Toward practical implementation of predictive models: cost-effective thresholds in machine learning models for chronic kidney disease screening.
Saved in:
| Title: | Toward practical implementation of predictive models: cost-effective thresholds in machine learning models for chronic kidney disease screening. |
|---|---|
| Authors: | Cao, Yiwen1 (AUTHOR) ycao0253@usc.edu, Suen, Sze-chuan1 (AUTHOR) |
| Source: | IISE Transactions on Healthcare Systems Engineering. Apr-Jun2025, Vol. 15 Issue 2, p167-182. 16p. |
| Subjects: | Machine learning, Cost effectiveness, Chronic kidney failure, Medical screening, Prediction models, Health outcome assessment, Risk assessment |
| Geographic Terms: | United States |
| Abstract: | Chronic kidney disease (CKD) presents both a significant health and economic burden to the US. Current screening policies may result in missed CKD diagnoses due to the use of unnuanced risk categorizations. An accurate, personalized risk assessment that can be easily performed on the general population to prompt medical screening among high-risk patients may result in a more cost-effective screening policy. A machine learning (ML) predictive tool may be able to provide such risk estimates, but the cost-effectiveness of using such a tool will depend on which sensitivity/specificity threshold is used to prompt medical screening. This study introduces a process for identifying cost-effective thresholds for machine learning (ML)-based risk assessment tools. We train a user-friendly ML-based prescreening tool for CKD requiring only a limited amount of patient information using nationally representative data to increase ease of use and applicability to diverse subpopulations. We then employ a simulation model to capture lifetime health outcomes and medical costs associated with using various ML risk thresholds to prompt medical screening. We identify the cost-effective region for the ML-based prescreening tool to implement in practice. We find sensitivity and specificity thresholds that could reduce lifetime costs, improve health outcomes, and reduce the risk of progression to end-stage renal disease (ESRD) compared to the status quo policy. We conclude that implementing our tailored, ML-based prescreening tool in the cost-effective region can enhance early CKD detection and health outcomes. This study underscores the importance of evaluating cost-effectiveness outcomes when implementing predictive models for disease management. [ABSTRACT FROM AUTHOR] |
| Copyright of IISE Transactions on Healthcare Systems Engineering is the property of Taylor & Francis Ltd and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.) | |
| Database: | Engineering Source |
|
Full text is not displayed to guests.
Login for full access.
|
|
| Abstract: | Chronic kidney disease (CKD) presents both a significant health and economic burden to the US. Current screening policies may result in missed CKD diagnoses due to the use of unnuanced risk categorizations. An accurate, personalized risk assessment that can be easily performed on the general population to prompt medical screening among high-risk patients may result in a more cost-effective screening policy. A machine learning (ML) predictive tool may be able to provide such risk estimates, but the cost-effectiveness of using such a tool will depend on which sensitivity/specificity threshold is used to prompt medical screening. This study introduces a process for identifying cost-effective thresholds for machine learning (ML)-based risk assessment tools. We train a user-friendly ML-based prescreening tool for CKD requiring only a limited amount of patient information using nationally representative data to increase ease of use and applicability to diverse subpopulations. We then employ a simulation model to capture lifetime health outcomes and medical costs associated with using various ML risk thresholds to prompt medical screening. We identify the cost-effective region for the ML-based prescreening tool to implement in practice. We find sensitivity and specificity thresholds that could reduce lifetime costs, improve health outcomes, and reduce the risk of progression to end-stage renal disease (ESRD) compared to the status quo policy. We conclude that implementing our tailored, ML-based prescreening tool in the cost-effective region can enhance early CKD detection and health outcomes. This study underscores the importance of evaluating cost-effectiveness outcomes when implementing predictive models for disease management. [ABSTRACT FROM AUTHOR] |
|---|---|
| ISSN: | 24725579 |
| DOI: | 10.1080/24725579.2025.2463625 |
