Next Article in Journal
A Case of Hepatic Immunoglobulin G4-Related Disease Presenting as an Inflammatory Pseudotumor and Sclerosing Cholangitis
Next Article in Special Issue
Optimization of Patient Management in the Gynecology Emergency Department Using Point-of-Care Beta hCG
Previous Article in Journal
Diagnostic Coding Intensity among a Pneumonia Inpatient Cohort Using a Risk-Adjustment Model and Claims Data: A U.S. Population-Based Study
Previous Article in Special Issue
The Pulmonary Artery Catheter in the Perioperative Setting: Should It Still Be Used?
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Diagnostics
Volume 12
Issue 6
10.3390/diagnostics12061496
diagnostics-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Explainable Machine Learning-Based Risk Prediction Model for In-Hospital Mortality after Continuous Renal Replacement Therapy Initiation

by
Pei-Shan Hung
1,
Pei-Ru Lin
2,
Hsin-Hui Hsu
1,
Yi-Chen Huang
3,
Shin-Hwar Wu
1 and
Chew-Teng Kor
2,4,*
1
Division of Critical Care Internal Medicine, Department of Emergency Medicine and Critical Care, Changhua Christian Hospital, Changhua 500, Taiwan
2
Big Data Center, Changhua Christian Hospital, Changhua 500, Taiwan
3
Department of Nursing, Changhua Christian Hospital, Changhua 500, Taiwan
4
Graduate Institute of Statistics and Information Science, National Changhua University of Education, Changhua 500, Taiwan
*
Author to whom correspondence should be addressed.
Diagnostics 2022, 12(6), 1496; https://doi.org/10.3390/diagnostics12061496
Submission received: 2 June 2022 / Revised: 16 June 2022 / Accepted: 17 June 2022 / Published: 19 June 2022
(This article belongs to the Special Issue Diagnostic Modalities in Critical Care)
Browse Figures
Review Reports Versions Notes

Abstract

In this study, we established an explainable and personalized risk prediction model for in-hospital mortality after continuous renal replacement therapy (CRRT) initiation. This retrospective cohort study was conducted at Changhua Christian Hospital (CCH). A total of 2932 consecutive intensive care unit patients receiving CRRT between 1 January 2010, and 30 April 2021, were identified from the CCH Clinical Research Database and were included in this study. The recursive feature elimination method with 10-fold cross-validation was used and repeated five times to select the optimal subset of features for the development of machine learning (ML) models to predict in-hospital mortality after CRRT initiation. An explainable approach based on ML and the SHapley Additive exPlanation (SHAP) and a local explanation method were used to evaluate the risk of in-hospital mortality and help clinicians understand the results of ML models. The extreme gradient boosting and gradient boosting machine models exhibited a higher discrimination ability (area under curve [AUC] = 0.806, 95% CI = 0.770–0.843 and AUC = 0.823, 95% CI = 0.788–0.858, respectively). The SHAP model revealed that the Acute Physiology and Chronic Health Evaluation II score, albumin level, and the timing of CRRT initiation were the most crucial features, followed by age, potassium and creatinine levels, SPO2, mean arterial pressure, international normalized ratio, and vasopressor support use. ML models combined with SHAP and local interpretation can provide the visual interpretation of individual risk predictions, which can help clinicians understand the effect of critical features and make informed decisions for preventing in-hospital deaths.

1. Introduction

Acute kidney injury (AKI) is common in critically ill patients in the intensive care unit (ICU). Multinational cross-sectional studies have reported that more than 50% of patients in the ICU have AKI regardless of the severity of AKI [1], and these patients are more likely to have higher mortality and long-term adverse outcomes, including cardiovascular complications, chronic kidney disease, and end-stage renal disease [2]. Continuous renal replacement therapy (CRRT) is the predominant form of renal replacement therapy (RRT), especially for patients with unstable hemodynamic status [3]. CRRT is a blood purification technology that removes excess water and toxins from blood to ensure fluid balance and electrolyte regulation required for organ support [4,5]. Despite advances in CRRT, the prognosis of patients receiving CRRT remains poor, with a mortality rate of 30% to 70% [6,7]. Furthermore, 25% of survivors receiving CRRT require dialysis upon hospital discharge [8]. Therefore, a tool for accurate risk prediction should be developed to determine the outcome of CRRT because a prediction tool provides valuable information that can help physicians make more effective clinical judgments and formulate appropriate treatment strategies as well as help family members make medical decisions.
The outcomes of CRRT depend on various clinical conditions, including patients’ underlying diseases, the severity and duration of renal impairment, renal function at baseline, and the presence of oliguria at the beginning of RRT [3,8,9,10]. Disease severity scores are commonly calculated using various tools, such as the Sequential Organ Failure Assessment (SOFA), Acute Physiology and Chronic Health Evaluation (APACHE) II, and Simplified Acute Physiology Score (SAPS), and are used to predict mortality in critically ill patients; however, these scores have limitations in patients with CRRT [11,12]. With advances in computational capacity in recent years, machine learning (ML) with the modeling of complex mathematical functions for many clinical variables has been used for accurate disease diagnosis and prognosis prediction. Two studies have used ML combined with electronic medical records (EMRs) to predict the mortality and RRT-free survival of critically ill patients receiving CRRT [7,13]. These models exhibited a higher prediction ability than did conventional disease severity scores. However, prediction accuracy of approximately 70% is not sufficient for clinical implementation, and interpretable risk prediction models are not yet available. Although an ML algorithm can indicate the probability of the prognosis of a disease, the algorithm cannot help physicians explain the logic of the decision to family members. Black-box artificial intelligence solutions in medicine without transparency and interpretability are still barriers [14]. To the best of our knowledge, no study has developed and examined an ML-based risk prediction model for CRRT.
Family members are concerned about whether the impaired renal function of patients with AKI admitted to the ICU can be restored or whether other adverse outcomes would occur after CRRT initiation. Even if patients are not optimistic, making the decision to discontinue CRRT is considerably challenging for patients and family members. Family members and physicians aim to provide the best possible care to prolong patients’ survival. When life-sustaining therapies cannot meet treatment goals, many clinical decisions, including those related to palliative care, should be discussed by physicians, patients, and families. The development of personalized and interpretable prediction models is crucial for identifying clinically relevant information hidden in large amounts of data and for creating a notification system that supports the work of physicians. In this study, we established an explainable and personalized risk prediction model for in-hospital mortality after CRRT initiation. This is the first study to use an ML algorithm combined with SHapley Additive Explanations (SHAP) and a local explanation method to provide clinical information to help physicians make better clinical judgments and customize treatment strategies.

2. Materials and Methods

2.1. Study Participants

This retrospective cohort study was conducted at Changhua Christian Hospital (CCH) in central Taiwan, which is a tertiary medical center with 130 ICU beds in five wards. We identified 3303 consecutive ICU patients who received CRRT between 1 January 2010, and 30 April 2021, from the CCH Clinical Research Database (CCHRD), which contains all electronic medical records including daily CRRT records; data on inpatient care, prescriptions, and clinical visits; laboratory results; and death records. AKI was diagnosed on the basis of Kidney Disease Improving Global Outcomes (KDIGO) criteria for serum creatinine elevation, which was determined by comparing baseline serum creatinine levels at admission with those before CRRT.
We excluded patients with preexisting end-stage renal disease receiving CRRT (n = 283), those aged <20 years (n = 15), and those with missing laboratory data (n = 73). Finally, 2932 eligible patients were included in this analysis (Figure 1). The data of the patients were randomly divided into the training set (80%, n = 2345) for developing ML models and the test set (20%, n = 587) for examining the performance of each model. The Institutional Review Board of CCH waived the requirement of informed consent and approved the study (IRB No: 210509). The CCHRD contains deidentified data. The researchers of this paper followed the Computer-Processed Personal Data Protection Law and privacy regulations in Taiwan.

2.2. Outcomes

The primary endpoint was in-hospital mortality after CRRT initiation that was determined by reviewing death records. The secondary endpoint was mortality at 28 and 90 days after CRRT initiation. Mortality risk was determined in the patients receiving CRRT, and the data of the patients were censored at the end of the respective follow-up periods (at discharge or at 28 and 90 days of follow-up) or at the last date for which data were available.

2.3. Feature Engineering

In our study, all features were performed according to the diagnostic criteria for AKI and risk factors associated with ICU mortality [6,7,13,15]. Data were selected based on the data available in our clinical research database, electronic medical records, and our areas of expertise. We collected information on 113 clinical features before CRRT initiation from the CCHRD, including demographics (age, sex, and body mass index [BMI]), APACHE II scores at admission to the ICU, diagnostic criteria of AKI, the timing of CRRT initiation, major diagnostic categories (MDCs), vital signs within 24 h after CRRT initiation, fluid balance, medication use, multiple organ support, serum biochemical data before CRRT, and comorbidities. Laboratory data were used as features for predicting CRRT mortality in the data mining phase, including complete blood count, blood gas analysis, white blood cell (WBC) and differential counts, and all biochemical data such as renal function, lipid profile, nutrition-related tests, etc. We excluded 24 features with a prevalence of <5% from the analysis to ensure variability and improve prediction precision [16]. To select the most favorable subset of features for predicting hospital mortality, we used the recursive feature elimination method with 10-fold cross-validation, repeated it five times, and finally selected 29 features. Table 1 presents the statistical descriptions of the selected features between the training and test sets.

2.4. Statistical Analysis and ML Algorithm

Categorical and continuous variables are expressed as numbers (proportions) and medians and interquartile ranges, respectively. The chi-square test was used to compare categorical variables, and the Mann–Whitney U test was used to compare continuous variables.
Figure 2 presents the study framework used for developing the prediction model for in-hospital mortality after CRRT initiation, which included data preprocessing, feature engineering, ML models, and models in global and local interpretation. Four ML algorithms, namely support vector machine (SVM), random forest (RF), gradient boosting machine (GBM), and extreme gradient boosting (XGB) [17], were used to develop a model for predicting in-hospital mortality after CRRT. To select the most satisfactory combination of model hyperparameters, the exhaustive grid search algorithm was used as a hyperparameter tuning tool, and a 5-fold cross-validation procedure was used for the training set. The most favorable hyperparameter for each ML model was determined on the basis of the highest area under the receiver operating characteristic curve (AUROC). ML models were developed using the one-hot encoding of categories and the normalization of continuous features. The Youden index was used to determine the optimal threshold to classify in-hospital deaths that occurred after CRRT initiation. Six evaluation metrics, namely sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), F1 score, and accuracy of the test set, were used to compare the prediction ability of each model. Because the study aim was predicting the occurrence of in-hospital mortality after CRRT initiation, we mainly focused on achieving a higher prediction accuracy and F1 score.
A raincloud plot including individual data points (horizontally jittered), density distributions, and statistical inferences of box plots with medians and interquartile ranges was created; this approach enabled the visualization of the distribution of in-hospital mortality after CRRT initiation and predicted scores obtained using ML models. We evaluated the agreement between the predicted probability and observed in-hospital mortality after CRRT initiation by using calibration belts. Kaplan–Meier curves were used to predict the survival rate at 28 days and 90 days.
SHAP is an explainable artificial intelligence technique that helps clinicians understand the results of ML models. Two advantages of SHAP increases the transparency of the model in providing global and local interpretability. Global interpretability lists the most important features in descending order, with the top-ranked features contributing more to the predictive model and having high predictive power. The SHAP value plot can further show the positive and negative relationship between the feature and the outcome variable. Local interpretability provides feature contributions to the model prediction output for an individual patient, which reveal the impact of input features on individual predictions. For clinical practice applications, local interpretability enables physicians to understand key features that affect a patient’s condition and accordingly implement treatment strategies to save the lives of patients receiving CRRT. In this study, all statistical analyses were performed using SPSS, and the ML model was established using R software (version 3.6.2; The Comprehensive R Archive Network: http://cran.r-project.org, accessed on 12 December 2019). All two-sided p values of <.05 indicated statistical significance.

3. Results

3.1. Study Population Characteristics

This study included a total of 2932 patients receiving CRRT, of whom 2024 (60.03%) died in the hospital. The patients who died in the hospital were older and had lower BMI, higher APACHE-II scores, a higher frequency of receiving CRRT > 24 h after AKI, more unstable vital signs (lower blood pressure and SPO2 and higher respiratory rate), and more fluid imbalance before CRRT initiation. Moreover, the patients who died in the hospital were more likely to require multiple organ support before CRRT initiation (including invasive mechanical ventilation, vasopressor support, and vasopressin) and medications (corticosteroids, parenteral nutrition, antibiotics, and furosemide). Moreover, these patients had higher serum potassium, lactate, sodium, and magnesium levels; red blood cell distribution width (RDW); international normalized ratio (INR); and activated partial thromboplastin time (APTT). However, these patients had lower serum creatinine and albumin levels, platelet counts, pH values, and O2 saturation. The prevalence of in-hospital death and the distributions of the features were similar between the training and test sets, except for RDW and APTT. Table 1 lists the features used to develop the ML model.

3.2. Model Prediction of In-Hospital Death after CRRT Initiation

As presented in Table 2, the SVM model exhibited a moderate discrimination ability (AUC = 0.750, 95% CI = 0.708–0.792 for the radial kernel; AUC = 0.756, 95% CI = 0.745–0.821 for the sigmoid kernel; and AUC = 0.784, 95% CI = 0.745–0.822 for the polynomial kernel). The RF, XGB, and GBM models exhibited a higher discrimination ability (AUC = 0.816, 95% CI = 0.781–0.851; AUC = 0.806, 95% CI = 0.770–0.843; and AUC = 0.823, 95% CI = 0.788–0.858, respectively). Among the 3 ML models, the RF model exhibited the highest sensitivity (74.69%); however, its specificity and PPV for predicting in-hospital death (75% and 86.38%, respectively), F1 score (80.11%), and accuracy (74.79%) were lower than those of the GBM and XGB models. By contrast, the XGB and GBM models more accurately predicted in-hospital death, with sensitivity of 73.43% and 74.19%, specificity of 80.32% and 78.72%, and PPV of 88.79% and 88.10%, respectively; their F1 scores (80.38% and 80.55%, respectively) and prediction accuracy were the highest (75.64% for both, respectively). Figure 3 presents the discrimination performance of each of the six ML models, as represented by the receiver operating characteristic curve.
The raincloud plot depicted in Figure 4 summarizes the distribution of the predicted scores for the in-hospital death and non–in-hospital death groups. Significant differences were noted in the predicted scores between the groups in the six ML models (all p < 0.001, Kolmogorov–Smirnov test), and the median score of in-hospital death was higher than that of non-hospital death (all p < 0.001, Wilcoxon rank-sum test). Furthermore, the XGB model overlapped on the 45° dotted line, indicating favorable agreement between predicted probabilities in ML and the actual results determined using the calibration plots (p = 0.122 for the XGB model; Figure 5). The Kaplan–Meier curves showed poor survival at 28 and 90 days in the high-risk group (Figure 6).

3.3. Model Explanations

Figure 7 presents the plot of the features vital for the XGB model in order of importance according to average absolute SHAP values, which were useful for determining the contribution of each feature to individual predictions. The APACHE II score, albumin level, and timing of CRRT initiation were the top three main features, with SHAP values of 0.300, 0.297, and 0.276, respectively (Figure 7a), followed by age, potassium levels, SPO2, mean arterial pressure (MAP), INR, creatinine levels, and vasopressor use. Higher APACHE II scores; potassium, magnesium, lactate, and sodium levels; MDC scores; RDW; APTT; fluid balance; respiratory rate; and older age as well as lower SPO2, MAP, BMI, platelet count, creatinine levels, and O2 saturation were positively correlated with in-hospital death (Figure 7b). Multiple organ support use (vasopressor, vasopressin, and invasive mechanical ventilation) and medication use (corticosteroids, parenteral nutrition, antibiotics, and furosemide) were positively associated with in-hospital death. The presence of diabetes mellitus was negatively associated with in-hospital death. A nonlinear relationship was observed between pH and in-hospital mortality. Although the distribution of SHAP values was highly dispersed, the correlation of the features with in-hospital death was still consistent with the domain knowledge of most of the features.
Figure 8 depicts four local explanation graphs of randomly chosen patients. Our proposed model correctly predicted the risk of in-hospital death for patients A and B but incorrectly predicted the risk for patients C and D.
Patient A died in the hospital, and the probability of in-hospital death predicted by the ML model for patient A was 91%. Patient A had lower SPO2 (88.93), higher pH (7.21), CRRT initiated >24 h after AKI, a higher sodium level (158), lower O2 saturation (85.3), a lower creatinine level (0.83), and a higher MDC score (24); these factors were significantly positively associated with in-hospital death (Figure 8a). Patient B did not die in the hospital, and the probability of in-hospital death predicted by the ML model for patient B was 36.54%. Patient B was aged 63 years; did not require parenteral nutrition; and had a higher albumin level (3.8), higher MAP (124.36), CRRT initiated within 24 h after AKI, higher BMI (25.14), higher SPO2 (95.64), and normal laboratory values (RDW; APTT; INR; platelet count; pH; and sodium, magnesium, and potassium levels); these factors were associated with a lower risk of in-hospital death (Figure 8b).
The probabilities of in-hospital death predicted using the ML model for patients C and D were inaccurate. Local explanations are presented in Figure 8c,d. The probabilities of in-hospital death for patients C and D were 48.2% and 74.4%, respectively. Despite inaccurate predictions, local interpretations can help clinicians effectively understand patients’ conditions and address key clinical features early on to modify treatment strategies and save lives. Similar results were obtained for the GBM and RF model (presented in Figure A1, Figure A2, Figure A3 and Figure A4), indicating that both the global and local interpretations of the ML model can help in making personalized care recommendations to prevent in-hospital death.

4. Discussion

In this study, we developed and validated an interpretable ML-based model to predict the in-hospital death of ICU patients receiving CRRT. Our findings revealed that the XGB, GBM, and RF models exhibited higher discrimination performance with an AUC of 0.80. The result indicated that ML can improve the prediction of in-hospital mortality in ICU patients receiving CRRT. Among the 3 ML models, the XGB model exhibited the most favorable performance and calibration; therefore, we used the XGB model to develop the interpretable ML-based in-hospital death risk prediction model. The XGB algorithm that combines SHAP and the local interpretation framework provides a visual feature importance score that can help physicians intuitively understand the key features of patients’ condition and accordingly modify treatment strategies and explain the condition to the patients’ families.
CRRT is a primary life-support technique used to maintain hemodynamic stability and aid in renal recovery in ICU patients [18]. However, higher mortality rates, ranging from 40% to 70%, have been reported in patients receiving CRRT [6,7]. Therefore, the main objective of CRRT is to prevent mortality and adverse outcomes. APACHE and SOFA scores, which are calculated using clinical and laboratory parameters, are widely used to estimate the risk of mortality in ICU patients [19,20]. However, the efficacy of these scores has not been adequately evaluated in patients receiving CRRT [21]. The accuracy of the newly developed mortality scoring system for AKI with CRRT in predicting mortality in patients receiving CRRT was only evaluated in a small study [12]. A large comprehensive medical database study from Taiwan validated these scoring systems in ICU patients receiving CRRT. The results indicated that the prediction performance of these scoring systems based on scores on days 1 and 3 of CRRT in terms of AUC ranged from 0.55 to 0.67. The average AUC of the discrimination ability was 0.74 on day 7 after CRRT initiation [21].
The use of ML models for data analysis has recently become popular due to a large amount of data in electronic hospital records. Furthermore, ML models can accurately predict mortality in ICU patients receiving CRRT. For example, Kang and colleagues developed an ML-based model to predict mortality in patients receiving CRRT [7]. They reported that the RF model was the most favorable prediction model with an AUROC of 0.782 and observed that the accuracy of an ML-based model in predicting mortality was higher than that of other scoring models. The same research team used an ML algorithm to predict hypotension after CRRT initiation and reported satisfactory performance of the XGB model, with an AUROC of 0.828 [22]. Pattharanitima et al. used ML and deep learning to build a model for predicting renal replacement therapy-free survival and demonstrated that the long short-term memory model with multilayer perceptron architecture exhibited high discrimination performance with an AUC of 0.70 [13]. Studies have investigated the accuracy of other ML-based mortality prediction models in intensive care settings, including in patients with lactic acidosis [23], mechanically ventilated patients [24], and those with COVID-19 and AKI [25]. These findings emphasize the importance of ML in predicting outcomes in critical care settings, especially for patients receiving CRRT.
In this study, we combined ML-based models with electronic medical record data to retrieve information on various clinical characteristics that affect in-hospital mortality, and we determined that the RF, XGB, and GBM models exhibited the most favorable discriminative ability with AUROCs of 0.816, 0.806, and 0.823, respectively. The accuracy of the ML models used in this study in predicting in-hospital mortality in the patients receiving CRRT is comparable to that reported in previous studies [7,10,22,23,24,25]. In particular, our model exhibited higher discriminative power than did the prediction model developed by Kang et al. [7]. In our study, the AUC of the XGB model was not higher than that of the GBM model; however, the PPV, F1 score and accuracy of the XGB model were comparable to those of the GBM model. Moreover, the XGB model was more well-calibrated than the GBM model. Thus, the GBM and XGB models may be suitable to predict in-hospital mortality in patients receiving CRRT.
Currently, ML models used to predict mortality in patients receiving CRRT still employ the black box methodology, which involves determining the relationship between given data (inputs) and outcomes (outputs) instead of creating rules based on knowledge. This study provides new insights into the stumbling block of black-box logic and an explainable approach to artificial intelligence. SHAP reported by Lundberg and Lee (2016) explains how different features affect the results of ML models [26]. SHAP is based on the theoretical optimal Shapley value of a game, a widely used concept in cooperative game theory. The feature values of data instances act as players in the coalition, and Shapley values are the average marginal contributions of features across all possible coalitions. Artificial intelligence should be applied in clinical practice to help clinicians understand the rationale of the predictions of ML models. In addition, providing detailed information can enable doctors to gain better insights and make informed decisions regarding preventing in-hospital deaths instead of blindly trusting or not trusting predictions. Moreover, this information can help physicians explain patients’ condition to the family to ensure transparent decision-making.
We used SHAP values and local interpretation to achieve the best prediction performance and interpretability. SHAP technology has been applied to many clinical problems, including acute exacerbations of chronic obstructive pulmonary disease, venous thrombosis of osteoarthritis, and coronary artery calcification [27,28,29]. Recent studies have applied explainable machine learning to predict hospital mortality in publicly available databases, such as the eICU and MIMIC-III databases [30,31]. To our knowledge, this is the first study to use SHAP in EMRs data to predict in-hospital mortality in ICU patients receiving CRRT. We graphically demonstrated the interpretability of the complicated XBG and GBM models by plotting individual risk prediction.
In our study, the model revealed that the APACHE II score, albumin level, and the timing of CRRT initiation are the three most crucial global features with the highest SHAP values, followed by age, potassium and creatinine levels, SPO2, MAP, INR, and vasopressor support use. Furthermore, the results revealed that a higher APACHE II score, potassium levels, lactate levels, sodium levels, and respiratory rate; multiple organ support use and medication use; and older age as well as lower SPO2, MAP, platelet count, BMI, and O2 saturation were positively correlated with in-hospital death. The findings regarding the importance of variables are consistent with those of a previous study [7,11,12,32,33,34]. In this study, an inverse relationship was noted between diabetes mellitus and in-hospital mortality. A lower creatinine level was associated with in-hospital mortality, whereas a nonlinear relationship was noted between pH and in-hospital mortality. These findings are in line with those of previous studies. The reason for lower mortality in patients with preexisting DM remains unclear. Patients with diabetes might have improved tolerance to acute hyperglycemic episodes during critical illness [35]. Although an elevated serum creatinine level indicates the severity of AKI, a meta-analysis suggested that patients with a lower serum creatinine level at the beginning of CRRT may have an increased risk of in-hospital death [18]. In surgical ICU, mortality was observed to have a U-shaped relationship with pH, with acidemia and alkalemia being independently associated with high mortality [36]. A U-shaped association of pH was observed in patients with CRRT. Future studies should investigate the effect of biochemical laboratory data, including creatinine and pH levels, on in-hospital mortality in patients receiving CRRT.
This study has some limitations. First, the use of data from a single-center electronic database may limit the generalizability of our model, and our findings cannot be applied directly to other medical institutions. Second, we used retrospective data for model construction and validation in this study; thus, some potential risks were not considered. Additional prospective experiments should be performed to validate the prediction model. Third, all models included static features, including demographics, medication use, and multiorgan support use, obtained from the time of admission to the start of CRRT. Time-series data, such as laboratory values, vital signs, and urine output, from ICU admission to CRRT initiation were included in the model as mean values. Future studies should use deep learning including various time-series data for predicting in-hospital mortality. Fourth, the sample size of this study population is moderate; however, the sample size of previous studies on related topics using machine learning to analyze CRRT data is approximately 684 to 2349 [7,13,22], which is smaller than our sample size.
In this study, the XGB and GBM models were used to accurately evaluate the risk of in-hospital mortality in ICU patients receiving CRRT. The model involving the use of SHAP and local interpretation provided visually interpretable individual risk predictions, which may help clinicians understand the effect of key features and make informed decisions regarding preventing in-hospital deaths. In addition, explaining the reasons for the treatment decision to the patient’s family can increase transparency. In conclusion, our model provides objective and interpretable predictions that can help clinicians implement appropriate treatment for patients receiving CRRT based on their specific prognosis.

Author Contributions

Conceptualization, P.-S.H. and C.-T.K.; methodology, C.-T.K.; software, P.-R.L. and C.-T.K.; validation, P.-R.L. and C.-T.K.; formal analysis, P.-R.L. and C.-T.K.; investigation, Y.-C.H.; resources, Y.-C.H. and H.-H.H.; data curation, P.-R.L.; writing—original draft preparation, P.-S.H. and C.-T.K.; writing—review and editing, S.-H.W., H.-H.H., and C.-T.K.; visualization, P.-R.L.; supervision, C.-T.K.; project administration, C.-T.K.; funding acquisition, C.-T.K. All authors have read and agreed to the published version of the manuscript.

Funding

This research was funded by Changhua Christian Hospital, grant number 111-CCH-IRP-016.

Institutional Review Board Statement

The study was conducted according to the guidelines of the Declaration of Helsinki and approved by the Institutional Review Board (or Ethics Committee) of Changhua Christian Hospital (protocol code 210509) at 4 June 2021.

Informed Consent Statement

Not applicable.

Data Availability Statement

Not applicable.

Acknowledgments

The authors would like to thank the Changhua Christian Hospital Big Data Center for their research collaboration and the staff working in the field.

Conflicts of Interest

The authors declare no conflict of interest.

Appendix A

Diagnostics 12 01496 g0a1
Figure A1. Summary SHapley Additive exPlanations (SHAP) plot. (a) Global feature importance in final GBM model output. (b) Relationship between features and in-hospital mortality in GBM model. Diversion on x-axis represents effect on model output, with colors used to represent low (yellow) to high (purple) value of predictors.
Figure A1. Summary SHapley Additive exPlanations (SHAP) plot. (a) Global feature importance in final GBM model output. (b) Relationship between features and in-hospital mortality in GBM model. Diversion on x-axis represents effect on model output, with colors used to represent low (yellow) to high (purple) value of predictors.
Diagnostics 12 01496 g0a1

Appendix B

Diagnostics 12 01496 g0a2aDiagnostics 12 01496 g0a2b
Figure A2. Local explanation plots for individuals with various in-hospital mortality statuses and GBM model predictions. (a) In-hospital death and AI predicted in-hospital death; (b) non-in-hospital death and AI predicted non-in-hospital death; (c) in-hospital death but AI predicted non-in-hospital death; (d) non-in-hospital death but AI predicted in-hospital death.
Figure A2. Local explanation plots for individuals with various in-hospital mortality statuses and GBM model predictions. (a) In-hospital death and AI predicted in-hospital death; (b) non-in-hospital death and AI predicted non-in-hospital death; (c) in-hospital death but AI predicted non-in-hospital death; (d) non-in-hospital death but AI predicted in-hospital death.
Diagnostics 12 01496 g0a2aDiagnostics 12 01496 g0a2b

Appendix C

Diagnostics 12 01496 g0a3aDiagnostics 12 01496 g0a3b
Figure A3. Summary SHapley Additive exPlanations (SHAP) plot. (a) Global feature importance in final random forest model output. (b) Relationship between features and in-hospital mortality in random forest model. Diversion on x-axis represents effect on model output, with colors used to represent low (yellow) to high (purple) value of predictors.
Figure A3. Summary SHapley Additive exPlanations (SHAP) plot. (a) Global feature importance in final random forest model output. (b) Relationship between features and in-hospital mortality in random forest model. Diversion on x-axis represents effect on model output, with colors used to represent low (yellow) to high (purple) value of predictors.
Diagnostics 12 01496 g0a3aDiagnostics 12 01496 g0a3b

Appendix D

Diagnostics 12 01496 g0a4aDiagnostics 12 01496 g0a4b
Figure A4. Local explanation plots for individuals with various in-hospital mortality statuses and random forest model predictions. (a) In-hospital death and AI predicted in-hospital death; (b) non-in-hospital death and AI predicted non-in-hospital death; (c) in-hospital death but AI predicted non-in-hospital death; (d) non-in-hospital death but AI predicted in-hospital death.
Figure A4. Local explanation plots for individuals with various in-hospital mortality statuses and random forest model predictions. (a) In-hospital death and AI predicted in-hospital death; (b) non-in-hospital death and AI predicted non-in-hospital death; (c) in-hospital death but AI predicted non-in-hospital death; (d) non-in-hospital death but AI predicted in-hospital death.
Diagnostics 12 01496 g0a4aDiagnostics 12 01496 g0a4b

References

  1. Hoste, E.A.; Bagshaw, S.M.; Bellomo, R.; Cely, C.M.; Colman, R.; Cruz, D.N.; Edipidis, K.; Forni, L.G.; Gomersall, C.D.; Govil, D.; et al. Epidemiology of acute kidney injury in critically ill patients: The multinational AKI-EPI study. Intensive Care Med. 2015, 41, 1411–1423. [Google Scholar] [CrossRef] [PubMed]
  2. Hoste, E.A.J.; Kellum, J.A.; Selby, N.M.; Zarbock, A.; Palevsky, P.M.; Bagshaw, S.M.; Goldstein, S.L.; Cerdá, J.; Chawla, L.S. Global epidemiology and outcomes of acute kidney injury. Nat. Rev. Nephrol. 2018, 14, 607–625. [Google Scholar] [CrossRef] [PubMed]
  3. Griffin, B.R.; Liu, K.D.; Teixeira, J.P. Critical Care Nephrology: Core Curriculum 2020. Am. J. Kidney Dis. 2020, 75, 435–452. [Google Scholar] [CrossRef] [PubMed]
  4. Macedo, E.; Mehta, R.L. Continuous Dialysis Therapies: Core Curriculum 2016. Am. J. Kidney Dis. 2016, 68, 645–657. [Google Scholar] [CrossRef]
  5. Karkar, A.; Ronco, C. Prescription of CRRT: A pathway to optimize therapy. Ann. Intensive Care 2020, 10, 32. [Google Scholar] [CrossRef]
  6. Kee, Y.K.; Kim, D.; Kim, S.J.; Kang, D.H.; Choi, K.B.; Oh, H.J.; Ryu, D.R. Factors Associated with Early Mortality in Critically Ill Patients Following the Initiation of Continuous Renal Replacement Therapy. J. Clin. Med. 2018, 7, 334. [Google Scholar] [CrossRef]
  7. Kang, M.W.; Kim, J.; Kim, D.K.; Oh, K.H.; Joo, K.W.; Kim, Y.S.; Han, S.S. Machine learning algorithm to predict mortality in patients undergoing continuous renal replacement therapy. Crit. Care 2020, 24, 42. [Google Scholar] [CrossRef]
  8. Palevsky, P.M.; Zhang, J.H.; O’Connor, T.Z.; Chertow, G.M.; Crowley, S.T.; Choudhury, D.; Finkel, K.; Kellum, J.A.; Paganini, E.; Schein, R.M.; et al. Intensity of renal support in critically ill patients with acute kidney injury. N. Engl. J. Med. 2008, 359, 7–20. [Google Scholar] [CrossRef]
  9. De Corte, W.; Dhondt, A.; Vanholder, R.; De Waele, J.; Decruyenaere, J.; Sergoyne, V.; Vanhalst, J.; Claus, S.; Hoste, E.A. Long-term outcome in ICU patients with acute kidney injury treated with renal replacement therapy: A prospective cohort study. Crit. Care 2016, 20, 256. [Google Scholar] [CrossRef]
  10. Singbartl, K.; Kellum, J.A. AKI in the ICU: Definition, epidemiology, risk stratification, and outcomes. Kidney Int. 2012, 81, 819–825. [Google Scholar] [CrossRef]
  11. da Hora Passos, R.; Ramos, J.G.; Mendonça, E.J.; Miranda, E.A.; Dutra, F.R.; Coelho, M.F.; Pedroza, A.C.; Correia, L.C.; Batista, P.B.; Macedo, E.; et al. A clinical score to predict mortality in septic acute kidney injury patients requiring continuous renal replacement therapy: The HELENICC score. BMC Anesthesiol. 2017, 17, 21. [Google Scholar] [CrossRef] [PubMed]
  12. Kim, Y.; Park, N.; Kim, J.; Kim, D.K.; Chin, H.J.; Na, K.Y.; Joo, K.W.; Kim, Y.S.; Kim, S.; Han, S.S. Development of a new mortality scoring system for acute kidney injury with continuous renal replacement therapy. Nephrology 2019, 24, 1233–1240. [Google Scholar] [CrossRef] [PubMed]
  13. Pattharanitima, P.; Vaid, A.; Jaladanki, S.K.; Paranjpe, I.; O’Hagan, R.; Chauhan, K.; Van Vleck, T.T.; Duffy, A.; Chaudhary, K.; Glicksberg, B.S.; et al. Comparison of Approaches for Prediction of Renal Replacement Therapy-Free Survival in Patients with Acute Kidney Injury. Blood Purif. 2021, 50, 621–627. [Google Scholar] [CrossRef] [PubMed]
  14. Kundu, S. AI in medicine must be explainable. Nat. Med. 2021, 27, 1328. [Google Scholar] [CrossRef]
  15. Jiang, Z.; Bo, L.; Xu, Z.; Song, Y.; Wang, J.; Wen, P.; Wan, X.; Yang, T.; Deng, X.; Bian, J. An explainable machine learning algorithm for risk factor analysis of in-hospital mortality in sepsis survivors with ICU readmission. Comput. Methods Programs Biomed. 2021, 204, 106040. [Google Scholar] [CrossRef]
  16. Ogundimu, E.O.; Altman, D.G.; Collins, G.S. Adequate sample size for developing prediction models is not simply related to events per variable. J. Clin. Epidemiol. 2016, 76, 175–182. [Google Scholar] [CrossRef]
  17. Chen, T.; Guestrin, C. XGBoost: A Scalable Tree Boosting System. In Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, San Francisco, CA, USA, 13–17 August 2016; pp. 785–794. [Google Scholar]
  18. Lee, H.J.; Son, Y.J. Factors Associated with In-Hospital Mortality after Continuous Renal Replacement Therapy for Critically Ill Patients: A Systematic Review and Meta-Analysis. Int. J. Environ. Res. Public Health 2020, 17, 8781. [Google Scholar] [CrossRef]
  19. Knaus, W.A.; Draper, E.A.; Wagner, D.P.; Zimmerman, J.E. APACHE II: A severity of disease classification system. Crit. Care Med. 1985, 13, 818–829. [Google Scholar] [CrossRef]
  20. Vincent, J.L.; Moreno, R.; Takala, J.; Willatts, S.; De Mendonça, A.; Bruining, H.; Reinhart, C.K.; Suter, P.M.; Thijs, L.G. The SOFA (Sepsis-related Organ Failure Assessment) score to describe organ dysfunction/failure. On behalf of the Working Group on Sepsis-Related Problems of the European Society of Intensive Care Medicine. Intensive Care Med. 1996, 22, 707–710. [Google Scholar] [CrossRef]
  21. Yen, C.L.; Fan, P.C.; Kuo, G.; Lee, C.C.; Chen, J.J.; Lee, T.H.; Tu, Y.R.; Hsu, H.H.; Tian, Y.C.; Chang, C.H. Prognostic Performance of Existing Scoring Systems among Critically Ill Patients Requiring Continuous Renal Replacement Therapy: An Observational Study. J. Clin. Med. 2021, 10, 4592. [Google Scholar] [CrossRef]
  22. Kang, M.W.; Kim, S.; Kim, Y.C.; Kim, D.K.; Oh, K.H.; Joo, K.W.; Kim, Y.S.; Han, S.S. Machine learning model to predict hypotension after starting continuous renal replacement therapy. Sci. Rep. 2021, 11, 17169. [Google Scholar] [CrossRef] [PubMed]
  23. Pattharanitima, P.; Thongprayoon, C.; Kaewput, W.; Qureshi, F.; Qureshi, F.; Petnak, T.; Srivali, N.; Gembillo, G.; O’Corragain, O.A.; Chesdachai, S.; et al. Machine Learning Prediction Models for Mortality in Intensive Care Unit Patients with Lactic Acidosis. J. Clin. Med. 2021, 10, 5021. [Google Scholar] [CrossRef] [PubMed]
  24. Kim, J.H.; Kwon, Y.S.; Baek, M.S. Machine Learning Models to Predict 30-Day Mortality in Mechanically Ventilated Patients. J. Clin. Med. 2021, 10, 2172. [Google Scholar] [CrossRef] [PubMed]
  25. Ponce, D.; de Andrade, L.G.M.; Claure-Del Granado, R.; Ferreiro-Fuentes, A.; Lombardi, R. Development of a prediction score for in-hospital mortality in COVID-19 patients with acute kidney injury: A machine learning approach. Sci. Rep. 2021, 11, 24439. [Google Scholar] [CrossRef]
  26. Lundberg, S.M.; Lee, S.-I. A unified approach to interpreting model predictions. In Proceedings of the 31st International Conference on Neural Information Processing Systems, Long Beach, CA, USA, 4–9 December 2017; pp. 4768–4777. [Google Scholar]
  27. Kor, C.T.; Li, Y.R.; Lin, P.R.; Lin, S.H.; Wang, B.Y.; Lin, C.H. Explainable Machine Learning Model for Predicting First-Time Acute Exacerbation in Patients with Chronic Obstructive Pulmonary Disease. J. Pers. Med. 2022, 12, 228. [Google Scholar] [CrossRef]
  28. Oh, T.R.; Song, S.H.; Choi, H.S.; Suh, S.H.; Kim, C.S.; Jung, J.Y.; Choi, K.H.; Oh, K.H.; Ma, S.K.; Bae, E.H.; et al. Predictive Model for High Coronary Artery Calcium Score in Young Patients with Non-Dialysis Chronic Kidney Disease. J. Pers. Med. 2021, 11, 1372. [Google Scholar] [CrossRef]
  29. Lu, C.; Song, J.; Li, H.; Yu, W.; Hao, Y.; Xu, K.; Xu, P. Predicting Venous Thrombosis in Osteoarthritis Using a Machine Learning Algorithm: A Population-Based Cohort Study. J. Pers. Med. 2022, 12, 114. [Google Scholar] [CrossRef]
  30. Viton, F.; Elbattah, M.; Guérin, J.L.; Dequen, G. Heatmaps for Visual Explainability of CNN-Based Predictions for Multivariate Time Series with Application to Healthcare. In Proceedings of the 2020 IEEE International Conference on Healthcare Informatics (ICHI), Oldenburg, Germany, 30 November–3 December 2020; pp. 1–8. [Google Scholar]
  31. Stenwig, E.; Salvi, G.; Rossi, P.S.; Skjærvold, N.K. Comparative analysis of explainable machine learning prediction models for hospital mortality. BMC Med. Res. Methodol. 2022, 22, 53. [Google Scholar] [CrossRef]
  32. Mao, I.-C.; Lin, P.-R.; Wu, S.-H.; Hsu, H.-H.; Hung, P.-S.; Kor, C.-T. First 24-Hour Potassium Concentration and Variability and Association with Mortality in Patients Requiring Continuous Renal Replacement Therapy in Intensive Care Units: A Hospital-Based Retrospective Cohort Study. J. Clin. Med. 2022, 11, 3383. [Google Scholar] [CrossRef]
  33. Sheng, S.; Zhang, Y.H.; Ma, H.K.; Huang, Y. Albumin levels predict mortality in sepsis patients with acute kidney injury undergoing continuous renal replacement therapy: A secondary analysis based on a retrospective cohort study. BMC Nephrol. 2022, 23, 52. [Google Scholar] [CrossRef]
  34. Schork, A.; Moll, K.; Haap, M.; Riessen, R.; Wagner, R. Course of lactate, pH and base excess for prediction of mortality in medical intensive care patients. PLoS ONE 2021, 16, e0261564. [Google Scholar] [CrossRef] [PubMed]
  35. Hu, C.; Hu, B.; Wang, J.; Peng, Z.; Kashani, K.B.; Li, J. Diabetes mellitus is associated with lower mortality in intensive care unit: A multicenter historical cohort study. Res. Sq. 2020. [Google Scholar] [CrossRef]
  36. Rajawat, M.S.; Rathore, S.; Choudhary, M. pH disorders and mortality in surgical intensive care unit patients. Int. Surg. J. 2016, 3, 905–907. [Google Scholar] [CrossRef][Green Version]
Diagnostics 12 01496 g001
Figure 1. Flow chart of study patients.
Figure 1. Flow chart of study patients.
Diagnostics 12 01496 g001
Diagnostics 12 01496 g002
Figure 2. Study framework and feature engineering.
Figure 2. Study framework and feature engineering.
Diagnostics 12 01496 g002
Diagnostics 12 01496 g003
Figure 3. Receiver operation characteristic curves of the models for predicting in-hospital mortality.
Figure 3. Receiver operation characteristic curves of the models for predicting in-hospital mortality.
Diagnostics 12 01496 g003
Diagnostics 12 01496 g004aDiagnostics 12 01496 g004b
Figure 4. The raincloud plot of in-hospital mortality predicted score in machine learning methods. (a) GBM, XGB and random forest, (b) SVM with radial kernel, SVM with polynomial kernel, and SVM with sigmoid kernel.
Figure 4. The raincloud plot of in-hospital mortality predicted score in machine learning methods. (a) GBM, XGB and random forest, (b) SVM with radial kernel, SVM with polynomial kernel, and SVM with sigmoid kernel.
Diagnostics 12 01496 g004aDiagnostics 12 01496 g004b
Diagnostics 12 01496 g005aDiagnostics 12 01496 g005b
Figure 5. Calibration of machine learning models for predicting in-hospital mortality with calibration belts. (a) GBM, (b) XGB, (c) random forest, (d) SVM with radial kernel, (e) SVM with polynomial kernel, and (f) SVM with sigmoid kernel.
Figure 5. Calibration of machine learning models for predicting in-hospital mortality with calibration belts. (a) GBM, (b) XGB, (c) random forest, (d) SVM with radial kernel, (e) SVM with polynomial kernel, and (f) SVM with sigmoid kernel.
Diagnostics 12 01496 g005aDiagnostics 12 01496 g005b
Diagnostics 12 01496 g006
Figure 6. The Kaplan–Meier curve for predicting the 28-day and 90-day mortality. (a) 28-day mortality, (b) 90-day mortality.
Figure 6. The Kaplan–Meier curve for predicting the 28-day and 90-day mortality. (a) 28-day mortality, (b) 90-day mortality.
Diagnostics 12 01496 g006
Diagnostics 12 01496 g007aDiagnostics 12 01496 g007b
Figure 7. Summary SHapley Additive exPlanations (SHAP) plot. (a) Global feature importance in final XGBoost model output. (b) Relationship between features and in-hospital mortality in XGBoost model. Diversion on x-axis represents effect on model output, with colors used to represent low (yellow) to high (purple) value of predictors.
Figure 7. Summary SHapley Additive exPlanations (SHAP) plot. (a) Global feature importance in final XGBoost model output. (b) Relationship between features and in-hospital mortality in XGBoost model. Diversion on x-axis represents effect on model output, with colors used to represent low (yellow) to high (purple) value of predictors.
Diagnostics 12 01496 g007aDiagnostics 12 01496 g007b
Diagnostics 12 01496 g008aDiagnostics 12 01496 g008b
Figure 8. Local explanation plots for individuals with various in-hospital mortality statuses and XGB model predictions. (a) In-hospital death and AI predicted in-hospital death; (b) non-in-hospital death and AI predicted non-in-hospital death; (c) in-hospital death but AI predicted non-in-hospital death; (d) non-in-hospital death but AI predicted in-hospital death. Green and red bars correspond to the contribution of the features to the prediction. Green represents a negative value, which decreases the predicted value; red represents a positive value, which increases the predicted value. X-axis represents model prediction value; y-axis lists the features and their observed values.
Figure 8. Local explanation plots for individuals with various in-hospital mortality statuses and XGB model predictions. (a) In-hospital death and AI predicted in-hospital death; (b) non-in-hospital death and AI predicted non-in-hospital death; (c) in-hospital death but AI predicted non-in-hospital death; (d) non-in-hospital death but AI predicted in-hospital death. Green and red bars correspond to the contribution of the features to the prediction. Green represents a negative value, which decreases the predicted value; red represents a positive value, which increases the predicted value. X-axis represents model prediction value; y-axis lists the features and their observed values.
Diagnostics 12 01496 g008aDiagnostics 12 01496 g008b
Table 1. Patients’ features in overall cohort and split datasets.
Table 1. Patients’ features in overall cohort and split datasets.
FeaturesNo in-Hospital DeathIn-Hospital Deathp-ValueTrain DatasetTest Datasetp-Value
Sample size9082024 2345587
Characteristics
  Age, yr66 (55–76)72 (60–81)<0.00170 (59–80)70 (57–81)0.768
  BMI25.6 (22.5–29)24 (21.1–27.3)<0.00124.5 (21.4–27.9)24.7 (21.5–28.1)0.303
  MDC5 (4–8)5 (4–7)0.9795 (4–7)5 (4–8)0.198
  APACHE II at admission26 (20–32)30 (24–37)<0.00129 (22–35)29 (22–35)0.571
  Timing of initiated CRRT
    Early-strategy group §518 (57%)962 (47.5%)<0.0011200 (51.2%)280 (47.7%)0.132
    Delayed Strategy390 (43%)1062 (52.5%) 1145 (48.8%)307 (52.3%)
Vital Sign at CRRT—no. (%)
  Systolic BP (mmHg)115.7
(105.4–130.3)		108.2
(98.2–119.7)		<0.001110.6
(100.3–122.8)		109.6
(100.8–123)		0.738
  Diastolic BP (mmHg)60.5 (53.5–70)57.7 (50.3–65.5)<0.00158.6 (51.4–66.8)58.1 (50.7–66.4)0.480
  Respiratory rate (/min)18.8 (16.3–21.4)20 (17–23.3)<0.00119.7 (16.7–22.7)19.7 (16.8–22.8)0.884
  SPO297.8 (96–99.1)96.6 (93.5–98.6)<0.00197.1 (94.4–98.8)96.9 (94.4–98.7)0.302
Fluid balance before CRRT—ml/24 hr1663 (500–2992)2221 (1000–3750)<0.0012000 (817–3550)2000 (835–3429)0.942
Coexisting conditions—no. (%)
  Diabetes mellitus368 (40.5%)642 (31.7%)<0.001812 (34.6%)198 (33.7%)0.683
Multiple organ support before CRRT—no. (%)
  Invasive mechanical ventilation719 (79.2%)1832 (90.5%)<0.0012043 (87.1%)508 (86.5%)0.709
  Vasopressors support with norepinephrine or epinephrine632 (69.6%)1792 (88.5%)<0.0011931 (82.3%)493 (84%)0.348
  Vasopressin178 (19.6%)843 (41.7%)<0.001801 (34.2%)220 (37.5%)0.131
Medication use before CRRT—no. (%)
  Corticosteroids429 (47.2%)1247 (61.6%)<0.0011347 (57.4%)329 (56%)0.542
  Parenteral Nutrition678 (74.7%)1769 (87.4%)<0.0011959 (83.5%)488 (83.1%)0.813
  Antibiotics828 (91.2%)1948 (96.2%)<0.0012219 (94.6%)557 (94.9%)0.800
  Furosemide438 (48.2%)1104 (54.5%)0.0021218 (51.9%)324 (55.2%)0.158
Laboratory data before CRRT
  Serum creatinine (mg/dL)2.6 (1.4–5.2)2 (1.2–3.6)<0.0012.1 (1.3–4)2.2 (1.3–4.2)0.640
  Serum potassium (mmol/L)3.9 (3.4–4.4)4 (3.4–4.8)<0.0013.9 (3.4–4.6)4 (3.4–4.7)0.699
  Serum albumin (g/dL)2.6 (2.1–3.1)2.2 (1.7–2.7)<0.0012.3 (1.8–2.8)2.3 (1.8–2.8)0.448
  Lactate, mmol/L2.9 (1.4–6.6)4.8 (2.2–10)<0.0014.1 (1.9–8.9)4 (1.9–9.5)0.906
  Platelet count126 (75–199)96 (53–165)<0.001106 (59–175)105 (56–177)0.930
  pH7.4 (7.3–7.4)7.3 (7.2–7.4)<0.0017.3 (7.2–7.4)7.3 (7.2–7.4)0.931
  Serum sodium (mmol/L)138 (134–141)139 (134–144)<0.001138 (134–143)138 (134–143)0.269
  RDW15.4 (14.3–17)16.1 (14.8–18.3)<0.00115.8 (14.6–17.8)16.1 (14.6–18.6)0.039
  Mg2 (1.8–2.3)2.1 (1.8–2.4)<0.0012.1 (1.8–2.4)2.1 (1.8–2.5)0.090
  INR1.2 (1.1–1.4)1.3 (1.1–1.7)<0.0011.3 (1.1–1.6)1.3 (1.1–1.7)0.785
  APTT36.5 (30.5–54.1)41 (32.8–69.6)<0.00139.2 (31.8–61.5)40.8 (32.4–69.7)0.027
  O2 Saturation98.8 (96.6–99.8)98 (95.1–99.5)<0.00198.3 (95.6–99.7)98.3 (95.7–99.6)0.955
Outcome
  In-hospital mortality0 (0%)2024 (100%)--1625 (69.0%)399 (67.0%)0.535
  28 days mortality2 (0.2%)1733 (85.6%)<0.0011385 (59.1%)350 (59.6%)0.804
  90 days mortality20 (2.2%)1984 (98%)<0.0011607 (68.5%)397 (67.6%)0.676
Abbreviations: BMI, body mass index; MDC, major diagnostic category; APACHE, Acute Physiology and Chronic Health Evaluation; CRRT, continuous renal replacement therapy; BP, blood pressure; RDW, red blood cell distribution width; INR, international normalized ratio; APTT, activated partial thromboplastin time. § Early-strategy group was defined by the renal-replacement therapy was initiated within 24 h of documenting failure-stage acute kidney injury.
Table 2. Comparison of various models’ performance for predicting in-hospital mortality using test data.
Table 2. Comparison of various models’ performance for predicting in-hospital mortality using test data.
ModelAUCThresholdSensitivitySpecificityPPVNPVF1 ScoreAccuracy
Support Vector Machine (SVM) with radial kernel0.75000.674962.66%78.72%86.21%49.83%72.57%67.80%
Support Vector Machine (SVM) with polynomial kernel0.78360.658867.67%77.13%86.26%52.92%75.84%73.59%
Support Vector Machine (SVM) with Sigmoid kernel0.75630.689771.43%75.53%86.10%55.47%78.08%72.74%
Random Forest (RF)0.81610.658774.69%75.00%86.38%58.26%80.11%74.79%
Extreme Gradient Boosting (XGBoost)0.80640.723473.43%80.32%88.79%58.75%80.38%75.64%
Gradient boosted machines (GBMs)0.82270.721674.19%78.72%88.10%58.96%80.55%75.64%
Abbreviations: PPV, positive predictive value; NPV, negative predictive value. The sensitivity, specificity, PPV, and NPV were calculated using Youden’s index.
Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Share and Cite

MDPI and ACS Style

Hung, P.-S.; Lin, P.-R.; Hsu, H.-H.; Huang, Y.-C.; Wu, S.-H.; Kor, C.-T. Explainable Machine Learning-Based Risk Prediction Model for In-Hospital Mortality after Continuous Renal Replacement Therapy Initiation. Diagnostics 2022, 12, 1496. https://doi.org/10.3390/diagnostics12061496

AMA Style

Hung P-S, Lin P-R, Hsu H-H, Huang Y-C, Wu S-H, Kor C-T. Explainable Machine Learning-Based Risk Prediction Model for In-Hospital Mortality after Continuous Renal Replacement Therapy Initiation. Diagnostics. 2022; 12(6):1496. https://doi.org/10.3390/diagnostics12061496

Chicago/Turabian Style

Hung, Pei-Shan, Pei-Ru Lin, Hsin-Hui Hsu, Yi-Chen Huang, Shin-Hwar Wu, and Chew-Teng Kor. 2022. "Explainable Machine Learning-Based Risk Prediction Model for In-Hospital Mortality after Continuous Renal Replacement Therapy Initiation" Diagnostics 12, no. 6: 1496. https://doi.org/10.3390/diagnostics12061496

APA Style

Hung, P.-S., Lin, P.-R., Hsu, H.-H., Huang, Y.-C., Wu, S.-H., & Kor, C.-T. (2022). Explainable Machine Learning-Based Risk Prediction Model for In-Hospital Mortality after Continuous Renal Replacement Therapy Initiation. Diagnostics, 12(6), 1496. https://doi.org/10.3390/diagnostics12061496

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop