โก Quick Summary
A recent study developed a machine learning tool to predict the risk of chronic kidney disease (CKD) using health examination data from over 30,000 participants. The models demonstrated high predictive accuracy, particularly for predicting CKD onset within 1 year with an AUROC greater than 0.9.
๐ Key Details
- ๐ Dataset: 30,273 participants, including 1,372 with CKD
- ๐งฉ Features used: Demographics, BMI, blood pressure, lab results, questionnaire responses
- โ๏ธ Technology: Logistic regression, neural networks, recurrent neural networks
- ๐ Performance: AUROC > 0.9 for 1-year predictions, AUROC ~0.65 for proteinuria predictions
๐ Key Takeaways
- ๐ Early detection of CKD is crucial for effective treatment.
- ๐ค Machine learning models can significantly enhance risk prediction for CKD.
- ๐ eGFR is a critical factor in predicting CKD onset.
- ๐ Predictive performance declines when eGFR is excluded from the analysis.
- ๐งช Further research is needed to improve predictions for proteinuria.
- ๐ Study conducted from 2017 to 2022, highlighting the importance of health examinations.
- ๐ Long-term predictions (5 years) showed lower specificity compared to 1-year predictions.
๐ Background
Chronic kidney disease (CKD) is a significant global health issue, often progressing silently until severe symptoms arise. Early detection through regular health examinations can lead to timely interventions, potentially slowing disease progression and improving patient outcomes. The integration of machine learning into health data analysis presents a promising avenue for enhancing CKD risk prediction.
๐๏ธ Study
This study utilized health examination data collected from patients between 2017 and 2022. The researchers aimed to develop predictive models for CKD risk over 1 and 5 years using various health metrics, including demographic information, body mass index (BMI), blood pressure, and laboratory test results. The study’s comprehensive approach involved multiple machine learning techniques to assess their effectiveness in predicting CKD onset.
๐ Results
The findings revealed that all models achieved predictive values, sensitivities, and specificities exceeding 0.8 for predicting CKD onset within 1 year, with an impressive AUROC greater than 0.9. However, the predictive performance for outcomes related to proteinuria was notably lower, with AUROCs around 0.65. The exclusion of eGFR from the models resulted in a significant drop in predictive accuracy, underscoring its importance in CKD risk assessment.
๐ Impact and Implications
The development of these machine learning models represents a significant advancement in the early detection of CKD. By leveraging health examination data, healthcare providers can identify at-risk individuals more effectively, leading to timely interventions. This approach not only has the potential to improve patient outcomes but also to reduce the overall burden of CKD on healthcare systems globally. The study highlights the need for ongoing research to refine these models and explore additional predictive factors.
๐ฎ Conclusion
This study illustrates the transformative potential of machine learning in predicting chronic kidney disease risk. The strong performance of the models, particularly in the short term, emphasizes the critical role of eGFR in CKD prediction. As we move forward, further research is essential to enhance our understanding of CKD progression and to develop more robust predictive tools that can aid in clinical decision-making.
๐ฌ Your comments
What are your thoughts on the use of machine learning for predicting chronic kidney disease? We would love to hear your insights! ๐ฌ Share your comments below or connect with us on social media:
Development of a machine learning tool to predict the risk of incident chronic kidney disease using health examination data.
Abstract
BACKGROUND: Chronic kidney disease (CKD) is characterized by a decreased glomerular filtration rate or renal injury (especially proteinuria) for at least 3โmonths. The early detection and treatment of CKD, a major global public health concern, before the onset of symptoms is important. This study aimed to develop machine learning models to predict the risk of developing CKD within 1 and 5โyears using health examination data.
METHODS: Data were collected from patients who underwent annual health examinations between 2017 and 2022. Among the 30,273 participants included in the study, 1,372 had CKD. Demographic characteristics, body mass index, blood pressure, blood and urine test results, and questionnaire responses were used to predict the risk of CKD development at 1 and 5โyears. This study examined three outcomes: incident estimated glomerular filtration rate (eGFR) <60โmL/min/1.73โm2, the development of proteinuria, and incident eGFR <60โmL/min/1.73โm2 or the development of proteinuria. Logistic regression (LR), conditional logistic regression, neural network, and recurrent neural network were used to develop the prediction models.
RESULTS: All models had predictive values, sensitivities, and specificities >0.8 for predicting the onset of CKD in 1โyear when the outcome was eGFR <60โmL/min/1.73โm2. The area under the receiver operating characteristic curve (AUROC) was >0.9. With LR and a neural network, the specificities were 0.749 and 0.739 and AUROCs were 0.889 and 0.890, respectively, for predicting onset within 5โyears. The AUROCs of most models were approximately 0.65 when the outcome was eGFR <60โmL/min/1.73โm2 or proteinuria. The predictive performance of all models exhibited a significant decrease when eGFR was not included as an explanatory variable (AUROCs: 0.498-0.732).
CONCLUSION: Machine learning models can predict the risk of CKD, and eGFR plays a crucial role in predicting the onset of CKD. However, it is difficult to predict the onset of proteinuria based solely on health examination data. Further studies must be conducted to predict the decline in eGFR and increase in urine protein levels.
Author: [‘Yoshizaki Y’, ‘Kato K’, ‘Fujihara K’, ‘Sone H’, ‘Akazawa K’]
Journal: Front Public Health
Citation: Yoshizaki Y, et al. Development of a machine learning tool to predict the risk of incident chronic kidney disease using health examination data. Development of a machine learning tool to predict the risk of incident chronic kidney disease using health examination data. 2024; 12:1495054. doi: 10.3389/fpubh.2024.1495054