On 2017 Oct 13, John Tucker commented:

In this article, Dr. Goldstein et. al raise the pressing issue of drug pricing. They conduct several regression analyses to determine the influence of market changes on oncology drug prices, including labeling changes, off-label use, and the approval of competing drugs on cumulative price increases. Most notably, however, the article serves as a vehicle for the price charts shown in Figures 1 and 2, showing cumulative price increases that are asserted to greatly outstrip the BLS Medical Care Consumer Price Index. The implication of these figures is that drug prices have risen even faster than healthcare prices overall, and thus play a unique role in the rapidly rising cost of care.

Unfortunately, the paper contains several errors that undercut support for this claim. The key figures show a 39% increase in the BLS Medical Care Price Index from 2005 to 2017, but reference to the primary source shows the increase was 49%. The authors appear to have confused the Medical Commodities CPI, which mainly addresses the costs of pharmaceuticals and medical devices, with the more broadly based Medical Care CPI.

Simultaneously, drug price increases are overstated by reliance on Average Sales Price data from CMS uncorrected for 340B discounts. The 340B program mandates discounts of 30 to 50% to hospitals meeting certain criteria related to providing services to low income population, and now includes approximately 1/3 of U.S. hospitals. The program has grown at double digit per annum rates over the time period covered by the Goldstein analysis. Genentech recently disclosed that 17% of its worldwide drug sales were conducted under 340B contracts, which would be approximately 35% of U.S. sales.

Correcting for the effects of the 340B program using publicly available data on the size of the program and hospital/clinic drug expenditures, one finds that the great majority of the price increases discussed in the paper fall below the medical inflation rate described by the BLS Medical Services CPI, and most fall below the general inflation rate as well.

Overall, there is inarguably a pressing need to control healthcare costs in the U.S. To achieve this we need to look at all sectors of spending, and should expect significant sacrifices across the board. Drug prices will be one part of that, but unnecessary and low value care provision, medical salaries, and insurer overhead will surely play an important role as well.

On 2017 Oct 13, John Tucker commented:

In this article, Dr. Goldstein et. al raise the pressing issue of drug pricing. They conduct several regression analyses to determine the influence of market changes on oncology drug prices, including labeling changes, off-label use, and the approval of competing drugs on cumulative price increases. Most notably, however, the article serves as a vehicle for the price charts shown in Figures 1 and 2, showing cumulative price increases that are asserted to greatly outstrip the BLS Medical Care Consumer Price Index. The implication of these figures is that drug prices have risen even faster than healthcare prices overall, and thus play a unique role in the rapidly rising cost of care.

Unfortunately, the paper contains several errors that undercut support for this claim. The key figures show a 39% increase in the BLS Medical Care Price Index from 2005 to 2017, but reference to the primary source shows the increase was 49%. The authors appear to have confused the Medical Commodities CPI, which mainly addresses the costs of pharmaceuticals and medical devices, with the more broadly based Medical Care CPI.

Simultaneously, drug price increases are overstated by reliance on Average Sales Price data from CMS uncorrected for 340B discounts. The 340B program mandates discounts of 30 to 50% to hospitals meeting certain criteria related to providing services to low income population, and now includes approximately 1/3 of U.S. hospitals. The program has grown at double digit per annum rates over the time period covered by the Goldstein analysis. Genentech recently disclosed that 17% of its worldwide drug sales were conducted under 340B contracts, which would be approximately 35% of U.S. sales.

Correcting for the effects of the 340B program using publicly available data on the size of the program and hospital/clinic drug expenditures, one finds that the great majority of the price increases discussed in the paper fall below the medical inflation rate described by the BLS Medical Services CPI, and most fall below the general inflation rate as well.

Overall, there is inarguably a pressing need to control healthcare costs in the U.S. To achieve this we need to look at all sectors of spending, and should expect significant sacrifices across the board. Drug prices will be one part of that, but unnecessary and low value care provision, medical salaries, and insurer overhead will surely play an important role as well.