Development of a Rule-Based Natural Language Processing Algorithm to Extract Sleep Information in Pediatric Primary Care Patients with a Sleep Diagnosis
Listed in
This article is not in any list yet, why not save it to one of your lists.Abstract
Introduction
Retrospective analysis of sleep health among pediatric patients can enable important care and condition related discoveries. Often, sleep health is only encoded in a patient’s structured data after formal diagnosis. However, their unstructured clinical text often contains many detailed sleep health mentions prior to diagnosis. These mentions are numerous and cannot reasonably be identified manually, thus computer assisted tools must be developed. We present a novel, low-resource sleep vocabulary that can be applied to identify notes containing sleep mentions automatically.
Methods
Using a combination of existing sleep ontologies, interviews with clinicians, and examination of clinical note narratives, we develop a novel vocabulary of sleep health terms and phrases that cover both technical terms, abbreviations, and colloquial keywords used in describing sleep health. We compare our vocabulary against a set of manually annotated clinical notes to determine the effectiveness of our vocabulary for identifying notes with sleep health mentions.
Results
Our vocabulary was able to correctly identify clinical notes with sleep health mentions with a precision of 0.838 and recall of 0.869.
Conclusion
Our vocabulary showed excellent performance for identifying sleep health mentions at the clinical note level. The vocabulary was not able to accurately identify the specific text spans containing the mentions, which likely would require a more high-resource model. Thus, our low-resource vocabulary, which can be deployed in almost any compute environment, can serve as an identifying first pass over clinical notes to identify which notes should be further processed by more advanced models or manual review to identifying sleep health mentions.
