Related Experiment Videos
Detecting hidden relations between time series of mortality rates
1Department of Medicine, University of Zurich, Switzerland.
Methods of Information in Medicine
|January 1, 1990
Summary
This study introduces a novel method using autoregressive integrated moving average (ARIMA) models to detect environmental factors influencing diseases. The approach identifies synchronized fluctuations between population subgroups, aiding public health research.
Area of Science:
- Environmental epidemiology
- Biostatistics
- Time series analysis
Background:
- Identifying environmental influences on disease requires robust statistical methods.
- Autocorrelation within population subgroups can obscure disease-environment relationships.
Purpose of the Study:
- To present a novel time series analysis method for detecting time-varying environmental factors affecting disease.
- To demonstrate the utility of autoregressive integrated moving average (ARIMA) models in epidemiological research.
Main Methods:
- Applied ARIMA models to filter time-series data from two population subgroups (e.g., males, females).
- Analyzed the cross-correlation function between filtered series to detect synchronized fluctuations at time lag 0.
- Utilized yearly mortality rate data for the elderly to exemplify the procedure.
Main Results:
- The described method effectively removes autocorrelation within individual time series.
- A marked peak in the cross-correlation function at time lag 0 suggests a shared environmental influence.
- The procedure was successfully demonstrated on elderly mortality data.
Conclusions:
- The ARIMA-based cross-correlation method offers a sensitive approach to identifying environmental factors impacting disease trends.
- This technique can enhance our understanding of disease etiology and inform public health interventions.