Notable Takeaways From SCOPE 2022

Clinical data strategy and analytics, wearables and digital biomarkers, and risk-based quality management highlight notable takeaways.

SCOPE 2022 was packed with many topics ranging from clinical biomarkers, outsourcing, enrollment, and protocol development, to name a few. However, in particular, three topics stood out to me as relevant and breakthrough; these topics included clinical data strategy and analytics, wearables and digital biomarkers, and risk-based quality management. I will summarize notable takeaways from these topics in this article.

eClinical analytics and predictive modeling

There was much talk around eClinical analytics and predictive modeling. Jennifer Price, Executive Director of Data and Analytics at THREAD Research, showcased some of the first predictive models and visualizations ever introduced to the clinical trials industry using data from a decentralized clinical trial setting. For instance, analytics, such as study tasks and training, are used to predict patient dropout rates, and evaluating compliance with study procedures helps predict patients at risk of non-compliance. Knowing this information helps study staff identify and conduct early engagement activities to change the future course of the study positively. "Using predictive models to predict out study visits, compliance and completion are key to study success," said Price.

Artificial intelligence (AI), machine learning (ML), and natural language processing (NLP) approaches were featured by Xiaoying Wu, Vice President, Data Science Platforms & Privacy, Data Science at Janssen. Wu presented a case study on Janssen's COVID trial; Janssen developed a disease forecasting model that leveraged machine learning and real-world evidence data to forecast COVID hotspots with a 90% accuracy rate, which helped reduce the study's duration by 6-8 weeks. To do this, Janssen took a scalable approach by building a system that allowed access to data scientists and quantitative clinicians to conduct analyses and share the analyses across the organization. "It is important to monitor and analyze the model's performance and feed that back into the machine learning models to continue to improve the predictions," said Wu.

Digital and wearables update

Digital endpoints was also a topic at SCOPE. While there was much discussion on the subject, a presentation by Carrie Northcott, Director and Project Lead at Pfizer, was interesting, as she discussed studies that compared the validation and accuracy of novel digital endpoints against traditional endpoints at the site. By using digital endpoints, researchers can obtain a better picture of patient outcomes through high-frequency data collection that provides more comprehensive information about patients. "Going for a study visit, you may not get the full picture; you're getting points in time evaluation, so by using digital health technologies, we can now get a real picture of efficacy and safety of our drugs, and we get quantitative measurements," said Northcott. To elaborate, Pfizer conducted a study that evaluated patient gait in the clinic on a gait mat versus lumbar sensors the patients took with them at home, and the study's endpoint was to assess gait speeds between a variety of age ranges. Pfizer discovered a clinic bias that occurred; in the clinic, older and younger patients tended to have similar gait speeds; however, when patients were sent home with the lumbar sensor, the older group exhibited lower gate speeds compared to the younger group. Essentially, older patients walked faster in the clinic setting than in their everyday environments.

Updates on clinical quality

Several experts, including FDA, discussed the ICH-E8-R1 draft and updated views on clinical trial quality. The definition of quality seems to be moving away from perfection and more towards a quality culture within the organization. This stems from both profound experiences with quality and the intent to design reproducible processes and results and starts with collecting data and drilling down to identifying root causes. "There needs to be an open dialogue on quality issues (within the organization); how can you learn if you shut off (communications on quality events and resolution) and keep them in the quality management group?" said Andy Lawton, Director and Consultant at Risk Based Approach, Ltd.

David Burrow, Director Office of Scientific Investigations at FDA CDER, also discussed quality expectations. Burrow discussed that clinical quality involves four facets. The first includes the scientific question; whether high-quality evidence will inform decision-making through the use of a preventative, diagnostic or therapeutic intervention. Second, whether the trial design is adequate to answer the scientific question. Third, whether the data produced are sufficiently accurate and reliable. Last, ensuring that patients' safety, rights, and welfare are protected. Burrow said, "quality in clinical research is how we ensure the data supporting drug approvals are reliable and interpretable in support of the indication in the labeling." Burrow also discussed several critical risk assessment and management methodologies to form an ecosystem, including risk identification, risk impact assessment, risk prioritization and analysis, and risk mitigation planning, implementation, and monitoring. Burrow described this process as an ecosystem that goes beyond a trial and applies to the organization. "We're talking about the ecosystem as a whole, organization as a whole, the structure, the technology, the data systems, the portfolio of products that you have across the organization and all of the other vendors and contractors that are involved… but what's fundamental to clinical trial quality is people; the people who are critical in making sure that the trials work," said Burrow.

Summary

In summary, we are seeing tremendous advances in data analytics and predictive modeling in clinical trials, especially with decentralized clinical trial technologies, new studies are proving that digital biomarkers can be superior to traditional data collection methods and will likely become an expectation for collecting data in many trials, and the risk and quality management perspectives continue to shift, with them moving out of a singular function within the clinical quality department to a holistic, cultural and organizational level.