The average Phase III study generates a staggering 3.4 million data points1—more than any human could ever hope to analyze in a lifetime. The advent of artificial intelligence (AI) is presenting researchers with a multitude of ways to extract valuable insights from this goldmine of information.
During a recent webinar, Laura Trotta, Vice President of Research at CluePoints, explored some of the best ways to leverage AI to encourage high-quality clinical trials.
What are the Traditional Challenges in Clinical Trials?
Clinical trials generate huge volumes of data, often in excess of 3 million data points.1 Yet traditional processing and analysis methods, such as data listing reviews and 100% source data verification (SDV), are slow and burdensome, and their manual nature can leave them prone to human error. All this makes them a compounding factor of modern drug development challenges.
Why Use AI for Clinical Trials?
Traditional processing and analysis approaches are often inefficient, contributing to the huge amount of time and money it takes to bring a new product to market. Currently, the development of a new medicine takes more than a decade2 and costs anywhere from less than $1 billion to $2 billion,3 leaving much room for improvement. Previous analyses have shown that only 1.1% of eCRF data entries are corrected by SDV,4 while less than 1.7% of data entries are impacted by (automatic/manual) queries.5 This contrasts with the large expenses of those activities knowing that the estimated cost of a single query ranges from $28 to $225 and that Phase III studies typically generate tens of thousands of them.6 This calls for more effective methods to oversee data quality—gaps that AI can help fill.
How Is AI Used in the Clinical Space?
Recent years have seen the emergence of advanced statistical and machine learning methods with the ability to address several challenges.
Centralized monitoring (CM)
Centralized monitoring refers to the monitoring of all clinical trial data from participating sites in a centralized location. CM allows for the near-real-time monitoring of information as it accumulates, streamlining processes and enabling sponsors to identify and respond to potential issues before they have a chance to impact data quality.
Centralized statistical monitoring (CSM)
Centralized statistical monitoring uses statistical tests to analyze all clinical and operational data to identify anomalies and discrepancies that previously may have remained hidden. This proactive method, used to support risk-based quality management (RBQM), accelerates the drug development pathway and can be further enhanced with the use of AI technologies.
Machine learning (ML)
Machine learning is a set of methods designed to allow a computer to learn without being explicitly programmed. ML is used to extract knowledge from large datasets to make predictions on new or unseen data. Deep learning (DL) is a set of advanced ML methods that interrogates structured and unstructured data using deep neural networks. DL is particularly powerful in dealing with large volumes of complex data, such as clinical and operational data collected during study conduct.
How Can AI Improve Clinical Trials?
At CluePoints, we are working on a variety of ML and DL tools, all designed to improve the quality and efficiency of clinical trials and enable sponsors and CROs to bring safe, efficacious new treatments to the people that need them most.
These tools can support rapid, accurate medical coding, improve risk detection and prioritization, and identify clinical safety signals.
Medical Coding
Previously, mapping adverse events and concomitant medications recorded in case report forms to MedDRA or WHODrug dictionaries was a manual, time-consuming task. Now, our deep learning model—which is already more than 90% accurate—guides researchers to the correct corresponding term in mere seconds.
Risk Detection & Prioritization
Currently, RBQM analysis processes raise “risk signals” when they spot a potential issue within the data. These are used to monitor and track any resulting investigations and corrective actions, meaning they contain realms of free text entered by various users as they document their findings. With more than 1,200 clinical trials monitored using the CluePoints RBQM platform, the team has access to a substantial database of clinical trial issues for analysis. This is being used as a foundation for developing a model that flags signals more likely to represent real issues. This model will help trials prioritize signal review and ensure effective follow-up and documentation of findings.
Clinical Safety Signal Detection
CluePoints is also investigating the use of deep learning to detect potential clinical safety signals. This requires the DL model to make medical hypothesis as if it were a clinical reviewer, which will open new opportunities as it further explores how deep learning can be used to acquire clinical knowledge.
The Future of AI in Clinical Research
At CluePoints, we’re excited by the potential that new technologies have for improving clinical trial conduct and the impact they could have on people’s lives. Clinical trials collect high volumes of data, and as the custodians of that information, we have a responsibility to make the very best of it.
Discover the value of cloud-based digital tools for clinical trial management and clinical trial analytics by diving deeper into CluePoints solutions.
For more information on using AI to improve clinical trial processes, contact CluePoints today.
References:
- Getz, K., Smith, Z., & Kravet, M. (2023). Protocol design and performance benchmarks by phase and by oncology and rare disease subgroups. Therapeutic Innovation & Regulatory Science, 57(1), 49-56.
- Mohs, R. C., & Greig, N. H. (2017). Drug discovery and development: Role of basic biological research. Alzheimer’s & Dementia: Translational Research & Clinical Interventions, 3(4), 651-657.
- Congressional Budget Office (2021). Research and Development in the Pharmaceutical Industry. https://www.cbo.gov/publication/57025.
- Wouters, O. J., McKee, M., & Luyten, J. (2020). Estimated research and development investment needed to bring a new medicine to market, 2009-2018. Jama, 323(9), 844-853.
- Sheetz, N., Wilson, B., Benedict, J., Huffman, E., Lawton, A., Travers, M., Nadolny, P., Young, S., Given, K., & Florin, L. (2014). Evaluating Source Data Verification as a Quality Control Measure in Clinical Trials. Therapeutic Innovation & Regulatory Science, 48(6), 671–680.
- Stokman, P. G., Ensign, L., Langeneckhardt, D., Mörsch, M., Nuyens, K.,Herrera, D., Hochgräber, G., Cassan, V., Beineke, P., Kwock, R Voortman, A., Vogelgesang, S., Boussetta, S. & Bitzer, B.(2021). Risk-based Quality Management in CDM An inquiry into the value of generalized query-based data cleaning. Journal of the Society for Clinical Data Management,1.
