ICER’s UPI Report: Hard to Be Reliable When So Many Numbers Are Wrong

Today, the Institute for Clinical and Economic Review published their third “Unsupported Price Increase” report, purporting to show the collective impact of price hikes on medicine that have no basis in new evidence (as ICER defines it).

ICER often finds itself in opposition to industry, largely because of philosophical or economic disputes related to how the group does its math. And, certainly, some of those disputes – such as ICER’s tendency to dismiss evidence – also drive the deep antipathy toward the Unsupported Price Increases work.

But the UPI effort is a uniquely flawed effort for reasons that go beyond the judgment calls that ICER makes: the underlying numbers are simply not reliable.

The report is built on a noble foundation: ICER tries to use net prices (the prices that drugmakers actually receive) rather than list prices (which bear little relation to what is actually paid for most medicine). That, in theory, should give a more accurate view of the world.

Unfortunately, net prices are also difficult to calculate. That makes it easy to make incorrect assumptions. Put more bluntly: is easy to go wrong.

And ICER went wrong. The UPI effort went wrong more than a dozen times.

That’s not my take from the pharma peanut gallery: the report itself acknowledges those errors. ICER looked at 250 medicines and determined that 34 deserved closer scrutiny, 32 because of 2020 list price increases above the group’s arbitrary cutoff of 6.11% and two because of public input.

Of those 34 drugs that ICER put under the microscope for high prices:

• Eleven medicines were excluded because – in the words of the report – “pricing information was deemed unreliable”
• Three companies (at least) had to correct ICER’s math. In two cases, ICER determined that net prices rose when they actually fell. Another company, after successfully lobbying ICER to correct net sales numbers, lamented that it “never received an explanation process that led to the original error.”
• Two drugs selected after “public input” couldn’t be assessed via the established methodology because ICER was “not able to estimate generalizable annual net price changes.” In one case, ICER said that the price paid by federal government payers had gone down. It didn’t matter that ICER couldn’t estimate net price changes or that there was evidence prices were falling: the report put both drugs in the “price increase” category.
• Five medicines saw their net price decline (that’s not an error, just an interesting side note). In two other cases, net prices remained unchanged.

If you’re counting, that’s 14 times ICER’s effort yielded unreliable or objectively wrong data on price or sales, seven times where the drugs being assessed actually had flat or falling net prices, plus two therapies where the pricing data were lousy but ICER declared there had been price increases anyway.

That’s not to say that price increases ought to be beyond scrutiny, or that working out net prices is a fool’s errand. It’s just a process so fraught with uncertainty and caveats that ICER’s goal – “providing the public and policymakers with information they can use to advance the public debate on drug price increases” – seems ill-advised, if not dangerous.?