A study of the average run length characteristics of the National Notifiable Diseases Surveillance System.

This study examines the statistical properties (that is, false positive and negative signals) in detecting unusual patterns of reported cases of diseases from the Centers for Disease Control and Prevention's National Notifiable Diseases Surveillance System. Control charts are applied to the residuals of one-step ahead forecasts based on Box-Jenkins models of reported cases of disease. Simulation and analytical techniques are used to study the average run length characteristics of these control charts for various types of changes in the levels of the series, including spike, trend and step changes. The average run lengths for the highly correlated disease series are much longer than for the usual independent data case. This increase in the average run lengths is strongly influenced by the type of change in the level of the series and by the type of control chart. Understanding the average run length characteristics of the control charts can lead to timely detection of changes in the levels of disease series, and subsequent timely public health actions to decrease unnecessary morbidity and mortality.

[1]  S. W. Roberts.  Control chart tests based on geometric moving averages , 2000 .

[2]  George C. Runger,et al.  Average run lengths for cusum control charts applied to residuals , 1995 .

[3]  Marion R. Reynolds,et al.  EWMA and CUSUM control charts in the presence of correlation , 1997 .

[4]  Layth C. Alwan,et al.  Time-Series Modeling for Statistical Process Control , 1988 .

[5]  E. S. Page.  CONTINUOUS INSPECTION SCHEMES , 1954 .

[6]  Herbert Moskowitz,et al.  Run Length Distributions of Residual Control Charts for Autocorrelated Processes , 1994 .

[7]  Douglas C. Montgomery,et al.  Some Statistical Process Control Methods for Autocorrelated Data , 1991 .

[8]  Benjamin M. Adams,et al.  An evaluation of forecast-based quality control schemes , 1994 .

[9]  J. Stuart Hunter,et al.  The exponentially weighted moving average , 1986 .

[10]  Gwilym M. Jenkins,et al.  Time series analysis, forecasting and control , 1972 .

[11]  Herbert Moskowitz,et al.  Run-Length Distributions of Special-Cause Control Charts for Correlated Processes , 1994 .

[12]  Herbert Moskowitz,et al.  Control Charts in the Presence of Data Correlation , 1992 .

[13]  G. D. Williamson,et al.  A monitoring system for detecting aberrations in public health surveillance reports. , 1999, Statistics in medicine.

[14]  Benjamin M. Adams,et al.  Combined Control Charts for Forecast-Based Monitoring Schemes , 1996 .

[15]  Richard A. Johnson,et al.  The Effect of Serial Correlation on the Performance of CUSUM Tests II , 1974 .