Intelligent Alert Suppression for CPOE Systems
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Alert fatigue in Computerized Physician/Provider Order Entry (CPOE) systems, where excessive alerts lead to providers ignoring vital warnings, compromising patient care and safety, as there is no effective solution to control alert frequency or suppress alerts without negative impact on patient safety.
Innovation Solution
A system that evaluates alerts based on specific parameters such as patient age, diagnosis, alert severity score, and override rate using an alert suppression engine to determine whether to present or suppress alerts, providing non-interruptive alerts that do not require immediate action, thus reducing unnecessary workflow interruptions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If alerts are presented to providers in CPOE systems, then patient safety monitoring is improved, but alert fatigue occurs causing providers to ignore vital warnings
Solution Approach 1:
The system changes parameters of alert presentation by evaluating multiple factors including alert severity score, patient-specific parameters (age, diagnosis, problems), historical override rates, and temporal patterns. Based on this evaluation, the system dynamically adjusts whether to suppress or present alerts, transforming the binary alert presentation into a nuanced decision based on multiple varying parameters
Solution Approach 2:
The alert suppression system acts as an intermediary between the CPOE system and providers. It receives alerts from the CPOE system, evaluates them against multiple criteria including patient context and historical data, and then decides whether to present or suppress the alert. This intermediary layer filters alerts before they reach providers, reducing alert fatigue while maintaining patient safety monitoring
2Reliability
If alert frequency is increased to ensure all critical warnings are noticed, then patient safety is improved, but alert fatigue increases causing providers to ignore alerts
Solution Approach 1:
The system applies different qualities of alert presentation to different alerts based on their specific characteristics. Each alert is evaluated individually against patient-specific parameters, alert severity, and historical data. Critical alerts that match current patient conditions are presented with high priority, while less relevant alerts are suppressed, creating localized quality in alert presentation rather than uniform treatment of all alerts
Solution Approach 2:
The system performs partial action by selectively suppressing certain alerts while presenting others. Rather than presenting all possible alerts (excessive action) or no alerts (insufficient action), the system evaluates each alert against multiple criteria and performs partial presentation - only showing alerts that meet specific thresholds for severity and relevance to current patient conditions, thereby reducing alert fatigue while maintaining safety
3Productivity
If alerts are suppressed to reduce workflow interruptions, then provider efficiency is improved, but critical warnings may be missed compromising patient safety
Solution Approach 1:
The system performs preliminary evaluation of alerts before they are presented to providers. It pre-assesses alerts against patient-specific parameters, historical override rates, and alert severity scores. This preliminary action filters out alerts that are unlikely to be relevant to current patient conditions, allowing providers to focus on pre-evaluated critical warnings without missing important safety alerts
Solution Approach 2:
The system incorporates feedback mechanisms by analyzing historical alert override rates and provider responses. It uses this feedback to continuously refine its suppression decisions, learning from past interactions to improve future alert presentation. This feedback loop ensures that suppression decisions are based on actual provider behavior patterns and patient outcomes, maintaining safety while improving efficiency
Data Source
AI summary
Methods, systems, and computer-readable media are provided for intelligent suppression of alerts. Alerts may be evaluated prior to presentation to determine whether to present the alert or suppress the alert to reduce alert fatigue. The evaluation includes analysis of the alert including specific alert parameters. Alert parameters include, but are not limited to, an age of the patient, one or more diagnoses of the patient, one or more problems of the patient, etc. Additional parameters that may be evaluated include an alert severity score, a threshold value, and an override rate. Alerts that qualify may be suppressed and the suppressed alerts may be provided as non-interruptive alerts (i.e., alerts that may be presented but do not require action in order to continue with a workflow).