Automated Phenotyping Tool for Identifying Developmental Language Disorder Cases in Health Systems Data (APT-DLD): A New Research Algorithm for Deployment in Large-Scale Electronic Health Record Systems.
Courtney E Walters, Rachana Nitin, Katherine Margulis and 8 others
PMID 32791019WHAT IT FOUND
An automated algorithm accurately identifies Developmental Language Disorder cases in electronic health records by filtering for specific diagnosis codes.
It achieved 95% positive predictive value in a discovery sample and 90% in a larger replication sample, allowing researchers to find large patient cohorts without manual chart review.
Key findings
01The algorithm achieved a positive predictive value of 95% in the discovery sample and 90% in the replication sample when compared to speech-language pathologist manual chart reviews.
02The tool successfully classified 6,013 cases as Developmental Language Disorder in a replication sample of 13,652 pediatric records.
03The algorithm is conservative, with a high false-negative rate in the replication sample, meaning it may miss some true cases of Developmental Language Disorder.
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 algorithm cannot access individual IQ scores, language test results, or educational outcomes, relying solely on diagnostic codes. The study was conducted using data from a single institution (Vanderbilt University Medical Center), which may limit generalizability to other populations. The algorithm is conservative and misses some true DLD cases (false negatives), particularly those where language issues are noted in text but not coded with specific ICD codes. The phenotype definition is based on English-speaking countries and may require validation for international health systems.
Declared interests
Research was supported by the National Institute on Deafness and Other Communication Disorders and the National Institutes of Health. Data sets were obtained from Vanderbilt University Medical Center's BioVU, supported by institutional funding and NIH grants.
The easy way to misread this
Do not assume this algorithm is a diagnostic replacement for clinical evaluation. It is designed for research cohort identification and has a notable miss rate (false negatives), meaning it will fail to identify some patients who truly have Developmental Language Disorder.