In Reply Dr Kitchell provides us an opportunity to explain the methods used in the IOM report on geographic variation more fully. The report used Medicare’s GPCI to adjust physician spending for geographic variation in input prices. We know of no serious concerns about the measurement of the price of the nonphysician inputs in the GPCI, such as the wages of billing clerks.
Questions about the GPCI center on the weight given to its physician work component and how well the geographic units it uses approximate local labor markets. Fortunately, the effects of inaccuracies in these 2 dimensions are necessarily limited. Physician work accounts for about half of physician costs (slightly more or less depending on the year), but by statute it is only given a weight of one-quarter of that amount in the actual index.1
Newhouse JP, Garber AM. Geographic Differences in US Health Care Spending—Reply. JAMA. 2014;311(6):624-625. doi:10.1001/jama.2013.284333