RxNorm coding of EHR medication orders using an NLP-based approach.

摘要

The main goal of this study was to evaluate the impact of incomplete EHR medication orders on the automation of RxNorm coding/mapping. Specifically, the research question is: how to develop a semi-automated mapping process for incomplete EHR medication orders using existing RxNorm API?;To answer the research question, the following specific objectives were addressed: Describe EHR issues that affect the ability to automate mapping of medication records to RxNorm. Map EHR medication records that have missing National Drug Codes (NDC) to RxNorm. Create a list of rules to enhance automated RxNorm coding (help finding the most accurate target RxNorm concept without manual review) of EHR medication based on observations from the manual review step. Validate the set of rules by applying the set of rules to RxNav API for all EHR medications and compare results to Medi-SpanRTM coding tool. Describe data quality issues that may affect mapping medication orders to RxCUIs.