Times are displayed in (UTC-07:00) Pacific Time (US & Canada) Change
11/9/2024 |
8:30 AM – 12:00 PM |
Imperial A
W04: Interpretable temporal data mining for clinical decision-making
Presentation Type: Workshop/Tutorial
Interpretable temporal data mining for clinical decision-making
Presentation Time: 08:30 AM - 12:00 PM
Abstract Keywords: Clinical Decision Support, Data Mining, Information Extraction, Knowledge Representation and Information Modeling
Primary Track: Foundations
Programmatic Theme: Clinical Research Informatics
The healthcare system is a complex, data-rich environment. Current clinical database systems have the capability to store vast amounts of data. Consequently, physicians have access to a large quantity of data for each patient collected in the electronic medical records. The analysis of such healthcare/medical data collections could significantly help to gain a deeper insight into the population's health conditions and extract useful information that can be exploited in the assessment of healthcare/medical processes. As a huge amount of medical data is inherently temporal, modeling temporal features of electronic health record (EHR) data is of paramount importance.
In this tutorial, we will introduce some temporal data mining approaches to support clinical decision-based tasks. The techniques we will describe allow a fast interpretation of derived results by the clinical stakeholders.
We will consider some main techniques supporting different perspectives and knowledge derived from the data mining task. Both patient-oriented and population-oriented hidden information may be discovered through the introduced techniques.
We will discuss temporal association rules and temporal pattern discovery, predictive functional dependencies, together with temporal phenotyping. The considered techniques, while general in the proposed approach, will be applied to specific clinical contexts, such as Acute Kidney Injury (AKI) detection in the Intensive Care Unit (ICU), Type 2 Diabetes management, social determinants of health, and COVID-19 Patients' Trajectories. Some final hands-on activities with web-based tools will be proposed to the participants, allowing them to evaluate the discussed techniques on real-world clinical data in an interactive way.
Speaker(s):
Lucia Sacchi, PhD
University of Pavia
Beatrice Amico, Ph.D.
Dept. of Computer Science - University of Verona
Carlo Combi, PhD
Università di Verona
Arianna Dagliati
University of Pavia
John Holmes, PhD, FACE, FACMI, FIAHSI
University of Pennsylvania School of Medicine
Author(s):
Presentation Time: 08:30 AM - 12:00 PM
Abstract Keywords: Clinical Decision Support, Data Mining, Information Extraction, Knowledge Representation and Information Modeling
Primary Track: Foundations
Programmatic Theme: Clinical Research Informatics
The healthcare system is a complex, data-rich environment. Current clinical database systems have the capability to store vast amounts of data. Consequently, physicians have access to a large quantity of data for each patient collected in the electronic medical records. The analysis of such healthcare/medical data collections could significantly help to gain a deeper insight into the population's health conditions and extract useful information that can be exploited in the assessment of healthcare/medical processes. As a huge amount of medical data is inherently temporal, modeling temporal features of electronic health record (EHR) data is of paramount importance.
In this tutorial, we will introduce some temporal data mining approaches to support clinical decision-based tasks. The techniques we will describe allow a fast interpretation of derived results by the clinical stakeholders.
We will consider some main techniques supporting different perspectives and knowledge derived from the data mining task. Both patient-oriented and population-oriented hidden information may be discovered through the introduced techniques.
We will discuss temporal association rules and temporal pattern discovery, predictive functional dependencies, together with temporal phenotyping. The considered techniques, while general in the proposed approach, will be applied to specific clinical contexts, such as Acute Kidney Injury (AKI) detection in the Intensive Care Unit (ICU), Type 2 Diabetes management, social determinants of health, and COVID-19 Patients' Trajectories. Some final hands-on activities with web-based tools will be proposed to the participants, allowing them to evaluate the discussed techniques on real-world clinical data in an interactive way.
Speaker(s):
Lucia Sacchi, PhD
University of Pavia
Beatrice Amico, Ph.D.
Dept. of Computer Science - University of Verona
Carlo Combi, PhD
Università di Verona
Arianna Dagliati
University of Pavia
John Holmes, PhD, FACE, FACMI, FIAHSI
University of Pennsylvania School of Medicine
Author(s):
W04: Interpretable temporal data mining for clinical decision-making
Description
