What can medical records reveal about subgroups of autism spectrum disorders (ASD)? Lots, say researchers from researchers at Harvard Medical School. A recent study analyzed the electronic medical records of children diagnosed with autism and identified the most common co-morbid conditions. Based on this data, they identified four subgroups of ASD. The researchers explain that their study does have limitations, but their work lays the groundwork for understanding autism’s subtypes.
The research team selected 14,000 medical records of children who had at least one entry for an autism diagnosis at Boston Children’s Hospital. They narrowed the sample to 4,934 by eliminating records of children who were under 15 years-old at the time of the study. The researchers searched for patterns in the records’ ICD-9 codes—used in medical billing to identify diagnoses, symptoms, etc. Then, they utilized a statistical clustering analysis to identify patterns over time, looking for similar symptoms at various ages.
Although they found 45 categories of symptoms appearing in at least five percent of cases, they only categorized about 10%, which only reinforces the fact that ASD consist of a diverse range of symptoms. The results led to the establishment of three specific subgroups and a fourth with “no distinguishing patterns.”
- Subgroup one consisted of 120 children with seizures, but without other prominent symptoms.
- Subgroup two is characterized by multi-system disorders; the 197 children in this subgroup often have gastrointestinal issues, asthma, or ear infections.
- Subgroup three includes children who have been diagnosed with a psychiatric disorder in addition to an ASD. This group had a higher prevalence of higher functioning children—only 28% had intellectual disabilities.
While these subgroups are distinct, there is significant overlap. For example, one-third of children in subgroup three also fell into subgroup one.
The researchers acknowledge that this study does have its limitations, such as the lack of a control group. Furthermore, the data were based on the ICD-9 codes, which are more informative for insurance companies than clinicians and scientists. However, this research is an important first step in identifying subgroups in autism and the researchers hope that other hospitals will perform their own analyses.
“Now it’s possible for many other health centers to do these studies, because many of them have electronic health records. We’re looking forward to collaborating with many of them to do some finer-grained work,” says Isaac Kohane, lead researcher and professor of pediatrics at Harvard Medical School.
This research is published in the journal Pediatrics.
Previous news in autism:
