I think a good idea could be to create a table with the treatment cost per stage (or cycle) eg. in the first line of the table $0, first cycle $100, second cycle $0, third cycle $0, fourth cycle $100 and so on.

Then on the incremental cost of the Markov info you can reference the table instead of using a value or a formula eg.: tablename[_stage], (don't forget to use the [_stage] command!!), and the software will use the cost of each cycle specified in the table.

This will count the cost every 3 weeks and is less work. Note that the '3' in the formula can be replaced with a variable if you want that to change over time.

I also use the modulo() function. Another example would be costs for an annual screening process where the cohort (or the individual in a microsimulation) is assigned a cost for a screening test every 12 months:

## Comments

Ruben RojasHi Tatsunori,

I think a good idea could be to create a table with the treatment cost per stage (or cycle) eg. in the first line of the table $0, first cycle $100, second cycle $0, third cycle $0, fourth cycle $100 and so on.

Then on the incremental cost of the Markov info you can reference the table instead of using a value or a formula eg.: tablename[_stage], (don't forget to use the [_stage] command!!), and the software will use the cost of each cycle specified in the table.

Hope this helps!,

Ruben

Harpreet SinghHi,

Another way is to use the modulo command. e.g.

If(modulo(_stage;3)=0;Costvariable;0)

This will count the cost every 3 weeks and is less work. Note that the '3' in the formula can be replaced with a variable if you want that to change over time.

David NaimarkI also use the modulo() function. Another example would be costs for an annual screening process where the cohort (or the individual in a microsimulation) is assigned a cost for a screening test every 12 months:

if(modulo(_stage;12/CycleLength) = 0;DR*cAnnualScreening;0)

where 'CycleLength' is the cycle length in months and 'DR' is the discount function (if global discounting is not used)

Andrew MunzerI like the Modulo approach suggested by Harpreet Singh and David Naimark.

This provides more flexibility by allowing you to run sensitivity analysis on the cost parameter (e.g., cAnnualScreening).

Please sign in to leave a comment.