AN INTRODUCTION TO MICROSIMULATION IN HEALTH-STATE TRANSITION MODELS

Overview: This course is meant to offer students a bridge between introductory courses on heath-state transition (HST) modeling, which focus on expected value cohort (EVC) simulation, and advanced courses which deal with probabilistic sensitivity analysis and 2- and 3-dimensional simulation. The focus of the course is on the theory and application of microsimulation (MS), an alternative to the discrete-time health state transition (HST) models commonly used in a variety of clinical and health policy decision analysis scenarios. Microsimulation methods can be used for a number of practical modeling tasks which will be covered in the course.

Background: Traditional HSTs lack ‘memory’. In other words, the probability that a hypothetical individual would transfer out of a health state is entirely dependent on the fact that the individual inhabits that particular state together with the various path-probabilities associated with that state: the events that occurred prior to the individual entering the state are not considered. This also implies that the members of the cohort inhabiting a state at a particular time are homogeneous with respect to the risk of subsequent events. Representing the subtleties and nuances of clinical or health policy problems often requires a plethora of health states. Microsimulation is an alternative modeling paradigm that allows one hypothetical person to traverse the model at a time. That person can be assigned attributes that depend on prior history and allow for the memory that is lacking in traditional HST models. The latter allows for substantial simplification model structure.

Format Requirements: The course will be divided into three one-hour sections - corresponding to the sections described above. Within each there will be 30 minutes of didactic teaching followed by a 30 minute exercise. The exercises will involve the use of TreeAge as the modelling environment. Students will need to have the latter installed on a laptop with either a full or student license. Knowledge of the implementation of EVC health-state transition models in TreeAge will be assumed.

Description and Objectives: This course will offer students a bridge between introductory courses on heath-state transition (HST) modeling, which describe expected value cohort (EVC) simulation, and advanced courses which deal with probabilistic sensitivity analysis and 2- and 3-dimensional simulation. The focus of the course is on the theory and application of microsimulation (MS).

Section 1 (background): limitations of EVC models, how MS models function and overcome the former; Section 2 (tracking variables & memory in MS models): properties and types of tracking variables, when to use regular vs. tracking variables, use of tracking variables in probability/utility expressions; section 3 (application of MS models): simplifying model structure, representing individual level uncertainty, dynamic modeling, validation and calibration of models.

Objectives: participants will gain an understanding of the theory and implementation of MS-based HST models, will understand how variable tracking and memory can increase the sophistication of HST models, they will be able to build dynamic models and use MS to validate and calibrate HST models.