Spatial-Cluster Signal Detection in Medical Devices Using Likelihood Ratio Test Method.

Area of Science:

• Medical device safety
• Spatial epidemiology
• Biostatistics

Background:

• Growing number of medical device databases necessitates understanding geographical patterns of adverse events (AEs).
• Detecting spatial clusters of AEs is crucial for targeted safety interventions.
• Existing methods may not fully leverage geographical and exposure data.

Purpose of the Study:

• To develop and validate a statistical method for detecting spatial clusters of medical device-related adverse events (AEs).
• To incorporate geographical and exposure information for more accurate AE rate assessment.
• To provide a framework for enhanced medical device safety surveillance.

Main Methods:

• Development of a likelihood ratio test (LRT) method.
• Incorporation of geographical region and exposure information.
• Application to Poisson-modeled count data, such as AE counts.

Main Results:

• The proposed LRT method demonstrates strong performance in simulations, with high power, sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV).
• The method was successfully applied to hypothetical case studies involving a medical device for end-stage heart failure.
• Validation through simulation confirms the method's reliability for spatial-cluster signal detection.

Conclusions:

• The developed statistical method provides a robust framework for identifying geographical hotspots of medical device adverse events.
• This approach enhances medical device safety surveillance by enabling targeted investigations in high-risk regions.
• The method is applicable to various medical device registries and databases with patient-level geographical information.