How to use Big Data and AI to return pharma productivity to sustainable levels
In 2018, the mean projected return on new drug research and development (R&D) investments by a dozen large cap biopharma firms fell to 1.9 percent, from 10.1 percent in 2010, according to an ongoing analysis by Deloitte (1). That return is already well below the benchmark 10-year US Treasury bond, and on pace to plunge below zero by 2020, an unsustainable trend by any definition.
While declining projected sales contribute, R&D expense is the biggest factor. The mean cost of bringing a new asset to market exploded to $2.1 billion in 2018 from $1.1 billion in 2010 – with clinical trials, especially late trials, making up a large and growing share (1, 2).
The Mean Cost of bringing a New Asset to Market
$1.1 billonIn 2010
$2.1 billionIn 2018
Based on industry trends over the past decade, we project that an overall productivity increase of 20 percent to 25 percent is needed to restore R&D returns to sustainable levels by 2030. What can be done to achieve this?
Many industry experts see a need to transform the way clinical trials are conceived, designed and conducted. This transformation will rely heavily on harnessing the power of digital technologies. The current wave of emerging digital technologies offers real opportunity to significantly disrupt pharma business operating models and improve R&D productivity and ROI in a variety of ways, including automating processes, making efficient use of massive new data sets, and supporting early decision-making with increasingly powerful predictive analytics and statistical models.
This digital transformation is already underway and likely to accelerate, according to an ICON survey of more than 300 executives, managers and professionals in biopharma and medical device development firms. Nearly 80 percent of respondents said their firm plans to use, or is using, AI or Big Data approaches to improve R&D performance. Within five years, two-thirds of survey respondents said they will pilot or use these analytic technologies in select programmes, and another 20 percent plan to use them in all development programs.
Here we discuss the potential of Big Data and AI to improve R&D productivity and multiply ROI on R&D.
Unquestionably, Big Data is diverse in its sources and quality, and massive in its volume. As a result, it takes considerable effort to evaluate, normalise and structure it so that it can be reliably used for analysis. Below are some of the major data sources, with a brief analysis of their potential value in improving clinical R&D performance.
- Structured clinical data – These include data from current and past clinical trials, real-world evidence from registries and peer-reviewed studies. This data can improve clinical study efficiency in several ways, such as enabling go/no-go decisions and helping close out studies faster. This data can also be helpful in streamlining current trial protocols by predicting potentially high-performing study sites, further shortening study timelines.
- Traditional clinical data – These include data from clinical electronic health records (EHRs), as well as from labs, pharmacies and insurance claims. EHR data are increasingly valuable for guiding study design and identifying promising study sites. They are also proving powerful for identifying patients at high risk of developing chronic diseases, particularly when merged with genetic data.
- Wearables and sensors – These sources range from commercial devices, such as Fitbits and cell phone accelerometers, to medical-grade heart, blood pressure and glucose monitors. The volume and granularity of data from mobile devices can increase the statistical power of subject data, allowing shorter periods to establish efficacy.
- Genomic data sources – Genomic, proteomic and imaging studies provide detail on an unprecedented level that can be used for diagnosis, monitoring and therapy development. The potential for these technologies for improving the efficiency of new molecule development is difficult to overstate, and has the potential to dramatically increase approval rates, multiplying R&D efficiency.
If Big Data is the raw material of digital transformation, AI is the engine that sponsors rely on to make use of it. AI-powered capabilities, including pattern recognition and evolutionary modelling, are essential to gather, normalise, analyse and harness the growing masses of data that fuel modern therapy development.
But what, exactly, is AI? And how can it be developed and used to transform clinical trials? Here we outline different approaches to AI and the benefits of each:
- Expert systems – Some of the earliest and most widely used AI applications are expert systems that use rules-based algorithms to mimic specific human expertise. One example is decision-support trees for routine diagnostic tasks, such as differentiating between bacterial and viral respiratory infections for prescribing antibiotics, which are built into virtually every EHR drug-ordering module.
- Robotic process automation – Variations on this approach have significant value in improving clinical trial efficiency in the realm of robotic process automation, which simply means designing trial processes that allow machines to do anything that a machine can reliably do. Of itself , RPA has no ‘intelligence’ – however, increasingly it is typically integrated with other AI technologies to create faster automation, and it’s organisational impact is proving to be significant. In clinical trials the immediate efficiency benefits of robotic process automation include reduced manpower, reduced errors and delays and reduced data loss. Moreover, robotic process automation lays the groundwork for incorporating massive data sets from EHRs, mobile devices, automated image scanning, and individual patient genomic and molecular data.
- Linking trial stages – Another way to leverage robotic process automation is linking processes across study stages. This involves considering the final outputs – which are data supporting regulatory approval and commercial payment – in the design of every study step and automatically adjusting those steps when a change occurs. Automatically linking study requirements from end-to-end can significantly reduce the delays and manual effort required to fully implement a protocol amendment. Moreover, adopting a linked process automation approach facilitates portfolio management decisions by allowing developers to model how specific changes in study protocols might affect development timelines.
The power of AI for revolutionising clinical practice is already evident. Some of the near- and long-term applications for improving clinical research returns include patient identification, site selection, patient monitoring, and support and cohort composition.
Harnessing digital technology to transform clinical trials will require sponsors to develop or acquire a wide range of capabilities. Beyond that, it may fundamentally change the way sponsors are organised and integrate R&D into the overall enterprise. Critical steps for moving forward include:
- Identifying and developing operational and IT expertise and capacity
- Developing statistical expertise
- Developing global reach
- Managing change
Digital Disruption in Biopharma
For more information on harnessing these capabilities and others, read our whitepaper.