Distributed as Quote of the Day on August 1, 2018
This morning the RAND study on the single payer New York Health Act was sent out primarily to demonstrate another study showing that the single payer model is effective for ensuring that everyone is covered while containing health care costs. However, a critique of the RAND analysis was not distributed, even though there are some significant deficiencies in the study.
One of the most important policies that characterizes a well designed single payer system is the recovery of a tremendous amount of administrative waste. Many of the studies minimize the amount of waste that exists and that can be recovered, in spite of the abundance of evidence published in the academic policy literature. This is crucial because that concept is often omitted from reports in the lay literature, and so the public is not well informed on this exceedingly important feature of the single payer model. In your single payer advocacy, always be sure that the magnitude of the recoverable administrative waste is not omitted from the discussions.
David Himmelstein and Steffie Woolhandler, distinguished professors of health policy at the City University of New York at Hunter College and co-founders of PNHP, have provided the following brief critique of the RAND study. In it they show that the benefit of single payer is far greater that that reported by the RAND authors, especially regarding the administrative savings.
***
Comments on the Rand Estimates of the Costs of the New York Health Act
By David U. Himmelstein, M.D. and Steffie Woolhandler, M.D., M.P.H.
1) The projected net changes in utilization given supply constraints are in the ballpark, although the portrayal of the supply constraints as implying waits is unduly negative. Supply constraints limited utilization increases with the implementation of Medicare and the ACA, yet there was no perceptible increase in waits for care. Rather, there was probably a reduction in unnecessary care for the previously well insured.
2) The estimate of reductions in drug prices is too low given the experience in other nations. In particular, the assumption that prices of drugs delivered in physicians’ offices would be reduced by a smaller percentage than other drug prices seems unjustified.
3) The estimates of administrative savings are absurdly low.
a) The assumption that health plan administrative overhead would only be reduced to 6% is completely unjustified. FFS Medicare’s administrative overhead is under 3%, while Canada’s single payer programs averaged overhead of 1.6% in 2017 – a total of $2.7 billion (Canadian). (Canada’s overhead costs are lower, at least in part, because it pays hospitals global budgets rather than FFS, as Medicare does. It seems likely that the NYH act would also globally budget hospitals, and hence achieve savings comparable to those in Canada). As the Rand authors note, the overhead rates for Medicare as a whole (7%), and for Medicaid (7%) are inflated because they include the extremely high overhead of managed care plans in those programs.
Rand claims that health plan administrative costs would be driven up by the need to verify residency, deal with out-of-state services, and track Medicaid and Medicare eligibility in order to draw down federal matching funds. The Canadian provincial programs perform the first two of these functions. Tracking Medicaid eligibility should take trivial resources – essentially just verifying income, which can largely be gleaned from income tax and Social Security records. The Social Security Administration should be able to provide information on Medicare eligibility at virtually no cost to the state.
In sum, there is no justification for the claim that it would cost $16.6 billion to administer coverage for 20 million New Yorkers – $830 per person, more than 8-fold higher than Canada’s cost. A more reasonable estimate would be 2%, or at most 3%.
b) It is hard to decipher exactly how the Rand authors estimated provider administrative savings. In Table 4.1 they indicate that at present, administrative costs account for 13% of physician and clinical service expenditures, 12% of hospital expenditures in New York and 10% of other service expenditures. They then estimate that provider administrative spending could be reduced by 13%, i.e. equivalent to a savings of less than 2% of total revenues.
While they list several references to support this estimate (and provide some further description in footnote 16 on page 33), it is unclear how they arrived at these figures, which are at odds with the actual findings of several of the references they cite. For instance, the Blanchfield paper they cite estimated that “excess” administrative costs in physician practices accounted for 11.9% of total net revenues, or about $50,250 per physician. The Himmelstein, Campbell and Woolhandler paper that they cite, found that administrative costs consumed about 26.9% of physicians’ revenues and 24.3% of hospital revenues, while these costs were 16.1% and 12.9% respectively in Canada. (Himmelstein et al’s more recent analysis of hospital administrative costs in 8 nations, which appeared in Health Affairs in 2014 reached virtually identical conclusions regarding hospital administrative costs and savings.) Both the Jiwani study and the Pozen and Cutler papers reached virtually identical conclusions as the Himmelstein, Campbell, Woolhandler study regarding administrative savings under single payer. The Mora et al study they cite concluded that physicians’ administrative costs totaled about $83,000 per physician in the US vs. about $27,000 per doctor in Canada, far larger savings than Rand assumes.
They reference Rand hospital data, which apparently relies on the same data source as the Himmelstein analyses of hospital administrative costs (which have undergone extensive peer review), but provide no information on how they derived administrative cost estimates from these data, and as far as we can tell their methods have neither been published nor peer reviewed. They cite their analysis of hospital administrative costs in Maryland’s all-payer system, which is irrelevant to an analysis of single payer, then use this paper as a “low estimate” in order to justify excluding a high estimate from a study that is clearly relevant.
In sum, according to the documentation in the report, it appears that the Rand authors have made an error in interpreting previous studies’ figures for administrative costs and savings for providers under single payer. These studies generally estimate that administrative SAVINGS under single payer would amount to 12%-13% of providers’ TOTAL REVENUES. The Rand authors appear to have mistakenly concluded that these figures are estimates of the total costs of administration (rather than potential savings), and have then assumed that 12% of administrative costs (rather than 12% of total revenues) would be saved. As a result they estimate that providers’ would achieve administrative savings of about 1% of total revenues, rather than about 12% – a very large difference.
Stay informed! Visit www.pnhp.org/qotd to sign up for daily email updates.