Concept Architecture
Concept
Theoretically, Beta Distribution is a continuous probability distribution defined on the interval from 0 to 1 and characterised by two positive shape parameters. It provides a flexible framework for modelling probabilities, proportions and bounded random variables, making it particularly suitable for representing uncertainty in health economic parameters such as transition probabilities, utilities and event risks.
Mathematically, the Beta distribution is represented by a probability density function parameterised by the shape parameters � and ?. The distribution can assume a wide range of forms, including uniform, symmetric, skewed and U-shaped distributions, depending on the values of these parameters. Its mathematical properties allow uncertainty in bounded parameters to be quantified and propagated through probabilistic sensitivity analysis.
In practice, Beta distributions are estimated from observed proportions, expert elicitation or method-of-moments estimation using sample means and variances. They are routinely assigned to probability and utility parameters in probabilistic sensitivity analyses within decision trees, Markov models and microsimulation models to preserve the natural constraint that values remain between 0 and 1.
Purpose
Used to model uncertainty in probabilities, proportions and utility values that are naturally bounded between 0 and 1 within health economic evaluations and probabilistic sensitivity analysis.
Mathematical Formulae
Primary Formula
f(x; �, ?) = �(� + ?) / (�(�)�(?)) ? x^(� ? 1) ? (1 ? x)^(? ? 1),?0 � x � 1
Supporting Formulae
Mean:
? = � / (� + ?)
Variance:
Var(X) = �? / ((� + ?)�(� + ? + 1))
Method-of-moments parameter estimation:
� = ? ? ((?(1 ? ?) / ��) ? 1)
? = (1 ? ?) ? ((?(1 ? ?) / ��) ? 1)
Related Mathematical Methods
- Method of Moments
- Maximum Likelihood Estimation
- Bayesian Analysis
- Probabilistic Sensitivity Analysis
- Monte Carlo Simulation
- Dirichlet Distribution
Example
A Markov model estimates the annual probability of disease progression as 0.30 with an estimated variance of 0.005.
? = 0.30
�� = 0.005
Estimated Beta parameters:
� = 12.3
? = 28.7
The transition probability is therefore sampled from Beta(12.3, 28.7) during each Monte Carlo simulation, ensuring all sampled probabilities remain between 0 and 1.
Excel Implementation
| Function | Example Formula | Health Economics Application |
|---|---|---|
| BETA.DIST | =BETA.DIST(A2,12.3,28.7,FALSE) | Calculate the Beta probability density for probabilistic sensitivity analysis. |
| BETA.INV | =BETA.INV(RAND(),12.3,28.7) | Generate random transition probabilities or utility values for Monte Carlo simulation. |
| AVERAGE | =AVERAGE(A2:A101) | Estimate the sample mean for Beta parameter estimation. |
| VAR.S | =VAR.S(A2:A101) | Estimate the sample variance for method-of-moments parameter estimation. |
VBA (Optional)
VBA can automate repeated sampling from Beta distributions during probabilistic sensitivity analyses and export simulation results for decision uncertainty analysis.
Sources
- Briggs A, Claxton K, Sculpher M. Decision Modelling for Health Economic Evaluation. Oxford University Press.
- Drummond MF, et al. Methods for the Economic Evaluation of Health Care Programmes. Oxford University Press.
- ISPOR-SMDM Modeling Good Research Practices Task Force Reports.
- NICE. Health Technology Evaluation Manual.
Related Concepts (3)
Library
Publications
1
Parameter Estimation and Uncertainty: A Report of the ISPOR-SMDM Modeling Good Research Practices Task Force-6 — Briggs, Weinstein, Fenwick, Karnon, Sculpher & Paltiel, Task Force Report 6 ed., 2012 (Value in Health / Medical Decision Making)
Best-practice guidance on parameter estimation and the characterisation of uncertainty in decision models, covering probabilistic sensitivity analysis, distributional choices, and correlation between parameters.
Journal ArticleView source →
Frequently Asked Questions (6)
What is the beta distribution?
A flexible probability distribution bounded between zero and one, commonly used to represent uncertainty in probability parameters.
Source: Briggs A, Claxton K, Sculpher M. Decision Modelling for Health Economic Evaluation. Oxford University Press; 2006. doi:10.1093/oso/9780198526629.001.0001.
What range of values does the beta distribution allow?
The beta distribution is confined to values between zero and one, which is exactly the range a probability can take, so it cannot produce the impossible values a less restricted distribution might. Within that range it can take many shapes, from symmetric to strongly skewed, letting it represent uncertainty about a probability that is thought to be low, high, or middling. Its two parameters can be set directly from the counts of events and non-events observed. This bounded flexibility is why it is the usual choice for probability parameters. Briggs and colleagues (2006) describe this use.
Source: Briggs et al. 2006
Why is the beta distribution used for probabilities?
The beta distribution is used for probabilities because it is bounded between zero and one, matching the range a probability must lie in, so sampled values are always valid, unlike a normal distribution that could give values below zero or above one. Its two parameters give flexibility to represent a range of shapes reflecting the uncertainty. This makes the beta distribution the natural choice for representing the uncertainty in probability and proportion parameters in probabilistic sensitivity analysis.
Source: Vose 2008
How are the parameters of a beta distribution set?
The parameters of a beta distribution are set to reflect the estimated probability and its uncertainty, often derived from the data behind the estimate: for a proportion from a sample, the two shape parameters can be set from the numbers of events and non-events, so the distribution reflects the sample size and hence the precision. A larger sample gives a tighter distribution. This lets the beta distribution represent both the central estimate and the uncertainty in a probability parameter based on its evidence.
Source: Vose 2008
Where is the beta distribution used in modelling?
In health economic modelling, the beta distribution is used to represent the uncertainty in probability and proportion parameters, such as transition probabilities, event risks, and utility values bounded between zero and one, in probabilistic sensitivity analysis. Values are sampled from the beta distribution for each such parameter, ensuring the samples stay within the valid range. This makes the beta distribution a standard tool for characterising and propagating the uncertainty in probability-type inputs through a model.
Source: Briggs, Claxton & Sculpher 2006
What are the limitations of the beta distribution?
The beta distribution is bounded between zero and one, so it suits probabilities and proportions but not quantities outside that range, such as costs, for which distributions like the gamma are used. Its two parameters, while flexible, may not capture every shape of uncertainty, and setting them requires appropriate data or judgement. For utilities that can be negative, representing states worse than death, the standard beta distribution does not apply directly without transformation. These limitations mean the beta distribution is used specifically for parameters bounded in the zero-to-one range.
Source: Briggs, Claxton & Sculpher 2006
Trust Record
Verified by Dr Darrin Baines
British health economist
Professional identity: darrinbaines.org
Verification date: 24 Oct 2025
Content version: 1.0.0
Canonical Identity
- Persistent URI
- https://healtheconomics.wiki/concept/beta-distribution
- Term code
- HE-EM-UA-008
Stable URI · Machine-readable · Resolvable · CC BY 4.0