From risk-based quality management’s (RBQM) inclusion in ICH E6 (R3) to the roll-out of numerous AI-driven solutions, 2023 has been a year of huge advance for clinical data management – and could mark a tipping point in the adoption of Research 2.0.
Here, Andy Cooper, CEO at CluePoints highlights the milestones and breakthroughs of the previous 12 months, and how they are shaping the future of our industry.
Driving RBQM adoption
May 2023 saw the publication of the final draft version of ICH E6 (R3). It tied the concepts of quality by design (QbD) and RBQM together to create a holistic, end-to-end clinical trial methodology.
The guidance was a big moment in the evolution of RBQM, which started life as the concept of risk-based monitoring (RBM), first recommended by the FDA in 2013. RBM focused primarily on applying risk-based approaches to site monitoring, shifting emphasis from resource-intensive, human error-prone 100% source data verification (SDV) to the centralized monitoring of the data most critical to trial quality. In the decade that has followed, it has become clear that the approaches’ quality- and efficiency-boosting advantages can be applied to all areas of study operations, from setup to execution.
Despite this, there is still much work to be done in terms of adoption. In June, the Tufts Center for the Study of Drug Development (CSDD), in collaboration with CluePoints and PwC, shared a first look at data from an industry-wide survey on RBQM adoption and attitudes at the DIA Global Annual Meeting. Conducted in collaboration with PwC, the work revealed that that while 78% of respondents believed RBQM would increase the quality of their studies and 53% trusted it would reduce timelines, just 56% of trials used it to support planning and design, and 52% for trial execution1.
New trends in clinical data management
As ever, SCDM Live, held in San Diego, California in October, proved to be a highlight of the clinical data management calendar. Generative AI, a technique which can create new content based on its original input data and includes large language models (LLM) such as ChatGTP, was certainly a hot topic.
Over the last 12 to 18 months, pharmaceutical and biotech companies have been building LLMs to assist in tasks such as writing protocols and designing risk assessments. Such applications can shave months off study set up timelines, thus accelerating the overall development pathway.
In the future, it is thought that generative AI could also be useful in developing simulated control groups, or “synthetic arms.” This could significantly streamline study initiation, particularly in indications where a control group is hard to recruit or retain, such as rare diseases or those carrying imminent mortality risks.
SCDM Live 2023 also illustrated that challenges remain, and the technology has its limitations, but the sector has shown its appetite to overcome the obstacles to embrace the opportunities. Throughout the year, webinars on topics such as leveraging machine and deep learning (ML/DL) for clinical trials and harnessing the power of natural language processing (NLP), for example, have been extremely well attended, prompting much discussion.
Real-world experience of AI
The last 12 months have seen the industry move from “talking the talk” to “walking the walk” of AI adoption.
One example is AI-assisted medical coding. Previously, mapping adverse events and concomitant drugs recorded in case report forms to MedDRA or WHODrug dictionaries was a manual, time consuming task. Now, DL NLP models can guide researchers to the correct corresponding term in seconds, with up to 90% accuracy2.
AI is also being used to improve risk detection. Currently, RBQM analyses raise “risk signals” when they spot a potential issue within the data. They are used to monitor and track any resulting investigations and corrective action, meaning they contain realms of free text, entered by various users as they document their findings. Now, NLP algorithms can screen all this information and flag signals that either lack the required documentation, or for which the root cause selected by the user is unreliable. This allows study teams to prioritize signal review and ensure effective follow-up and documentation of findings.
The 2024 research 2.0 toolkit
The emergence of new techniques and technologies presents us with new opportunities to streamline data management. They have provided a new paradigm that includes replacing 100% SDV with centralized monitoring and data analytics, using RBQM to focus on critical to success data, and utilizing AI to drive efficiency and quality.
As set out in ICH E6 (R3) this year, embracing this holistic approach will not only protect study success and patient safety during the trial, but also result in a higher level of submission readiness once the last patient has completed the last visit.
Yet with 69% of respondents to the CluePoints, Tufts, and PwC survey citing a lack of cross-functional awareness as their top challenge1, there is a clear need for education. We already have the computer power to build this bright new future, as has been demonstrated during 2023. Now, we need the people power. Achieving Research 2.0 will rely on the industry’s ability to understand, accept, and manage change throughout 2024 and beyond.
References:
