By Piotr Maślak, director, Operational Data Intelligence, member of Warsaw AI Centre of Excellence Governance, and Joanna Sasin-Kurowska associate director, clinical programmer II member of Warsaw AI Centre of Excellence Governance
astrazeneca_pole

 

Early days of digital transformation

Highly regulated industries, such as pharmaceutical R&D may seem to stay behind the technological developments we see every day, but the reality is that the industry has been at the tip of the spear for almost 50 years. The first digital revolution we experienced was the transition from paper to computer. While it seems natural today, it was quite the opposite in the 1980s. One of the first challenges with implementing systems like Electronic Data Capture or Trial Master File was to provide a personal computer to every hospital and clinic and creating first database systems, but the benefits, such as ability to implement data checks or centralise the data collection far outweighed the shortcomings. The internet revolution in the 1990s allowed remote data checking and other advances, which continue to this day.

Navigating complexity of regulatory environment

Advances in the digitalisation of the environment introduced many challenges, such as evolution of Clinical Data Management in response to the changes in the regulatory landscape with the introduction of Good Clinical Practice (GCP)[1] by the International Council for Harmonisation of Technical Requirements for Pharmaceuticals for Human Use (ICH)[2], along with country specific regulatory bodies such as Medicines and Healthcare products Regulatory Agency (MHRA)[3] in the UK. This creates a complex landscape to navigate when introducing innovative technologies, as the systems must be validated according to the regulations, and data must be collected in line with legitimate interest and informed consent of the patient.

Clinical Trials as a key to healthcare evolution

R&D in the pharmaceutical industry is inherently tied to clinical trials, an intricate process that involves healthy volunteers and patients. Every clinical trial design has the safety of patients in mind and every data point is collected to ensure that the benefits of the investigational product outweigh the risk for the patient. With the advancement of technology, we collect more data than ever before, following robust security protocols and ensuring patients’ privacy in line with regulations like ICH GCP. This introduces a holistic view of the patient population and development of targeted treatment, which has better efficacy and less risk related to the therapy. This enables medical professionals to tailor treatment to the patient’s needs, significantly improving outcomes.

Increasing volume of data

While the increasing volume of collected data enables more intricate research and introduces advanced therapies to global healthcare, it also comes with its own conundrums. Analysing the data, ensuring safety and wellbeing of patients becomes increasingly challenging, which requires pharmaceutical companies to adapt approach to data. Complex processes require appropriate data and analytics governance to ensure compliance with the regulations and scaling the data operations requires adaptive approaches, such as risk-based quality management and automated data checks.

The rise of data science and AI

Recent years have seen a surge in interest in data science and artificial intelligence developments, with experts heralding it as the fourth industrial revolution[4],[5]. While the full impact of AI on the economy is yet to be seen, pharmaceutical R&D has leveraged its potential in all stages of drug development. Artificial intelligence aids scientists in identifying promising compounds with a treatment potential; genomics and proteomics research is on the rise, but every company across all the industries can benefit from the use of advanced digital technologies. There are multiple initiatives in AstraZeneca that promote their use, such as cross-industry collaboration with Cambridge Centre for AI in Medicine and sponsoring the Warsaw AI Centre of Excellence, which brings together data science and AI professionals in Poland and promotes the use of innovative technology. The advancements are not limited to scientific research – AI was democratised by ChatGPT, advanced analytics are used in company operations, aiding in business development by identification of trends and creating statistical models that simulate future outcomes. Increased use in everyday activities has led to regulations that aim to increase trust in technology and improve the safety of its users, such as The EU Artificial Intelligence Act[6].

Data Science transforming business strategy

The use of data science methods can revolutionise the process of business strategy development in company operations. Intricate statistical models predict the number and geographical distribution of patients and clinics within the conducted trials, which ensure that the supply chain delivers the medicines to the patients without disruptions. Complex algorithms predict which experts will support medical professionals, equipping physicians with state-of-the-art technology that tailors treatment to patients and the devices required for the delivery of the trial and subsequent data collection. Predictive modelling prevents and identifies fraud and misconduct by partners and vendors to ensure that patients receive appropriate care. Advanced analytics are not only influencing business strategies but are foundational to business development and are at the very core of data-driven decision making.

Data science, machine learning and AI in patient care

There are multiple digital solutions that support experts in increasing patient safety, such as electronic Patient Reported Outcomes (ePRO), which allows patients to report their health outcomes digitally via their computer or even smartphone. Implementing this solution to clinical care improves communication between patient, caregiver, and physician by enabling identification of risk groups, unmet needs, tracking the treatment, adverse events, and disease course. With every technology there are challenges, such as users’ digital literacy and design of the tool itself that can prohibit patients from entering the required data. Experts address them by simplifying the design and limiting data collection to absolute minimum using machine-learning algorithms, while providing data which allows life and healthcare quality improvements. AI is augmenting the communication interfaces, enhancing patients’ satisfaction, and reducing digital barriers.

Synthetic data reducing patient burden

Limiting data collection to the absolute minimum requires complex approaches to protect sensitive medical information by pseudonymising or anonymising personal data wherever possible, careful design of access protocols, and restrictions which augment robust security measures. These processes allow the creation of synthetic data[7] – high quality, realistic and artificial data created without any real patient information. This approach can generate data for control arms in the clinical trials, which reduces the burden on patient populations. Synthetic control arms mitigate the ethical dilemma by providing modern treatments to all patients in the study by design.

Summary

While AI is at the tip of the spear of the digital revolution, technology is rapidly evolving, and we are yet to see the full potential, challenges, and regulatory response to the innovative solutions. AI introduces a promise to increase productivity and simplify our daily activities – we are already seeing that tools developed by experts and governed by seasoned professionals are incorporated into daily activities, becoming yet another asset in our digital toolbox – and an increasingly useful one.

[1] https://database.ich.org/sites/default/files/E6_R2_Addendum.pdf

[2] https://www.ich.org/

[3] https://www.gov.uk/government/organisations/medicines-and-healthcare-products-regulatory-agency

[4] https://courier.unesco.org/en/articles/fourth-revolution

[5] https://www.mckinsey.com/capabilities/operations/our-insights/adopting-ai-at-speed-and-scale-the-4ir-push-to-stay-competitive

[6] https://artificialintelligenceact.eu/high-level-summary/

[7] https://www.edps.europa.eu/press-publications/publications/techsonar/synthetic-data_en