PURPOSE

: To demonstrate how machine learning and data mining techniques can enhance health economics and outcomes research (HEOR) studies through practical examples leveraging real-world data and publicly available biomedical literature.

DESCRIPTION

: Artificial intelligence (AI) and machine learning (ML) are terms often used for a range of technologies aiming to improve on traditional data analytic approaches. While many efforts exist to explore their potential applications and impact, it is difficult to find practical examples demonstrating how ML/AI can tangibly benefit HEOR studies. In addition, specific concepts under the umbrella of machine learning – such as natural-language processing (NLP) and predictive modeling – are often discussed separately, and less discussed is how they can be used in combination to boost research.

In this workshop, we will describe an NLP platform that mines publicly available literature for associations among biological concepts and how the insights generated can add value to comparative effectiveness research (CER) and patient identification in real-world data. Furthermore, we will describe how ML techniques encompass a wide range of related concepts, which can be used to complement traditional statistical methods in healthcare research.

Dr. Yeh will open the workshop by providing an overview of some challenges in HEOR and the opportunities that can be addressed by AI/ML. Mr. Lopez-Marquez will describe an NLP platform called nferX and how it synthesizes information from publicly available biomedical literature. Dr. Sheinson will provide concrete examples of applying associations discovered by nferX to control for confounders in CER and to identify patients in rare diseases. Lastly, Dr. Lo-Ciganic will conclude with a broader picture of how ML can be best utilized in healthcare research moving forward.