1
|
Lee WT, Fang YW, Chang WS, Hsiao KY, Shia BC, Chen M, Tsai MH. Data-driven, two-stage machine learning algorithm-based prediction scheme for assessing 1-year and 3-year mortality risk in chronic hemodialysis patients. Sci Rep 2023; 13:21453. [PMID: 38052875 PMCID: PMC10698192 DOI: 10.1038/s41598-023-48905-9] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 09/30/2023] [Accepted: 12/01/2023] [Indexed: 12/07/2023] Open
Abstract
Life expectancy is likely to be substantially reduced in patients undergoing chronic hemodialysis (CHD). However, machine learning (ML) may predict the risk factors of mortality in patients with CHD by analyzing the serum laboratory data from regular dialysis routine. This study aimed to establish the mortality prediction model of CHD patients by adopting two-stage ML algorithm-based prediction scheme, combined with importance of risk factors identified by different ML methods. This is a retrospective, observational cohort study. We included 800 patients undergoing CHD between December 2006 and December 2012 in Shin-Kong Wu Ho-Su Memorial Hospital. This study analyzed laboratory data including 44 indicators. We used five ML methods, namely, logistic regression (LGR), decision tree (DT), random forest (RF), gradient boosting (GB), and eXtreme gradient boosting (XGB), to develop a two-stage ML algorithm-based prediction scheme and evaluate the important factors that predict CHD mortality. LGR served as a bench method. Regarding the validation and testing datasets from 1- and 3-year mortality prediction model, the RF had better accuracy and area-under-curve results among the five different ML methods. The stepwise RF model, which incorporates the most important factors of CHD mortality risk based on the average rank from DT, RF, GB, and XGB, exhibited superior predictive performance compared to LGR in predicting mortality among CHD patients over both 1-year and 3-year periods. We had developed a two-stage ML algorithm-based prediction scheme by implementing the stepwise RF that demonstrated satisfactory performance in predicting mortality in patients with CHD over 1- and 3-year periods. The findings of this study can offer valuable information to nephrologists, enhancing patient-centered decision-making and increasing awareness about risky laboratory data, particularly for patients with a high short-term mortality risk.
Collapse
Affiliation(s)
- Wen-Teng Lee
- Division of Nephrology, Department of Internal Medicine, Shin-Kong Wu Ho-Su Memorial Hospital, No. 95, Wen-Chang Rd, Shih-Lin Dist., Taipei, 11101, Taiwan
| | - Yu-Wei Fang
- Division of Nephrology, Department of Internal Medicine, Shin-Kong Wu Ho-Su Memorial Hospital, No. 95, Wen-Chang Rd, Shih-Lin Dist., Taipei, 11101, Taiwan
- Department of Medicine, Fu Jen Catholic University, No. 510, Zhongzhen Rd., Xinzhuang Dist., New Taipei City, 24205, Taiwan
| | - Wei-Shan Chang
- Artificial Intelligence Development Center, Fu Jen Catholic University, No. 510, Zhongzhen Rd., Xinzhuang Dist., New Taipei City, 24205, Taiwan
- Graduate Institute of Business Administration, College of Management, Fu Jen Catholic University, No. 510, Zhongzhen Rd., Xinzhuang Dist, New Taipei City, 24205, Taiwan
| | - Kai-Yuan Hsiao
- Artificial Intelligence Development Center, Fu Jen Catholic University, No. 510, Zhongzhen Rd., Xinzhuang Dist., New Taipei City, 24205, Taiwan
- Graduate Institute of Business Administration, College of Management, Fu Jen Catholic University, No. 510, Zhongzhen Rd., Xinzhuang Dist, New Taipei City, 24205, Taiwan
| | - Ben-Chang Shia
- Artificial Intelligence Development Center, Fu Jen Catholic University, No. 510, Zhongzhen Rd., Xinzhuang Dist., New Taipei City, 24205, Taiwan
- Graduate Institute of Business Administration, College of Management, Fu Jen Catholic University, No. 510, Zhongzhen Rd., Xinzhuang Dist, New Taipei City, 24205, Taiwan
| | - Mingchih Chen
- Artificial Intelligence Development Center, Fu Jen Catholic University, No. 510, Zhongzhen Rd., Xinzhuang Dist., New Taipei City, 24205, Taiwan.
- Graduate Institute of Business Administration, College of Management, Fu Jen Catholic University, No. 510, Zhongzhen Rd., Xinzhuang Dist, New Taipei City, 24205, Taiwan.
| | - Ming-Hsien Tsai
- Division of Nephrology, Department of Internal Medicine, Shin-Kong Wu Ho-Su Memorial Hospital, No. 95, Wen-Chang Rd, Shih-Lin Dist., Taipei, 11101, Taiwan.
- Department of Medicine, Fu Jen Catholic University, No. 510, Zhongzhen Rd., Xinzhuang Dist., New Taipei City, 24205, Taiwan.
| |
Collapse
|
2
|
Yang S, Diao Z, Liu W, Guo W. Effect of dialytic phosphate reduction rate on mortality in maintenance hemodialysis patients: a matched case-control study. BMC Nephrol 2023; 24:167. [PMID: 37308828 DOI: 10.1186/s12882-023-03199-x] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 12/01/2022] [Accepted: 05/14/2023] [Indexed: 06/14/2023] Open
Abstract
BACKGROUND Phosphates, similar to urea, are small molecular substances that can be cleared during dialysis. Dialytic phosphate reduction rate (PRR) may, to some extent, be related to the relative amount of phosphates cleared during dialysis. However, few studies have evaluated the associations between PRR and mortality in maintenance hemodialysis (MHD) patients. In this study, we investigated the association between PRR and clinical outcomes in MHD patients. METHODS This was a retrospective, matched case-control study. Data were collected from the Beijing Hemodialysis Quality Control and Improvement Center. Patients were divided into four groups according to PRR quartile. Age, sex, and diabetes were matched between the groups. The primary outcome was all-cause death, and the secondary outcome was cardiocerebrovascular death. RESULTS The study cohort comprised 4063 patients who were divided into four groups according to the PRR quartile: group PRR1 (< 48.35%), group PRR2 (48.35% - 54.14%), group PRR3 (54.14% - 59.14%), and group PRR4 (≥ 59.14%). We enrolled 2172 patients (543 in each study group) by case-control matching. The all-cause death rates were as follows: group PRR1: 22.5% (122/543), group PRR2: 20.1% (109/543), group PRR3: 19.3% (105/543), and group PRR4: 19.3% (105/543). No significant differences in all-cause and cardiocerebrovascular death rates according to the Kaplan-Meier survival curves were found between the groups (log-rank test, P > 0.05). Multivariable Cox regression analysis revealed no significant differences in all-cause and cardiocerebrovascular death rates between the four groups (P = 0.461; adjusted hazard ratio, 0.99; 95% confidence interval, 0.97 - 1.02 versus P = 0.068; adjusted hazard ratio, 0.99; 95% confidence interval, 0.97 - 1.00, respectively). CONCLUSIONS Dialytic PRR was not significantly associated with all-cause death and cardiocerebrovascular death in MHD patients.
Collapse
Affiliation(s)
- Shuixiu Yang
- Department of Nephrology, Beijing Friendship Hospital, Capital Medical University, Beijing, 100050, China
- Blood Purification Center, Guiyang Public Health Clinical Center, Guiyang, 550001, Guizhou, China
| | - Zongli Diao
- Department of Nephrology, Beijing Friendship Hospital, Capital Medical University, Beijing, 100050, China
| | - Wenhu Liu
- Department of Nephrology, Beijing Friendship Hospital, Capital Medical University, Beijing, 100050, China.
| | - Wang Guo
- Department of Nephrology, Beijing Friendship Hospital, Capital Medical University, Beijing, 100050, China.
| |
Collapse
|
3
|
Thongprayoon C, Cheungpasitporn W, Hansrivijit P, Thirunavukkarasu S, Chewcharat A, Medaura J, Mao MA, Kashani KB. Impact of serum phosphate changes on in-hospital mortality. BMC Nephrol 2020; 21:427. [PMID: 33028266 PMCID: PMC7542949 DOI: 10.1186/s12882-020-02090-3] [Citation(s) in RCA: 6] [Impact Index Per Article: 1.5] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 10/03/2019] [Accepted: 10/05/2020] [Indexed: 12/25/2022] Open
Abstract
Background Fluctuations in serum phosphate levels increased mortality in end-stage renal disease patients. However, the impacts of serum phosphate changes in hospitalized patients remain unclear. This study aimed to test the hypothesis that serum phosphate changes during hospitalization were associated with in-hospital mortality. Methods We included all adult hospitalized patients from January 2009 to December 2013 that had at least two serum phosphate measurements during their hospitalization. We categorized in-hospital serum phosphate changes, defined as the absolute difference between the maximum and minimum serum phosphate, into 5 groups: 0–0.6, 0.7–1.3, 1.4–2.0, 2.1–2.7, ≥2.8 mg/dL. Using serum phosphate change group of 0–0.6 mg/dL as the reference group, the adjusted odds ratio of in-hospital mortality for various serum phosphate change groups was obtained by multivariable logistic regression analysis. Results A total of 28,149 patients were studied. The in-hospital mortality in patients with serum phosphate changes of 0–0.6, 0.7–1.3, 1.4–2.0, 2.1–2.7, ≥2.8 mg/dL was 1.5, 2.0, 3.1, 4.4, and 10.7%, respectively (p < 0.001). When adjusted for confounding factors, larger serum phosphate changes were associated with progressively increased in-hospital mortality with odds ratios of 1.35 (95% 1.04–1.74) in 0.7–1.3 mg/dL, 1.98 (95% CI 1.53–2.55) in 1.4–2.0 mg/dL, 2.68 (95% CI 2.07–3.48) in 2.1–2.7 mg/dL, and 5.04 (95% CI 3.94–6.45) in ≥2.8 mg/dL compared to serum phosphate change group of 0–0.6 mg/dL. A similar result was noted when we further adjusted for either the admission or mean serum phosphate during hospitalization. Conclusion Greater serum phosphate changes were progressively associated with increased in-hospital mortality.
Collapse
Affiliation(s)
- Charat Thongprayoon
- Division of Nephrology and Hypertension, Department of Medicine, Mayo Clinic, 200 First St SW, Rochester, MN, 55905, USA
| | - Wisit Cheungpasitporn
- Division of Nephrology, Department of Internal Medicine, University of Mississippi Medical Center, Jackson, MS, USA
| | - Panupong Hansrivijit
- Department of Internal Medicine, University of Pittsburgh Medical Center Pinnacle, Harrisburg, PA, USA
| | - Sorkko Thirunavukkarasu
- Division of Nephrology and Hypertension, Department of Medicine, Mayo Clinic, 200 First St SW, Rochester, MN, 55905, USA
| | - Api Chewcharat
- Division of Nephrology and Hypertension, Department of Medicine, Mayo Clinic, 200 First St SW, Rochester, MN, 55905, USA
| | - Juan Medaura
- Division of Nephrology, Department of Internal Medicine, University of Mississippi Medical Center, Jackson, MS, USA
| | - Michael A Mao
- Division of Nephrology and Hypertension, Mayo Clinic, Jacksonville, FL, 32224, USA
| | - Kianoush B Kashani
- Division of Nephrology and Hypertension, Department of Medicine, Mayo Clinic, 200 First St SW, Rochester, MN, 55905, USA. .,Division of Pulmonary and Critical Care Medicine, Department of Medicine, Mayo Clinic, Rochester, MN, USA.
| |
Collapse
|