Anyone at all interested in the US health debate will already have read Atul Gawande's excellent New Yorker piece, which compares health spending in McAllen, Texas ($15,000 per Medicare enrollee in 2006) with spending in El Paso (only half as much, even though its relevant characteristics seem similar). His recent commencement address at the Pritzker School of Medicine is also very good. Both pieces serve to shift the focus away from the preoccupation with public v private insurance, because the enormous variations in health spending he discusses are within the public system. They also slightly dampen one's hopes for comparative effectiveness research, all the rage just now, as a way to save money--unless it is used alongside changes in financial incentives. Changes in payment delivery are the key thing. How are doctors and hospitals reimbursed? That is what matters.
The Health Affairs blog has two posts for supplementary reading. One is about the Dartmouth research that provides the data for this kind of comparison (and which is facing criticism of late). The other is a roundtable on the Gawande article.