Clinical Researcher—August 2019 (Volume 33, Issue 7)

DATA-TECH CONNECT

David Bennett

 

The life sciences industry currently finds itself facing a perfect storm of challenges—from society’s rising concern over health costs to changes in the physician’s role. Artificial intelligence (AI) and machine learning are being rapidly adopted to transform existing business processes and unlock additional value and insights, but the required data science talent is in desperately short supply. The next generation of accessible machine learning platforms will be crucial in helping departments working from the research and development (R&D) stages through to product commercialization to unlock the full value of their data.

Deloitte research from 2018 shows that productivity and R&D returns in biopharmaceutical companies have dropped to their lowest levels in nine years. The conundrum for these companies is where the R&D burden should fall, and they are continually evaluating whether to move efforts in-house, outsource to smaller companies, or involve academia in the process with a view to pursuing automation.

All this comes at a time of an emerging and shifting dynamic of rising payer—or formulary—power while physicians’ prescribing influence decreases and the cost of cutting-edge healthcare begins to exceed society’s willingness to pay.

Meanwhile, larger, more agile, and tech-focused companies such as Google and Amazon are sizing up the life sciences space with an eye to discontinuous disruption of the established order. These disruptors bring extensive financial clout and proven expertise with emerging technologies as a key enabler and differentiator, but technology also holds the key to the ability of “traditional” life sciences companies to fight back.

Machine Learning to the Rescue

How can the life sciences sector as a whole boost productivity, reduce the time to market, and unlock the full value of its data? The answer lies in the ability to successfully internalize and operationalize the promise of AI and machine learning, and to move it beyond the current ivory towers of data science.

Released this year, the 22nd Annual Global CEO Survey from PwC on healthcare and pharmaceutical trends revealed the stark contrast between data abundance and quality. C-level executives are hungry for data on brand and reputation, financial forecasts, and customer demands, but they simply do not have access to data that are fit for purpose or tools that are capable of deriving comprehensive business insights from the data that they do have. This is at a moment when the industry generates more data than ever before.

New developments in applied machine learning offer the opportunity to quickly explore data and identify complex patterns from vast datasets, including on patient health measurements, clinical trial feedback, and research outcomes.

Solving AI Pain Points for the Industry

Pharma businesses are already seeing return on investment from initial projects. In the United Kingdom, the Medicines Catapult 2019 State of the Discovery Nation report revealed that 90% of small and medium enterprises (SMEs) in the pharma industry required data science as part of their drug discovery operations, with half of these SMEs requiring AI and machine learning. However, there are still issues associated with AI in the life sciences industry.

Capabilities for data discovery are not clear and curation and preparation are still limited—all significantly lengthening the average project timeframe. There are also transparency considerations. Is the selected machine learning model reproducible across other datasets and business problems? Is the prediction accuracy visible, and can output easily be understood without ongoing reference to specialist data scientists?

Many of these pain points will be resolved by turning to platforms that automate significant amounts of the data preparation process, that are truly end-to-end and transparent in their operations, and that ensure the user is kept fully in the loop.

Humanized Machine Learning Empowers the Citizen Data Scientist

With talented data scientists in scarce supply, the skills gap is continuing to pose challenges to life sciences organizations. Existing data science departments do not have a wealth of data scientists, so their talents—and workloads—are reserved solely for the most business-critical and time-sensitive tasks, particularly in the R&D space. This means that other business units (e.g., medical, commercial) enjoying an equally vast although different wealth of data are unable to harness this expertise to generate insights and refine their operations with any velocity.

New applied machine learning technologies enable these life sciences organizations to bring machine learning and other advanced technology within the remit of employees of all skill levels, helping these problem owners become “citizen data scientists” in their own right. The ideal platforms for such technologies put the ability to prepare, manipulate, and visualize data for creating, managing, and optimizing deployable machine learning models within minutes into the hands of every employee, effectively coaching the user from data preparations right through to model deployment and management.

Such platforms are designed with accessibility in mind, eliminating the need for extensive training or a background in data science. A business or science problem owner can quickly harness the full power of advanced machine learning, intuitively augmenting his or her existing expertise and problem knowledge.

The bottlenecks of a limited data scientist talent pool are avoided, and projects can be completed quickly—without adding weeks or even months to the timeframe of a project that is waiting to be resourced.

Far-Reaching Applications Unlock Business Value Across the Enterprise

Beyond all the promises that have been made for AI in drug discovery, the real transformation in productivity in life science companies value chain will be wrought by augmenting the existing workforce with AI and moving beyond the realm of the specialist data scientist. Machine learning can be harnessed to find and enroll patients in the most suitable trials and facilitate the entire patient journey. Market access, sales, and marketing teams can make better decisions faster, their productivity while using scarce resources such as medical science liaisons can be transformed, and patient-centric, real-world evidence can be made truly useful.

Transforming Every Step of the Life Sciences Value Chain

While we already talk about the applications of AI and machine learning in life sciences, the next generation of cloud-based solutions is now poised to bring these advanced capabilities into the hands of every department and employee with a dataset and the desire to extract greater business insights and value.

These solutions can be easily deployed to rapidly tackle specific business problems, empowering pharmaceutical companies and other players in the life sciences sector to unlock the full value of their data.

David Bennett is Life Sciences Advisor at Mind Foundry in Oxford, United Kingdom.