Screening model - entering a cohort within an age range and setting proportions for screening

Hi there,

I am working on a screening model with two arms: 1) screening arm where all patients are offered a blood test and 2) non-screening arm where patients non-systematically get offered a blood test (currently about 60% of people get offered this test). It's currently a decision node with 2 Markov nodes.

I had two questions I hoped I could get help with. I've attached an image of the first part of my model.

1. Is it possible for a cohort of men to enter within a range of age groups. I want men with 5-year age groups (e.g. 49-44, 45-50). I'm currently using the CurrentAge = StartAge+_stage approach so all men enter at the same age.


2. What would be the best approach to enable modelling of different proportions of the population being screened in each arm? I thought I would add an additional branch to the right of the 'Cancer free' state, with 'Screening' and 'No screening' and put the probabilities there? The transition probabilities for a diagnosis of cancer used in the screening arm are taken from an RCT where there was 90% uptake of screening. So I'm not sure if it's correct to change the screening proportions this way. 

Many thanks for your time.



1 comment
  • Cindy

    I can provide some quick answers to your question through this Community Forum - indeed more people may have feedback too.

    1. In a cohort model, there is no way to set up different age ranges within the same model. By nature, the cohort model is dealing with averages if the population. You could run groups through the model one at a time, each group would have a different state age. Or you could create a patient simulation model and patients of various ages in the model.

    2. If you are looking to model different proportions of the cohort being screened in each arm, it would just depend on where the screening happens. Your suggestion of adding the screening to the Cancer Free node could work - and then you can just set a variable which defines the proportion of the cohort to be screened or not. You can then do sensitivity analysis on this variable and see how to proportion screened impact the model outcomes.

    My only other comment is in relation to the other branches from the Cancer Free state. Are these Low/Imtermediate...etc risks found through the screening? If so, they may need to be Health States as opposed to where you have them now.

    If you want to discuss further details, you may want to contact Support directly through

    Thanks, Belinda


    Comment actions Permalink

Please sign in to leave a comment.

Didn't find what you were looking for?

New post