Developing new pharmaceuticals is ever more expensive, with these costs reflected in drug prices. Plenty of viable drugs are not even profitable enough to justify development.
Part of the reason is the exorbitant costs created by the high percentage of failures at late stage clinical trials. The pharma industry has a huge opportunity to reduce these failures., and it involves being smarter with data.
Pharma spends fortunes trialling, recording and interrogating information about different permutations of molecules. The result is a portfolio of data which represents the opportunity to licence a drug.
That data is rarely exploited further. This is a shame, since it offers huge opportunities to direct new research – their own or others – and reduce time wasted pursuing dead ends.
Firstly, by spotting missed opportunities. Pfizer was close to abandoning a disappointing angina treatment, when an interesting side-effect was noticed. It became Viagra, the fastest selling drug ever. How many similar opportunities are missed? How many could be discovered by revisiting data?
Secondly, though personalised drugs. Currently drugs are tested against diverse populations, so medicines that are most effective in a sub-population can be missed. By matching historical trial data to new data – eg from genome sequencing, healthcare records, and lifestyle measurements from smart devices - we may find failed drugs are actually highly effective for certain groups – and worth reinvestigating.
One company found their drug was less effective in the US due to high aspirin usage there. This was a huge data analytics task, but resulted in a clear licensing case for several regions. There may be hundreds of trials that missed such anomalies and dropped perfectly good drugs.
Finally, data helps make research more efficient. Being able to review past trial data would allow pharma companies to cross-reference their research plans, avoiding repeating failures and directing research in the most promising directions. Our own conservative estimates suggest that finding 70% of the bad drugs before clinical trials, instead of 50%, would add at least $1bn annual profit to the global pharma industry.
Wherever possible, pharma companies should do more to share past clinical trial data with their own researchers to guide drug developments plans – a huge opportunity that is currently missed.
But we can go further, by finding secure ways to share data externally. Academics with a boarder view than pharma teams may spot clues in the data to treating completely different diseases.
In an ideal world there would be a safe space for sharing all past trial data. In reality, we live in world where companies must protect their IP. Nevertheless companies such as GSK, AstraZeneca and Sanofi are making promising moves here – for example agreeing to exchange chemical compounds libraries. There are plenty of opportunities for more agreements like these.
There are of course challenges. We need to convince data’s custodians to share it, create secure spaces to do so, and trust those doing the analysis. But it is possible to overcome these obstacles. And the benefits they will bring are huge: faster development cycles, new discoveries, more personalised drugs and cheaper end products. A win for pharma companies and patients alike.