Kangho Suh wrote an excellent article on the use of instrumental variables in healthcare. I agree this technique strengthens causal inference for analyses of observational data. This is even more important when a randomized controlled trial is not feasible or ethical.
