Overlooking the vast amount of available data that’s generated during downstream processing may be a significant hurdle to pharmaceutical businesses wishing to digitise their manufacturing platforms, suggests Katrin Wiederhold, Director of Life Sciences, Whitehat Analytics
For every 5000 compounds discovered in the laboratory, five are tested in humans and only one makes it to market. Moreover, it takes approximately 10 years and an average cost of $2–3 billion to develop each new drug.1
During this development process, a vast amount of molecular and clinical data are created and subsequently stored in proprietary networks.
Significant effort is currently being focused on data management in the drug discovery phase. However, the substantial impact that data science can have further downstream goes largely neglected, meaning that pharmaceutical businesses are missing out on opportunities to digitally transform their manufacturing and commercialisation operations.
As the sector continues to outsource various aspects of production to third-party suppliers, the need for better end-to-end visibility and auditability increases.
The transformational impact that data science can have on Big Pharma is increasingly being recognised, with significant efficiency gains now recognised as the core objective of digital strategies.
With rapid advances in research technology, developing new therapies has never been more complex … or costly. Research and development (R&D) spend in the global pharmaceutical industry continues to rise, valued at £188 billion in 2020 (an increase of £39 billion since 2015).2
The focus of large pharmaceutical organisations is on the clinical development of drug candidates, including costs such as manufacturing, clinical trials and the processes of commercialisation, production and marketing.
A study of the top 10 US-based drug makers by CSRxP and GlobalData found that, although drug companies spend identical revenue percentages on R&D and on operation and production (22% each), they also spend nearly as much (19% — or $47 billion) on marketing, advertising and promotion.3
The extensive use of data science within drug discovery at the clinical trial stage is well established or, is at least under intense investment. The real — and as yet untapped — opportunity for drug developers to implement an effective data science programme lies within manufacturing and commercialisation.
Bringing a new pharmaceutical product to market is a lengthy process that can encounter many bottlenecks, diversions and U-turns along the way. Trials regularly fail to meet their initial objectives, which can add significant delays and increase the costs of an already expensive investment.
The vast amount of data generated from the start and throughout the drug development process makes implementing a cohesive data science strategy at every stage the next logical step.
The application of cutting-edge data technologies, machine learning and continuous deployment infrastructures are starting to bring technology closer to matching the dynamic nature of the drug development process, helping to translate digital initiatives into long-promised operational performance improvements.
Many of the Big Pharma players have begun major initiatives to implement these technological advancements in the drug discovery and clinical phases, wherein they have proven to extract hidden insights from collected data.
The drug commercialisation process is one of the most challenging steps in drug development and the most successful strategies are founded in data from multiple sources. Drug marketing and promotion is another area of significant opportunity for data science.
With increasing competition from generics, Big Pharma is getting smarter about analysing and driving effectiveness in its sales and marketing operations — which is evidenced by the high spend.
Layering data by combining and analysing information from social media, demographics, electronic medical records and other sources can help to uncover new, niche and under-served markets.
Pharmaceutical manufacturers are increasingly shifting from blockbuster drugs to shorter lifecycle biological and genomic medicines, which must be delivered according to tightly regulated, validated quality processes within shorter time frames.
Data science can help to support quality control throughout pharma manufacturing in many ways, such as by providing automated and reproducible analytical pipelines that utilise predictive models to identify adverse events.
The power of in-memory computing technology and interconnected and automated systems allow large amounts of quality, environmental and Internet of Things (IoT)-generated data to be analysed.
Tapping into this big data allows pharma companies to build AI-driven end-to-end process controls, resulting in higher quality products, more predictability, more efficient manufacturing and faster times to market.
There is considerable pressure on pharmaceutical companies to find ways to either identify successful new products quicker or to fail faster. Integrating clinical data with other sources supports machine learning to speed up data analysis, leading to real-time insights into the clinical trials that facilitate critical decision making.
Data science can also help identify an appropriate patient cohort to participate in a clinical trial by analysing demographic and historical clinical data, monitoring patients remotely, reviewing previous clinical trial events and helping to identify potential side-effects.
Integrating patient data will help pharmaceutical companies to consider more factors, such as genetic information, to help companies identify niche patient populations using sources such as social media that, in turn, will accelerate and reduce the cost of trials.
Data science can also build data aggregation models across multiple clinical trials that look for patterns beyond the specific use case to create predictable foresights. Without these predictive models, performing cross-trial analysis is a daunting and costly task that requires programmers and developers to integrate data manually, which is a time-consuming endeavour prone to human error.
AI and machine learning methods consistently deliver more accurate outcomes in less time than conventional assessments. In the short-term, this translates into a competitive advantage and better sales results; and, with time, these automated solutions deliver extended value by continuously monitoring trends and optimising results as new data is generated.
There is huge potential to make better use of the vast amount of data across a wide range of areas affecting drug commercialisation, including
It is clear that the diverse fields of life sciences, computer science and data science are converging. Within the next 20 years, being driven by data will be the norm, particularly in the pharma space where efficiency gains are significant.
The future of pharma organisations will depend largely on their ability to bring data to the forefront of business decision making. Research from McKinsey suggests that companies who are digitising their supply chains can expect to boost annual growth by 3.2%.4
Although technical and cultural challenges undoubtedly lie ahead, advanced data analytics represent the biggest growth opportunity of the 21st century.