Prefiltered component-based greedy (PreCoG) scan method

Area of Science:

• Epidemiology
• Biostatistics
• Spatial Analysis

Background:

• Understanding disease spatial distribution is crucial for identifying spread patterns and risk factors.
• Accurate detection of disease clusters aids in discovering novel risk factors and implementing timely interventions.
• Existing scan methods may face limitations in detecting irregularly shaped disease clusters.

Purpose of the Study:

• To introduce a novel scan method, Prefiltered Component-based Greedy (PreCoG), for efficient and accurate detection of disease clusters.
• To evaluate the performance of the PreCoG scan method in identifying both regular and irregular cluster shapes.
• To provide a flexible and powerful tool for disease surveillance systems.

Main Methods:

• Development of the Prefiltered Component-based Greedy (PreCoG) scan algorithm.
• Utilizing a prefiltered component-based approach for cluster detection.
• Comparative analysis of PreCoG against existing scan methods.

Main Results:

• The PreCoG scan method demonstrates high efficiency and accuracy in detecting irregularly shaped disease clusters.
• PreCoG exhibits flexibility in detecting both regular and irregularly shaped clusters.
• The method offers high power, sensitivity, and positive predictive value compared to other scan methods.
• The PreCoG method has been implemented in the publicly available smerc R package.

Conclusions:

• The PreCoG scan method offers a unique and innovative approach to disease cluster detection.
• This method can significantly enhance the accuracy and effectiveness of disease surveillance systems.
• The availability of the smerc R package facilitates broader research and application of the PreCoG method.