Topic

Discrete event and individual-level simulation

Individual-level simulation follows patients one at a time, so that each person's history can shape what happens next. Discrete event simulation moves patients through events in continuous time and can represent queues and limited resources, while microsimulation and agent-based models track individual characteristics or interactions between people.

Concepts in this topic

• Agent-Based ModelAn agent-based model (ABM) simulates individuals as agents that follow rules and interact, so outcomes such as infections, costs and QALYs emerge.
• Arrival ProcessArrival process is the pattern in time, random or scheduled, by which patients or other entities enter a queueing model or simulation of a health service.
• Departure ProcessThe pattern by which entities, such as patients, leave a discrete event simulation after completing the services or processes being modelled.
• Discrete Event SimulationDiscrete event simulation (DES) is a modelling method that follows each patient from event to event over time, tracking their history and use of resources.
• Event QueueAn event queue is the time-ordered collection of pending events that a discrete-event simulation processes to advance its clock and update system states.
• Individual Patient SimulationIndividual patient simulation is a modelling approach that generates separate patient histories under specified rules and aggregates their costs and health outcomes for a target population.
• Markov MicrosimulationMarkov microsimulation is an individual-level state-transition modeling method that simulates each person's sequence of health states over time using transition probabilities that may depend on current state, personal characteristics, time, and explicitly retained history.
• MicrosimulationMicrosimulation models individuals one at a time, allowing their characteristics and event histories to influence future outcomes.
• Monte Carlo SimulationMonte Carlo simulation uses repeated random sampling to show how uncertainty in model inputs produces uncertainty in model outputs.
• Pseudo-Random NumberA pseudo-random number is a value produced by a deterministic algorithm whose sequence is designed to exhibit specified statistical properties of random draws.
• Queuing ModelA queuing model is a mathematical representation of how arrivals, service times, capacity, and service rules generate waiting, congestion, and throughput in a service system.
• Random Number GenerationRandom number generation supplies the uncertain draws used in probabilistic sensitivity analysis, microsimulation, discrete-event simulation, bootstrap procedures, and other stochastic analyses.
• Stochastic SimulationStochastic simulation represents a system in which at least one event, transition, time or outcome is generated probabilistically.
• Time-to-EventOutcome data recording the duration until a specific event of interest occurs, such as disease progression or death, possibly censored.