Related Experiment Videos
Use of the scan statistic to detect time-space clustering
S Wallenstein1, M S Gould, M Kleinman
1School of Public Health, Columbia University, New York, NY.
American Journal of Epidemiology
|November 1, 1989
Summary
A new scan statistic method detects time-space disease clustering using event dates and locations. This approach enables calculation of attributable risk and effect size, improving epidemiological analysis.
Area of Science:
- Epidemiology
- Biostatistics
- Spatial Analysis
Background:
- Time-space clustering is a critical concern in public health, requiring robust statistical methods for detection.
- Existing methods for time-space clustering often lack the ability to quantify risk or effect size.
Purpose of the Study:
- To introduce a novel scan statistic for identifying time-space disease clustering.
- To provide a method that allows for the calculation of attributable risk and effect size.
Main Methods:
- The proposed method utilizes a scan statistic based on the maximum number of events within a 365-day period across geographic units.
- The statistic is calculated as the ratio of excess events to the square root of the sum of variances.
- Data requires exact event dates and geographic units over several years with constant disease risk.
Main Results:
- The scan statistic is demonstrated to be effective in identifying time-space clustering.
- The method allows for the computation of attributable risk and effect size, offering deeper insights than traditional methods.
- Illustrative data from adolescent suicide cases are used to showcase the procedure's application.
Conclusions:
- The developed scan statistic offers a valuable tool for epidemiological research, enhancing the detection of time-space disease clustering.
- The ability to calculate attributable risk and effect size makes this method particularly useful for public health interventions and policy.
- The provided tables and formulas support the evaluation of both temporal and time-space clustering.