A toad is more accurate
Investigator Robert Bendesky reports that, far more frequently than customer service representatives, a toad produced the correct answer to physicians' questions about Medicare policy. Using a system in which a jump to the left meant "yes" and a jump to the right meant "no," the toad, not surprisingly, scored 50% correct. Unfortunately, the customer service representatives were incorrect 96% of the time, according to a 2004 GAO study described here (warning: link leads to a PDF file).
OK, perhaps it wasn't a fair test, given that the toad only had to choose between "yes" or "no" answers, but sadly I doubt the result would be much better even if the same questions had been given the CSRs. On the toad's side, the report notes that this test was entirely "open book." The managers of the call service centers knew in advance on what days the test calls would be made, and the questions had been intentionally designed to be simple and straightforward questions based on the insurer's own policies as published on their own websites. Indeed, the call center managers had even been informed of the questions ahead of time.
No word about whether the questions for the toads had been designed to be simple or whether the toad had been informed in advance of the test.
The conclusion: Medicare regulations are so ridiculously complex that neither doctors nor even service representatives (whose job it is to explain them) understand them. So, not only does Medicare reimburse physicians at a rate that barely covers their expenses (and sometimes not even that), but it makes it very difficult for the physician to collect even that!