The promising signs and the worrying trends of modern R&D

What is happening to R&D productivity in pharma? It partly depends on how you measure it.

When AstraZeneca warded off a bid from Pfizer four years ago, its chief executive Pascal Soriot pinpointed 2018 as the year when the company would start to grow again. Until now the UK company has been weighed down by patent expiry: along with the US tax changes and other factors, this led to a fall in profits and revenue for 2017. Yet the company claims that its efforts to generate new revenues continue to prosper, with revenue from new products and R&D productivity steadily rising.

For evidence the company points to a recent research report, published in early February by Mene Pangalos, executive vice president of AstraZeneca’s Innovative Medicines and Early Development (IMED) Biotech Unit. It found that the proportion of pipeline molecules at the unit that moved from preclinical investigation to completion of Phase III clinical trials had risen from four percent to 19 percent between 2012 and 2016. The report compares this with an industry average of 6 percent, as cited in CMR International’s 2016 Global R&D Performance Metrics Programme. For this success, AstraZeneca credits the introduction in 2010 of a 5R framework for R&D (right target, right patient, right tissue, right safety, right commercial potential).

In other words, AstraZeneca has become more rigorous about the molecules it investigates and quicker to pull the plug if research is misfiring. As a result, the number of molecules under investigation at the IMED unit has dropped sharply, but the number going into Phase III trials has stayed the same since 2010 (hence the rising proportion). This means the IMED is now a leaner but more efficient R&D unit, and that is important because this is only one of several AstraZeneca research arms. Overall the company spent 26 percent of its revenue on R&D in 2017, a high share compared with many of its peers, but managed to pare back the total by three percent compared with 2016.

End of the blockbuster

That, anyway, is the positive way of looking at the AstraZeneca data. The less positive way is to point out that the number of molecules emerging from the other end of this leaner process has not risen, and the likely returns from them are probably lower than in the blockbuster days of pharma. The company’s chances of replacing the billions in revenues lost due to patent expiry on products such as Crestor and Nexium are therefore not good.

This problem is not unique to AstraZeneca, of course. A recent study by Deloitte shows that returns on R&D investment at the world’s 12 biggest pharma companies dropped to 3.2 percent in 2017, from 10.1 percent in 2010. There was an uptick in returns in 2017, partly because the number of assets in late-stage development has dropped. However, Deloitte estimates that the cost of getting each of these assets to market is now nearing $2 billion, a sum that few of them will make back. Moreover, a disquieting share of the total returns (37 percent) is now coming from expensive cancer drugs – a trend that pharmaceutical payers are keen to clamp down on.

To be clear, this is no longer a problem with approval backlogs, as it was a decade ago. Thanks to the introduction of accelerated, breakthrough or priority review processes, approvals have now been buoyant, albeit uneven, for several years. At the US Food and Drug Administration (FDA), they soared again to 46 approvals in 2017, after a sharp dip to 22 in 2016. At the European Medicines Agency (EMA) there were 92 approvals in 2017, up from 81 in 2016 and far ahead of the 57 reported in 2012.

AstraZeneca’s 5R lessons are certainly part of an industry trend whereby many pharmaceutical companies are being more hard-headed about the drugs they push through the trial process. One indication of that is the sheer number of orphan drugs that are now being developed – 39 percent of the FDA approvals in 2017. This points directly to the tax breaks on research spending, the longer market exclusivity periods, the often higher prices and the comparative lack of competition that companies expect for such drugs.

Arguably, however, by being more rigorous about targeting research, AstraZeneca and its peers have narrowed down the chances of the kind of serendipitous discovery that spawned Viagra or Ventolin. Indeed a 2012 paper found that 24 percent of drugs on the market at that point were wholly or partly the result of serendipity. When it came to anti-cancer drugs, the share rose to 35 percent. Fortunately that gap is starting to be filled by artificial intelligence, as computers scrutinise those abandoned projects for the opportunities that were missed.

Ana Nicholls is chief healthcare analyst at the Economist Intelligence Unit