Automated phonological analysis and treatment target selection using AutoPATT.
Philip Combiths, Ray Amberg, Gregory Hedlund and 2 others
PMID 34085574WHAT IT FOUND
AutoPATT generated speech inventories with 98 to 100 percent accuracy, compared to 78 to 96 percent for manual analysis.
The software missed one complex target due to a coding error, so clinicians must still review its suggestions before selecting treatment goals.
Key findings
01AutoPATT's automated inventories were 98 to 100 percent accurate, while manual inventories were 78 to 96 percent accurate.
02Most disagreements between the two methods came from humans omitting sounds that the software correctly identified.
03The software missed a three-element cluster target for one child because it misread a two-element cluster with a diacritic, showing that automated suggestions still need clinical review.
STILL TO COME
How it was doneWhat they foundWhat it means for SLPs
Read the rest of this summary
You get three full summaries a month, free, and we do not ask for a card. Search, the TL;DRs and your library stay unlimited either way.
What it does not show
The study only looked at young, monolingual English-speaking children with phonological disorder, so the results do not apply to older children, bilingual speakers, or children with other speech conditions. The researchers did not measure how much time the automated method actually saved compared to the manual method. The study used a small sample of 25 children, and the software was tested in a controlled research setting rather than a busy clinic.
Declared interests
The study was supported by a grant from the National Institutes of Health.
The easy way to misread this
Do not assume the software is perfect because it beat the human researchers. It made a systematic error that missed a valid treatment target for one child. Always check its suggestions against your own assessment of the patient.