This story is about using brain magnetic resonance imaging (MRI) to predict which young children with autism will go on to develop normal language skills and which ones won’t. Although it provides a nice general overview on the challenges of treating children with autism spectrum disorder, it ignores many critical details including the cost and accuracy of the test for predicting future language abilities, the limitations of the research, and the potential harms of an inaccurate scan.
A 2014 report by the Centers for Disease Control and Prevention estimated that 1 out of 68 children in the United States have been identified as having autism spectrum disorder. The symptoms of autism spectrum disorder can vary widely in character and severity and may or may not include significant language deficits. Finding a way to determine at a very early age which children are at risk of developing language deficits might help parents and clinicians choose more appropriate therapies.
There was no discussion of costs. It would have been worth mentioning in the story that the out-of-pocket cost of a brain MRI ranges from $1,000 to several thousand dollars depending on the type of test and where it is done. However, most insurance policies do cover some or all of the cost of imaging tests if they’re considered medically necessary.
The article did not quantify what the researchers found in their study. According to the study, among 60 youngsters diagnosed with autism and 43 controls who took part, the MRI when combined with a behavior test was 80 percent accurate in predicting which children would have normal or poor language outcomes. When either test was performed separately, the prediction accuracy dropped to 68 percent.
The story also doesn’t have much to say about what parents and clinicians might do with these scans to improve treatment for autistic children, beyond noting that “numerous studies have found that children’s outcomes improve with early therapy.” We’d offer a San Diego Union-Tribune coverage as an example of the context the story could have provided here by reaching out to an independent expert.
“The combination of brain imaging and behavioral measures was somewhat better than either alone at predicting outcome, but whether this improvement was statistically significant was not tested, said Dr. Bradley Scott Peterson, who directs the Institute for the Developing Mind at The Saban Research Institute of Children’s Hospital Los Angeles. “Moreover, even if the combination was better at a statistically significant level, the important question is whether this improvement in predictive accuracy from the combination of measures would be clinically meaningful in terms of altering treatment planning and in justifying the costs of the brain imaging procedures,” Peterson said by email. “This level of clinical significance of the combined measures was not assessed in the paper, and should be built into future studies using these measures.”
The story provided no description of potential harms from this scanning approach. The chief potential harm would be a false positive or false negative scan. If the scan is wrong, that may trigger unnecessary additional anxiety or false reassurance for parents regarding their child’s likely outcome. It may also lead parents and doctors to implement the wrong kind of therapy for autistic children. The researchers on this study have suggested that non-verbal therapies may be more appropriate for the children whose scans suggest they are not “language-ready” (and vice-versa). But if the scan is wrong, these children may be offered the wrong kind of treatment — possibly leading to a worse outcome than if they hadn’t been scanned at all.
The story also could have noted some of the challenges involved with getting MRI images on young children. MRI scanners make loud banging noises which can be frightening to children and sedation is sometimes required. The researchers in this study were able to do the scans while the children were asleep, which didn’t require sedating the children.
The article provided little detail on the study methodology and didn’t caution readers as to the limitations of this preliminary study. The number of subjects in the study was small and the results may not be reproducible in a larger series. Additional studies will need to be conducted before anything definitive can be said about the reliability of the scanning technique.
In our view, instead of squandering its limited word count on pie-in-the-sky speculation about the future uses of brain scans (Example: “Scientists at a growing number of research centers are pinning their hopes that these images will help them unlock some of the secrets of everything from what causes childhood disorders like Attention Deficit Hyperactivity Disorder to how neurons in the brain process vision and memory.”), the story could have more profitably used those words to add context and depth to the study at hand.
The article skirts the line when it talks about the “soaring” rates of autism diagnosis, which has reached 1 in 68 children. But it does provide some context to that number when it says, “The spectrum is wide and some are destined to be on the mild end and be very talkative, sometimes almost indistinguishable from those without the disorder in some settings, while others will suffer from a more severe form and have trouble being able to speak basic words.”
The source quotes only one source — a study co-author. We see this as a drawback since there are numerous autism experts available, many of them holding differing views on autism diagnosis and treatment. As a service to reporters, HealthNewsReview.org maintains a list of independent experts who are willing to comment on studies and related news on health interventions.
While noting that there are currently no blood tests or biological markers available to autism specialists, the story shares that specialists who treat children with autism currently rely on “parent and teacher reports, observations and play assessment” to make a diagnosis.
The story doesn’t state when or if clinicians will adopt the testing protocol for newly diagnosed autism patients. Because the research was only recently published in a journal, one must assume clinicians will not be implementing the protocol anytime soon. But the story doesn’t make this clear or provide any sort of time frame for readers.
We’ll give the benefit of the doubt on this one. While the article doesn’t clearly articulate what’s novel about this research — the fact that it’s the first study to use magnetic resonance imaging (MRI) in very young children diagnosed with autism as a way to predict their future level of language abilities — it does establish the diagnosis problem and suggests that the study represents a new approach to dealing with it. A San Diego Union-Tribune story on the same study expertly stated the novelty this way: “If the results hold up, this would be the first time that brain imaging of living children has been used successfully to guide autism diagnosis and help choose appropriate therapy. Currently, autism assessments are based on behavior.”
The story appears to rely heavily on a news release, since the single quote attributed to the source was lifted word for word from the news release. The article might have been improved by using more data on the study protocol available from the news release, or better yet — the study and an additional expert or two. However, it’s noted that the article has some useful information that the news release didn’t include, namely, the ages of the test subjects when they were initially tested and when they had their followup MRIs (12-29 months for the initial test and followup at 30 to 48 months).