How to constrain the constant term using command constraints

[xuemin zhu]

Following a guideline, I am using a regression to calculate the treatment effect on cost while controling for other baseline characteristics. The constant of the regression is interpreted as the average cost for control group (in mean age and being a female) and the TreatmentDummy coefficient is the average treatment effect (in mean age and being a female), and the sum of constant and TreatmentDummy coeffecient is the average cost for the treated group (in mean age and being a female).

reg cost  TreatmentDummy age_centred gender 

However, my cost data is extremly sparsed and the estimated cost for treatment could sometimes be negative. I want to constrain 0<=TreatmentDummy coeffecient<= constant, and regress the model with cnsreg, but dont know how to write this constraint

[George Ford]

can you post the guideline? this seems very spooky.

[xuemin zhu]

can you post the guideline? this seems very spooky.

Hello George, this is to adjust for the potential baseline imbalance from recruitment in randomised control trials. here is a citation for similar circumstance.
Manca A, Hawkins N, Sculpher MJ. Estimating mean QALYs in trial-based cost-effectiveness analysis: the importance of controlling for baseline utility. Health Econ. 2005 May;14(5):487-96. doi: 10.1002/hec.944. PMID: 15497198.

[George Ford]

Thanks. I’ll look it over.