Should I Give An Expensive Drug To A Patient That Will Likely Be Non-Adherent

This is a tough but interesting question.  If I know that a patient is not likely to take their medication for very long (i.e., be non-adherent) for whatever reason, should I pay for their drug?  For a long-term maintenance medication, if they only take it for 3 months and then stop, I’ve just wasted money. 

For a generic drug, this isn’t much of an issue.  My cost is low.

For an average priced brand drug, this is probably a calculated risk.

BUT, for an expensive specialty medication that might cost $10,000 (or more) a month, this is an interesting ethical question.

• Can I really predict their likelihood of being adherent?  If I’m sure their not going to be adherent, why pay for the medication?
• Can I put some burden on the patient (e.g., your copay is $50 per month, but if you stop taking it within 6 months, we’ll bill you $1,000)?  In theory, that creates an interesting stick / carrot, but since I’m not going to sit there and monitor them (unless it’s an injectible which is provided in the physician’s office), that would likely lead to some manipulation?

I was looking at an article this morning about how FICO (credit scores) can be used for predicting adherence, and it made me think of this issue. 

Another reason to keep your credit scores up?  I doubt it will ever be used this way, but it does show how information crosses from one area of our live to another.  On the flipside, it also indicates that behaviors in one area may be good indicators of behaviors in another area.