Improving EWMA Plans for Detecting Unusual Increases in Poisson Counts

摘要

Automated public health records provide the necessary data for rapid outbreak detection. Anadaptive exponentially weighted moving average (EWMA) plan is developed for signallingunusually high incidence when monitoring a time series of nonhomogeneous daily disease counts.A Poisson transitional regression model is used to fit background/expected trend in counts andprovides "one-day-ahead" forecasts of the next day's count. Departures of counts from theirforecasts are monitored. The paper outlines an approach for improving early outbreak data signalsby dynamically adjusting the exponential weights to be efficient at signalling local persistent highside changes. We emphasise outbreak signals in steady-state situations; that is, changes that occurafter the EWMA statistic had run through several in-control counts.