Tutorials

Every tutorial, in the order a reader learns the package, from the standard analysis workflow through the model engines to calibration and replications.

Keywords

cost-effectiveness modeling tutorial, cost-effectiveness analysis, value of information, probabilistic sensitivity analysis, Markov model, microsimulation, discrete-event simulation, Bayesian calibration, health economics, Python

This page lists the tutorials in the order a reader learns the package: the standard analysis workflow first, then one tutorial per model engine, then sensitivity and value-of-information analysis, calibration, and replications of published models. Each group builds on the ones before it, and within a group the prerequisite comes first.

Getting started

The standard probabilistic sensitivity analysis runs in four steps: specify parameters, sample draws, evaluate a model, and analyze cost-effectiveness and value of information. The three tutorials here cover that workflow, and how to supply results or parameters computed outside the package.

Full pipeline

Run the standard probabilistic sensitivity analysis end to end: parameters, sampled draws, run_psa, then cost-effectiveness and value of information.

Bring your own outputs

Take a costs-and-effects table computed elsewhere through cost-effectiveness, value of information, plots, and a reproducibility record, without building a model engine.

Parameter inputs from data

Turn point values, an exported draw matrix, or a weighted posterior sample into the parameter draw matrix run_psa expects.

Model engines

Each engine builds a different kind of model and then runs the same analysis, so only the model-building step changes between them. The cohort state-transition model is the standard engine; the microsimulation, discrete-event, and compartmental engines add the patient history, resource contention, and disease transmission a cohort average cannot represent.

Cohort state-transition model

Build a cohort state-transition model with MarkovModel and validate it against a published Sick-Sicker result.

Microsimulation engine

Build an individual-level model with MicrosimModel, for outcomes that depend on patient history or on heterogeneity a cohort average removes.

Markov vs microsimulation models

Build one model as both a cohort trace and an individual simulation from the same rates, to cross-validate the engines and show what heterogeneity adds.

Discrete-event simulation engine

Build a resource-constrained model with DESModel, for a scarce shared resource that makes patients queue.

Compartmental transmission model

Build an infectious-disease model with ODEModel, a susceptible-exposed-infectious-recovered model of a vaccination program.

Sensitivity and value of information

Deterministic sensitivity analysis asks which parameters move the result; value-of-information analysis asks what resolving the remaining uncertainty is worth. Both read the outcomes an engine produces, so they apply after any of the model tutorials.

Deterministic sensitivity analysis

Run the three deterministic sensitivity designs, one-way, one-at-a-time, and grid, and read them as tornado and heatmap reports.

Value of information

Run the value-of-information workflow: expected value of perfect, partial perfect, and sample information.

Calibration

Calibration fits model parameters to observed data and carries the fitted uncertainty into the analysis. The workflow tutorial comes first, the next three compare inference methods on one shared model, and the last calibrates a stochastic microsimulation.

Calibration workflow

Mix calibrated parameters with literature parameters in one probabilistic analysis using mix_draws.

Calibrating with ABC

Calibrate a parameter to observed data with approximate Bayesian computation, the first of four calibration tutorials that share one disease model.

Calibrating with simulation-based inference

Calibrate the same disease model with neural posterior estimation, a form of simulation-based inference, and recover the same posterior as approximate Bayesian computation.

Surrogate-accelerated calibration

Fit a fast surrogate of an expensive model on a small design, then calibrate against the surrogate instead of the model.

Calibrating a stochastic microsimulation

Calibrate a stochastic microsimulation, carrying its simulation noise into the posterior with a Gaussian process surrogate.

Replications

Each replication reproduces a published cost-effectiveness analysis, matching its deterministic result before extending it to a probabilistic analysis. Together they check the package against known answers and show complete worked models.

Replication gallery

Four published Sick-Sicker cost-effectiveness tutorials reproduced with heormodel, one per model type.

Time-dependent cohort model

Extend the cohort replication with age-varying mortality and one-time transition rewards.

Microsimulation replication

Reproduce the Sick-Sicker microsimulation of Krijkamp and others (2018) with MicrosimModel.

Discrete-event simulation replication

Reproduce the continuous-time Sick-Sicker discrete-event simulation of Lopez-Mendez and others (2026).

Keywords: cost-effectiveness modeling tutorial, cost-effectiveness analysis, value of information, probabilistic sensitivity analysis, Markov model, microsimulation, discrete-event simulation, Bayesian calibration, health economics, Python