I think it's not so complicated, once you understand better what you want to measure and what you can measure. Often, the data is not what you want, so you use proxies. But then you have "complications" and disagreements.
Much closer study of what hospitals actually do, more what business consultants might do, would pay off. But this puts Big Data at a disadvantage. Economic research really needs some of those anthropological types if it is to be a real policy contributor, rather than a group of people with one main methodological focus. Economics can be, and is practiced through fieldwork, albeit by a very small number of economists.