Pressure to move toward a risk-based approach in clinical trials has intensified. In 2016, quality standards were changed to the International Council for Harmonisation of Technical Requirements for Pharmaceuticals for Human Use (ICH) Good Clinical Practice (GCP). Since that time, maintaining quality and safety standards using the traditional model — on-site monitoring with high source document verification (SDV) — has become more challenging and expensive. Regulators insist on making better use of advancements such as centralized monitoring and analytics to more efficiently maintain quality and safety standards.
Implementing data-driven, risk-based monitoring into clinical trials — an approach called risk-based quality management (RBQM) — involves more than updating some paperwork, however. It requires a fundamental shift in how sponsors design and manage clinical trials.
As with most major undertakings, adoption has been slow. Pharmaceutical companies that try to implement RBQM by retrofitting siloed systems may find themselves with an expensive, inefficient program as time goes on. A scalable solution that adapts as the company adapts is the approach regulators recommend.
Implementing RBQM effectively requires extensive planning around people, processes and technology. The change does come at a cost; however, cloud-based technology allows sponsors to operationalize costs for what they use instead of paying a large annual capital expense. More importantly, the long-term benefits — more efficient and safe clinical trials, with less risk of failure and more accurate data — are worth it.
RBQM at Scale
Regardless of a clinical trial's size, duration or amount of data generated, sponsors can implement RBQM. The European Medicines Agency, in a reflection paper on RBQM, recommends a scalable approach to cover the needs of academic researchers, pharmaceutical companies of varying sizes and CROs.¹ To do so requires stakeholder buy-in, thoughtful processes and smart technology, but it does not need to be elaborate.
"We heard comments such as 'my study is only three months' or 'this is a huge study and we have to do 100 percent SDV,'" said Rhonda Roberts, senior data scientist and biostatistician for Remarque Systems, a quality-management platform provider for clinical trials. "Regardless of the size of the trial and the therapeutic area, risk-based quality management can overlay on anything."
Further, RBQM is adaptive: The concept is to focus monitoring and activities on areas with the greatest potential to affect subject safety and data quality. Sponsors can tailor or scale RBQM to suit the risk assessment and study protocol.
Why Ad Hoc Systems Won't Work
Some pharmaceutical companies have tried to address RBQM by retrofitting their existing siloed system. An ad hoc process may suffice for a pilot program, but as a company becomes more sophisticated in its RBQM process, it will need to create a new system that scales as the company grows and as it gains experience in RBQM.
"A large majority of data is integrated manually, and visualizations are created by business intelligence tools," says Roberts. "Large pharma companies would need a data warehouse. The issue then becomes building the data warehouse, standardizing data feeds, and building and maintaining a large capital expense. This solution is not scalable. A cloud-based software application scales up and out as needs change."
People, Processes, Technology
Technology for risk-based monitoring has become more sophisticated to meet the needs of data-intensive, global studies. This sophistication, which includes the use of artificial intelligence (AI) and machine learning, is necessary to successfully implement RBQM throughout the course of a clinical trial.
To take full advantage of technology and meet the efficiency and risk-based monitoring goals regulators look for, they must align processes and people with that technology. Consider the following:
- People. Implementing any new technology, including risk-based monitoring, requires buy-in from the top down. Sponsors will want support from executives as well as clinical operations, data management, safety, IT, training, statistics and regulatory departments. All stakeholders must understand what will become clearly defined roles in a shared process.
"Under the current model, there are multiple departmental SOPs [standard operating procedures]," Roberts said. "Risk-based quality management asks that we all, an entire study team, own these processes."
- Processes. Implementation starts with defining the project scope and continues with reviewing existing procedures and plans and creating new plans and processes as needed. Sponsors should develop a plan to document and address any issues identified during the study, as well as a communication plan for the internal team and vendors. Sponsors may discover they need to hire central monitors or new vendors.
As the plan becomes final, many sponsors benefit from launching a pilot study to test the new system and technology on a small scale. This gives sponsors an opportunity to refine the process before using it in a large, complex study.
- Technology. RBQM handles diverse data from multiple systems and sources. A shift in technology starts with understanding who owns the data, where it is stored, who has authority to share that data and how it can be shared. It continues with determining how to transfer that data. Many risk-based monitoring vendors require data in a specific format. Do you have the personnel and technical resources to collect and transfer that data?
How AI and Machine Learning Help You Scale
The goal of RBQM is to monitor, prevent and mitigate risk throughout the course of a trial — from planning to regulatory submission. Technology that incorporates artificial intelligence has the power to collect data from multiple sources, identify issues as they arise and generate insights that help operational staff take appropriate steps to mitigate those risks. Machine learning searches that data for trends, patterns and anomalies that not only help mitigate risk in the existing trial but also predict future risk.
"With AI, we can identify a site that's not performing at the same rate as others," Roberts said. "We can also go down to the patient level and start predicting what groups of patients will perform a certain way and really leverage the power of having so much data."
With an advanced technology solution that can handle electronic data capture (EDC), interactive response technology (IRT) and clinical trial management system (CTMS) data all on one platform, sponsors can use advanced machine learning and data analytics to catch critical errors early and often. This ensures that sponsors have the most accurate data available to know whether their drug is safe and effective.
"When you follow RBQM, you hopefully reduce the number of protocol amendments, because protocol amendments are often the result of a misunderstanding of how to operationalize the protocol at the site level," Roberts said. "Monitoring risk from the outset also helps reduce the number of database changes and helps allocate resources in a specific way. It's not about taking away oversight, it's about allocating it differently so at the end of the trial there are no surprises."
The ICH GCP amendments have had, and will continue to have, a significant effect on drug development. To implement RBQM with as few hiccups as possible, sponsors of clinical trials must plan to develop the people, processes and technology that will support RBQM goals. A scalable approach paired with scalable technology gives emerging treatments the best shot at success.
Reflection paper on risk-based quality management in clinical trials. European Medicines Agency, August 4, 2011.