Skip to Main Content

As insulin becomes a poster child for the growing outcry over the cost of medicines, a new analysis suggests that Medicare could have saved $4.4 billion in 2017 for the diabetes treatment if the program negotiated the same prices as the Department of Veterans Affairs, which already bargains for discounts.

For more than two dozen different insulin products covered by Medicare Part D that year, spending was $7.8 billion, based on estimated rebates of 41% received from drug makers. However, if the federal health care program had negotiated as the VA does — and used the same formulary, or list, of covered medicines — spending would have dropped to nearly $3.4 billion.

Unlock this article by subscribing to STAT+ and enjoy your first 30 days free!

  • The numerical precision of this paper creates an false illusion of accuracy.

    The researchers used an average 41% rebate from ALL drugs to come up with the sizzling $4.4 billion headline figure. Most reporters dutifully reported this figures, which doesn’t reflect anything at all except an obviously false assumption.

    The authors then conducted a “sensitivity” analysis that used a 61% rebate estimate for insulin. The savings now drop to $1.1 billion. Everyone ignored this lower figure, including the paper’s sponsor (John Arnold). To Ed’s credit, he reported the lower figure. Kudos, Ed!

    In reality, neither figure reflects any actual data, because Part D rebates for insulin are unknown. The false precision of the $4.4 billion figure gets reported as “truth,” when there was no actual research behind this figure.

    This paper is no more than a hypothetical illustration. The authors also conveniently neglect to mention that nearly all Part D plans have already shifted to closed formularies with a single product for insulin. So, actual Part D plan rebates are likely to be pretty close to the 74% reported for the VA formulary, i.e., very limited savings.

    Just because someone’s suppositions are published in a peer-reviewed journal doesn’t make them accurate.

Comments are closed.