Detecting Hidden Relations between Time Series of Mortality Rates

Abstract

In the present report a method is described which may help to decide if a disease is influenced by an environmental factor which fluctuates in time: For each of two naturally arising subgroups of a population (such as males and females) an ARI MA model (autoregressive integrated moving average model) is identified. These models are used as filters to remove the autocorrelation in each series. If the resulting crosscorrelation function between the two filtered series shows a marked peak at time lag 0 this may indicate that such an environmental factor is present. The procedure is demonstrated using yearly data of mortality rates among the elderly.

Key-Words

ARIMA Models - Crosscorrelation Function - Prewhitening - Mortality Rates

• REFERENCES

• 1 Box G E P, Jenkins G M. Time Series Analysis, Forecasting and Control (revised ed.). San Francisco: Holden Day; 1976
• 2 Abraham B, Ledolter J. Statistical Methods for Forecasting. New York: John Wiley; 1983
• 3 Bartlett M S. Some aspects of the time correlation problem in regard to tests of significance. J Royal Statist Soc 1935; 98: 536-43.
• 4 Box G E P, Newbold P. Some comments on a paper of Cohen, Gomme and Kendall. J Royal Statist Soc 1971; 134: 229-40.
• 5 Haugh L D. Checking the independence of two covariance-stationary time series: A univariate residual cross-correlation approach. J Amer Statist Ass 1976; 71: 378-85.
• 6 National Center for Health Statistics. Changes in Mortality Among the Elderly, United States, 1940-1978. Vital and Health Statistics . Washington D C: US Government Printing Office; Series 3 No. 22, 1982
• 7 Ljung G M, Box G E P. On a measure of lack of fit in time series models. Biometrika 1978; 65: 297-303.
• 8 Pierce D A. Relationships and the lack thereof between economic time series, with special reference to money and interest rates. J Amer Statist Ass 1977; 72: 11-6.