Developing Natural Language Processing to Extract Complementary and Integrative Health Information from Electronic Health Record Data
Huixue Zhou · 2022 IEEE 10th International Conference on Healthcare Informatics (ICHI) · 2022
Complementary and Integrative Health (CIH) has grown in popularity over the last few decades. The goal of this research project is to: 1) create a data model to represent information about both psychological and physical CIH approaches documented in unstructured clinical notes in the EHR; 2) develop an NLP extraction model to extract CIH information; and 3) use the extracted information to better understand the role of CIH in diseases. To develop the annotation guide, A total of 300 notes (100 notes per CIH approaches) were randomly selected and manually annotated. Annotations were made for status, symptom, and frequency of each approach. In 300 notes, 115 sentences pertaining to music therapy, 128 sentences pertaining to chiropractic and 136 sentences pertaining to aquatic exercise were identified. Status was documented the most among three entities. Our recent investigation discovered a wealth of CIH information in EHR clinical notes, as well as the ability to represent CIH data. The project's next stage will be to train NLP models using our manually annotated data to extract CIH information and investigate the role of CIH techniques in various diseases.