Leveraging Clinical Laboratory Data for a Learning Health System Though Advanced AI and Data Science

Navigating the healthcare process is similar to piecing together an intricate puzzle, where every medical test is a crucial piece, and the completed image reflects a patient's comprehensive care pathway. With clinical laboratory data representing an estimated 70% of all healthcare data and influencing 70% of medical decisions, this work aims to exploit this data more effectively within a learning health system framework, enhancing patient care through the innovative application of artificial intelligence (AI) and data science.

The research identifies two main challenges in the utilization of clinical laboratory data in the intricate puzzle of healthcare. The first challenge involves the secondary use of laboratory data, including clinical research and the development of clinical machine learning (ML) models, which is often compromised by issues of data quality. The second challenge pertains to the primary clinical use of laboratory data, where the inappropriate selection and utilization of laboratory tests can hinder the delivery of optimal patient care, underscoring the difficulty of ensuring the right test for the right patient at the right time.

To address these challenges, the research proposes a two-pronged AI-driven approach. Firstly, by developing “lab2clean” novel algorithms designed to automate and standardize the data cleaning process, we aim to improve laboratory data quality for accurate research and effective clinical ML models. Secondly, we aim to research how AI can drive a “Clinical Laboratory Intelligent Test Engine (CLINTE)” that can support clinicians in making more informed decisions regarding laboratory test utilization, aiming to optimize patient care.

By integrating the disciplines of laboratory medicine with cutting-edge AI and data science techniques, this research seeks to refine the clarity and fit of each piece of clinical laboratory data, ensuring its meaningful contribution to the comprehensive healthcare puzzle and the overarching goal of a more adaptive and learning health system that supports personalized and evidence-based patient care.