I was delighted to receive an e-mail from a neuropsychologist colleague I haven't had the pleasure of meeting in person, but for whom I've developed a great deal of respect based on her contributions to a list serve we both belong to: Dr. Donna Henderson. She is part of the highly-respected Stixrud Group in Silver Spring, Maryland. Donna asked me if I'd ever considered writing an article on one of the most common (and perhaps overly touted) autism diagnostic tools we have - the ADOS-2. I told her I was swamped, so she took it on - and brilliantly! Here's what she wrote:

There is no doubt that the ADOS-2 has been a useful tool in both research and clinical practice. However, at some point someone named it the “gold standard” for autism diagnosis, and the name stuck, even though this has not necessarily been supported by the research. So, is the ADOS-2 the gold standard? To answer this, let’s consider (a) whether the ADOS-2 is always necessary for a diagnosis, (b) whether the ADOS-2 is sensitive to different populations, (c) whether the ADOS-2 is always reliable in clinical practice and (d) other factors that may make the ADOS-2 less than an ideal choice as part of your child’s evaluation.

Is the ADOS-2 necessary for a diagnosis of autism? No. As far as I can tell, it has never been suggested by a respected journal, institution, or autism expert (including Cathy Lord) that the ADOS-2 is necessary for a diagnosis of autism. If your school system or mental health care professional suggests this, please ask them for research or any definitive statement by a reputable source (such as the American Psychological Association).

Are there populations that the ADOS-2 is less sensitive to? Absolutely. Research indicates that the ADOS-2 can be less sensitive (that is, it can fail to identify the autism) in females, because they can have more subtle symptoms and also because they can camouflage their symptoms far better than males can. The ADOS-2 can also be less sensitive in individuals who are high functioning or have a subtler presentation of autism. For instance, one study of high functioning individuals with autism found that the ADOS-2 only identified 33% of them. For these two groups, there can be a lot of “false negatives” with the ADOS-2 (meaning the people actually have autism but the ADOS-2 score incorrectly suggested that they didn’t). In contrast, people with complicated psychiatric profiles can have “false positives.” That is, they can get an “autism score” on the ADOS-2 even if they don’t have autism.

Is the ADOS-2 always reliable in clinical practice? Reliability means that a measure is consistent, so you would get more or less the same results over time and with different testers. To understand reliability with the ADOS-2, you have to know that there are two levels of ADOS-2 training. The research reliability training is quite extensive, as the researchers have to go through three full days of training and have their skills evaluated. In contrast, becoming clinically trained to do the ADOS-2 is far less rigorous. So, while the research level of training can result in great reliability, the clinical level of training may or may not. Overall, the ADOS-2 is reliable under optimal conditions but may or may not always be as reliable in clinical settings.

Are there other reasons not to use the ADOS-2 for every patient? Yes. The ADOS-2 takes time to administer and score, and most of our clients are not simply looking for a “yes” or a “no” in regards to the possibility of autism. They are really looking for rich, detailed information about all aspects of their child’s cognitive, academic, social, and emotional functioning. The ADOS-2 does not provide this. To get a wholistic and detailed understanding of a child and their needs, the clinician must spend time interviewing the patient, parents, teachers, and other health professionals involved, as well as administering a variety of other objective measures of functioning in each domain. This is particularly important because so many of the issues involved co-exist with each other (i.e., social problems, ADHD, anxiety, depression, learning disabilities, fine motor issues, etc.). A comprehensive evaluation results in a wholistic and deep understanding of a child, and it guides individualized recommendations, rather than simply providing a score that is above or below a cutoff for one diagnosis.

Are you saying that the ADOS-2 is not a useful clinical tool? No! While the ADOS-2 is not necessarily the gold standard in diagnosing autism, it is important to note that there are many useful components within the measure that can assist in our evaluations. For example, many practitioners (including me) will use the questions about relationships, the conversation starters, and/or pretend play components from the ADOS-2 in their assessments.

Thank you Donna for this great contribution to our knowledge!