Optimizing the Digital Opportunity in Biopharmaceutical R&D

Article

If advanced technologies hold the key to at least some of the productivity issues R&D organizations need to overcome, Nicholas Lakin asks what are the best strategies and operating models for exploiting these aids to their fullest potential.

If advanced technologies hold the key to at least some of the productivity issues R&D organizations need to overcome, what are the best strategies and operating models for exploiting these aids to their fullest potential? Dr. Nicholas Lakin explores the use cases and distils some practical advice.

Dr. Nicholas Lakin

In 2017, only one in 20 biopharmaceutical CEOs attending the annual JP Morgan healthcare conference referred to ‘digital’ activities during their presentations. By contrast, these now feature on almost every leading player’s strategic agenda, while significant investments are being seen in purchases of digital technology startups, to money being put up for ambitious projects based on machine intelligence. GSK’s investment in Verily Health; spin-out innovation centers such as BI-X or Sanofi’s 39Bis; and partnership programs with AI startups, including GSK and Sanofi’s recently signed engagements with Exscientia, add to the momentum. A survey by the Pistoia Alliance, meanwhile, suggests that 44% of life science professionals are using or experimenting with AI and 94% expect an increase in the use of machine learning, in particular, within two years[1].

So what has changed? Digitization’s growing appeal to life sciences is directly related to its potential to address the well-known and extensively documented challenges around R&D productivity. The prospect of simultaneously tackling regulatory complexity, delivering much sought-after patient access and intimacy, and inspiring new product innovation are further reasons for the mounting interest. Vas Narasimhan, CEO of Novartis, has gone so far as to claim that the technologies it is now integrating into every aspect of its approach to R&D form the basis of its repositioning as a 'medicines and data science’ organization. At the top of the market, R&D firms do not lack ambition.

But there is work to do to get from here to there. Firstly, R&D organizations must be clear what they are aiming for: what they mean by digitization and how they might apply this to address current and emerging challenges. Secondly, companies must coordinate projects properly. It is not uncommon for R&D organizations to have 50 digital initiatives in flight, involving different technologies and with separate sponsorship - spending more than they need to, and developing overlapping capabilities.

Thirdly, companies need measurable business cases to drive timely decisions. R&D leaders and teams must help set the aims, driving digital initiatives as business-technology partnerships. This approach should be extended to external partnerships too, so that companies priorities ‘innovation’ when choosing service and/or technology providers. Finally, to distance practices from the manual, document-centric processes firms have relied on in the past, R&D organizations will need to bring effective business change management into play, to encourage the design of processes that capture, harness and are informed by rich data from an increasingly broad range of sources.

From vision to a coordinated plan for change

Making progress across all of these areas requires clear vision from R&D leaders to inspire change; better understanding of what the digital technology options are and how they can be applied; and a consistent value framework for benefit evaluation which drives strategic decisions.

It is the last two points which need particular attention now. Although organizations’ vision is developing, most R&D groups do not yet fully appreciate the types of change which digital technologies can enable and how to translate these into a practical roadmap supported by robust business cases.

There are three converging developments which are driving the digital opportunity. The first is something we define as the ‘Automation to Artificial Intelligence’ continuum. It starts with human tasks being accelerated by robots, moves through machines that learn from data and extends towards more intelligent strategic problem-solving using neural networks. These opportunities in turn are being fed by a proliferation of new sources of data – in particular non-traditional, real-world data (RWD) from patient communities, electronic health records, and connected devices or sensors. Combined, these possibilities are giving rise to new user experiences, for example enhanced personalization of products and services through the provision of direct and contextual information to patients and healthcare professionals.

With a greater understanding of all of this, R&D groups next need to draw up a cohesive roadmap and plan for digitally-enabled change, with tangible milestones and wins along the way. Knowing that, for instance, at least 50% of safety processing and signal detection effort could be automated, is leading to focused efforts in this area as an early example. This could free up skilled professionals to work on transforming patient insight and engagement models for drug safety using more complex AI. If similar efficiencies could be extrapolated from the same approach and technology reinvestments to other operational processes, the business case becomes more than compelling.

Benefits evaluation

Establishing a benefits framework is a valuable step in enabling companies to articulate and quantify benefits; to align investments with ROI; and to avoid fragmented, overlapping initiatives. Taking three major strategic priorities that will be common to most biopharmaceutical R&D companies, benefits can be expressed in line with the following clear categories:

Operational excellence

Improving productivity, a priority for all biopharmaceutical R&D organizations, starts with finding more efficient and repeatable ways of doing things. Digital process redesign should be treated as an overarching initiative, with common tools and approaches employed from the start. Establishing common and consistent assessment criteria can help with prioritizing what to automate. Robotic Process Automation (RPA) has huge potential for stable processes where accuracy or frequency are high, for instance, or where cross-department or system barriers need to be traversed. Machine Learning (ML) has greater potential to drive more flexible and data-driven processes, but the investment in algorithm development is higher. Establishing an Automation Architecture can enable RPA and ML strategies to co-exist as companies look to use RPA for rapid benefits and ML for more sustained and scalable data processing.

It’s also vital to consider process design and sourcing strategies as heavily connected with automation ambitions. Biopharmaceutical companies need to decide whether they should own and drive the technology for automation themselves or partner with BPO / technology vendors to provide this.

Business process management (BPM) tools are now well integrated with automation approaches and can help with standardization, to arrive at more consistent document management, data quality, master data and reference data to support AI algorithms and data-driven processes. One practical use case is in improving the control of decentralized data entry involving affiliate regulatory and safety operations, using guided workflows and pre-filled content. This would reduce the effort expended on centralized data correction in Regulatory Information Management (RIM) systems, and shorten preparation cycles. Similar efficiency gains can be extended to labelling and submission document assembly.

Product innovation

Expanded data insights offer the potential to drive product innovation - the core mandate of any R&D organization. (Ameet Nathwani, Chief Medical Officer at Sanofi, has described data as the new healthcare currency.)

Digital initiatives to date have been concerned largely with generating insight during earlier phases in the development process, through improved access to real-world data (RWD), mined using AI. The opportunity is to be able to profile a disease, candidate molecule or patient population using unprecedented volumes and combinations of data to build predictive models – to identify likely winners and losers in the pipeline much earlier in the development process. Although many of these investments are highly speculative, the returns are potentially high as they address key scientific challenges - from improving trial feasibility and recruitment; to improving drug safety and efficacy; to arriving at a better understanding of the value of the medication in the real world.  

The choice of which data and where to apply it will be important, and technology must be applied consistently. Between data feeds from social networks, call centers, spontaneous adverse event reports, and patient focus groups, there are plenty of interesting new RWD options. The challenge then becomes how to integrate these diverse sources to enable exploitation through AI.

Developing a real-world evidence ‘playbook’ will help highlight traditional and evolving data sources, questions and case studies, as a means to socialize and optimize new types of real-world evidence with product teams. Companies should perform continuous assessments of their sourcing approaches for new data types too. As companies build AI into their existing data lakes, being open to dual sourcing strategies presents a good way to accelerate access to innovation, and/or make this more viable financially. To access the best AI capabilities, while containing risk, companies are tending towards joint ventures and acquisitions to get up and running.

Customer intimacy

Customer intimacy from an R&D perspective relies on creating sustainable relationships during the R&D lifecycle – not just with patients, but also with investigators and other healthcare professionals (HCPs). Digital initiatives offer the opportunity to make these interactions more informed, reciprocal and transparent, boosting trust and engagement on all sides.

The shift towards digitization of clinical, medical and regulatory documentation has created the opportunity to increase patient understanding and manage compliance with clinical study protocols. Biopharmaceutical companies have started to invest in eConsent tools as they seek to improve the effectiveness and responsiveness of the informed consent process. Digitization can also make the process more intimate, user-friendly and reassuring to subjects – by employing interactive suggestions and tips, or creating more engaging multimedia representations of the study journey.

The rise of patient rights is another critical consideration for biopharmaceutical companies. Patient transparency regulations, such as EMA Policy 0070 and the EU General Data Protection Regulation (GDPR), are driving the application of digital technologies to manage data anonymization and access rights. Blockchain-based databases have substantial potential here, providing a distributed, decentralized data infrastructure which, in principle, is immutable and insists on trust being established between the data providers (patients, investigators and other HCPs) and data collectors (sponsors). R&D companies will need to consider patient access as part of future data management systems and processes, too.

More ambitiously, the rise of new digital engagement models raises the possibility of biopharmaceutical companies moving towards long-promised ‘beyond the pill’ business models, combining medicines with digital capabilities to offer sensor-enabled medications which enable improvements in dosing regimen and patient adherence.

The impetus for change

From improving R&D productivity and cutting costs, to improving pipeline and managing intensifying regulatory complexity, biopharmaceutical companies have much to contend with, so the impetus for change is considerable. To unleash the full potential of new sources of data, and exploit new insights, improve patient engagement and create digitally-enabled products, R&D strategists need a clear focus of where they want to be and a cross-functional approach to change that will maximize ROI. This will be their best chance to gain real competitive advantage through digitization.

Dr. Nicholas Lakin is a VP in Kinapse’s advisory practice in London, with experience across discovery, clinical and regulatory functions in biopharmaceutical R&D. www.kinapse.com/contact-us.

• This topic is explored further in the white paper, The Digitization of Biopharmaceutical R&D - Critical Considerations for Strategists & the Operating Model, available here https://kinapse.com/wp-content/uploads/2018/06/Kinapse-Digital-RD-Operating-Model.pdf.

 

[1] 44% of Life Science Professionals Already Using or Experimenting with AI and Deep Learning, Finds Survey from The Pistoia Alliance: http://www.pistoiaalliance.org/pa-ai-survey/

Related Videos
Related Content