Modelling approaches: the case of schizophrenia

Area of Science:

• Health Economics
• Computational Biology
• Psychiatry

Background:

• Schizophrenia is a chronic, heterogeneous disease with variable patient trajectories.
• Accurate modeling is crucial for understanding disease progression and treatment effectiveness.

Purpose of the Study:

• To critically assess decision trees, Markov models, and discrete event simulation for schizophrenia research.
• To guide the selection of appropriate modeling techniques based on research objectives and data characteristics.

Main Methods:

• Comparative analysis of three modeling techniques: decision trees, Markov models (cohort and micro-simulation), and discrete event simulation.
• Evaluation criteria included building time, data needs, clinical representation, and ability to handle patient variability and event timing.

Main Results:

• Each modeling technique has distinct advantages and disadvantages.
• Discrete event simulation models are superior for capturing complex interdependencies and patient heterogeneity in schizophrenia, especially for cost-effectiveness analyses.

Conclusions:

• The choice of modeling technique for schizophrenia should align with the specific research question, considering factors like covariate interactions and simulation time.
• Discrete event simulation is recommended for cost-effectiveness analyses requiring detailed representation of patient heterogeneity and event dynamics in schizophrenia.