How Machine Learning Can Automate Data Anomaly Detection, Streamlining Query Management to Accelerate Database Lock and Regulatory Submission

With increasing data sources and pressure to speed up clinical trial development, we need to explore how to streamline processes without compromising quality. Traditional methods of manual data query are inefficient, costing an average of $6.2M per study with only 42% resulting in effective data changes. At the same time, 45% of senior pharmaceutical and CRO executives cite data quality as one of their top three operational issues. Supervised machine learning (ML) can be used to automate data anomaly detection with improved accuracy and consistency, streamlining query management to accelerate database lock and regulatory submission. By eliminating the noise of false positives, an ML approach can also free up data managers to focus on critical queries, allowing earlier and easier identification of potential medical or safety issues. This session will explore the key pain points of data management and how harnessing supervised ML can help to overcome them. Key challenges discussed will include: • Inefficient, manual query processes • High volumes of false positives • Non-programmable conditions • The cost of manual queries • Standardizing query descriptions • Data reconciliation Solutions explored will include: • Real-world data training of ML models to enable real-time updates and simplify the handling of both new and previously identified data issues. • Study-agnostic models which seamlessly identify data anomalies across different trials, effectively handling various naming conventions and protocols, without requiring specific programming. • Using ML models to detect true data anomalies, ensuring fewer unnecessary or redundant queries are raised. At the end of this session, delegates will have an increased understanding of: 1. The challenges facing data managers in a modern clinical trial environment. 2. The opportunities offered by new technologies like ML. 3. How an integrated approach to data review accelerates drug development, improves regulatory compliance and ultimately contributes to the success of clinical trials.