Concept Architecture
Concept
Theoretically, Probability-to-Rate Conversion is the mathematical transformation of an event probability observed over a specified time interval into a continuous event rate. The conversion is based on the exponential survival model, which assumes that events occur according to a constant hazard over the interval. In health economics, this transformation allows probabilities reported in clinical studies to be incorporated into models that require transition rates, particularly continuous-time Markov models and state-transition models with differing cycle lengths.
Mathematically, probability-to-rate conversion is derived from the relationship between the exponential survival function and the cumulative probability of an event. Assuming a constant hazard, the event rate is obtained by taking the negative natural logarithm of the survival probability divided by the observation period. This preserves consistency between probabilities and rates under the exponential distribution and enables conversion across different model time horizons.
In practice, reported probabilities from clinical trials, epidemiological studies or registries are routinely converted into transition rates before model implementation. The resulting rates can subsequently be converted back into probabilities for alternative cycle lengths or incorporated directly into continuous-time models. The method is widely applied when harmonising evidence sources measured over different follow-up periods.
Purpose
Used to convert observed event probabilities into continuous event rates for implementation in continuous-time models and for rescaling transition parameters across different model cycle lengths.
Mathematical Formulae
Primary Formula
For an event probability observed over time t,
r = ?ln(1 ? p) / t
where:
- r = continuous event rate
- p = event probability over time interval t
- t = observation period
Supporting Formulae
Relationship with the exponential survival function:
S(t) = e???
Probability derived from the survival function:
p = 1 ? e???
Related Mathematical Methods
- Hazard-to-probability conversion
- Rate-to-probability conversion
- Exponential survival modelling
- Survival analysis
- Continuous-time Markov modelling
- State-transition modelling
Example
A clinical study reports a one-year probability of disease progression of 0.30.
The corresponding continuous annual transition rate is
r = ?ln(1 ? 0.30) = ?ln(0.70) = 0.3567
This annual rate can subsequently be converted into probabilities for monthly or quarterly Markov model cycles while preserving the underlying hazard assumption.
Excel Implementation
| Function | Example Formula | Health Economics Application |
|---|---|---|
| LN | =-LN(1-B2)/C2 | Convert an observed probability into a continuous event rate |
| EXP | =1-EXP(-B2*C2) | Verify the converted rate by recovering the original probability |
| IF | =IF(B2=1,""Undefined"",-LN(1-B2)/C2) | Prevent calculation when probability equals one |
| ROUND | =ROUND(-LN(1-B2)/C2,6) | Round converted rates for model input tables |
| POWER | =1-EXP(-B2*(1/12)) | Calculate monthly probability from an annual rate |
VBA (Optional)
Automate conversion of probabilities reported over varying follow-up periods into continuous transition rates for health economic models.
Sources
- Briggs AH, Claxton K, Sculpher MJ. Decision Modelling for Health Economic Evaluation. Oxford University Press; 2006.
- Fleurence RL, Hollenbeak CS. Rates and Probabilities in Economic Modelling: Transformation, Translation and Appropriate Application. Pharmacoeconomics. 2007;25(1):3-6.
- Miller DK, Homan SM. Determining Transition Probabilities: Confusion and Suggestions. Medical Decision Making. 1994;14(1):52-58.
- Collett D. Modelling Survival Data in Medical Research. 3rd ed. CRC Press; 2015.
- NICE. Health Technology Evaluation Manual.
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 probability-to-rate conversion?
The mathematical transformation converting a probability of an event over a discrete time interval into its corresponding continuous-time hazard rate.
Source: Miller & Homan 1994
What does converting a probability to a rate make possible?
Expressing an event's chance as a continuous-time rate rather than a probability over a fixed period lets it be re-expressed for any other period consistently. A rate can be scaled to a shorter or longer cycle and then turned back into a probability appropriate to that cycle, which a probability cannot undergo directly without error. This is why probabilities drawn from studies over one period are converted to rates before being adapted to a model's cycle length. The rate is the transferable form. Fleurence and Hollenbeak (2007) explain this step.
Source: Fleurence & Hollenbeak 2007
How is a probability converted to a rate?
A probability over an interval is converted to a rate using the relationship that the rate equals the negative natural logarithm of one minus the probability, divided by the length of the interval. This inverts the exponential relationship linking rates and probabilities. The resulting continuous-time rate can then be scaled to a different interval and converted back to a probability for that period. Working through the rate ensures probabilities are re-expressed correctly across time periods, avoiding the errors of scaling probabilities directly.
Source: Miller & Homan 1994
Why convert a probability to a rate?
Converting a probability to a rate is necessary when a probability estimated over one time period must be re-expressed for a different period, because probabilities do not scale linearly with time and cannot simply be added or multiplied across intervals. By converting to the underlying continuous-time rate, which can be scaled to any interval, and then back to a probability, the transformation is done correctly. This is common when evidence gives a probability over one period but a model requires it over a different cycle length.
Source: Miller & Homan 1994
What errors arise if probabilities are rescaled directly?
Rescaling probabilities directly across time periods, such as doubling a one-year probability to get a two-year probability, produces errors because probabilities do not add or scale linearly: the correct combination accounts for the event not having occurred in the earlier period. Direct rescaling can also give probabilities above one, which are meaningless. These errors, highlighted by Miller and Homan, bias model results, so the conversion through the underlying rate is used to re-express probabilities across periods correctly.
Source: Miller & Homan 1994
How does probability-to-rate conversion support modelling?
Probability-to-rate conversion supports modelling by allowing event probabilities from evidence, which may be reported over various periods, to be re-expressed for a model's cycle length correctly. By converting a probability to its rate, the rate can be scaled to the required interval and converted back, ensuring the transition probabilities used match the cycle. This is important because model results depend on probabilities applied over the right period, so proper conversion, through the rate, is a basic step in deriving valid inputs.
Source: Briggs, Claxton & Sculpher 2006
Trust Record
Verified by Dr Darrin Baines
British health economist
Professional identity: darrinbaines.org
Verification date: 10 Oct 2025
Content version: 1.0.0
Canonical Identity
- Term code
- HE-EM-MP-035
Stable URI · Machine-readable · Resolvable · CC BY 4.0