After our annual PharmaTec Series events, consisting of three congresses on Pharma IT and Data, A.I (Artificial Intelligence) In Drug Development and SmartLabs Laboratory Informatics, we met with an advisory group of key industry leaders to discuss present trends and critical issues currently affecting the industry. The transition to digital workflows in the laboratory and the use of advanced computing tools in medicine and manufacturing has increased exponentially over the last decade and there are several exciting industry trends currently developing.
Integrating Real-World Data in Research and Design
Randomised controlled trials (RCTs) has long been the standard process for generating clinical data. Over the past few years there has been a shift of interest towards also including Real-World data (RWD) and multi-source data integration. Dr Richardus Vonk highlights RWD as an area that “is becoming increasingly important. If you look at the development area, people are talking about virtual trials, or intelligent clinical trials, including Real-World Data into submission packages.” RWD in medicine is data sourced from avenues associated with outcomes in a heterogeneous patient population in real-world settings. This includes patient surveys, clinical trials, and observational cohort studies and increasingly, connected devices and the Internet of Things (IoT). Smart watches connected medical devices and even our phones can generate huge volumes of data that can be analysed and interpreted. As an example of this the latest apple watch can track heart rate, blood oxygen and irregular cardiac events, information which can be analysed anonymously and in a non-invasive manner.
The huge volumes of data gathered has caused some discussion over how to ensure high standards of data privacy. While the healthcare industry has had a long relationship, a lot of RWD is sourced from tech and marketing companies, which can make it more difficult to ensure data privacy. These issues are likely to lead to increased legislation that is stricter than GDPR and CCPA. The proliferation of propriety devices and multiple apps is causing many in the Pharma industry to take a step back and consider a more standardized, centralized method of collaborative data collection.
Despite concerns over privacy, RWD is becoming more popular due to its ability to generate actionable Real-World Evidence (RWE). RWE is defined by FDA (Food and Drug Administration) as “clinical evidence regarding the usage and potential benefits or risks of a medical product derived from analysis of RWD.”
As emphasized by our experts, RWD has multiple usages. Pre-launch, RWD can be used to create evidence of effect variation in specific subpopulations, which in turn can be used to target clinical trials and make the path to approval more efficient. RWD can also be used to identify unmet clinical needs, improve patient centricity, and collect adherence data. In the post-launch phase, RWD can provide evidence of long-term drug safety and is increasingly being utilized to satisfy regulatory requirements agendas. Additionally, data gathered from real world settings are useful in building a case for new forms of delivery devices as well as drug reformulation and repositioning.
FAIR Principles: Well Managed Data Provides a Competitive Advantage
Pharmaceutical companies are amassing larger and larger volumes of data, and it is becoming increasingly difficult to manage. Data is often gathered from numerous sources and devices and stored separately, making it difficult for collaborative teams to share data and integrate findings. Data is the driving force of digitalisation, but poorly managed data can hinder performance and hamper innovation.
FAIR Data Principles were created to provide guidance on how to efficiently manage data. Many argue that these principles have the power to transform the industry and accelerate faster drug discovery and development, but the pharma industry has been slow to adopt it. Implementing FAIR Principles can be a daunting task; it is rarely a quick and straightforward process, as it requires a shift in how data created and used within an organization. It is, however, almost always worthwhile.
Engaging with Fair Data Principles opens the door to multidimension analysis, gives access to data analytics to more people, and allows additional automation and connectivity, improving productivity. Dr Richardus Vonk also mentions that the public are increasing requesting for better “availability of data around submissions and approvals” and highlights fair rate as a route to meet this demand. New “Fairifcation” tools are appearing on the market with several solution providers marketing streamlined processes and guidance for applying FAIR principles or evaluating existing data. As more companies understand the competitive advantage that comes with professionally managed data, this market is highly likely to expand in the coming years.
Machine Learning: Predicting the Course of Cancer
The third market trend was identified by Dr Colin Campbell (University of Bristol). He highlighted machine learning as an area of interest, particularly for its usage in prediction of cancer drivers as well as “predicting clinical outcomes or finding subtypes of disease.” The importance of understanding the likely course of disease in cancer patients many research teams to investigate machine learning as a tool for more accurate prognosis and prediction. Identifying whether patients are low or high risk and allocated appropriate resources and treatments to them is one of the most crucial steps in cancer care. Because cancer is a heterogeneous disease with many subtypes this can be exceedingly difficult with our current tools and knowledge.
Machine Learning is a form of AI that uses numerous repeated procedures to complete tasks, learning and improving after every attempt. Machine Learning tools can analyse complex data sets far more quickly than traditional programming or human assessment. The average time for a trained pathologist to perform a biopsy is ten days, while computers utilizing machine learning can do hundreds per second. Predicting the likelihood of recurrence of cancer is even more challenging; fortunately, machine learning can help here too with recent studies showing extremely elevated levels of accuracy in reoccurrence prediction.
While there is an impressive breadth of evidence emerging showing the benefits of machine learnings, more research needs to be done for wider adoption in everyday clinical practice. There are multiple AI/ML models currently in testing and experimental phases and as more results are published the likelihood of machine learning becoming an integral part of cancer prognosis and prediction becomes ever more likely.
This article provided a brief overview of three trends identified by our experts. The advancements in data utilization are opening new doors; helping pharma companies identify new markets, lower costs, and improve workflows R&D processes. If the topics discussed here interest you, consider taking a look at our upcoming PharmaTec events here.