Commonly used decision-analytic models for cost-effectiveness analysis simulate time in discrete steps. Use of discrete-time steps can introduce errors when calculating cumulative outcomes such as costs and quality-adjusted life years. There are a number of myths or misconceptions concerning the need for and how to correct these errors. This tutorial shows that by neglecting to apply within-cycle (or sometimes referred to as half-cycle or continuity) correction methods to the results of discrete-time models the analyst may arrive at the wrong recommendation regarding the use of a technology.
The tutorial provides examples of different within-cycle correction methods including trapezoidal and Simpson’s 1/3 rule. We demonstrate that it is straightforward to implement these rules within commonly used software such as Excel and TreeAge