|
Abstract:
|
The MITRE Corporation designed algorithms to examine opioid prescribing behaviors and identify potentially abnormal prescribing patterns. These algorithms, based on healthcare claims data, examined opioid prescriptions from different perspectives including the context in which they were written or filled. These algorithms used a variety of analytical techniques to analyze a patient’s medical history, context in which the prescription was written, and the prescribing habits of the prescriber.
Using machine learning and statistics techniques, these algorithms aim to extend the traditional "outlier" analysis to identify abnormal prescribing patterns. The pattern of life analysis uses the timing of prescriptions and other events to examine the potential adverse event risk when receiving another opioid prescription. The prescriber of last resort analysis aims to understand which prescribers give prescriptions to patients who were previously denied prescriptions. Prescriber normalization analyzes the context in which prescribers write prescriptions compared to their peers, with regards to both the individual patient and the types of patients seen at their practices.
|
