Modeling complex HTA problems easily and transparently: Discretely Integrated Condition Event (DICE) Simulation

Abstract: As clinical knowledge expands and interventions become more sophisticated and individualized, the modeling to inform HTA decisions grows increasingly complex; yet everyone wants transparent, flexible, easy-to-use but realistic simulations. In this seminar, a modeling approach expressly designed for HTA will be introduced. The two core concepts of DICE are: conditions, which hold the information used in the model; and events, which reflect the instantaneous occurrences where information is modified. These will be explained in the context of several example to show how DICE simulation remains very transparent while not imposing any assumptions or limitations. The entire model—regardless of complexity--is specified in a small set of MS Excel tables and one simple open-source macro executes the model. It provides an intuitive way to represent disease pathways, is straightforward to implement, debug and, most of all, to communicate and review as there is nothing hidden. Discrete event simulations are readily implemented as well as cohort Markov models and hybrids—indeed several structures they can be combined in one model.  Regardless of the complexity of the problem, DICE simulation can handle it while allowing the modeler to focus on formulating the decision problem properly and on obtaining and analyzing the required data, rather than on the modeling technique itself.