Adding Events to a Markov Model Using DICE Simulation.

Area of Science:

• Health economics and outcomes research
• Mathematical modeling in healthcare
• Decision analysis

Background:

• Health care decisions frequently involve uncertainty, necessitating robust modeling to evaluate choices and outcomes.
• State-transition (Markov) models are standard but primarily focus on states, neglecting explicit event considerations.

Purpose of the Study:

• To introduce Discretely Integrated Condition Event (DICE) simulation as an extension of Markov models.
• To demonstrate how DICE simulation can explicitly incorporate events into health care decision models.

Main Methods:

• DICE simulation models aspects that persist over time (conditions) and discrete occurrences (events).
• Markov models are specified in DICE by treating states as conditions with transition events and adding other explicit events.

Main Results:

• DICE specifications are compact, transparent (tabular), and flexible for alternative structures.
• Events can represent clinical occurrences, treatments, or health care activities, coinciding with or independent of transitions.
• Varying cycle times and sensitivity analyses are easily implemented.

Conclusions:

• DICE simulation expands the Markov formulation to explicitly include numerous events occurring at various times.
• This approach offers a more comprehensive framework for modeling complex health care scenarios.