Automated Interpretation of Medical Prescription Text, 10-R9642Printer Friendly Version
Inclusive Dates: 07/01/06 Current
Background - Computerized Physician Order Entry (CPOE) systems are becoming more prevalent. CPOE systems support the data entry of medical orders, preferably by the physicians themselves. These systems have been shown to reduce the number of serious medication errors when implemented and used properly. Even though CPOE systems have been shown to improve safety, many physicians have been reluctant to adopt CPOE systems, citing a steep learning curve and a perceived lack of efficiency. This perceived lack of efficiency is often attributed to the user interface, which is invariably a form-like screen.
Approach - The objective of this project is to explore an alternative in which the physician enters traditional prescription text, which is then automatically interpreted into standardized data. For the physician, this process would be more analogous to manually writing prescriptions and have advantages that include ease of use, fewer clinical and legal differences from past practices, and more complete expressiveness in the data conveyed. Challenges need to be addressed to retain the capabilities of pre-existing systems using form-style data entry. These challenges include using natural language processing to parse the text into specific machine-recognizable fields appropriate to American National Standards Institute standards and addressing the human issues of clarity, confidence, efficiency, and ease.
Accomplishments - To gain a better understanding of the prescription writing habits, a survey was conducted of volunteer physicians. These physicians were asked to write sample, typical prescriptions to treat hypothetical diagnoses. Based on the information gathered in the survey, a formal grammar was designed that would allow physicians to use any one of many different ways to express the same pharmaceutical order.
An initial prototype has been developed to explore possible implementation of a system that provides the desired ease of use and flexibility. The prototyping process has provided a deeper understanding of the problems in natural language processing, user interface concepts, and proof-of-concept software that this research will explore to provide a basis for future marketable projects.