White Paper

Pharma's Digital Awakening: Research-Ready Health Information And AI To Reduce Cost And Deliver Better Treatments

Source: Ciox Life Sciences

By Ciox Real World Data Leaders

Block Chain, AI, Internet Of Things: Future Of PharmaTech?

Digital innovations and their emerging technologies, such as artificial intelligence (AI), advanced analytics, and cloud-based computing, are transforming industries and markets across the world as they offer novel ways to boost R&D, increase product quality and safety, and ultimately improve customer satisfaction. For example, the digitization of the automotive industry now allows manufacturers to decentralize production and distribute their supply chain, leading to partner systems’ integration via cloud connectivity.1 Giving all stakeholders the ability to analyze the same data increases productivity and lowers costs through fluid design, manufacturing, and delivery, while being increasingly closer to the consumer and their personalized requirements. This “digital awakening” offers consumer-centric products and service innovations that stand to transform not just the products that are built, but how and for whom they are built.2

The pharmaceutical industry is striving for a similar type of revolution, where drug companies determine the best course of treatment for a disease based on a unique patient’s physiology and their ability to metabolize targeted drugs. This customized approach to healthcare leading to personalized medicine engages with a precise patient population, rather than a large, undifferentiated market of patients. Using real-world data to represent the evolving demands and requirements of targeted markets—as demonstrated in the automotive industry—would create drug manufacturing pipelines that are living ecosystems, ushering flexible production into the pharmaceutical industry. Specifically, an agile, digitally driven supply chain uses AI to sift through the ocean of information to support the engagement of target patients, identify early efficacy indicators and potential safety issues, leading to adaptive design, manufacturing, and distribution. This ability to go from real world data to real world evidence could dramatically reduce investments overheads, shortening time to production and increasing positive patient outcomes by creating a holistic improvement in their health.

According to a recent article from McKinsey3, “In increasingly cost-constrained global healthcare markets, pharmacos that leverage analytics for advanced data-driven decision making over the next one to three years will gain a decisive advantage over their peers.” Vertically-integrated pharma supply chains are often challenged in understanding and adapting to the complex social, economical, and behavioral habits of large patient cohorts. Therefore, the key to success is making sure the millions of real-world data points acquired from the continuum of care are assembled in an accurate representation of patient characteristics and needs; otherwise, it is not fit for purpose. Reliable data opens up a digital revolution at the clinical trial level, as researchers form hypotheses, identify patterns, and then iteratively validate assumptions in-silico to produce drugs tailored for specific patient outcomes.

This insight becomes even more valuable as the possibilities with personalized medicine grow in the evolving landscape of drug development and manufacturing. A transparent and linkable ecosystem of data from information providers striving to re-establish trust with the pharma community strengthens the network of partnerships necessary to drive a faster execution of the drug manufacturing process and ultimately save lives.

Reestablishing Pharma’s Data Trust

When you look at the businesses of automotive and drug manufacturing, the steps to product delivery are comparable: design, manufacturing, supply, distribution, and consumption. In the automotive industry, customer feedback, such as from incident reports, maintenance issues, or even social media, gives automobile manufacturers valuable insights into how their products perform once they leave the factory as well as how they influence their customers’ experiences. Their research teams can inject these signals in the simulation environments they have built to test adjusted hypotheses based on what did or did not meet customer needs. This is like the pharmaceutical industry, where patients’ experience across visits, tests, treatments and outcome are curated to demonstrate how a drug participated in treating a particular disease. These factors help paint an accurate picture for clinical researchers developing new drugs.

While the similarity between the two industries exist, there is often a lack of reliable system integration for the various data points collected about a patient’s medical history in the healthcare industry. This limits the level of information depth researchers have access to, oftentimes requiring sourcing from numerous information brokers using different standards and structure, in turn limiting the ability to correlate an outcome to a hypothesis. Should a pharma market a drug that has, for example, an unforeseen, significant side effect, it not only risks the longevity of a brand but, more importantly, the safety of the patient. In order to address this uncertainty, the industry and its regulators have gone to great lengths to create best practices and local, national, and global regulations intended to ensure only the safest and most effective drugs possible are supplied to patients. Yet, if products were developed based on a patient’s historical health context along with a collection of relevant “-omics” (socio-economics, metabolomics, genomics, proteomics, etc.), it would increase confidence about safety and efficacy, potentially reducing the need for sometimes over caution regulations and accelerating the delivery of a cure. Similar to how the automotive industry injects customer feedback to its design process, gathering early indicators from patients and those who interact with them would allow access to applicable data on a just-in-time basis.

Understandably, frustration has grown across the industry when it comes to gaining access to the “right” information. A considerable effort is required to access, digitize, clean, structure, and deliver the complex levels of data in healthcare, and, in the past, data providers have not always given clear expectations about the depth or quality of data they can or cannot provide. The result is a lack of trust across the community, due to expectations with data delivery not matching reality. Hence, researchers are extremely cautious when using third-party data that can influence their development process, as biased data could lead to faulty hypotheses and eventually ineffective drugs or—worse—public health risk.

This creates an environment that limits what we could achieve through a more cohesive bond of trust.

To renew the industry’s confidence in data, there must be a consistent understanding among partners of not just what data is needed but especially what data can be provided. In addition, a future where the power of AI unveils the future of healthcare requires the collection of data that is truly representative of the patient population, as well as a gateway to deliver it in a way that has never been available before.

Using AI To Collect Accurate And Cumulative Patient Data

Inherently, clinical trials are the mechanism by which hypotheses are put to the test, either prospectively or retrospectively. They require access to a vast amount of information, potentially over a significant period of time, and each has its own benefits and shortcomings. With retrospective clinical trials, you need a network of providers from which you can source data and then run numbers to confirm a hypothesis. This type of trial costs less; however, you only have access to what has been gathered during the time of treatment. Prospective clinical trials allow researchers to influence the type of information that is captured and the precision in the notes. Yet, designing, initiating, and managing a cohort of patients over time is expensive, as it might require a larger team to coordinate the patients and care teams. Additionally, the onus falls on caregivers to be information-gatherers in the agreed-upon structure, velocity, and density for each specific study, often taking away their ability to focus on the patient.

When applicable, hybrid clinical trials are a best-of-breed of these two worlds: first, they leverage the historical information available at sites where a statistically relevant target population is known to have been treated. Then, researchers can work with those sites to engage with patients, manage the cohort throughout the duration of the trial, coordinate with sites to access a continued flow of data points, and establish a protocol industry peers can systematically review and validate. This is where AI can play a key role in clinical trial management. Using the DataFit Platform™, a decentralized, virtual data set of patients and specimens that meet real-world data requirements for the study, along with AI’s advanced computation and regression models, genetic algorithms, and machine learning at scale could facilitate the overall governance of analyzing the data for any signals that relate to the protocol being tested. This data can then be aggregated, so the research team can make required adjustments to their hypotheses and more quickly navigate the care ecosystem to effectively reach a conclusion as quickly as possible. A digital strategy for clinical trials also limits the amount of variability from doctor to doctor, hospital to hospital, etc. by standardizing health information in terms of time horizon, breadth, depth, and structure, so records are consumable and executable across the network.

Data-driven, hypothesis-first approaches enabled by a digital strategy allow researchers to break down problems in such a way that smaller questions can start the journey toward validating an assumption. Researchers can then continuously grow a confirmed body of knowledge and iterate on a series of easily traceable adjustments to arrive at a conclusion that instills confidence in each stage of their exploration.

A Call To Action For Pharma

Personalized medicine holds the promise to advance patient treatment and elevate the standard of care but only with proper insight into the mechanisms causing or contributing to a disease. Advanced analytics achieved through the implementation of AI with machine learning and large-scale information and signals processing, at all levels of drug production, could hold the advantage the pharma industry needs to achieve this. The ability it offers to engage in safe, ethical design protocols that provide early indications of efficacy and safety may be the key factor in driving the industry to focus on individual needs, rather than those of an entire population. It creates a shortcut of years of costly post-production studies, helps recoup the high costs of drug development, and generates high-profit margins.

Doing so requires a trusted information provider to acquire and prepare the data without insertion of bias and offers a line of traceability that effectively demonstrates the accuracy and sustainability of the data delivery. By establishing and maintaining trust with information consumers, they can operate an information gateway that is logical, as it is based on what is being seen in the healthcare ecosystem for a targeted patient cohort. It is important, however, that pharma companies review this relationship by running parameters or scenarios of known outcomes, so the ecosystem can be tested on specific questions to see what process is used to arrive at the expected answers.

Restored trust across the health information supply chain could lead to a tremendous collaboration between all actors. It would also further establish bridges across interconnected research communities and eventually position patients at the center of the health information exchanging, rewarding them for proactively taking part in this complex research effort. In the end, designing and developing drugs with a more accurate understanding of not just what the patient needs but what medication is most effective opens the pharmaceutical industry up to new opportunities that could lower the cost of drug development, transform patient care, and save lives, ultimately reshape the future of medicine.

  1. World Economic Forum, Digital Transformation Of Industries: Automotive Industry — https://www.accenture.com/t20170116T084448__w__/us-en/_acnmedia/Accenture/Conversion-Assets/WEF/PDF/Accenture-Automotive-Industry.pdf
  2. Capgemini, The Automotive Industry’s Digital Awakening — https://www.capgemini.com/2018/01/automotive-industrys-digital-awakening/
  3. McKinsey & Company, How Pharma Can Accelerate Business Impact From Advanced Analytics — https://www.mckinsey.com/industries/pharmaceuticals-and-medical-products/our-insights/how-pharma-can-accelerate-business-impact-from-advanced-analytics