Cost savings–a bleg
April 1, 2014 2 Comments
My small brain is turned in knots on the seemingly simple question laid out below.
Say I have a clinical trial where the intervention is stopping the drug, and there are no mortality differences between treatment and control groups, and I want to estimate the cost savings of stopping the drug, based on the assumption that it is a good thing to do since it didn’t reduce mortality.
- Option 1. cost savings are defined after randomization as: incurred costs of the continued group (assume +$10) + the avoided costs of the discontinue group (assume -$10). Savings = $20/person
- Option 2. cost savings are defined after randomization as: the avoided costs of the discontinued group alone (assume =$10). Savings = $10/person
Option 1 highlights the counterfactual of continued use of the drug, and shows what could be achieved if you managed to get lots of folks to follow the results of the trial (I think). Option 2 is more conservative and only focuses on reduced costs/person who stops.
Which is correct?