Digitalising pharma R&D

Getting your R&D digitally enabled is not a matter of improving your competitive edge – it’s a question of life or death. This may sound like a hyperbole, so let’s get real. Here, I discuss some avenues to trigger your R&D digital transformation in order to stay relevant in the near future. 

Developing drugs is a tedious and costly journey, averaging 12-15 years and setting you back roughly 1-2 billion dollars. Leveraging digital and data to ease this pressure sounds like a walk in the park, especially when you read posts just like this every other day. Reality, however, will strike with the force of the tiger. Scepticism, cynicism, reluctance, disbelief, legacy systems, rusty processes and change rebels are just a few hurdles you will have to overcome in your R&D digital transformation. The few success stories share one point in common: resilience. They do this by pulling champions together to distil the essence and power of digital in their own business before building on the ‘quick wins’. Simple, right? Not quite. The million-dollar question is: where to start?

1. In silico discovery and optimisation of target identification

• In silico target identification is probably one of the first pharma use case of computational biology. In the world of computer science, this problem is what is known as NP-hard. In other words, it would take an infinite amount of time to find the optimal answer to how two molecules interact. Why? Because they can interact in countless ways – and we don’t know of all of them. So why bother down this path? Because this is a problem quantum computing will be fantastic at solving – or approximating. So while investing in AI and artificial neural networks to improve in silico discovery might never yield the expected results in the short run, developing the capabilities and underlying thought processes will support the transition to the quantum paradigm shift – and this will produce blockbusters. 

2. Using the past to predict the future – maximise the value of your data 

• The existence of a data desert in the public domain, between the time a compound enters a portfolio to the first steps of clinical development is well known, and is not likely to change in the years to come. However, there is a lot of internal data that companies can use to inform their strategies, portfolios and decisions. In this instance, the main challenge is around data interoperability. While this may be the least ‘sexy’ of the cases presented here, it may well be the one with the highest potential. The reason is that efficient data interoperability implies a clear data strategy, data governance and data management. Once these pillars are implemented, the use cases will be abundant. From independent project decision-making to portfolio management and risk optimisation, functions will work more closely together and will feel more empowered through increased data visibility and decision transparency.

3. Making protocols easy

• One of the main hurdles in clinical operations is the disconnect between clinical development plans and clinical trial protocols. In other words, between strategy and the operations. Trial protocols can be the magic glue to bridge this gap, assuming the authoring process ‘breaks the silos’ and becomes more data-driven. This means, for example, using smart tools to understand the impact of inclusion-exclusion criteria on the overall targetable patient population of your trial. Another example is to use such ‘intelligent’ algorithms to minimise protocol amendments, which cost between 100 and 500.000 dollars a pop, two thirds of which are thought to be avoidable. In other words, your trial design and clinical operations processes must become more data-driven, fluid and circular, with one feeding into the other – and vice versa.

4.  Maximising patient recruitment through avatars

• With 80% of trials failing to start or end on time, and 30% of trial time spent on it, patient recruitment is being scrutinised more than ever. Despite the increase in competition due to more trials ran simultaneously, there is room for improvement when it comes to in silico modelling. Defining your ideal trial avatar will inform you on the feasibility of the study and help you better plan the duration of recruitment, as well as the trial length. For example, uncovering indirect indicators to recruit patients with multiple sclerosis who will suffer a relapse in the next three months rather than the next year will be a considerable advantage if this is what your primary endpoint relies on.

5. Making patients partners

• Digital maturity and pervasiveness – at least in the high-income countries – provides an unprecedented opportunity to become more patient-centric. Few companies have had the courage to go down this route because they focus on the negative effects and the increased burden it entails. Patient reported outcomes (PROs) are a good example. The simple fact that PROs can become adverse events, and that adverse events are subjected to very tight legal processes and procedures, makes it tempting to take the ‘don’t ask don’t tell’ strategy – the easy way. An alternative strategy is to see PROs as a source of knowledge that can be fed back into the R&D process. In order to succeed in this approach, process automation and contextual language processing are key, especially given the fact that the more you ask, the more people will tell you.

Pharma R&D is at a crossroads. On the one hand, it needs to keep churning the wheel and advance molecules through the pipeline. On the other hand, society is becoming more digitally aware, and the repercussions are very real. Patients spend more time googling their symptoms than talking to doctors – and doctors find more answers online than in books. Reported adverse effects are growing exponentially, just like the costs of healthcare. And soon, there will be more patients than the systems can handle. The COVID19 dilemma will become a daily struggle. Digitisation – and digitalisation – are not something companies can debate on. It isn’t about building a competitive advantage or a core differentiator – it is about staying relevant in the very near future. To make it successful, it is less about where you start than about where you want to go. So make sure you have a vision, and that the pilots and initiatives you run tie back to it and can help it scale.