Machine learning outperforms traditional statistical methods in addressing missing data in electronic health records
2025-01-15
(Press-News.org)
Researchers from the National Institute of Health Data Science at Peking University and the Department of Clinical Epidemiology and Biostatistics at Peking University People's Hospital have conducted a comprehensive systematic review evaluating strategies for addressing missing data in electronic health records (EHRs). Published in Health Data Science, the study highlights the growing importance of machine learning methods over traditional statistical approaches in managing missing data scenarios effectively.
Electronic health records have become a cornerstone in modern healthcare research, enabling analysis across clinical trials, treatment effectiveness studies, and genetic association research. However, missing data remains a persistent challenge, potentially introducing bias and undermining the reliability of findings. This study reviewed 46 research papers published between 2010 and 2024, systematically comparing the performance of traditional statistical methods, such as Multiple Imputation by Chained Equations (MICE), with modern machine learning approaches like Generative Adversarial Networks (GANs) and k-Nearest Neighbors (KNN).
The findings reveal that machine learning techniques, particularly GAN-based methods and context-aware time-series imputation (CATSI), consistently outperformed traditional statistical approaches in handling both longitudinal and cross-sectional datasets. For longitudinal data, Med.KNN and CATSI showed superior performance, while probabilistic principal component analysis (PCA) and MICE were more effective for cross-sectional datasets.
"Machine learning methods show significant promise for addressing missing data in EHRs," said Dr. Huixin Liu, Associate Professor at Peking University People's Hospital. "However, no single approach offers a universally applicable solution, underscoring the need for standardized benchmarking analyses across diverse datasets and missingness scenarios".
The study also identifies key challenges, including the heterogeneity of EHR datasets, the opacity of machine learning models, and the lack of universal benchmarks for assessing methodology performance. Future research aims to establish a standardized protocol for handling missing EHR data and develop benchmarking datasets for comprehensive evaluation.
"Our ultimate goal is to create a universally accepted protocol for handling missing data in electronic health records, ensuring more reliable and reproducible findings across medical research," added Dr. Shenda Hong, Assistant Professor at the National Institute of Health Data Science at Peking University.
This research marks a significant step toward addressing one of the most pressing challenges in digital healthcare research, offering insights that can help bridge the gap between data scarcity and robust analysis.
END
ELSE PRESS RELEASES FROM THIS DATE:
2025-01-15
About The Study: In this multicenter validation study, trained health care professionals with artificial intelligence (AI) assistance achieved lung ultrasound images meeting diagnostic standards compared with lung ultrasound experts without AI. This technology could extend access to lung ultrasound to underserved areas lacking expert personnel.
Corresponding Author: To contact the corresponding author, Cristiana Baloescu, MD, MPH, email cristiana.baloescu@yale.edu.
To access the embargoed ...
2025-01-15
About The Study: This survey study documents increasingly prevalent poor mental health from 2011 to 2022 across multiple U.S. health surveys, with notable prevalence differences in Behavioral Risk Factor Surveillance System and National Survey on Drug Use and Health vs National Health Interview Survey. Inequities in these outcomes by age, sex, and racial and ethnic group were often sizeable and changed over time in distinct ways, consistent with findings in prior literature.
Corresponding Author: To contact the corresponding ...
2025-01-15
About The Study: In this cohort study including 38 attending surgeons and 793 patients, increased surgeon stress at the beginning of a procedure was associated with improved clinical patient outcomes. The results are illustrative of the complex relationship between physiological stress and performance, identify a novel association between measurable surgeon human factors and patient outcomes, and may highlight opportunities to improve patient care.
Corresponding Author: To contact the corresponding author, Jake Awtry, MD, email jawtry@bwh.harvard.edu.
To access the embargoed study: Visit our For The Media ...
2025-01-15
According to the United Nations, soil salinization affects between 20% and 40% of arable land globally, with human activity and climate change – especially rising sea levels – largely responsible for this process. While the human body needs sodium to function, this is not the case for most plants. In fact, excess salt around plants’ roots gradually blocks their access to water, stunting their growth, poisoning them and hastening their death. Ten million hectares of farmland are destroyed by soil salinization every year, posing a threat to global food security.
Scientists at EPFL, ...
2025-01-15
While most known types of DNA damage are fixed by our cells’ in-house DNA repair mechanisms, some forms of DNA damage evade repair and can persist for many years, new research shows. This means that the damage has multiple chances to generate harmful mutations, which can lead to cancer.
Scientists from the Wellcome Sanger Institute and their collaborators analysed family trees of hundreds of single cells from several individuals. The team pieced together these family trees from patterns of shared mutations between the cells, indicating common ancestors.
Researchers uncovered unexpected ...
2025-01-15
Researchers have discovered a biological mechanism that makes plant roots more welcoming to beneficial soil microbes.
This discovery by John Innes Centre researchers paves the way for more environmentally friendly farming practices, potentially allowing farmers to use less fertiliser.
Production of most major crops relies on nitrate and phosphate fertilisers, but excessive fertiliser use harms the environment.
If we could use mutually beneficial relationships between plant roots and soil microbes to enhance nutrient uptake, ...
2025-01-15
CAMBRIDGE, MA -- Nearly 50 years ago, neuroscientists discovered cells within the brain’s hippocampus that store memories of specific locations. These cells also play an important role in storing memories of events, known as episodic memories. While the mechanism of how place cells encode spatial memory has been well-characterized, it has remained a puzzle how they encode episodic memories.
A new model developed by MIT researchers explains how those place cells can be recruited to form episodic memories, even when there’s no spatial component. According to this model, place cells, along with grid cells found in the entorhinal cortex, act as a scaffold ...
2025-01-15
Arizona State University and a team of its collaborators have received $11.2 million in funding from the U.S. Department of Energy to begin developing a regional Direct Air Capture (DAC) Hub for removing carbon dioxide (CO2) from the atmosphere. The team will prepare to build a multi-site Direct Air Capture Hub located in the Four Corners area of the Southwestern United States. Additionally, the project will receive $11.2 million in matching funds from the project partners.
In May of 2022, the Biden administration announced the Bipartisan Infrastructure Law’s $3.5 billion DOE program to establish large-scale Direct Air Capture Hubs for removing carbon ...
2025-01-15
Isotretinoin, commonly referred to as Accutane, is the only approved medical treatment capable of inducing long-term remission of severe acne. Although highly effective, some individuals experience recurrence of acne after a course of treatment. A new study from researchers at Mass General Brigham examined how often acne recurs after isotretinoin and what factors might put patients at risk of acne coming back. They found that acne recurrence necessitating treatment with an oral medication such as oral antibiotics, spironolactone, or another ...
2025-01-15
New proteins not found in nature have now been designed to counteract certain highly poisonous components of snake venom. The deep learning, computational methods for developing these toxin-neutralizing proteins offer hope for creating safer, more cost-effective and more readily available therapeutics than those currently in use.
Each year more than 2 million people suffer snakebites. More than 100,000 of them die, according to the World Health Organization, and 300,000 suffer severe complications and lasting disability ...
LAST 30 PRESS RELEASES:
[Press-News.org] Machine learning outperforms traditional statistical methods in addressing missing data in electronic health records
