Contributed Commentary Steve Young, CluePoints

October 6, 2023 | At the recent DIA 2023 Global Annual Conference, I focused my attention on discussions about the need for clear processes for reviewing audit trails, overseeing decentralized clinical trial (DCT) data, the impact of ICH E6(R3) on these areas, and the need for an industry-wide definition of Risk-Based Quality Management (RBQM).  

At an RBQM expert dinner hosted by CluePoints and the Tufts Center for the Study of Drug Development’s (CSDD), Ken Getz shared the results of a global industry-wide survey on RBQM adoption and attitudes. The survey was conducted by Ken and his team at the Tufts CSDD in collaboration with CluePoints and Price Waterhouse Cooper. 

Industry-Wide RBQM Survey Findings 

Over 200 respondents, representing 125 distinct pharmaceutical companies and contract research organizations (CROs), participated in the Tufts CSDD survey. 

Sponsors and CROs report using RBQM components in 55% of clinical trials, on average, in their portfolios. The highest level of RBQM use is in the Documentation and Resolution stage. The lowest is in the Clinical Trial execution stage. 

Organizations are starting to recognize that RBQM and quality by design represent a single methodology that starts with a good risk-based design of clinical trials and continues to apply risk-based thinking throughout the planning and execution phases of research. This recognition calls for greater RBQM alignment and communication between the design phase of the trial and the planning and execution phases, which however remains a challenge in many organizations due to functional silos. This challenge is reflected in the survey results, with the top perceived RBQM adoption challenge (69%) being a lack of cross-functional awareness.  

There is also work to increase uptake among biotech and small pharma companies. Overall, 63% of survey respondents trusted that RBQM will enable increased efficiency and cost savings, and 78% believed RBQM would improve the overall quality of clinical research. However, this dropped to 43% and 69% for organizations with an annual trial volume of 25 or less, which could be due to a lack of cross-functional awareness within an organization; illustrating and educating on the benefits of RBQM adoption in all areas of the clinical trials process could aid in improving this.  

An RBQM Definition 

Within the 55% of clinical trials using RBQM, there are different levels of adoption, and this comes back to the issue of a common definition. RBQM for one company might be as simple as reduced source data verification (SDV). For another, it might be adopting a risk-based approach to data management (e.g., ML-enabled medical coding) or medical review. 

First and foremost, RBQM is about critical thinking—anticipating what could go wrong and how to mitigate those risks. In terms of implementation, always keep it simple. Think about why you are doing it—to improve quality and be more targeted and efficient in running and managing clinical trials. If you stick with that approach, success follows. 

A critical RBQM element that people are still underestimating is data analysis. This is where RBQM, as a discipline, needs tools and technology to support it. If we look at the challenge of increased data volume, varied data sources, ePRO and eCOA, a single person, or even team, cannot review all that data. Some thing must do the heavy lifting. This is where advances in statistical methods, machine learning (ML) and deep learning (DL) are supporting people to make the right decisions. 

Manual review of listing after listing, graph after graph, is not realistic or scalable. RBQM today is to think, act, and act upon what these new tools show you. 

AI and Machine Learning 

AI is everywhere, although the term is applied broadly and often misused. However, companies in the clinical trial space are beginning to find ways to leverage ML and DL algorithms to improve clinical research. 

One piece of the RBQM process, which has always been very manual, is centralized statistical monitoring – identifying data anomalies or risk signals. Analysis of those signals is a very refined skill set. Using data from more than 1,000 clinical trials, we have developed a machine learning (ML) algorithm which identifies which signals are most likely to be an actual issue, what that issue is expected to be, and some mitigation strategies that could be used. 

It is not about the ML learning algorithm replacing the human role. It provides suggestions to help streamline the process and to help guide skilled staff with the benefit of data that has been reviewed and analyzed through the algorithm. 

Realistically, even with solid regulatory support the adoption of RBQM methodologies is still not as high as we would like. We must emphasize how ML and DL techniques can support RBQM, but only time will tell exactly where AI can make the biggest impact.  

ICH E6(R3) 

The final draft of ICH E6(R3) is welcome because it reinforces that RBQM is here to stay and is an end-to-end concept. It ties together RBQM and quality by design, which will help organizations understand that we are in a risk-based paradigm now. The new guidance also includes a new level of flexibility, inviting more innovative study designs, more DCTs, more use of wearables, and other data collection technology. 

There are some areas where clarification will be useful, for example, quality tolerance limits (QTLs). That is an area where we are working with industry colleagues to compile some feedback for the FDA.  

We would encourage everyone to read the final draft of ICH E6(R3) and provide their feedback. 

Key Takeaways 

We have come a long way in terms of the sophistication of RBQM solutions and the adoption of these processes. But there is still much to do. 

We must come to a common understanding and agreement on what RBQM is, which will facilitate a greater level of information and best-practice sharing and accelerate the path to achieve its promise of highest quality clinical research executed with greater efficiency and speed. 

We must reinforce the importance of RBQM methods from end to end across the network. 

We need to better convince small pharma and biotech companies that RBQM is not just for big pharma. 

Hopefully, we will soon reach a point where we do not need to talk about defining RBQM because it is just how we are running clinical trials. 

As Chief Scientific Officer for CluePoints, Steve Young oversees the research and development of advanced methods for data analytics, data surveillance, and risk management, along with guiding customers in RBQM methodology and best practices. Steve worked for three biopharmaceutical companies over 15 years, where he assumed leadership positions in clinical data management and led the successful enterprise roll-out of EDC at both J&J and Centocor. He spent an additional six years with eClinical solution providers Medidata and OmniComm, leading the development of analytics and risk-based quality management (RBQM) solutions and providing RBQM consultation to many organizations. Steve also led a pivotal RBM-related analysis in collaboration with TransCelerate and is currently leading RBQM best-practice initiatives for several industry RBM consortiums. Steve holds a Master’s degree in Mathematics from Villanova University. He can be reached at steve.young@CluePoints.com.