Pharmacy LASA Drug Alerting Using Anomaly Scores
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Medication errors occur due to misinterpretation of drug names with similar-sounding or -looking names, leading to potential harm or fatality, and existing alert systems generate high false positives, causing alert fatigue among pharmacists.
Innovation Solution
A targeted alerting framework that identifies 'look-alike, sound-alike' (LASA) drug pairs using historical medication errors and patient purchase history, providing real-time probabilistic alerts to minimize errors while reducing unnecessary alerts.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If existing alert systems are used to identify LASA drug pairs, then medication errors can be detected, but false positive alerts increase causing alert fatigue among pharmacists
Solution Approach 1:
The system segments the alert generation process into multiple specialized modules: LASA pair identification module, anomaly detection module, and probabilistic alerting module. Each module handles a specific aspect of error detection, allowing for more precise control over alert generation and reducing false positives while maintaining high detection accuracy.
Solution Approach 2:
The system dynamically adjusts alerting parameters based on historical data and confidence scores. By changing the threshold parameters for alert generation based on learned patterns from historical medication errors, the system optimizes the balance between detecting true errors and avoiding false alarms, thereby reducing alert fatigue.
2Reliability
If traditional manual verification methods are used, then pharmacist attention is maintained, but productivity decreases due to time-consuming checks
Solution Approach 1:
The system performs self-verification by automatically comparing prescribed drugs against the LASA database and generating anomaly scores without requiring continuous pharmacist intervention. The probabilistic alerting system only flags cases that meet specific confidence thresholds, allowing pharmacists to focus their attention on high-risk cases while maintaining high dispensing speed.
Solution Approach 2:
The system performs preliminary verification of all prescriptions by automatically checking for LASA pairs before pharmacist review. This preliminary filtering action identifies potential errors in advance, so that pharmacists only need to verify cases that have already been flagged by the system, significantly improving productivity while maintaining verification accuracy.
3Reliability
If comprehensive alerting for all potential LASA drugs is implemented, then patient safety is maximized, but alert fatigue increases due to high volume of alerts
Solution Approach 1:
The system applies different alerting strategies to different LASA drug pairs based on their specific risk profiles and historical error rates. High-risk pairs generate alerts with higher confidence thresholds, while lower-risk pairs use more lenient thresholds. This localized approach to alerting maintains patient safety for high-risk drugs while reducing overall alert volume to improve workflow efficiency.
Solution Approach 2:
The system implements partial alerting by only generating alerts for cases that exceed specific anomaly score thresholds, rather than alerting on all potential LASA matches. This selective approach provides sufficient safety coverage for high-probability errors while avoiding the excessive alerting that causes fatigue, achieving an optimal balance between safety and operability.
Data Source
AI summary
A system comprising one or more processors and one or more non-transitory computer-readable media storing computing instructions that, when executed on the one or more processors, cause the one or more processors to perform functions comprising: receiving a drug profile for a drug identified as part of a high-risk look-alike-sound-alike (LASA) drug pair prior to filling a prescription for the drug; generating an anomaly score based on direction components of the drug profile; and transmitting an alert to a fulfilment screen when the anomaly score exceeds a predetermined threshold. Other embodiments are disclosed.