Concept Architecture
Concept
Theoretically, Hazard-to-Probability Conversion is the process of converting a continuous-time hazard rate into the probability that an event occurs over a specified discrete time interval. The conversion is necessary because survival analyses typically estimate hazards, whereas health economic models such as decision trees and Markov models usually require transition probabilities for fixed cycle lengths. The conversion ensures consistency between continuous-time event processes and discrete-time model implementation.
Mathematically, Hazard-to-Probability Conversion is based on the exponential survival function under the assumption of a constant hazard within the time interval. The probability of experiencing the event during a model cycle is obtained by subtracting the survival probability from one. This relationship preserves the underlying continuous-time hazard process while producing cycle-specific transition probabilities suitable for discrete-time simulation.
In practice, Hazard-to-Probability Conversion is applied when parameterising state-transition models, microsimulation and individual patient simulation using hazard rates obtained from clinical trials, survival models or epidemiological studies. The conversion is performed separately for each model cycle and may be repeated when hazards vary over time or differ between patient subgroups.
Purpose
Used to convert continuous-time hazard rates into discrete-time event probabilities for implementation in health economic decision models.
Mathematical Formulae
Primary Formula
Hazard-to-probability conversion:
p = 1 ? e^(?ht)
where:
- p is the probability of the event during the interval
- h is the constant hazard rate
- t is the length of the time interval.
Supporting Formulae
Survival function:
S(t) = e^(?ht)
Relationship between survival and probability:
p = 1 ? S(t)
Related Mathematical Methods
- Survival analysis
- Hazard modelling
- Parametric survival modelling
- State-transition modelling
- Markov modelling
- Microsimulation
- Individual patient simulation
Example
A Markov model uses annual hazard ratios obtained from a clinical trial but applies six-month model cycles.
The annual hazard rate is:
h = 0.24
The cycle length is:
t = 0.5
The six-month transition probability is:
p = 1 ? e^(?0.24 ? 0.5)
p = 1 ? e^(?0.12) = 0.113
Therefore, the probability of experiencing the event during each six-month cycle is 11.3%.
Excel Implementation
| Function | Example Formula | Health Economics Application |
|---|---|---|
| EXP | =1-EXP(-A2*B2) | Convert hazard rates into cycle-specific transition probabilities. |
| POWER | =1-POWER(EXP(1),-A2*B2) | Alternative implementation of hazard-to-probability conversion. |
| IF | =IF(A2>0,1-EXP(-A2*B2),0) | Prevent invalid probabilities when hazards equal zero. |
| XLOOKUP | =XLOOKUP(Age,HazardTable[Age],HazardTable[Hazard]) | Retrieve hazard rates before conversion. |
| INDEX | =INDEX(HazardTable[Hazard],MATCH(Stage,HazardTable[Stage],0)) | Retrieve subgroup-specific hazards for conversion. |
VBA (Optional)
Automate conversion of hazard rates to cycle-specific transition probabilities across multiple health states and model cycles.
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.
- Miller DK, Homan SM. Determining transition probabilities: confusion and suggestions. Medical Decision Making. 1994;14(1):52?58.
- National Institute for Health and Care Excellence (NICE). Health Technology Evaluation Manual. Latest edition.
- ISPOR-SMDM Modeling Good Research Practices Task Force. Modeling Good Research Practices. Value in Health. 2012.
Related Concepts (2)
Library
Publications
1
Decision Modelling for Health Economic Evaluation — Briggs, Claxton & Sculpher, 1st Edition ed., 2006 (Oxford University Press)
Foundational textbook on decision-analytic modelling for economic evaluation, covering decision trees, Markov models, handling parameter and structural uncertainty, probabilistic sensitivity analysis, and value of information. Volume 1 in the Handbooks in Health Economic Evaluation series.
BookView source →
Frequently Asked Questions (6)
What is hazard-to-probability conversion?
The mathematical transformation converting an instantaneous hazard rate into the probability of an event occurring over a specific discrete time interval.
Source: Miller & Homan 1994
Why can a hazard rate not be used directly as a probability?
A hazard rate measures events per unit of time and can exceed one, whereas a probability must lie between zero and one, so the two are different quantities that cannot be equated. A rate of two events per year, for instance, is not a probability of two, which is impossible. Converting between them uses an exponential relationship that accounts for the chance of the event recurring within the interval. Treating a rate as though it were a probability produces values that are wrong and sometimes impossible. Fleurence and Hollenbeak (2007) explain this conversion.
Source: Fleurence & Hollenbeak 2007
Why must hazard rates be converted to probabilities?
Hazard rates must be converted to probabilities because they are different quantities: a rate is an instantaneous frequency that can exceed one and does not directly give the chance of an event over a period, whereas a Markov model needs the probability of transition within a cycle. Using a rate as if it were a probability, or scaling it linearly with time, gives wrong results, since probabilities do not add across intervals. Proper conversion ensures the transition probabilities in the model are correct.
Source: Miller & Homan 1994
How is a hazard rate converted to a probability?
A hazard rate is converted to a probability over an interval using the relationship that the probability of the event equals one minus the exponential of the negative rate times the interval length. This accounts for the continuous nature of the rate and gives a probability between zero and one for the period. The formula ensures that converting a rate to a probability, or a probability for one period to another, is done correctly, since it works through the underlying rate rather than scaling probabilities directly.
Source: Miller & Homan 1994
What errors arise from incorrect conversion?
Incorrect conversion, such as treating a rate as a probability or scaling a probability linearly with time, produces wrong transition probabilities that bias a model's results. For example, doubling a one-year probability to get a two-year probability overstates it, since probabilities do not add, and using a rate above one as a probability is meaningless. These errors, discussed by Miller and Homan, can substantially distort outcomes, so the exponential conversion through the underlying rate is used to avoid them.
Source: Miller & Homan 1994
Why is hazard-to-probability conversion important in Markov models?
Hazard-to-probability conversion is important in Markov models because transition probabilities must match the model's cycle length, and evidence often comes as rates or as probabilities over different periods. Converting rates to cycle-specific probabilities correctly, through the exponential relationship, ensures that the model's transitions reflect the true risk over each cycle. Because errors in this conversion bias the movement of patients and hence the outcomes, applying it properly is a basic requirement for deriving valid transition probabilities in a Markov model.
Source: Briggs, Claxton & Sculpher 2006
Trust Record
Verified by Dr Darrin Baines
British health economist
Professional identity: darrinbaines.org
Verification date: 9 Oct 2025
Content version: 1.0.0
Canonical Identity
- Term code
- HE-EM-MP-017
Stable URI · Machine-readable · Resolvable · CC BY 4.0