Improving patient safety via automated laboratory-based adverse event grading

Area of Science:

• Clinical research informatics
• Pharmacovigilance and patient safety
• Medical data analysis

Background:

• Manual identification and grading of adverse events (AEs) in clinical trials is resource-intensive and prone to errors.
• Accurate AE assessment is critical for patient safety and regulatory compliance during drug development.

Purpose of the Study:

• To describe and evaluate a novel software tool developed to automate the assessment and grading of AEs using complex algorithms on laboratory results.
• To compare the performance of the automated AE grading system against manual assessment methods.

Main Methods:

• Development of a software tool employing complex algorithms for automated AE identification and grading from laboratory data.
• Comparative analysis of AEs identified by the automated system versus manual grading.
• Prospective paired time assessment comparing automated and manual AE assessment durations.

Main Results:

• The automated grading tool demonstrated a substantial improvement in accuracy and completeness, identifying an additional 17% of severe grade 3-4 AEs missed or misgraded manually.
• The automated system resulted in an average time saving of 5.5 minutes per treatment course compared to manual assessment.
• Significant implications for patient safety are highlighted given the scale of ongoing trials and laboratory data volume.

Conclusions:

• Automated AE identification and grading software offers a significant advancement over manual processes in clinical trials.
• The tool enhances patient safety by improving the accuracy of AE detection and grading.
• Implementation of this automated system can lead to considerable time and resource efficiencies in clinical trial management.