Clinical research requires that data be generated for knowledge. Machine learning, which develops algorithms to find patterns, has difficulty with data related to health records because this type of information is neither static nor regularly collected. In a new study, a transparent and reproducible tool for machine learning was developed to facilitate the analysis of health information. The tool can be used for clinical forecasting, which can be used to predict trends as well as outcomes in individual patients.

The study by a researcher at Carnegie Mellon University (CMU) appears in the Proceedings of Machine Learning Research.

“Temporal Learning Lite, or TL-Lite, is a visualization and forecasting tool designed to bridge the gap between clinical visualization and machine learning analysis,” said Jeremy Weiss, Assistant Professor of Health Informatics at CMU’s Heinz College, who authored the study. “While the individual elements of this tool are known, integrating it into an interactive clinical research tool is new and useful for healthcare professionals. With training, users can perform preliminary analysis in minutes.”

Time is an integral part of the clinical data collected in the delivery of health services. For example, when discussing patients in rounds where doctors visit hospital patients to see how they are doing, medical staff use visual aids that show measurements of progression and recovery. Since electronic health records became widespread, significant advances have been made in the visualization of clinical data as well as clinical prediction. Still, there remains a gap between the two.

TL-Lite starts with the visualization of information from databases and ends with the visual risk assessment of a time model. Along the way, users can see the impact of their design decisions through visual summaries at the individual and group level. This allows users to better understand their data and customize machine learning settings for their analysis.

To show how the tool can be used, Weiss demonstrated the model with three electronic health records that relate to three health issues: Predicting severe thrombocytopenia (with abnormally low platelet levels) while in the intensive care unit (ICU) in patients with sepsis , Predicting the survival of patients admitted to the ICU one day after admission and predicting microvascular complications of type 2 diabetes in those with the disease.

“The central goal of TL-Lite is to enable well-specified and well-designed forecasting forecasts, and this visualization tool is intended to simplify the process,” says Weiss. “At the same time, the organization of the clinical data stream in meaningful visualizations can be supported by the introduction of elements of machine learning. These approaches complement each other, so that the advantages of an approach in which another meets roadblocks lead to a better overall solution.”

The researchers are building models using machine learning technology to improve predictions of COVID-19 outcomes

More information:
Temporal visualization and learning for clinical prognosis. Provided by Carnegie Mellon University

Quote: New machine learning tool makes it easier to analyze health information and clinical prognoses (2021, February 25), accessed on February 26, 2021 from -clinical.html

This document is subject to copyright. Except for fair trade for the purpose of private study or research, no part may be reproduced without written permission. The content is provided for informational purposes only.


Please enter your comment!
Please enter your name here