Does a claims diagnosis of autism mean a true case?

The team asked a simple question. Do insurance claims that list autism really match a child’s chart?

They pulled claims that had at least two autism codes. Then they opened each child’s medical chart and checked if a clinician truly gave an autism diagnosis.

Eight out of every ten claims with two or more autism codes were right. The charts backed the diagnosis.

In plain words, the claims trick is good enough for big research studies.

Lemons et al. (2015) tried the same two-code rule in hospital records. They got almost the same hit rate. The rule works in both claims and EMRs.

Norris et al. (2010) warned that some paper checklists miss lots of kids. Their review says the SCQ works, but GARS fails often. The claims rule does not replace those tools; it just spots cases faster in huge data sets.

van den Broek et al. (2006) showed that toddler screens like M-CHAT miss higher-functioning kids. The claims study looked at all ages together, so it could still miss the same mild cases. Use both methods if you need the full picture.

If you run program evaluations or funding reports, you can trust the two-code rule to count autism cases. You do not need to read every chart. Just remember: the rule may skip mild or very young kids, so pair it with clinical tools when you plan services.