"It’s not FAIR": the data challenge facing pharmaceutical R&D

Data has never been as important in pharmaceutical research as it is today - but is big pharma really getting the most value from its data? 

In a world using increasingly digital methods for drug discovery, it has never been so important to leverage the data collected across pre-clinical and clinical research to maximise potential insights.  

There is a way that pharmaceutical companies can improve the reusability of this data, facilitating further scientific discoveries from their pre-existing data pool. From identifying new patient populations to discovering new uses for drugs already on the market, the value hidden in the vast quantities of data is waiting to be unlocked. 

Adopting something we call the FAIR set of guiding principles for data management and stewardship can improve data reusability. This is a mammoth task and will require not only frameworks and tools, but a cultural shift in attitudes where we begin to see data’s longer term reuse potential, not just its short-term use. 

What is FAIR data? 

FAIR data is Findable, Accessible, Interoperable and Reusable. These qualities can be used as a set of driving principles to define how we store and describe data to ultimately increase its reusability. By ensuring it is easy to find and access, and using a consistent approach to make it easy to combine with other datasets, we can unlock more value from our data. 

Through Oaklin’s work at a FTSE 100 global pharmaceutical company, we have seen the challenge of making data FAIR first-hand and understand the importance of creating a truly FAIR data ecosystem to maximise the return from research investment. 

How do we create a FAIR ecosystem?

There are two main focus areas in creating a FAIR ecosystem. 

The first is forward looking: ensuring the processes, education, training and standards are in place to guarantee new data is FAIR when generated. This includes storage and cataloguing practices, and application of consistent standards. 

The second is retrospective: applying these same standards to pre-existing data sets to elevate the ever-growing data pool, maximising its value. For large pharmaceutical companies with decades-worth of data, the key is prioritising data with the most potential for impact on their drug development pipeline.  

This may be less significant for smaller companies and startups. Provided processes are set up with robust FAIR practices, they will not face the arduous task of retroactively applying standards. 

Both considerations are underpinned by the need for agreed vocabularies and standards. This means creating terms with agreed definitions to be used for data and metadata so they can be mapped between different data sets. For example, understanding equivalences where different terms have been used to describe a patient’s home country (e.g. ‘UK’ or ‘United Kingdom of Great Britain and Northern Ireland’ or ‘United Kingdom’). This may seem simple, but it becomes increasingly complex when establishing standards for highly technical terminology.

How big is the challenge?

This is no small feat. It needs coordination from those generating and using data to guarantee a consistent approach. It relies on implementation of detailed processes, which may require changes to technology and the creation of tools to facilitate.  

With the plethora of different data collection techniques and systems, these need to be suitably simple for researchers to make data FAIR as it is generated, no matter the research method being employed. 

The work doesn’t stop there 

Nurturing a culture of FAIR data generation will require significant shifts in researcher attitudes. Investigators are now no longer collecting data only for their research but must also be convinced to make their data useful to others. For this, you need significant buy-in to the benefits of FAIR data and its potential to revolutionise the drug discovery process. 

The effort is worth it when considering it has the potential to reduce the need for clinical trials, the average cost of which is $33.8 million per drug (1). One of the largest factors driving cost is the patient numbers required to establish clinical effects. Cross trial analysis can provide an essential tool for demonstrating responses to treatment which may otherwise be overlooked. 

Consider a drug approved to treat a rare disease. A small patient population poses a challenge to pharmaceutical companies: how do you provide value to patients and recoup costs of drug development? Within siloed trial data, a marginal benefit can also be seen for diabetics, but the individual data sets cannot provide sufficient evidence for this.  

A FAIR ecosystem would allow data to be combined from multiple trials to analyse a larger population of diabetics. Providing the statistical power required to demonstrate the drug’s effect for diabetics unlocks a whole new, much larger patient population, without the requirement for additional, expensive, time-consuming trials.  

This is just one example of how adopting FAIR principles can maximise return on pharmaceutical R&D investment and accelerate the time to market for treatments. With the industry-wide shift away from expensive laboratory research towards data-driven insights, and with AI developments accelerating, FAIR data adoption has become pivotal for driving maximum value from a company’s portfolio. 

FAIR data beyond pharma 

At Oaklin we have seen first-hand how the pharmaceutical industry are leading the way in understanding the challenge of ‘FAIR’, but this lesson can be learned by any organisation wrestling with data. From customer and employee data to product performance metrics, every organisation can benefit from ensuring one of its most valuable assets can be reused. 

To learn more about this, please reach out to Dominic Hurndall or Miriam Grant.