Concept Architecture
Concept
Theoretically, Individual Patient Simulation is a patient-level decision-analytic modelling approach in which each simulated individual progresses independently through a disease pathway according to specified probability distributions, event rules and patient characteristics. It was developed to overcome limitations of cohort models when modelling patient heterogeneity, complex treatment pathways, competing risks, history-dependent events and individual variability. Each simulated patient represents one realisable disease trajectory, and population outcomes are obtained by aggregating results across many simulated individuals.
Mathematically, Individual Patient Simulation is represented as a stochastic simulation in which patient-specific state transitions, event times or outcomes are generated using probability distributions conditioned on individual characteristics and previous history. Individual trajectories are produced through repeated Monte Carlo sampling, with expected costs, health outcomes and uncertainty estimated by averaging results across a sufficiently large simulated population. There is no universally recognised canonical mathematical formula.
In practice, Individual Patient Simulation is implemented by defining baseline patient characteristics, clinical pathways, event probabilities, treatment effects, costs and health-related quality of life for each simulated individual. Patient histories are updated sequentially throughout the simulation, allowing future events to depend on previous events and accumulated characteristics. The model outputs include expected costs, life-years, quality-adjusted life-years (QALYs), incremental cost-effectiveness ratios and probabilistic measures of decision uncertainty.
Purpose
Used to evaluate healthcare interventions when patient heterogeneity, disease history, competing events, complex treatment pathways or individual variability cannot be adequately represented using cohort-based models.
Mathematical Formulae
Primary Formula
There is no universally recognised canonical mathematical formula.
Supporting Formulae
Expected population outcome:
E(Y) = (1/N) ????? Y?
where:
- Y? is the simulated outcome for patient i
- N is the number of simulated patients.
Incremental Net Monetary Benefit:
INMB = ??Q ? ?C
where:
- ? is the willingness-to-pay threshold
- ?Q is incremental QALYs
- ?C is incremental costs.
Related Mathematical Methods
- Monte Carlo simulation
- Probability sampling
- Survival analysis
- Time-to-event modelling
- Markov microsimulation
- Discrete event simulation
- Regression modelling
- Probabilistic sensitivity analysis
Example
A health technology assessment compares a new anticoagulant with standard therapy in 100,000 simulated patients with atrial fibrillation. Each patient is assigned age, sex, prior stroke history and bleeding risk. During each simulation, stroke, major bleeding and death are sampled according to patient-specific risks.
The average outcomes are:
- Mean lifetime cost (standard care): �19,450
- Mean lifetime cost (new treatment): �21,120
- Mean QALYs (standard care): 10.42
- Mean QALYs (new treatment): 10.88
Incremental values:
?C = �1,670
?Q = 0.46
ICER = 1670 / 0.46 = �3,630 per QALY
The intervention would generally be considered cost effective under a �20,000 per QALY threshold.
Excel Implementation
| Function | Example Formula | Health Economics Application |
|---|---|---|
| RAND | =RAND() | Generate uniform random numbers for Monte Carlo sampling. |
| NORM.INV | =NORM.INV(RAND(),65,10) | Sample patient age or continuous clinical parameters. |
| IF | =IF(RAND()<0.08,1,0) | Simulate occurrence of clinical events. |
| INDEX | =INDEX(Table1[Cost],State) | Retrieve state-specific costs for each patient. |
| SUM | =SUM(Costs) | Calculate total patient costs. |
| AVERAGE | =AVERAGE(QALYs) | Estimate expected population outcomes across simulated patients. |
VBA (Optional)
Automate the simulation of thousands of individual patient trajectories and aggregate lifetime costs, health outcomes and probabilistic cost-effectiveness results across repeated Monte Carlo iterations.
Sources
- Briggs A, Claxton K, Sculpher M. Decision Modelling for Health Economic Evaluation. Oxford University Press; 2006.
- Drummond MF, Sculpher MJ, Claxton K, Stoddart GL, Torrance GW. Methods for the Economic Evaluation of Health Care Programmes. 4th ed. Oxford University Press; 2015.
- ISPOR-SMDM Modeling Good Research Practices Task Force. Modeling Good Research Practices. Value in Health. 2012.
- National Institute for Health and Care Excellence (NICE). Health Technology Evaluation Manual. Latest edition.
- Husereau D, Drummond M, Augustovski F, et al. CHEERS 2022 Statement: Updated Reporting Guidance for Health Economic Evaluations.
Related Concepts (4)
Library
Publications
4
NICE DSU Technical Support Document 16: Adjusting survival time estimates in the presence of treatment switching — Latimer & Abrams, TSD 16 ed., 2014 (NICE Decision Support Unit (University of Sheffield))
Guidance on statistical methods (RPSFTM, IPCW, two-stage) for adjusting overall-survival estimates when patients in a trial switch from the control arm to the experimental treatment, a common problem in oncology economic evaluation.
Modeling Using Discrete Event Simulation: A Report of the ISPOR-SMDM Modeling Good Research Practices Task Force-4 — Karnon, Stahl, Brennan, Caro, Mar & Moller, Task Force Report 4 ed., 2012 (Value in Health / Medical Decision Making)
Best-practice guidance on discrete event simulation (DES) for health economic evaluation — when DES is preferable to cohort approaches, and how to structure, populate and validate such models.
Journal ArticleView source →Microsimulation Modeling for Health Decision Sciences Using R: A Tutorial — Krijkamp, Alarid-Escudero, Enns, Jalal, Hunink & Pechlivanoglou, Vol. 38, No. 3 ed., 2018 (Medical Decision Making)
The DARTH workgroup’s tutorial on implementing individual-level (microsimulation) models in R, including vectorisation techniques that dramatically reduce run time — the standard reference for patient-level simulation in R.
Journal ArticleView source →Patient-Level Health Economic Modeling in Excel Without VBA: A Tutorial — Mike Paulden, Tutorial ed., 2025 (PharmacoEconomics)
A step-by-step tutorial implementing an individual-level discrete event simulation entirely in native Excel using contemporary functions and no Visual Basic (VBA) code, demonstrating flexible patient-level modelling in familiar spreadsheet software.
Journal ArticleView source →
Tools & Resources
1
hesim — Health Economic Simulation Modeling and Decision Analysis (R package) — Devin Incerti & Jeroen P. Jansen, R package ed., 2024 (CRAN)
A modular, computationally efficient R package for building and analysing health economic simulation models — cohort state-transition, partitioned survival, and individual-level continuous-time models — with fast individual-patient simulation and PSA via C++.
Software (R package)View source →
Frequently Asked Questions (6)
What is individual patient simulation?
A modelling technique simulating a large number of individual patients one at a time, each following their own pathway from random sampling.
Source: Barton, Bryan & Robinson 2004
How does individual patient simulation differ from a cohort model?
Individual patient simulation follows one simulated patient at a time from start to finish, giving each their own characteristics and letting chance decide their path, whereas a cohort model moves a whole group together as proportions across states. Because each patient is tracked separately, the method can let outcomes depend on an individual's full history and represent variation between people. The cost is that many patients must be run to average out the randomness. Davis and colleagues (2014) contrast the two approaches.
Source: Davis et al. 2014
How does individual patient simulation work?
Individual patient simulation works by generating patients one at a time, assigning each characteristics sampled from input distributions, then simulating their progression through the model, with the events and transitions they experience determined by random sampling against the relevant probabilities. Each patient accrues costs and effects along their own path, which may depend on their characteristics and history. Many patients are simulated, and their outcomes are averaged to estimate expected results, so population outcomes emerge from the aggregation of individual pathways.
Source: Barton, Bryan & Robinson 2004
What can individual patient simulation capture?
Individual patient simulation can capture patient heterogeneity, since each patient can have distinct characteristics affecting their risks, costs, and outcomes, and patient history, since a patient can carry a memory of past events that influences future transitions, which a memoryless cohort model cannot. It can also represent complex interactions between characteristics and events. These features let it model situations where individual variation and history matter, which is its main advantage over the aggregate cohort approach.
Source: Barton, Bryan & Robinson 2004
What are the costs of individual patient simulation?
Individual patient simulation is computationally intensive, requiring many patients, often many thousands, to be simulated to average out the random variation in individual paths, and its results carry Monte Carlo sampling error that a deterministic cohort model avoids. It is more complex to build, check, and communicate, and it requires data on the distribution of patient characteristics. These costs mean it is used when the heterogeneity and history it captures materially affect the results, rather than as a default.
Source: Barton, Bryan & Robinson 2004
When is individual patient simulation preferred?
Individual patient simulation is preferred when outcomes depend on patient heterogeneity or on history beyond the current state, which a cohort Markov model cannot represent without impractically many states. For example, where future risks depend on the accumulation of past events, or where the population is heterogeneous in ways that affect results, individual simulation captures this. Where such features do not materially affect the outcome, a simpler cohort model suffices, so individual simulation is chosen when its added realism is needed.
Source: Barton, Bryan & Robinson 2004
Trust Record
Verified by Dr Darrin Baines
British health economist
Professional identity: darrinbaines.org
Verification date: 8 Oct 2025
Content version: 1.0.0
Canonical Identity
- Term code
- HE-EM-MS-001
Stable URI · Machine-readable · Resolvable · CC BY 4.0