations and users. Importantly, the validity of the instrument must also be established by evaluating criteria like convergent validity, divergent/discriminant validity, known groups validity, and predictive validity through comparisons with existing diagnostic measures. Ensuring proper validation gives professionals and families confidence in the meaningfulness of assessment results and provides evidence that scores relate to a true manifestation of autism symptoms rather than statistical noise or other influences. This is especially critical as the field seeks to improve assessment accessibility through remote administration whilst maintaining accuracy. Utilisation of well-validated instruments can aid tremendously in research and care regarding this complex diagnosis.