In early 2020, I researched why medical innovation was falling. I meant to put this together into one grand policy paper, but I have decided to serialise some of my findings here. Perhaps at the end I will pull it all together.
The pharmaceutical sector is pretty complex, and somewhat different to other sectors. Most drugs are simple chemicals that it’s extremely cheap to produce and distribute, but difficult to come up with, and extraordinarily expensive to get approved.
You might use science to filter down a database of ideas, then test your favourite handful of thousands in a highly stylised scenario, like a petri dish. Those that perform best might be tested on a bunch of genetically modified clone mice, and then your absolute best performers might be tested on some humans, to see if they’re safe. Then you might test to see if any surviving drug actually works. Then finally you will try and prove it works, looking at thousands or tens of thousands of people over several years. You might fail to actually get any drugs through all these stages for years.
So it might be reassuring then to find that in pharmaceuticals, as elsewhere, we get what we pay for.
Right now we have a system where, roughly, firms are incentivised to develop ‘blockbuster’ drugs—drugs that lots of patients want, lots of insurance companies want to fund, and even government bureaucracies think are valuable to fund. The more valuable those drugs are to patients, the more the firms can charge for each unit, and the more units they can ship. Of course, the system is a simplification—for example, the pharmaceutical R&D system is built on top of an infrastructure of government-funded basic and applied research. And not all of the purchasing decisions of ‘payers’ that fund healthcare are rational. But this is roughly the system we in the developed world have, and it more or less works.
Broadly, the numbers support this judgement—more money spent on drugs means more drugs developed. A RAND paper in 2015 found for one extra new drug to be developed, the world’s market for pharmaceuticals would need to be around $2.5 billion larger. This reflected an average cost (then) of around $1 billion to bring a new drug to market, and then costs of production and distribution of around 50 percent of revenues. Other research on market size finds broadly similar results.
There are a lot of kinks in the system. Firms, for example, aren’t incentivised to develop drugs that will be the most useful to society over the long term, just the ones that will sell the most during a short window of exclusivity—perhaps as little as five years. And Big Pharma has little incentive to find alternative uses for their products once exclusivity has been lost—or for traditional or natural products where no one has ever held any exclusivity. Further, various reforms that have been introduced to reduce costs have themselves distorted incentives and produced unforeseen and unintended negative consequences—especially the introduction of surrogates as trial endpoints for approval.
This stuff mostly matters if it happens in America. In the USA, not only is far more spent on healthcare in total compared to anywhere else, but the unit price of most new drugs is more expensive. This means that despite the fact that US consumers buy a lot of cheap generic drugs, the US market is of outsized importance, even relative to its large share of the world economy. For example, in 2018 GlaxoSmithKline, the UK’s biggest drugmaker, earned 39 percent of its revenues in the USA. The USA has about four percent of the world’s population and under a quarter of its GDP. Europe as a whole was worth 26 percent to GSK. And GSK is representative of the market at large: overall, the US buys about 46 percent of the OECD’s in-patent drug revenues. Britain is about 2.5 percent of the world pharmaceutical market
These overall figures overstate the contribution of non-US sources to profits, since unit prices are higher in the US. Something like 77% of all pharmaceutical profits earned across OECD countries and 70% of world profits come from sales in the US. This in turn still understates the US contribution, since the contribution comes earlier on—dr
gs are released there more quickly. Calculated for net present value—and therefore the value of revenues accounting for time, risk, and uncertainty—the US contribution is even higher.
So you get the drugs you incentivise. And US policy is considerably more important than the policies of every other country in the world put together. Thus, we are getting the drug development that US policy tells pharmaceutical firms to deliver. Is it the drug development we want? I shall spend the rest of the series trying to answer that question.