A risk-adjusted Sets method for monitoring adverse medical outcomes

Abstract

The Sets method has been advocated in previous work as a method for monitoring adverse medical outcomes where the adverse event rate is low. Here, a risk-adjusted version of the refined Sets method is presented and an example is given to demonstrate its advantage over the unadjusted method. The method is suitable for any risk distribution and does not assume that changes in risk will be small. A graphical representation, referred to as the Grass plot, of the original and risk-adjusted methods is also given.