search

pierucci / heemod

:chart_with_upwards_trend:Markov Models for Health Economic Evaluations
https://pierucci.org/heemod/
Other
41 stars 25 forks source link

count result is not what is expected #338

Closed mhrlittle closed 5 years ago

mhrlittle commented 5 years ago

Many thanks for creating this package. I've noticed the count produced by run_model to not be what would be expected. In the introduction example the initial count puts 1000 in state A with a transition probability of 0.1 to state B. Therefore, I would expect a count of 900 in the first cycle but the bottom the example shows a count of 950. I'm not sure if this is intended or not. Could you let me know how the count of 950 is arrived at.

jrdnmdhl commented 5 years ago

This is a result of the half-cycle correction method, which defaults to life-table (i.e. averaging the counts from adjacent times).

mhrlittle commented 5 years ago

Thanks