Comparison of spatial scan statistic and spatial filtering in estimating low birth weight clusters

Area of Science:

• Epidemiology
• Geographic Information Systems (GIS)

Background:

• Low birthweight (LBW) presents significant public health challenges.
• Understanding spatial and population factors influencing LBW is crucial for targeted interventions.

Purpose of the Study:

• To compare two cluster estimation techniques for analyzing LBW: Kulldorff's Spatial Scan Statistic (SaTScan) and Rushton's Spatial filtering.
• To examine how varying spatial filter sizes impact the identification of LBW clusters.

Main Methods:

• Utilized Kulldorff's Spatial Scan Statistic (SaTScan) for cluster detection.
• Employed Rushton's Spatial filtering technique with increasing circular filter sizes.
• Compared the spatial and population characteristics of clusters identified by both methods.

Main Results:

• Spatial filtering did not identify areas lacking statistical significance according to SaTScan.
• Persistent high LBW rates were observed as filter sizes increased, suggesting non-random variation.
• Maternal characteristics within clusters varied significantly between SaTScan and spatial filtering.
• Larger spatial filters smoothed local variations, leading to a more uniform LBW pattern.

Conclusions:

• SaTScan and spatial filtering yield different cluster results from identical birth data.
• Clusters identified by each method differ in geographic scope and associated population demographics.
• Integrating both SaTScan and spatial filtering enhances the detailed analysis of cluster features.