Latest Artificial Intelligence (AI) Research Uses Synthetic Patient Data To Simulate A Machine Learning Enabled Learning Health System (LHS)

In recent years, machine learning (ML) techniques in the medical field have grown exponentially. Due to the high performance of deep learning models, machines can diagnose and classify diseases better, sometimes even more than experts. To do this, the model will use data such as medical images by accessing personal information of patients. This use of personal data raises privacy concerns. One of the most important barriers to the research and development of Learning Health Systems (LHS) is the lack of access to EHR patient data.

Due to the advancement of synthetic patient data technology, synthetic patient data has recently been adopted as an alternative data to test new systems related to EHR data. It is now possible to generate synthetic patient data that can be used to create LHS with ML and shared among the research community without restriction. Using this method, synthetic datasets on heart disease, even cancer, have already appeared. In this case, a group of researchers from California proposed an iterative method using synthetic patients to develop LHS risk predictions based on ML data.

Also Read :  Rising Demand for Smart Robotics in Manufacturing Industries Presents Opportunities

Obviously, the authors of the article conducted research through simulation. In this study, LHS prediction is performed by performing the XGBoost baseline for various target diseases, such as lung cancer or stroke, from electronic health record (EHR) data. ). This simulation study follows two steps: In the first step, a new ML-enabled LHS method is proposed to construct LHS risk predictions for lung cancer and synthetic patients. In a second step, different strokes were used to check the effectiveness of the new LHS method for building LHS risk predictions and accurate risk predictions for different target diseases. The authors proposed an advanced data-centric scheme and LHS design for risk prediction. Initially, ML models were built from original EHR patient data. Next, the LHS learning process uses new patient data to improve ML models and quickly generate new models that clinicians can use to make risk predictions.

The LHS provided by ML was created using a dataset of 30,000 Synthea synthetic patients, and the XGBoost model was used to predict the risk of lung cancer. Then, four additional datasets of 30,000 patients were created. These four new records were added one by one to the first updated dataset to accommodate the addition of new patients, resulting in datasets of 60,000, 90,000, 120,000, and 150,000 patients. In each instance, a new version of XGBoost is built. The results show that the performance improves when the data size increases, reaching 0.936 recall and 0.962 AUC on a dataset of 150,000 patients. The effectiveness of the new ML-based LHS method was assessed by implementing the XGBoost model for predicting stroke risk in a cohort of Synthea patients.

Also Read :  Pensacola history lives on at Fort Pickens, Aviation Museum and more

This paper introduces the methodology of ML models based on synthetic health data for the first time. This study demonstrated the effectiveness of this new LHS method that can treat a variety of diseases from other EHR data. The proposed model continuously learns from a newly created patient to improve its performance until it reaches a risk prediction greater than 95% for metric recall and precision. Finally, the authors say that since real data is different from synthetic data, real data ML models can be improved by hyperparameter adjustment.

This Article is written as a research summary article by Marktechpost Staff based on the research paper 'Simulation of a machine learning enabled learning health system for risk prediction using synthetic patient data'. All Credit For This Research Goes To Researchers on This Project. Check out the paper.
Please Don't Forget To Join Our ML Subreddit


Also Read :  Europe’s digital ambitions won’t be realized without cloud – POLITICO

Mahmoud is a PhD researcher in machine learning. It also holds a
bachelor’s degree in physical science and master’s degree in
communication systems and networks. Its current area of
research related to computer vision, product prediction and depth
education. He came up with a lot of science fiction about someone else.
characteristics and studies are strong and stable in depth
network.


Source

Leave a Reply

Your email address will not be published.

Related Articles

Back to top button