Apremilast is compared to best supportive care (BSC), and response to treatment is assessed after a trial period of 16 weeks, which corresponds to the primary efficacy endpoint in the apremilast clinical trials. DES is a technique that is event-based and models a population at the individual level in terms of utility, costs, and resource use. The DES was implemented as a series of input, calculation, and output flowcharts in Arena®. The DES reported an incremental QALY per patient of 0.16 and incremental cost per patient of $17,325. The Markov model reported an incremental QALY of 0.18 and incremental cost per patient of $34,733. The differences can be explained when taking into account the differences in how age and life expectancy are handled in the models. The Markov model uses a single age: 50, whereas the DES uses an age distribution. The Markov model uses an actuarial table to calculate background mortality whereas the DES uses a distribution for life expectancy. These differences between Markov and DES in not just inputs, but how patient characteristics are handled, lead to different results. Markov models are structured in terms of states and transitions, and the main limitations are: using states to represent the problem; the need for homogenous populations; and the inability to take competing risks into account. Transitioning between states can account for the disease progression, but this method is unable to handle competing risks. Discrete event simulations do not have those limitations and also allow for structural flexibility and are easier to adapt if further data become available.
No audio presentation, just PowerPoint slides.
Recommended CitationClancy, PharmD, Zoe, "Direct Comparison of Apremilast and Best Supportive Care Using a Discrete Event Simulation" (2014). Master of Science in Applied Health Economics and Outcomes Research. Presentation 7.