My understanding of intention to treat analysis is that it is commonly used in a clinical trial situation. The way you manage your data is you treat everyone who was in treatment group 1 (the first antipsychotic medication)as if they took the medication exactly as instructed. The same for treatment group 2.
The reason you do this is there is no way you could factor in compliance or non-compliance in either group. You often have no idea how compliant particpants have been. So you analyse their data as if they receive treatment as per your initial intention; hence intention to treat analysis.
Thanks for helping me I randomized patients into two groups e.g one group recieving drug A and another drug B. v hav to follow both of these for 6 months and then vl analyze data regarding effectiveness. But cases of dropout are high so how ITT could b implemented and how statistically analyzed.
It sounds like you have a randomized open label/not blinded trial. That would mean that you may not only have drop out but also also cross-over where patients in group A start taking drug B and vice versa.
I think that I agree with David Schmidt but I am not certain where he says: "as if they took the medication exactly as instructed"
An intent to treat analysis follows patients split by trial arm according to the originally intended treatment at time of enrollment = according to randomization.
You measure the health effects as the differences between the trial arms and you also measure the cost impacts as differences between the trial arms.
If a patients stops taking the drug to which he or she was randomized, the the trial results reflect less health effects than when the patient would have taken the drug but because the cost of taking the drug is also not incurred, nor the costs associated with the decreased health effects so that you get a proper health economic analysis that follows the statistically valid analysis of the trial results.