Risk-adjusted continuous outcome monitoring with an EWMA chart: could it have detected excess mortality among intensive care patients at Bundaberg Base Hospital?

OBJECTIVE: To test whether applying a continuous riskadjusted charting method, using an exponentially weighted moving average (EWMA) chart, would have been useful for monitoring outcomes of patients admitted to the intensive care unit at Bundaberg Base Hospital, Queensland, between November 2000 and December 2005. DESIGN, SETTING AND PARTICIPANTS: An EWMA chart was constructed to show the change in observed compared with predicted mortality over time, using data submitted to the Australian and New Zealand Intensive Care Society Adult Patient Database. Limitations and practical implications of this monitoring technique were evaluated and compared with a routine monitoring technique using the annual standardised mortality ratio. MAIN OUTCOME MEASURE: In-hospital mortality.
RESULTS: Data were submitted on three occasions (August 2002, November 2002 and February 2006). In each year before 2005, the EWMA chart showed periods when observed mortality appeared higher than predicted. These periods were not detectable by analysing the data with an annual standardised mortality ratio. Comparison of aggregated data from peer group hospitals suggested that the mortality prediction model (APACHE III-j) was an appropriate risk adjustment model for this analysis.
CONCLUSIONS: Continuous monitoring of outcomes using an EWMA chart may have advantages over other techniques. Had data been available, analysis with an EWMA chart might have prompted review of processes and outcomes among patients at Bundaberg Base Hospital ICU.