Machine Learning-Based Risk Prediction Model for Fatigue in Chronic Heart Failure Patients

Machine Learning-Based Risk Prediction Model for Fatigue in Chronic Heart Failure Patients | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Machine Learning-Based Risk Prediction Model for Fatigue in Chronic Heart Failure Patients Min Zhou, Jingran Yang, Yimei Zhang, Yu Wang, Ruijie Yanglan, and 4 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8543031/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Background Accurately identifying high-risk individuals with fatigue among patients with chronic heart failure (CHF) is crucial for improving their quality of life. This study aimed to construct a risk prediction model for fatigue in patients with CHF based on machine learning (ML) algorithms. Method The study population consisted of patients diagnosed with CHF at two tertiary hospitals in Yunnan from May 10, 2024, to October 31, 2024. LASSO (Least Absolute Shrinkage and Selection Operator) and logistic regression were employed for variable selection. Prediction models were developed and validated using five ML algorithms, and the model’s performance was assessed using several metrics, including the area under the receiver operating characteristic curve (ROC AUC), accuracy, sensitivity, specificity, F1 score, and brier score. SHAP (SHapley Additive exPlanations) plots were utilized for model interpretation. Results A total of 1171 CHF patients were included. Among the five ML models, Random Forest (RF) had the best predictive performance and was the optimal prediction model for fatigue in CHF patients. The best predictors identified included New York Heart Association (NYHA) classification, anxiety, sleep quality, depression, and activities of daily living (ADL). Conclusion The RF model demonstrated robust performance in predicting fatigue risk in CHF patients, providing a valuable tool for healthcare professionals to identify high-risk individuals and implement timely interventions. Trial registration: Not applicable Chronic heart failure Fatigue Risk factors Machine learning Figures Figure 1 Figure 2 Figure 3 Figure 4 1. Introduction Chronic heart failure (CHF), as the persistent stage of heart failure, is the terminal stage of various cardiovascular diseases and the leading cause of death 1 . Patients with CHF experienced symptoms such as dyspnea, fatigue, and depression, leading to adverse health consequences characterized by frequent exacerbations and readmissions 2 . At present, for patients with CHF, alleviating the symptoms is the primary treatment goal. Heart failure guidelines also emphasize the importance of symptom monitoring and management in the long-term care of CHF patients, which can effectively improve the clinical symptoms of CHF patients, reduce medical costs and lower mortality rates 3 , 4 . Among CHF symptoms, fatigue is considered one of the most common symptoms, with a prevalence rate of 42% to 82% 5 . Fatigue in patients with CHF manifests in the early stage of the disease and persists throughout its course, significantly impairing their ability to perform daily activities. Furthermore, due to increased dependence on others and restricted social activities, patients' mental health and social functioning are also compromised, seriously affecting their quality of life 6 , 7 . Stanek et al. 8 found that CHF patients were more interested in relieving fatigue than in increasing their life expectancy. However, fatigue is rarely assessed or treated. Research indicated that healthcare professionals typically do not discuss the experience of fatigue and their impact with individuals or families living with chronic illness 9 . The reason for this may be the vague and relative nature of fatigue, which makes it difficult to identify and manage, posing a challenge to current symptom management of CHF patients 10 . Present research indicated that the early identification of fatigue in patients with CHF can facilitate timely medical intervention, reduce hospitalization rates, and enhance their quality of life 11 . Therefore, if medical staff can accurately predict the risk of fatigue in CHF patients at an early stage and take precise intervention measures, it can effectively reduce the incidence of fatigue and improve the clinical outcomes of CHF patients. Existing research primarily investigates potential risk factors for fatigue in patients with CHF, such as New York Heart Association (NYHA) classification 12 , Left ventricular ejection fraction (LVEF) 11 , gender 12 , and depression 13 . However, most studies have small sample sizes, posing a risk of statistical overfitting. Furthermore, current research on fatigue prediction has mainly focused on patients with malignant tumors 14 , 15 , with a notable lack of fatigue risk prediction models specifically for those with CHF. Therefore, there is an urgent clinical need to develop an accurate and practical fatigue risk prediction tool for CHF patients. In recent years, information technology has been widely applied in the medical field. As a core technology in the field of artificial intelligence, machine learning (ML) has demonstrated excellent predictive performance in the diagnosis and prognosis of cardiovascular diseases 16 . Compared with traditional risk prediction models, ML demonstrates superior sensitivity and specificity, and effectively addresses the limitations of conventional models in handling large-scale, high-dimensional, nonlinear, and incomplete data 17 , thereby enabling more accurate and efficient risk estimation for fatigue. In a comparison of various risk prediction models, risk prediction models based on ML outperformed models based on traditional regression in terms of predictive performance 14 , 18 . Therefore, this study aimed to construct a risk prediction model for fatigue in patients with CHF based on ML, providing healthcare professionals with an assessment tool for early identification of CHF patients at high risk of fatigue. 2 Method 2.1 Study design This study is a cross-sectional study. Data were collected using a questionnaire (Supplementary Material file) and divided into a training set and a validation set at a ratio of 7:3. Five ML algorithms, including Least Absolute Shrinkage and Selection Operator (LASSO)-Logistic regression (LR), Random Forest (RF), K-Nearest Neighbors (KNN), eXtreme Gradient Boosting (XGBoost), and Decision Tree (DT) were used for model construction, evaluation and validation. Finally, the optimal model is selected as a clinical assessment tool (Fig. 1 ). 2.2 Study participants In this study, by using the convenience sampling method, patients diagnosed with CHF at two tertiary hospitals in Yunnan Province between May 10, 2024, and October 31, 2024 were recruited as the research participants. The inclusion criteria were (1) aged ≥ 18 years; (2) met the diagnostic criteria for CHF of the 2022 AHA/ACC/HFSA Guidelines for the Management of Heart Failure 19 ; (3) willing to participate in this study and had normal comprehension and verbal ability. The exclusion criteria were (1) concurrent malignant tumors or severe chronic diseases; (2) mental illness or current consciousness impairment that prevent effective communication; (3) missing more than 30% of important data. 2.3 Study tools 2.3.1 Fatigue assessment This study used the Multidimensional Fatigue Inventory (MFI) to assess fatigue in CHF patients. This scale was developed by Smets 20 , consists of five dimensions (general fatigue, physical fatigue, reduced activity, decreased motivation, and mental fatigue) and 20 items. According to the literature, when the score of the general fatigue dimension reaches 12 points or higher (items 1, 5, 12, 16), fatigue is diagnosed 21 , 22 . The scale demonstrates good internal consistency, with a Cronbach's α coefficient of 0.882 23 . 2.3.2 Predictor variables After reviewing the literature 7 , 24 , 25 and combining it with clinical experience, we selected 29 potential risk factors, including (1) demographic data : age, gender, convenience from the place of residence to the hospital, educational level, marital status, occupation; (2) biological data : dyspnea, NYHA classification, LVEF, cerebrovascular disease, anemia, diabetes, number of comorbidities, hypertension, body mass index (BMI), sleep quality, activities of daily living (ADL), hemoglobin, hematocrit, N-terminal (NT)-pro brain natriuretic peptide (BNP), beta-blockers, statins, diuretics, nitrates, calcium channel blockers, psychotropics; (3) psychological data : anxiety, depression, and (4) sociological data : social support. We assessed the variables of sleep quality, activities of daily living, anxiety, depression, and social support using the following scales: (1) Sleep quality: The Pittsburgh Sleep Quality Index(PSQI) is mainly used to assess sleep patterns over the past month. The total score ranges from 0 to 21, with higher scores indicating poorer sleep quality. The score of PSQI ≤ 7 indicates normal sleep quality, while the score > 7 indicates poor sleep quality. The Cronbach's α coefficient is 0.84 26 . (2) Activities of daily living: The Activities of Daily Living (ADL) Scale was used in this study, which includes 10 dimensions of daily living activities. The total score ranges from 0 to 100 points. A score of ≥ 60 indicates good self-care ability, while a score of < 60 indicates functional impairment. The Cronbach's α coefficient was 0.887 27 . (3) Anxiety and depression: The Hospital Anxiety and Depression Scale (HADS) is primarily used to evaluate the anxiety and depression levels of hospitalized patients, consisting of two subscales for anxiety and depression, each with seven items. Each item is scored on a scale of 0 to 3. A total score of 0 to 7 indicates no anxiety or depression, 8 to 10 indicates mild anxiety or depression, and 11 to 21 indicates severe anxiety or depression. The Cronbach's α coefficients for the HADS total scale and the two subscales are 0.89, 0.82, and 0.80, respectively 26 . (4) Social support: The Social Support Rating Scale (SSRS) encompasses three dimensions and ten items. The total score ranges from 0 to 66 points, with 0ཞ22 points indicating low social support, 23ཞ44 points indicating moderate social support, and 45ཞ66 points indicating high social support. The Cronbach's α coefficient is 0.92 14 . 2.4 Data collection This study was conducted by two researchers who had undergone scientific research training. After obtaining the informed consent of the research participants, data was gathered through face-to-face methods. Some data was obtained by the investigators from the electronic medical record system. After the data collection was completed, two investigators verified the collected data to ensure the quality. 2.5 Statistical analysis 2.5.1 Data preprocessing Before developing the model, two researchers were responsible for data entry and management using Epidata 3.0. A cross-verification approach was employed to check for data issues such as duplicates, missing values, or errors that could compromise data accuracy. Duplicate data was removed. For noise values, we described the distribution of each variable using SPSS 26.0, including the maximum and minimum values. Values exceeding the reasonable threshold were removed and processed as missing values. When the missing value rate exceeded 30% and the missing pattern was random, the variable and patient data were removed. In our study, the missing data of NT-proBNP exceeded 40%, which was excluded. The mean was used to impute the missing values because the missing values for the other variables were less than 5% and showed a random missing pattern. 2.5.2 Data analysis SPSS 26.0 software was used for descriptive analysis and comparison of inter-group data. All statistical tests adopted a two-tailed test approach, with a p-value < 0.05 indicating statistically significant differences. Continuous data with normal distribution were presented as mean and standard deviation, while non-normally distributed data were described by the median (P 25 , P 75 ). Categorical data were described by frequency and percentage. In the comparison of inter-group data, for categorical data, the χ 2 test was employed. Continuous variables that follow a normal distribution, Student's t-test was utilized, while for non-normally distributed variables, the Mann-Whitney U rank-sum test was used for inter-group comparisons. 2.5.3 Model construction and validation R 3.0 was used for model construction and validation. Based on the “glmnet” package, LASSO was applied to screen predictive factors. Subsequently, the selected predictive factors were incorporated into a multivariate logistic regression analysis using the forward-backward stepwise method. The final variables entered into the model were determined according to the optimal results of the Akaike Information Criterion (AIC). Furthermore, we utilized software packages including “glmnet”, “class”, “random forest”, “xgboost”, and “rpart” to construct models. Optimal parameters for each model were selected using 10-fold cross-validation to ensure model performance and generalization capability. To assess the predictive performance of the models, this research employed the area under the receiver operating characteristic (ROC) curve (AUC) as the key metric. It reflects the classification performance of the model across various thresholds, and a value closer to 1 indicates a stronger discriminatory power of the model. Accuracy was computed to represent the overall predictive performance. Sensitivity was used to measure the ability to identify positive cases. Specificity was employed to reflect the accuracy of predicting negative cases. The F1-score was used to balance the model's performance between sensitivity and precision. Additionally, calibration curves and Brier scores were adopted to evaluate the calibration of the predictive models. 2.5.4 Model interpretation SHAP (SHapley Additive exPlanation) is a model-interpretability tool based on Shapley values. In this study, after selecting the optimal prediction model, the SHAP plot was used to interpret the model by determining the importance of each predictor in the decision-making process and visualizing the results. 3 Results 3.1 Characteristics of the study participants A total of 1186 questionnaires were distributed in this study. Seven patients were excluded due to concomitant malignant tumors, and eight patients were excluded due to missing variables > 30%. Ultimately, 1171 patients were collected, and 484 patients (41.3%) experienced fatigue. The dataset was divided into a training set (819 cases) and a validation set (352 cases) at a ratio of 7:3. The age range of the study participants was from 18 to 96 years old, and the number of people aged 66 to 80 was the largest, reaching 529 (45.2%). There were 680 male participants (58.1%) and 491 female participants (41.9%). Characteristics of patients with statistical differences between the fatigue group and the non-fatigue group were shown in Table 1 . The difference in baseline information between the training and test sets was not statistically significant (Supplementary file). Table 1 Baseline Characteristics of CHF Patients (n = 1171). Variables Overall (n = 1171) Non-fatigue (n = 687) Fatigue (n = 484) P value Age, n (%) 0.002 ≤ 30 31ཞ50 51 ~ 65 66 ~ 80 ≥ 81 17 (1.5) 108 (9.2) 330 (28.2) 529 (45.2) 187 (16) 15 (2.2) 77 (11.2) 193 (28.1) 304 (44.3) 98 (14.3) 2 (0.4) 31 (6.4) 137 (28.3) 225 (46.5) 89 (18.4) Gender, n (%) 0.022 male female 680 (58.1) 491 (41.9) 418 (60.8) 269 (39.2) 262 (54.1) 222 (45.9) Convenience from the place of residence to the hospital, n (%) < 0.001 highly inconvenient moderately inconvenient relatively convenient highly convenient 211 (18) 318 (27.2) 307 (26.2) 335 (28.6) 135 (19.7) 183 (26.6) 152 (22.1) 217 (31.6) 76 (15.7) 135 (27.9) 155 (32) 118 (24.4) Educational level, n (%) 0.341 primary school level and below junior middle school high school university and above 257 (53.1) 104 (21.5) 121 (25) 2 (0.4) 382 (55.6) 161 (23.4) 141 (20.5) 3 (0.4) 257 (53.1) 104 (21.5) 121 (25) 2 (0.4) Marital status, n (%) 0.028 single married divorced widowed 32 (2.7) 887 (75.7) 19 (1.6) 233 (19.9) 22 (3.2) 537 (78.2) 10 (1.5) 118 (17.2) 10 (2.1) 350 (72.3) 9 (1.9) 115 (23.8) Occupation, n (%) < 0.001 famer production and transportation workers enterprise staff medical personnel service industry workers civil servants retirees unemployed others 395 (33.7) 26 (2.2) 41 (3.5) 1 (0.1) 10 (0.9) 13 (1.1) 453 (38.7) 201 (17.2) 31 (2.6) 240 (34.9) 20 (2.9) 28 (4.1) 1 (0.1) 3 (0.4) 6 (0.9) 234 (34.1) 132 (19.2) 23 (3.3) 155 (32) 6 (1.2) 13 (2.7) 0 (0) 7 (1.4) 7 (1.4) 219 (45.2) 69 (14.3) 8 (1.7) Dyspnea, n (%) 0.069 yes no 498 (42.5) 673 (57.5) 277 (40.3) 410 (59.7) 221 (45.7) 263 (54.3) NYHA class, n (%) 50% 685 (58.5) 486 (41.5) 407 (59.2) 280 (40.8) 278 (57.4) 206 (42.6) Cerebrovascular disease, n (%) 0.048 yes no 134 (11.4) 1037 (88.6) 68 (9.9) 619 (90.1) 66 (13.6) 418 (86.4) Anemia, n (%) 0.453 yes no 91 (7.8) 1080 (92.2) 50 (7.3) 637 (92.7) 41 (8.5) 443 (91.5) Diabetes, n (%) 0.028 yes no 265 (22.6) 906 (77.4) 140 (20.4) 547 (79.6) 125 (25.8) 359 (74.2) Number of comorbidities, n (%) 0.035 5 483 (41.2) 565 (48.2) 123 (10.5) 304 (44.3) 318 (46.3) 65 (9.5) 179 (37) 247 (51) 58 (12) Hypertension, n (%) 0.954 yes no 577 (49.3) 594 (50.7) 339 (49.3) 348 (50.7) 238 (49.2) 246 (50.8) Table 1 (continued) Variables Overall (n = 1171) Non-fatigue (n = 687) Fatigue (n = 484) P value BMI, n (%) 0.051 <18.5 18.50ཞ23.99 24.00ཞ27.99 ≥ 28 124 (10.6) 568 (48.5) 333 (28.4) 146 (12.5) 60 (8.7) 334 (48.6) 209 (30.4) 84 (12.2) 64 (13.2) 234 (48.3) 124 (25.6) 62 (12.8) Sleep quality, n (%) < 0.001 normal bad 399 (34.1) 772 (65.9) 321 (46.7) 366 (53.3) 78 (16.1) 406 (83.9) Activities of daily living, n (%) < 0.001 normal disorder 1090 (93.1) 81 (6.9) 665 (96.8) 22 (3.2) 425 (87.8) 59 (12.2) Hemoglobin, n (%) < 0.001 160 80 (6.8) 905 (77.3) 186 (15.9) 36 (5.2) 521 (75.8) 130 (18.9) 44 (9.1) 384 (79.3) 56 (11.6) Hematocrit (median [IQR],vol%) 42 (37,47) 43 (38,48) 41 (36,45) < 0.001 Beta-blockers, n (%) 0.254 yes no 741 (63.3) 430 (36.7) 444 (64.6) 243 (35.4) 297 (61.4) 187 (38.6) Statins, n (%) 0.911 yes no 641 (54.7) 530 (45.3) 377 (54.9) 310 (45.1) 264 (54.5) 220 (45.5) Diuretics, n (%) 0.865 yes no 980 (83.7) 191 (16.3) 576 (83.8) 111 (16.2) 404 (83.5) 80 (16.5) Nitrates, n (%) 0.146 yes no 67 (5.7) 1104 (94.3) 45 (6.6) 642 (93.4) 22 (4.5) 462 (95.5) Calcium channel blockers, n (%) 0.617 yes no 122 (10.4) 1049 (89.6) 69 (10) 618 (90) 53 (11) 431 (89) Psychotropics, n (%) 0.048 yes no 103 (8.8) 1068 (91.2) 51 (7.4) 636 (92.6) 52 (10.7) 432 (89.3) Anxiety, n (%) < 0.001 None Mild severe 961 (82.1) 150 (12.8) 60 (5.1) 622 (90.5) 47 (6.8) 18 (2.6) 339 (70) 103 (21.3) 42 (8.7) Depression, n (%) < 0.001 None Mild severe 955 (81.6) 156 (13.3) 60 (5.1) 616 (89.7) 56 (8.2) 15 (2.2) 339 (70) 100 (20.7) 45 (9.3) Social support, n (%) 0.027 low level middle level hihg level 61 (5.2) 1082 (92.4) 28 (2.4) 26 (3.8) 646 (94) 15 (2.2) 35 (7.2) 436 (90.1) 13 (2.7) 3.2 Feature selection LASSO was used to perform dimensionality reduction analysis on 28 potential predictive variables. Some predictors with weak correlations to the outcome variable were automatically excluded as their coefficients were compressed to zero (Fig. 2 A). The model was further optimized using 10-fold cross-validation. To ensure the model’s fitting performance and simplicity, we selected the lambda value corresponding to the smallest standard error. Under this lambda value, the original 28 variables were simplified to 5 key variables (Fig. 2 B). The five variables identified by LASSO were incorporated into the multivariate logistic regression analysis. The results are presented in Table 2 . When five predictive variables, NYHA classification (OR = 1.688, 95% CI: 1.351–2.116, P < 0.001), sleep quality (OR = 3.929, 95% CI: 2.745–5.701, P < 0.001), ADL (OR = 2.508, 95% CI: 1.292–5.089, P = 0.008), anxiety (OR = 1.899, 95% CI: 1.196–3.060, P = 0.007), and depression (OR = 1.218, 95% CI: 0.786–1.896, P = 0.377), were combined, the corresponding AIC value was minimal, indicating optimal model performance. Table 2 Results of Logistic Regression Variables B SE OR CI ( 95%) Wald χ 2 P value NYHA class 0.524 0.11426 1.688 1.688 (1.351–2.116) 4.583 <0.001 Sleep quality 1.368 0.18608 3.929 3.929 (2.745–5.701) 7.354 <0.001 Activities of daily living 0.919 0.34783 2.508 2.508(1.292–5.089) 2.643 0.008 Anxiety 0.641 0.23849 1.899 1.898 (1.196–3.060) 2.689 0.007 Depression 0.198 0.22349 1.218 1.218 (0.786–1.896) 0.884 0.377 3.3 Model performance We trained and internally validated the model based on these five features, as shown in Fig. 3 . Table 3 summarizes the performance of the five ML models. RF, XGBoost, and LASSO-LR all exhibited favorable predictive performance in the training set and validation set. The AUC for RF was slightly higher compared to XGBoost and LASSO-LR, with AUC values of 0.761 (95% CI: 0.729–0.794) in the training set, and 0.721 (95% CI: 0.669–0.774) in the validation set, indicating its moderate to good classification accuracy and predictive performance. In addition, it also has better overall performance metrics, such as accuracy (reflecting the correctness of predictions), sensitivity (identifying positive cases), specificity (identifying negative cases), and F1 Score (balancing recall and precision). This suggested that RF has a slightly superior predictive ability for fatigue risk in CHF patients compared to other models. Table 3 A The performance of the model in the training set Model AUC Accuracy Sensitivity Specificity F1 Score Brier DT 0.736 0.707 0.662 0.739 0.65 0.198 RF 0.761 0.713 0.647 0.759 0.65 0.234 XGBoost 0.753 0.705 0.653 0.741 0.645 0.197 KNN 0.669 0.680 0.490 0.813 0.557 0.285 LR 0.759 0.707 0.659 0.741 0.649 0.194 Table 3 B The performance of the model in the validation set Model AUC Accuracy Sensitivity Specificity F1 Score Brier DT 0.684 0.673 0.626 0.707 0.615 0.216 RF 0.721 0.676 0.612 0.722 0.612 0.263 XGBoost 0.706 0.676 0.619 0.717 0.615 0.210 KNN 0.633 0.642 0.442 0.785 0.508 0.198 LR 0.718 0.676 0.626 0.712 0.617 0.208 3.4 Interpretation of the final model In this study, the contribution of each feature to the RF model was quantified using SHAP values. The features are ranked in descending order of their contribution as follows: NYHA classification, anxiety, sleep quality, depression and ADL, as depicted in Fig. 4 . Each point represents a sample, with the horizontal axis displaying the SHAP values for each feature. Colors indicate the direction and relative magnitude of the feature within the classification, with red corresponding to high values, blue to low values, and purple to the mean. Taking the first row as an example, a high NYHA functional class (red) has a positive impact on the prediction, while a low NYHA functional class (blue) has a negative impact. 4 Discussion As one of the landmark symptoms of CHF, fatigue is prevalent among CHF patients and significantly impacts their quality of life. In this study, a total of 1171 patients with CHF were included. Among them, 484 patients developed fatigue, with an incidence of 41.3%, which was consistent with the results of the study by Walke et al. 28 . Despite the relatively high incidence of fatigue, medical staff and patients often overlook the experience of fatigue and its implications 9 , which suggest that healthcare professionals should strengthen the screening of CHF patients who are at high risk for fatigue to reduce the incidence. Currently, there is still a lack of risk prediction tools for fatigue in CHF patients. With the advent of the big data era, an increasing number of scholars are using ML to construct risk prediction models, which play an important role in disease diagnosis and management, as well as personalized treatment 29 . These models not only effectively reduce human intervention in the evaluation process, but also make the prediction results more accurate 30 . Therefore, this study introduced five ML algorithms to construct a fatigue risk prediction model for CHF patients, including LASSO-LR, RF, KNN, XGBoost, and DT. Multiple ML algorithms were selected to construct predictive models because they have different computational characteristics and have been widely used in other studies 18 , 31 . By integrating multiple decision trees, the RF model effectively processes high-dimensional, non-linear cardiovascular data that may contain missing values. Furthermore, the training process of RF can be parallelized, significantly expediting model construction and enhancing data fitting compared to traditional regression methods. Additionally, RF is characterized by high prediction accuracy and robust anti-noise capabilities, which contribute to improving the overall robustness of the model 32 , 33 . In addition, the use of LASSO regression in ML for variable selection offers notable advantages. Through L1 regularization, it automatically shrinks the coefficients of insignificant variables to zero, thereby enabling efficient variable selection. Moreover, this approach effectively mitigates the risks of multicollinearity and overfitting 34 . As a result, the model can achieve a higher level of precision and sensitivity when predicting fatigue in patients with CHF. Another advantage of our research is the introduction of the SHAP method to explain the RF model. ML models are often referred to as black box models, lacking transparency in decision-making. While the SHAP algorithm can clearly explain the complex relationship between features and prediction results, and has significant advantages in model interpretability and visualization 35 . Through the output of SHAP, we found that the risk factors for fatigue of CHF patients were NYHA classification, anxiety, sleep quality, depression, and ADL. NYHA classification is generally regarded as the cornerstone for evaluating the functional status and the severity of symptoms in patients with HF 36 . The results of our study showed that NYHA classification has the greatest contribution to fatigue, and the higher the NYHA classification, the greater the risk of fatigue in CHF patients. Multiple studies have also demonstrated that NYHA classification is significantly associated with fatigue in patients with CHF 7 , 37 . Therefore, it is necessary to promptly identify changes in patients' cardiac function and conduct in-depth research to strengthen the understanding of the relationship between NYHA classification and fatigue. Anxiety is another strong predictor. Persistent mobility issues, palpitations, chest tightness, repeated hospitalizations, and uncertainty about the disease can all contribute to CHF patients’ prolonged anxiety 38 . Long-term anxiety, in turn, can exacerbate the burden on the body, leading to excessive physical exertion, weakened immunity and increased fatigue 39 . However, few studies have paid attention to the anxiety of patients with CHF, which indicated that medical staff may need to devote more attention to the anxiety of CHF patients and provide them with mental health support as well as management of symptom distress. Interestingly, in our study, the results of multivariate logistic regression indicated that depression was not significantly associated with fatigue in CHF patients, but it had the smallest AIC value when combined with the other four variables. This suggested that depression has a potential contribution to the prediction of fatigue and can improve the predictive performance of the model through the combined effect with other variables. Consequently, our research incorporated depression into the model as a predictive factor for fatigue. The relationship between depression and fatigue in CHF patients may be influenced by reduced cerebral blood flow, along with the effects of depression on executive dysfunction and individual independence 7 , 40 . Future studies could further analyze the association between fatigue and depression in CHF patients from the perspectives of variable interactions, dynamic relationships, and potential mechanisms. Sleep quality ranked third in the importance of risk factors. Research has found that the reduced cardiac output in patients with CHF leads to insufficient cerebral perfusion, which may cause changes in the function of the central nervous system related to sleep regulation, leading to alterations in the patients' sleep characteristics. Additionally, factors such as the use of diuretics disrupt the patients' circadian rhythms and negatively influence patients’ sleep quality, ultimately resulting in physical fatigue 41 , 42 . ADL is also a predictor of fatigue. On the one hand, impaired ADL indicates that the patient is in a poor state of bone marrow suppression and malnutrition, which leads to a reduction in muscle mass and the occurrence of physical fatigue. On the other hand, limited self-care ability further caused muscle atrophy, decreased cardiopulmonary function, and aggravated fatigue 43 . At present, data on the effects of exercise suggests that helping patients increase their activity levels may be the best way to manage fatigue 44 . Therefore, healthcare professionals should encourage CHF patients to develop personalized exercise plans based on their individual circumstances, gradually increasing their activity levels to improve physical fitness and alleviate fatigue. Limitations This study has several limitations. Firstly, the data collection for this study was limited to two tertiary hospitals in Yunnan Province, and the sample coverage has certain geographical limitations. Secondly, our study is a cross-sectional study, which can effectively reveal the correlation between variables, but cannot infer causal relationships, and it is difficult to conduct in-depth analysis of longitudinal clinical data, which may result in selection bias. Thirdly, the large amount of missing data for NT-proBNP affected the integrity of the data and limited our ability to conduct a more comprehensive analysis of the NT-proBNP indicator. Finally, this study did not conduct external validation, which might have limited the generalization ability of the model. Therefore, we fully verified the reliability and stability of the model through the internal validation method of K-fold cross-validation. To further enhance the model's clinical application value, future studies should expand the sample size and conduct multi-center external validation, incorporating additional objective measurement indicators to enhance the model's generalizability and applicability. Conclusion In summary, by comparing the performance of various ML algorithms in predicting fatigue risk, we ultimately selected RF as the optimal model. Through SHAP interpretation of the model, we ranked the contribution of five predictive variables, including NYHA classification, anxiety, sleep quality, depression, and ADL. This provides a basis for healthcare professionals to CHF patients at high risk for fatigue at an early stage, and provide personalized treatment. Declarations Conflicts of Interest The authors have no conflicts of interest in this study. Ethics approval The procedures followed in this study comply with the Declaration of Helsinki and were approved by the Ethics Committee of the First Affiliated Hospital of Kunming Medical University (Approval No.: 2024-L-8). All participants signed written informed consent forms. Funding This research was supported by the Reserve Talent Project of Academic and Technological Leaders for Young and Middle-aged People in Yunnan Province, China, No. 202205AC160017 and the Yunnan Fundamental Research Kunming Medical University, No 202301AY070001-153. Author Contribution Min Zhou (first author) contributed to the study design, data collection and analysis, and writing the manuscript. Fang Ma (corresponding author) was responsible for designing the study and revising the manuscript. Jingran Yang, Yimei Zhang, Yu Wang were responsible for data collection. Ruijie Yanglan, Qinlan Li were responsible for the data curation. Yangjuan Bai, Wei Wei contributed to the review of the manuscript. All authors read and approved the final manuscript. Data Availability The data that support the findings of this study are available on request from the corresponding author. References Hunt SA, American College of Cardiology, American Heart Association Task Force on Practice Guidelines (Writing Committee to Update the 2001 Guidelines for the Evaluation and Management of Heart Failure). ACC/AHA 2005 guideline update for the diagnosis and management of chronic heart failure in the adult: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (Writing Committee to Update the 2001 Guidelines for the Evaluation and Management of Heart Failure). J Am Coll Cardiol. 2005;46:e1–82. Kim JSR, Risbud R, Gray C, Banerjee D, Trivedi R. The Dyadic Experience of Managing Heart Failure A Qualitative Investigation. J Cardiovasc Nurs. 2020;35:12–8. Alpert CM, Smith MA, Hummel SL, Hummel EK. Symptom burden in heart failure: assessment, impact on outcomes, and management. Heart Fail Rev. 2017;22:25–39. McDonagh TA, Metra M, Adamo M, Gardner RS, Baumbach A, Böhm M, Burri H, Butler J, Čelutkienė J, Chioncel O, Cleland JGF, Coats AJS, Crespo-Leiro MG, Farmakis D, Gilard M, Heymans S, Hoes AW, Jaarsma T, Jankowska EA, Lainscak M, Lam CSP, Lyon AR, McMurray JJV, Mebazaa A, Mindham R, Muneretto C, Francesco Piepoli M, Price S, Rosano GMC, Ruschitzka F, Kathrine Skibelund A, ESC Scientific Document Group. 2021 ESC Guidelines for the diagnosis and treatment of acute and chronic heart failure. Eur Heart J. 2021;42:3599–726. Higa H, Lennie TA, Chung ML, Tsuchihashi-Makaya M. Associations of multidimensional fatigue with the physical, psychological, and situational factors in outpatients with heart failure: a cross-sectional study. Eur J Cardiovasc Nurs 2022:zvac117. Schjoedt I, Sommer I, Bjerrum MB. Experiences and management of fatigue in everyday life among adult patients living with heart failure: A systematic review of qualitative evidence. JBI Database Syst Reviews Implement Rep. 2016;14:68–115. Chen L-H, Li C-Y, Shieh S-M, Yin W-H, Chiou A-F. Predictors of fatigue in patients with heart failure: Fatigue in patients with heart failure. J Clin Nurs. 2009;19:1588–96. Stanek EJ, Oates MB, McGhan WF, Denofrio D, Loh E. Preferences for treatment outcomes in patients with heart failure: Symptoms versus survival. J Card Fail. 2000;6:225–32. Whitehead L. The Family Experience of Fatigue in Heart Failure. J Fam Nurs. 2017;23:138–56. Falk K, Swedberg K, Gaston-Johansson F, Ekman I. Fatigue is a prevalent and severe symptom associated with uncertainty and sense of coherence in patients with chronic heart failure. Eur J Cardiovasc Nurs 2007;6. Tang W-R, Yu C-Y, Yeh S-J. Fatigue and its related factors in patients with chronic heart failure. J Clin Nurs. 2010;19:69–78. Hägglund L, Boman K, Stenlund H, Lundman B, Brulin C. Factors related to fatigue among older patients with heart failure in primary health care. Int J Older People Nurs. 2008;3:96–103. Higa H, Lennie TA, Chung ML, Tsuchihashi-Makaya M. Associations of multidimensional fatigue with the physical, psychological, and situational factors in outpatients with heart failure: a cross-sectional study. Eur J Cardiovasc Nurs. 2023;22:647–54. Huang S-T, Ke X, Huang Y-P, Wu Y-X, Yu X-Y, Liu H-K, Liu D. A prediction model for moderate to severe cancer-related fatigue in colorectal cancer after chemotherapy: A prospective case–control study. Support Care Cancer. 2023;31:426. Courtier N, Gambling T, Enright S, Barrett-Lee P, Abraham J, Mason MD. A prognostic tool to predict fatigue in women with early-stage breast cancer undergoing radiotherapy. Breast. 2013;22:504–9. Jing L, Ulloa Cerna AE, Good CW, Sauers NM, Schneider G, Hartzel DN, Leader JB, Kirchner HL, Hu Y, Riviello DM, Stough JV, Gazes S, Haggerty A, Raghunath S, Carry BJ, Haggerty CM, Fornwalt BK. A Machine Learning Approach to Management of Heart Failure Populations. JACC Heart Fail. 2020;8:578–87. Kokori E, Patel R, Olatunji G, Ukoaka BM, Abraham IC, Ajekiigbe VO, Kwape JM, Babalola AE, Udam NG, Aderinto N. Machine learning in predicting heart failure survival: a review of current models and future prospects. Heart Fail Rev. 2024;30:431–42. Yang J, Xie Y, Wang T, Pu Y, Ye T, Huang Y, Song B, Cheng F, Yang Z, Zhang X. Machine learning-based risk prediction of mild cognitive impairment in patients with chronic heart failure: A model development and validation study. Geriatr Nurs. 2025;62:145–56. Heidenreich PA, Bozkurt B, Aguilar D, Allen LA, Byun JJ, Colvin MM, Deswal A, Drazner MH, Dunlay SM, Evers LR, Fang JC, Fedson SE, Fonarow GC, Hayek SS, Hernandez AF, Khazanie P, Kittleson MM, Lee CS, Link MS, Milano CA, Nnacheta LC, Sandhu AT, Stevenson LW, Vardeny O, Vest AR, Yancy CW. 2022 AHA/ACC/HFSA Guideline for the Management of Heart Failure: A Report of the American College of Cardiology/American Heart Association Joint Committee on Clinical Practice Guidelines. Circulation. 2022;145:e895–1032. Smets EM, Garssen B, Bonke B, De Haes JC. The Multidimensional Fatigue Inventory (MFI) psychometric qualities of an instrument to assess fatigue. J Psychosom Res. 1995;39:315–25. Aluzaite K, Al-Mandhari R, Osborne H, Ho C, Williams M, Sullivan M-M, Hobbs CE, Schultz M. Detailed Multi-Dimensional Assessment of Fatigue in Inflammatory Bowel Disease. Inflamm Intest Dis. 2019;3:192–201. Purcell A, Fleming J, Bennett S, Burmeister B, Haines T. Determining the minimal clinically important difference criteria for the Multidimensional Fatigue Inventory in a radiotherapy population. Support Care Cancer. 2010;18:307–15. Miao Y, Liu X, Liu Z, Xie H, Deng G. The preliminary revision of the Chinese version of the Multidimensional Fatigue Scale in military primary medical staff. Chin J Mental Health. 2008:658–60. Pavlovic NV, Gilotra NA, Lee CS, Ndumele C, Mammos D, Dennisonhimmelfarb C, AbshireSaylor M. Fatigue in Persons With Heart Failure: A Systematic Literature Review and Meta-Synthesis Using the Biopsychosocial Model of Health. J Card Fail. 2022;28:283–315. Williams BA. The clinical epidemiology of fatigue in newly diagnosed heart failure. BMC Cardiovasc Disord. 2017;17:122. Liao T, Gan Y, Liu L, Du Y, Li G, Guo M. The Association of Sleep Quality With Anorexia in Patients With Heart Failure: Do Anxiety and Depressive Symptoms Mediate This Association. J Cardiovasc Nurs 2025. Lawton MP, Brody EM. Assessment of older people: self-maintaining and instrumental activities of daily living. Gerontologist. 1969;9:179–86. Walke LM, Gallo WT, Tinetti ME, Fried TR. The burden of symptoms among community-dwelling older persons with advanced chronic disease. Arch Intern Med. 2004;164:2321–4. Iwasaki Y, Takeuchi I, Stanev V, Kusne AG, Ishida M, Kirihara A, Ihara K, Sawada R, Terashima K, Someya H, Uchida K, Saitoh E, Yorozu S. Machine-learning guided discovery of a new thermoelectric material. Sci Rep. 2019;9:2751. Lorenzoni G, Sabato SS, Lanera C, Bottigliengo D, Minto C, Ocagli H, De Paolis P, Gregori D, Iliceto S, Pisanò F. Comparison of machine learning techniques for prediction of hospitalization in heart failure patients. JCM. 2019;8:1298. Hu Y, Ma F, Hu M, Shi B, Pan D, Ren J. Development and validation of a machine learning model to predict the risk of readmission within one year in HFpEF patients. Int J Med Informatics. 2025;194:105703. Shi G, Liu G, Gao Q, Zhang S, Wang Q, Wu L, He P, Yu Q. A random forest algorithm-based prediction model for moderate to severe acute postoperative pain after orthopedic surgery under general anesthesia. BMC Anesthesiol. 2023;23:361. Mansoor H, Elgendy IY, Segal R, Bavry AA, Bian J. Risk prediction model for in-hospital mortality in women with ST-elevation myocardial infarction: A machine learning approach. Heart Lung. 2017;46:405–11. McNeish DM. Using Lasso for Predictor Selection and to Assuage Overfitting: A Method Long Overlooked in Behavioral Sciences. Multivar Behav Res. 2015;50:471–84. Lin Q, Zhao W, Zhang H, Chen W, Lian S, Ruan Q, Qu Z, Lin Y, Chai D, Lin X. Predicting the risk of heart failure after acute myocardial infarction using an interpretable machine learning model. Front Cardiovasc Med. 2025;12:1444323. Dickstein K, Cohen-Solal A, Filippatos G, McMurray JJV, Ponikowski P, Poole-Wilson PA, Strömberg A, van Veldhuisen DJ, Atar D, Hoes AW, Keren A, Mebazaa A, Nieminen M, Priori SG, Swedberg K, ESC Committee for Practice Guidelines (CPG). failure 2008: the Task Force for the diagnosis and treatment of acute and chronic heart failure 2008 of the European Society of Cardiology. Developed in collaboration with the Heart Failure Association of the ESC (HFA) and endorsed by the European Society of Intensive Care Medicine (ESICM). Eur J Heart Fail. 2008;10:933–89. ESC guidelines for the diagnosis and treatment of acute and chronic heart. Inamdar A, Inamdar A. Heart failure: Diagnosis, management and utilization. JCM. 2016;5:62. Rashid S, Qureshi AG, Noor TA, Yaseen K, Sheikh MAA, Malik M, Malik J. Anxiety and Depression in Heart Failure: An Updated Review. Curr Probl Cardiol. 2023;48:101987. Falk K, Patel H, Swedberg K, Ekman I. Fatigue in patients with chronic heart failure - a burden associated with emotional and symptom distress. Eur J Cardiovasc Nurs. 2009;8:91–6. Fink AM, Gonzalez RC, Lisowski T, Pini M, Fantuzzi G, Levy WC, Piano MR. Fatigue, inflammation, and projected mortality in heart failure. J Card Fail. 2012;18:711–6. Riegel B, Ratcliffe SJ, Sayers SL, Potashnik S, Buck HG, Jurkovitz C, Fontana S, Weaver TE, Weintraub WS, Goldberg LR. Determinants of excessive daytime sleepiness and fatigue in adults with heart failure. Clin Nurs Res. 2012;21:271–93. Redeker NS, Stein S. Characteristics of sleep in patients with stable heart failure versus a comparison group. Heart lung: J Crit care 2006;35. Gomes-Neto M, Durães AR, Conceição LSR, Roever L, Silva CM, Alves IGN, Ellingsen Ø, Carvalho VO. Effect of combined aerobic and resistance training on peak oxygen consumption, muscle strength and health-related quality of life in patients with heart failure with reduced left ventricular ejection fraction: a systematic review and meta-analysis. Int J Cardiol. 2019;293:165–75. Hsiung P, Lin P-C, Lin T-Y, Wu W-T, Sun J-L, Chou P-L. Effectiveness of Exercise on Fatigue for Patients With Heart Failure: A Systematic Review and Meta-Analysis. J Appl Gerontol. 2025;44:312–26. Additional Declarations No competing interests reported. Supplementary Files BaselineCharacteristicsoftrainingsetandvalidationset.docx Questionnaire.docx Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-8543031","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":578522284,"identity":"53a72379-9fc0-43c1-9e06-0877eb468e8e","order_by":0,"name":"Min Zhou","email":"","orcid":"","institution":"The First Affiliated Hospital of Kunming Medical University","correspondingAuthor":false,"prefix":"","firstName":"Min","middleName":"","lastName":"Zhou","suffix":""},{"id":578522286,"identity":"2c012b5c-9a6f-4e9c-8abf-51d33c7d898c","order_by":1,"name":"Jingran Yang","email":"","orcid":"","institution":"The First Affiliated Hospital of Kunming Medical University","correspondingAuthor":false,"prefix":"","firstName":"Jingran","middleName":"","lastName":"Yang","suffix":""},{"id":578522290,"identity":"41287f95-3275-4fa8-afce-12d5e942b5d2","order_by":2,"name":"Yimei Zhang","email":"","orcid":"","institution":"The First Affiliated Hospital of Kunming Medical University","correspondingAuthor":false,"prefix":"","firstName":"Yimei","middleName":"","lastName":"Zhang","suffix":""},{"id":578522298,"identity":"19101e38-da07-4858-af7a-490235fb43ce","order_by":3,"name":"Yu Wang","email":"","orcid":"","institution":"The First Affiliated Hospital of Kunming Medical University","correspondingAuthor":false,"prefix":"","firstName":"Yu","middleName":"","lastName":"Wang","suffix":""},{"id":578522303,"identity":"07dc8a60-9079-4ca8-b087-c70ecb97fcb6","order_by":4,"name":"Ruijie Yanglan","email":"","orcid":"","institution":"The First Affiliated Hospital of Kunming Medical University","correspondingAuthor":false,"prefix":"","firstName":"Ruijie","middleName":"","lastName":"Yanglan","suffix":""},{"id":578522304,"identity":"b29fdb70-30f5-4030-badf-e822470c1eda","order_by":5,"name":"Qinlan Li","email":"","orcid":"","institution":"The First Affiliated Hospital of Kunming Medical University","correspondingAuthor":false,"prefix":"","firstName":"Qinlan","middleName":"","lastName":"Li","suffix":""},{"id":578522306,"identity":"133432f0-3282-45df-8610-74d7ad10b3ae","order_by":6,"name":"Yangjuan Bai","email":"","orcid":"","institution":"The First Affiliated Hospital of Kunming Medical University","correspondingAuthor":false,"prefix":"","firstName":"Yangjuan","middleName":"","lastName":"Bai","suffix":""},{"id":578522309,"identity":"fffb6cae-3948-4e84-918f-1e551a30efa5","order_by":7,"name":"Wei Wei","email":"","orcid":"","institution":"The First Affiliated Hospital of Kunming Medical University","correspondingAuthor":false,"prefix":"","firstName":"Wei","middleName":"","lastName":"Wei","suffix":""},{"id":578522314,"identity":"9d5387cd-fb30-4852-8999-4126a7c26e77","order_by":8,"name":"Fang Ma","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA3klEQVRIiWNgGAWjYFACHijNzsBw4IOBjRwJWpgZGA7OKEgzJk0LM8+Hw4kENZi39x588KPGLpq/mfngYRsD5gQG9sNHN+DTInPmXLJhz7Hk3BmH2RIO5xiw5THwpKXdwKdFQiLHTJqxgTm34TCPAVALTzGDBI8Zfi3yb8x/MzbU584/zP/hsIWBRGIDQS1ABcyMDYdzNxzmYTjMYGBAhBaevGTJnmPHczceZjM42GOQYMxG0C/sZw9++FFTnTvvePPjDz/+/JfjZz98DK8WTMBGmvJRMApGwSgYBdgAAHJoSMhoKUk9AAAAAElFTkSuQmCC","orcid":"","institution":"The First Affiliated Hospital of Kunming Medical University","correspondingAuthor":true,"prefix":"","firstName":"Fang","middleName":"","lastName":"Ma","suffix":""}],"badges":[],"createdAt":"2026-01-07 15:08:53","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8543031/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8543031/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":101171493,"identity":"139fe0d1-4cee-4f4c-9df9-cd46ddc4c7bb","added_by":"auto","created_at":"2026-01-27 00:08:48","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":85012,"visible":true,"origin":"","legend":"","description":"","filename":"Manuscript.docx","url":"https://assets-eu.researchsquare.com/files/rs-8543031/v1/36a1279dca2a200e9bfebd55.docx"},{"id":101171558,"identity":"342f4f12-3237-4518-a492-627a5387d32b","added_by":"auto","created_at":"2026-01-27 00:08:58","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":676296,"visible":true,"origin":"","legend":"","description":"","filename":"Figures.docx","url":"https://assets-eu.researchsquare.com/files/rs-8543031/v1/a17fd6705cc21e492c1a6d0e.docx"},{"id":101171556,"identity":"1821550a-bc00-4ca7-8ffd-7c06d903b3c1","added_by":"auto","created_at":"2026-01-27 00:08:58","extension":"docx","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":40925,"visible":true,"origin":"","legend":"","description":"","filename":"Tables.docx","url":"https://assets-eu.researchsquare.com/files/rs-8543031/v1/32ece8acb018c46a5b078236.docx"},{"id":101206646,"identity":"d608d29d-1ebe-4a39-8c61-15ca99f7bea2","added_by":"auto","created_at":"2026-01-27 09:56:35","extension":"json","order_by":3,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":9739,"visible":true,"origin":"","legend":"","description":"","filename":"f1f4938ac8eb461d81e67490651659cc.json","url":"https://assets-eu.researchsquare.com/files/rs-8543031/v1/14f238e11e3d7848b9df511c.json"},{"id":101171550,"identity":"6694e662-d459-420d-a491-04b9749549e9","added_by":"auto","created_at":"2026-01-27 00:08:58","extension":"docx","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":28033,"visible":true,"origin":"","legend":"","description":"","filename":"BaselineCharacteristicsoftrainingsetandvalidationset.docx","url":"https://assets-eu.researchsquare.com/files/rs-8543031/v1/c8196f5ec536f4b4bb903d5d.docx"},{"id":101171640,"identity":"a0c24ebe-345d-4c49-bebe-478097159d96","added_by":"auto","created_at":"2026-01-27 00:09:02","extension":"docx","order_by":5,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":20546,"visible":true,"origin":"","legend":"","description":"","filename":"Questionnaire.docx","url":"https://assets-eu.researchsquare.com/files/rs-8543031/v1/6b824ba4c8046d5391c5e9dc.docx"},{"id":101171507,"identity":"6b6ce33e-e431-4a65-ba78-81be9ea3c257","added_by":"auto","created_at":"2026-01-27 00:08:50","extension":"xml","order_by":6,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":165044,"visible":true,"origin":"","legend":"","description":"","filename":"f1f4938ac8eb461d81e67490651659cc1enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-8543031/v1/599742b3910a8c39ba7c4944.xml"},{"id":101171559,"identity":"90d86a65-ca0f-4006-b4a4-7ed0b31c100e","added_by":"auto","created_at":"2026-01-27 00:08:59","extension":"png","order_by":7,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":93715,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-8543031/v1/0a839eb2a40b828fef1eb1e8.png"},{"id":101171560,"identity":"fcceac36-eb74-4af6-b7a5-64e094db1603","added_by":"auto","created_at":"2026-01-27 00:08:59","extension":"jpeg","order_by":8,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":245183,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage2.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-8543031/v1/a3a3c4ed53054fed61d3b214.jpeg"},{"id":101171543,"identity":"cb94a1c4-615a-46f5-be72-28e632c55bfe","added_by":"auto","created_at":"2026-01-27 00:08:57","extension":"jpeg","order_by":9,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":490035,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage3.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-8543031/v1/3160c4521db65b9b2ccfb4fa.jpeg"},{"id":101207007,"identity":"a9eaced2-0967-41b4-9a09-caf7fa2d8d1f","added_by":"auto","created_at":"2026-01-27 09:57:07","extension":"jpeg","order_by":10,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":273678,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage4.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-8543031/v1/ce6ca0addbb3e26d2dbb5386.jpeg"},{"id":101171539,"identity":"b73def47-52d9-45fb-afe5-1acf2e500f04","added_by":"auto","created_at":"2026-01-27 00:08:57","extension":"png","order_by":11,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":92971,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-8543031/v1/b0f2683ee6e1fe5196093067.png"},{"id":101207432,"identity":"20bf1481-702a-4dc0-9922-ed9cab08b252","added_by":"auto","created_at":"2026-01-27 10:04:05","extension":"png","order_by":12,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":43590,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-8543031/v1/725429c95115d4d641ce9ef1.png"},{"id":101171495,"identity":"f00befdb-2647-4296-8107-8eaf0579d08a","added_by":"auto","created_at":"2026-01-27 00:08:49","extension":"png","order_by":13,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":106701,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-8543031/v1/ffb99ebf1b77f36719bc55f3.png"},{"id":101171634,"identity":"ded600d1-b9ca-4a26-9c5b-b404960b18a4","added_by":"auto","created_at":"2026-01-27 00:09:01","extension":"png","order_by":14,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":60115,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-8543031/v1/371de23c3dbafa72c73221c7.png"},{"id":101171655,"identity":"1713ebf1-40c4-4967-b7d0-23c397969c7d","added_by":"auto","created_at":"2026-01-27 00:09:05","extension":"xml","order_by":15,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":161336,"visible":true,"origin":"","legend":"","description":"","filename":"f1f4938ac8eb461d81e67490651659cc1structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-8543031/v1/7b0f09bca40a584709a4da89.xml"},{"id":101171505,"identity":"f3a874d7-afa1-44b5-ab1f-8acf971c9cf6","added_by":"auto","created_at":"2026-01-27 00:08:50","extension":"html","order_by":16,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":177245,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-8543031/v1/bccd19bbca5939528f28425e.html"},{"id":101171643,"identity":"2af8c50f-e31b-454b-9d10-00dad89480a8","added_by":"auto","created_at":"2026-01-27 00:09:04","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":93715,"visible":true,"origin":"","legend":"\u003cp\u003eFlow chart of the study\u003c/p\u003e","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-8543031/v1/4f5419a0fa01d1749f474eff.png"},{"id":101206087,"identity":"059e7774-6ec8-4c0c-9e8d-dbfb13f54baa","added_by":"auto","created_at":"2026-01-27 09:53:31","extension":"jpeg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":148614,"visible":true,"origin":"","legend":"\u003cp\u003eLASSO was used for preliminary variable screening. \u003cstrong\u003eA.\u003c/strong\u003eThe shrinkage paths of feature coefficients and the trend of variation in the number of non - zero features under different regularization strengths (λ); \u003cstrong\u003eB.\u003c/strong\u003e Based on the results of 10-fold cross-validation, vertical lines were drawn at the point of minimum error and within one standard deviation (λ=0.0541) of the minimum error, corresponding to model feature subsets of different complexities.\u003c/p\u003e","description":"","filename":"floatimage2.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-8543031/v1/8d18112a225a0bd0f84d1e21.jpeg"},{"id":101171527,"identity":"bd575c68-f1cd-444e-8a91-02a2b3d56fe4","added_by":"auto","created_at":"2026-01-27 00:08:56","extension":"jpeg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":490035,"visible":true,"origin":"","legend":"\u003cp\u003eThe Performance of trainingand validation sets for machine learning models. \u003cstrong\u003eA.\u003c/strong\u003e ROC curves of machine learning models in training sets. \u003cstrong\u003eB.\u003c/strong\u003e Calibration Curve of machine learning models in training sets. \u003cstrong\u003eC.\u003c/strong\u003e ROC curves of machine learning models in validation sets. \u003cstrong\u003eD.\u003c/strong\u003e Calibration Curve of machine learning models in validation sets.\u003c/p\u003e","description":"","filename":"floatimage3.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-8543031/v1/9a9fbf6570293e2c3612be1e.jpeg"},{"id":101171522,"identity":"75a3b44d-110f-48cf-827c-2d7c08bbcb4a","added_by":"auto","created_at":"2026-01-27 00:08:52","extension":"jpeg","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":273678,"visible":true,"origin":"","legend":"\u003cp\u003eExplain the model features using SHAP values. \u003cstrong\u003eA.\u003c/strong\u003e Summary plot of SHAP features. \u003cstrong\u003eB.\u003c/strong\u003eFeatures contribution by RF model.\u003c/p\u003e","description":"","filename":"floatimage4.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-8543031/v1/0471216a98fd1afd0190d89e.jpeg"},{"id":108006563,"identity":"8acf7f4f-2247-41d0-afd8-a21f424a9d75","added_by":"auto","created_at":"2026-04-28 12:56:03","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1674739,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8543031/v1/145065fa-0e90-46cf-a602-3910a8440368.pdf"},{"id":101171503,"identity":"b7694a55-593e-43a0-84b3-5571863a251c","added_by":"auto","created_at":"2026-01-27 00:08:50","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":28033,"visible":true,"origin":"","legend":"","description":"","filename":"BaselineCharacteristicsoftrainingsetandvalidationset.docx","url":"https://assets-eu.researchsquare.com/files/rs-8543031/v1/49681ace9d8ab015e41fbb4f.docx"},{"id":101171548,"identity":"7dc40f0e-4066-4c78-906b-3b1dcb506113","added_by":"auto","created_at":"2026-01-27 00:08:58","extension":"docx","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":20546,"visible":true,"origin":"","legend":"","description":"","filename":"Questionnaire.docx","url":"https://assets-eu.researchsquare.com/files/rs-8543031/v1/074568bafbedc683c685286b.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Machine Learning-Based Risk Prediction Model for Fatigue in Chronic Heart Failure Patients","fulltext":[{"header":"1. Introduction","content":"\u003cp\u003eChronic heart failure (CHF), as the persistent stage of heart failure, is the terminal stage of various cardiovascular diseases and the leading cause of death\u003csup\u003e\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e\u003c/sup\u003e. Patients with CHF experienced symptoms such as dyspnea, fatigue, and depression, leading to adverse health consequences characterized by frequent exacerbations and readmissions\u003csup\u003e\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e. At present, for patients with CHF, alleviating the symptoms is the primary treatment goal. Heart failure guidelines also emphasize the importance of symptom monitoring and management in the long-term care of CHF patients, which can effectively improve the clinical symptoms of CHF patients, reduce medical costs and lower mortality rates\u003csup\u003e\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e,\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eAmong CHF symptoms, fatigue is considered one of the most common symptoms, with a prevalence rate of 42% to 82%\u003csup\u003e5\u003c/sup\u003e. Fatigue in patients with CHF manifests in the early stage of the disease and persists throughout its course, significantly impairing their ability to perform daily activities. Furthermore, due to increased dependence on others and restricted social activities, patients' mental health and social functioning are also compromised, seriously affecting their quality of life\u003csup\u003e\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e,\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e\u003c/sup\u003e. Stanek et al.\u003csup\u003e8\u003c/sup\u003e found that CHF patients were more interested in relieving fatigue than in increasing their life expectancy. However, fatigue is rarely assessed or treated. Research indicated that healthcare professionals typically do not discuss the experience of fatigue and their impact with individuals or families living with chronic illness\u003csup\u003e\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e\u003c/sup\u003e. The reason for this may be the vague and relative nature of fatigue, which makes it difficult to identify and manage, posing a challenge to current symptom management of CHF patients\u003csup\u003e\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e\u003c/sup\u003e. Present research indicated that the early identification of fatigue in patients with CHF can facilitate timely medical intervention, reduce hospitalization rates, and enhance their quality of life\u003csup\u003e\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e\u003c/sup\u003e. Therefore, if medical staff can accurately predict the risk of fatigue in CHF patients at an early stage and take precise intervention measures, it can effectively reduce the incidence of fatigue and improve the clinical outcomes of CHF patients.\u003c/p\u003e \u003cp\u003eExisting research primarily investigates potential risk factors for fatigue in patients with CHF, such as New York Heart Association (NYHA) classification\u003csup\u003e\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e\u003c/sup\u003e, Left ventricular ejection fraction (LVEF)\u003csup\u003e\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e\u003c/sup\u003e, gender\u003csup\u003e\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e\u003c/sup\u003e, and depression\u003csup\u003e\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e\u003c/sup\u003e. However, most studies have small sample sizes, posing a risk of statistical overfitting. Furthermore, current research on fatigue prediction has mainly focused on patients with malignant tumors\u003csup\u003e\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e,\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e\u003c/sup\u003e, with a notable lack of fatigue risk prediction models specifically for those with CHF. Therefore, there is an urgent clinical need to develop an accurate and practical fatigue risk prediction tool for CHF patients.\u003c/p\u003e \u003cp\u003eIn recent years, information technology has been widely applied in the medical field. As a core technology in the field of artificial intelligence, machine learning (ML) has demonstrated excellent predictive performance in the diagnosis and prognosis of cardiovascular diseases\u003csup\u003e\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e\u003c/sup\u003e. Compared with traditional risk prediction models, ML demonstrates superior sensitivity and specificity, and effectively addresses the limitations of conventional models in handling large-scale, high-dimensional, nonlinear, and incomplete data\u003csup\u003e\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e\u003c/sup\u003e, thereby enabling more accurate and efficient risk estimation for fatigue. In a comparison of various risk prediction models, risk prediction models based on ML outperformed models based on traditional regression in terms of predictive performance\u003csup\u003e\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e,\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e\u003c/sup\u003e. Therefore, this study aimed to construct a risk prediction model for fatigue in patients with CHF based on ML, providing healthcare professionals with an assessment tool for early identification of CHF patients at high risk of fatigue.\u003c/p\u003e"},{"header":"2 Method","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003e2.1 Study design\u003c/h2\u003e \u003cp\u003eThis study is a cross-sectional study. Data were collected using a questionnaire (Supplementary Material file) and divided into a training set and a validation set at a ratio of 7:3. Five ML algorithms, including Least Absolute Shrinkage and Selection Operator (LASSO)-Logistic regression (LR), Random Forest (RF), K-Nearest Neighbors (KNN), eXtreme Gradient Boosting (XGBoost), and Decision Tree (DT) were used for model construction, evaluation and validation. Finally, the optimal model is selected as a clinical assessment tool (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003e2.2 Study participants\u003c/h2\u003e \u003cp\u003eIn this study, by using the convenience sampling method, patients diagnosed with CHF at two tertiary hospitals in Yunnan Province between May 10, 2024, and October 31, 2024 were recruited as the research participants. The inclusion criteria were (1) aged\u0026thinsp;\u0026ge;\u0026thinsp;18 years; (2) met the diagnostic criteria for CHF of the 2022 AHA/ACC/HFSA Guidelines for the Management of Heart Failure\u003csup\u003e\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e\u003c/sup\u003e; (3) willing to participate in this study and had normal comprehension and verbal ability. The exclusion criteria were (1) concurrent malignant tumors or severe chronic diseases; (2) mental illness or current consciousness impairment that prevent effective communication; (3) missing more than 30% of important data.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003e2.3 Study tools\u003c/h2\u003e \u003cdiv id=\"Sec6\" class=\"Section3\"\u003e \u003ch2\u003e2.3.1 Fatigue assessment\u003c/h2\u003e \u003cp\u003eThis study used the Multidimensional Fatigue Inventory (MFI) to assess fatigue in CHF patients. This scale was developed by Smets\u003csup\u003e\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e\u003c/sup\u003e, consists of five dimensions (general fatigue, physical fatigue, reduced activity, decreased motivation, and mental fatigue) and 20 items. According to the literature, when the score of the general fatigue dimension reaches 12 points or higher (items 1, 5, 12, 16), fatigue is diagnosed \u003csup\u003e\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e,\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e\u003c/sup\u003e. The scale demonstrates good internal consistency, with a Cronbach's α coefficient of 0.882\u003csup\u003e23\u003c/sup\u003e.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec7\" class=\"Section3\"\u003e \u003ch2\u003e2.3.2 Predictor variables\u003c/h2\u003e \u003cp\u003eAfter reviewing the literature\u003csup\u003e\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e,\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e,\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e\u003c/sup\u003e and combining it with clinical experience, we selected 29 potential risk factors, including \u003cb\u003e(1) demographic data\u003c/b\u003e: age, gender, convenience from the place of residence to the hospital, educational level, marital status, occupation; \u003cb\u003e(2) biological data\u003c/b\u003e: dyspnea, NYHA classification, LVEF, cerebrovascular disease, anemia, diabetes, number of comorbidities, hypertension, body mass index (BMI), sleep quality, activities of daily living (ADL), hemoglobin, hematocrit, N-terminal (NT)-pro brain natriuretic peptide (BNP), beta-blockers, statins, diuretics, nitrates, calcium channel blockers, psychotropics; \u003cb\u003e(3) psychological data\u003c/b\u003e: anxiety, depression, and \u003cb\u003e(4) sociological data\u003c/b\u003e: social support. We assessed the variables of sleep quality, activities of daily living, anxiety, depression, and social support using the following scales:\u003c/p\u003e \u003cp\u003e(1) Sleep quality: The Pittsburgh Sleep Quality Index(PSQI) is mainly used to assess sleep patterns over the past month. The total score ranges from 0 to 21, with higher scores indicating poorer sleep quality. The score of PSQI\u0026thinsp;\u0026le;\u0026thinsp;7 indicates normal sleep quality, while the score\u0026thinsp;\u0026gt;\u0026thinsp;7 indicates poor sleep quality. The Cronbach's α coefficient is 0.84\u003csup\u003e26\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003e(2) Activities of daily living: The Activities of Daily Living (ADL) Scale was used in this study, which includes 10 dimensions of daily living activities. The total score ranges from 0 to 100 points. A score of \u0026ge;\u0026thinsp;60 indicates good self-care ability, while a score of \u0026lt;\u0026thinsp;60 indicates functional impairment. The Cronbach's α coefficient was 0.887\u003csup\u003e27\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003e(3) Anxiety and depression: The Hospital Anxiety and Depression Scale (HADS) is primarily used to evaluate the anxiety and depression levels of hospitalized patients, consisting of two subscales for anxiety and depression, each with seven items. Each item is scored on a scale of 0 to 3. A total score of 0 to 7 indicates no anxiety or depression, 8 to 10 indicates mild anxiety or depression, and 11 to 21 indicates severe anxiety or depression. The Cronbach's α coefficients for the HADS total scale and the two subscales are 0.89, 0.82, and 0.80, respectively\u003csup\u003e\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003e(4) Social support: The Social Support Rating Scale (SSRS) encompasses three dimensions and ten items. The total score ranges from 0 to 66 points, with 0ཞ22 points indicating low social support, 23ཞ44 points indicating moderate social support, and 45ཞ66 points indicating high social support. The Cronbach's α coefficient is 0.92\u003csup\u003e14\u003c/sup\u003e.\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003e2.4 Data collection\u003c/h2\u003e \u003cp\u003eThis study was conducted by two researchers who had undergone scientific research training. After obtaining the informed consent of the research participants, data was gathered through face-to-face methods. Some data was obtained by the investigators from the electronic medical record system. After the data collection was completed, two investigators verified the collected data to ensure the quality.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec9\" class=\"Section2\"\u003e \u003ch2\u003e2.5 Statistical analysis\u003c/h2\u003e \u003cdiv id=\"Sec10\" class=\"Section3\"\u003e \u003ch2\u003e2.5.1 Data preprocessing\u003c/h2\u003e \u003cp\u003eBefore developing the model, two researchers were responsible for data entry and management using Epidata 3.0. A cross-verification approach was employed to check for data issues such as duplicates, missing values, or errors that could compromise data accuracy. Duplicate data was removed. For noise values, we described the distribution of each variable using SPSS 26.0, including the maximum and minimum values. Values exceeding the reasonable threshold were removed and processed as missing values. When the missing value rate exceeded 30% and the missing pattern was random, the variable and patient data were removed. In our study, the missing data of NT-proBNP exceeded 40%, which was excluded. The mean was used to impute the missing values because the missing values for the other variables were less than 5% and showed a random missing pattern.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec11\" class=\"Section3\"\u003e \u003ch2\u003e2.5.2 Data analysis\u003c/h2\u003e \u003cp\u003eSPSS 26.0 software was used for descriptive analysis and comparison of inter-group data. All statistical tests adopted a two-tailed test approach, with a p-value\u0026thinsp;\u0026lt;\u0026thinsp;0.05 indicating statistically significant differences. Continuous data with normal distribution were presented as mean and standard deviation, while non-normally distributed data were described by the median (P\u003csub\u003e25\u003c/sub\u003e, P\u003csub\u003e75\u003c/sub\u003e). Categorical data were described by frequency and percentage. In the comparison of inter-group data, for categorical data, the χ\u003csup\u003e2\u003c/sup\u003e test was employed. Continuous variables that follow a normal distribution, Student's t-test was utilized, while for non-normally distributed variables, the Mann-Whitney U rank-sum test was used for inter-group comparisons.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section3\"\u003e \u003ch2\u003e2.5.3 Model construction and validation\u003c/h2\u003e \u003cp\u003eR 3.0 was used for model construction and validation. Based on the \u0026ldquo;glmnet\u0026rdquo; package, LASSO was applied to screen predictive factors. Subsequently, the selected predictive factors were incorporated into a multivariate logistic regression analysis using the forward-backward stepwise method. The final variables entered into the model were determined according to the optimal results of the Akaike Information Criterion (AIC). Furthermore, we utilized software packages including \u0026ldquo;glmnet\u0026rdquo;, \u0026ldquo;class\u0026rdquo;, \u0026ldquo;random forest\u0026rdquo;, \u0026ldquo;xgboost\u0026rdquo;, and \u0026ldquo;rpart\u0026rdquo; to construct models. Optimal parameters for each model were selected using 10-fold cross-validation to ensure model performance and generalization capability.\u003c/p\u003e \u003cp\u003eTo assess the predictive performance of the models, this research employed the area under the receiver operating characteristic (ROC) curve (AUC) as the key metric. It reflects the classification performance of the model across various thresholds, and a value closer to 1 indicates a stronger discriminatory power of the model. Accuracy was computed to represent the overall predictive performance. Sensitivity was used to measure the ability to identify positive cases. Specificity was employed to reflect the accuracy of predicting negative cases. The F1-score was used to balance the model's performance between sensitivity and precision. Additionally, calibration curves and Brier scores were adopted to evaluate the calibration of the predictive models.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section3\"\u003e \u003ch2\u003e2.5.4 Model interpretation\u003c/h2\u003e \u003cp\u003eSHAP (SHapley Additive exPlanation) is a model-interpretability tool based on Shapley values. In this study, after selecting the optimal prediction model, the SHAP plot was used to interpret the model by determining the importance of each predictor in the decision-making process and visualizing the results.\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e"},{"header":"3 Results","content":"\u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003e3.1 Characteristics of the study participants\u003c/h2\u003e \u003cp\u003eA total of 1186 questionnaires were distributed in this study. Seven patients were excluded due to concomitant malignant tumors, and eight patients were excluded due to missing variables\u0026thinsp;\u0026gt;\u0026thinsp;30%. Ultimately, 1171 patients were collected, and 484 patients (41.3%) experienced fatigue. The dataset was divided into a training set (819 cases) and a validation set (352 cases) at a ratio of 7:3. The age range of the study participants was from 18 to 96 years old, and the number of people aged 66 to 80 was the largest, reaching 529 (45.2%). There were 680 male participants (58.1%) and 491 female participants (41.9%). Characteristics of patients with statistical differences between the fatigue group and the non-fatigue group were shown in Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e1\u003c/span\u003e. The difference in baseline information between the training and test sets was not statistically significant (Supplementary file).\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eBaseline Characteristics of CHF Patients (n\u0026thinsp;=\u0026thinsp;1171).\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"5\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eVariables\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eOverall\u003c/p\u003e \u003cp\u003e(n\u0026thinsp;=\u0026thinsp;1171)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eNon-fatigue\u003c/p\u003e \u003cp\u003e(n\u0026thinsp;=\u0026thinsp;687)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eFatigue\u003c/p\u003e \u003cp\u003e(n\u0026thinsp;=\u0026thinsp;484)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003e\u003cem\u003eP\u003c/em\u003e value\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAge, n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.002\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u0026le;\u0026thinsp;30\u003c/p\u003e \u003cp\u003e31ཞ50\u003c/p\u003e \u003cp\u003e51\u0026thinsp;~\u0026thinsp;65\u003c/p\u003e \u003cp\u003e66\u0026thinsp;~\u0026thinsp;80\u003c/p\u003e \u003cp\u003e\u0026ge;\u0026thinsp;81\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e17 (1.5)\u003c/p\u003e \u003cp\u003e108 (9.2)\u003c/p\u003e \u003cp\u003e330 (28.2)\u003c/p\u003e \u003cp\u003e529 (45.2)\u003c/p\u003e \u003cp\u003e187 (16)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e15 (2.2)\u003c/p\u003e \u003cp\u003e77 (11.2)\u003c/p\u003e \u003cp\u003e193 (28.1)\u003c/p\u003e \u003cp\u003e304 (44.3)\u003c/p\u003e \u003cp\u003e98 (14.3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2 (0.4)\u003c/p\u003e \u003cp\u003e31 (6.4)\u003c/p\u003e \u003cp\u003e137 (28.3)\u003c/p\u003e \u003cp\u003e225 (46.5)\u003c/p\u003e \u003cp\u003e89 (18.4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGender, n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.022\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003emale\u003c/p\u003e \u003cp\u003efemale\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e680 (58.1)\u003c/p\u003e \u003cp\u003e491 (41.9)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e418 (60.8)\u003c/p\u003e \u003cp\u003e269 (39.2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e262 (54.1)\u003c/p\u003e \u003cp\u003e222 (45.9)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eConvenience from the place of residence to the hospital, n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ehighly inconvenient\u003c/p\u003e \u003cp\u003emoderately inconvenient\u003c/p\u003e \u003cp\u003erelatively convenient\u003c/p\u003e \u003cp\u003ehighly convenient\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e211 (18)\u003c/p\u003e \u003cp\u003e318 (27.2)\u003c/p\u003e \u003cp\u003e307 (26.2)\u003c/p\u003e \u003cp\u003e335 (28.6)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e135 (19.7)\u003c/p\u003e \u003cp\u003e183 (26.6)\u003c/p\u003e \u003cp\u003e152 (22.1)\u003c/p\u003e \u003cp\u003e217 (31.6)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e76 (15.7)\u003c/p\u003e \u003cp\u003e135 (27.9)\u003c/p\u003e \u003cp\u003e155 (32)\u003c/p\u003e \u003cp\u003e118 (24.4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eEducational level, n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.341\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eprimary school level and below\u003c/p\u003e \u003cp\u003ejunior middle school\u003c/p\u003e \u003cp\u003ehigh school\u003c/p\u003e \u003cp\u003euniversity and above\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e257 (53.1)\u003c/p\u003e \u003cp\u003e104 (21.5)\u003c/p\u003e \u003cp\u003e121 (25)\u003c/p\u003e \u003cp\u003e2 (0.4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e382 (55.6)\u003c/p\u003e \u003cp\u003e161 (23.4)\u003c/p\u003e \u003cp\u003e141 (20.5)\u003c/p\u003e \u003cp\u003e3 (0.4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e257 (53.1)\u003c/p\u003e \u003cp\u003e104 (21.5)\u003c/p\u003e \u003cp\u003e121 (25)\u003c/p\u003e \u003cp\u003e2 (0.4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMarital status, n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.028\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003esingle\u003c/p\u003e \u003cp\u003emarried\u003c/p\u003e \u003cp\u003edivorced\u003c/p\u003e \u003cp\u003ewidowed\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e32 (2.7)\u003c/p\u003e \u003cp\u003e887 (75.7)\u003c/p\u003e \u003cp\u003e19 (1.6)\u003c/p\u003e \u003cp\u003e233 (19.9)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e22 (3.2)\u003c/p\u003e \u003cp\u003e537 (78.2)\u003c/p\u003e \u003cp\u003e10 (1.5)\u003c/p\u003e \u003cp\u003e118 (17.2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e10 (2.1)\u003c/p\u003e \u003cp\u003e350 (72.3)\u003c/p\u003e \u003cp\u003e9 (1.9)\u003c/p\u003e \u003cp\u003e115 (23.8)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eOccupation, n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003efamer\u003c/p\u003e \u003cp\u003eproduction and transportation workers\u003c/p\u003e \u003cp\u003eenterprise staff\u003c/p\u003e \u003cp\u003emedical personnel\u003c/p\u003e \u003cp\u003eservice industry workers\u003c/p\u003e \u003cp\u003ecivil servants\u003c/p\u003e \u003cp\u003eretirees\u003c/p\u003e \u003cp\u003eunemployed\u003c/p\u003e \u003cp\u003eothers\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e395 (33.7)\u003c/p\u003e \u003cp\u003e26 (2.2)\u003c/p\u003e \u003cp\u003e41 (3.5)\u003c/p\u003e \u003cp\u003e1 (0.1)\u003c/p\u003e \u003cp\u003e10 (0.9)\u003c/p\u003e \u003cp\u003e13 (1.1)\u003c/p\u003e \u003cp\u003e453 (38.7)\u003c/p\u003e \u003cp\u003e201 (17.2)\u003c/p\u003e \u003cp\u003e31 (2.6)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e240 (34.9)\u003c/p\u003e \u003cp\u003e20 (2.9)\u003c/p\u003e \u003cp\u003e28 (4.1)\u003c/p\u003e \u003cp\u003e1 (0.1)\u003c/p\u003e \u003cp\u003e3 (0.4)\u003c/p\u003e \u003cp\u003e6 (0.9)\u003c/p\u003e \u003cp\u003e234 (34.1)\u003c/p\u003e \u003cp\u003e132 (19.2)\u003c/p\u003e \u003cp\u003e23 (3.3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e155 (32)\u003c/p\u003e \u003cp\u003e6 (1.2)\u003c/p\u003e \u003cp\u003e13 (2.7)\u003c/p\u003e \u003cp\u003e0 (0)\u003c/p\u003e \u003cp\u003e7 (1.4)\u003c/p\u003e \u003cp\u003e7 (1.4)\u003c/p\u003e \u003cp\u003e219 (45.2)\u003c/p\u003e \u003cp\u003e69 (14.3)\u003c/p\u003e \u003cp\u003e8 (1.7)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDyspnea, n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.069\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003cp\u003eno\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e498 (42.5)\u003c/p\u003e \u003cp\u003e673 (57.5)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e277 (40.3)\u003c/p\u003e \u003cp\u003e410 (59.7)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e221 (45.7)\u003c/p\u003e \u003cp\u003e263 (54.3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNYHA class, n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eⅠ\u003c/p\u003e \u003cp\u003eⅡ\u003c/p\u003e \u003cp\u003eⅢ\u003c/p\u003e \u003cp\u003eⅣ\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e59 (5)\u003c/p\u003e \u003cp\u003e546 (46.6)\u003c/p\u003e \u003cp\u003e452 (38.6)\u003c/p\u003e \u003cp\u003e114 (9.7)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e44 (6.4)\u003c/p\u003e \u003cp\u003e390 (56.8)\u003c/p\u003e \u003cp\u003e214 (31.1)\u003c/p\u003e \u003cp\u003e39 (5.7)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e15 (3.1)\u003c/p\u003e \u003cp\u003e156 (32.2)\u003c/p\u003e \u003cp\u003e238 (49.2)\u003c/p\u003e \u003cp\u003e75 (15.5)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLVEF, n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.537\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u0026le;\u0026thinsp;50%\u003c/p\u003e \u003cp\u003e\u0026gt;50%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e685 (58.5)\u003c/p\u003e \u003cp\u003e486 (41.5)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e407 (59.2)\u003c/p\u003e \u003cp\u003e280 (40.8)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e278 (57.4)\u003c/p\u003e \u003cp\u003e206 (42.6)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCerebrovascular disease, n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.048\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003cp\u003eno\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e134 (11.4)\u003c/p\u003e \u003cp\u003e1037 (88.6)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e68 (9.9)\u003c/p\u003e \u003cp\u003e619 (90.1)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e66 (13.6)\u003c/p\u003e \u003cp\u003e418 (86.4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAnemia, n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.453\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003cp\u003eno\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e91 (7.8)\u003c/p\u003e \u003cp\u003e1080 (92.2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e50 (7.3)\u003c/p\u003e \u003cp\u003e637 (92.7)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e41 (8.5)\u003c/p\u003e \u003cp\u003e443 (91.5)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDiabetes, n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.028\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003cp\u003eno\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e265 (22.6)\u003c/p\u003e \u003cp\u003e906 (77.4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e140 (20.4)\u003c/p\u003e \u003cp\u003e547 (79.6)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e125 (25.8)\u003c/p\u003e \u003cp\u003e359 (74.2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNumber of comorbidities, n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.035\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u0026lt;3\u003c/p\u003e \u003cp\u003e3ཞ5\u003c/p\u003e \u003cp\u003e\u0026gt;5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e483 (41.2)\u003c/p\u003e \u003cp\u003e565 (48.2)\u003c/p\u003e \u003cp\u003e123 (10.5)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e304 (44.3)\u003c/p\u003e \u003cp\u003e318 (46.3)\u003c/p\u003e \u003cp\u003e65 (9.5)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e179 (37)\u003c/p\u003e \u003cp\u003e247 (51)\u003c/p\u003e \u003cp\u003e58 (12)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHypertension, n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.954\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003cp\u003eno\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e577 (49.3)\u003c/p\u003e \u003cp\u003e594 (50.7)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e339 (49.3)\u003c/p\u003e \u003cp\u003e348 (50.7)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e238 (49.2)\u003c/p\u003e \u003cp\u003e246 (50.8)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003e(continued)\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"6\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003eVariables\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eOverall\u003c/p\u003e \u003cp\u003e(n\u0026thinsp;=\u0026thinsp;1171)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eNon-fatigue\u003c/p\u003e \u003cp\u003e(n\u0026thinsp;=\u0026thinsp;687)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eFatigue\u003c/p\u003e \u003cp\u003e(n\u0026thinsp;=\u0026thinsp;484)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u003cem\u003eP\u003c/em\u003e value\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBMI, n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.051\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"1\" nameend=\"c6\" namest=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u0026lt;18.5\u003c/p\u003e \u003cp\u003e18.50ཞ23.99\u003c/p\u003e \u003cp\u003e24.00ཞ27.99\u003c/p\u003e \u003cp\u003e\u0026ge;\u0026thinsp;28\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e124 (10.6)\u003c/p\u003e \u003cp\u003e568 (48.5)\u003c/p\u003e \u003cp\u003e333 (28.4)\u003c/p\u003e \u003cp\u003e146 (12.5)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e60 (8.7)\u003c/p\u003e \u003cp\u003e334 (48.6)\u003c/p\u003e \u003cp\u003e209 (30.4)\u003c/p\u003e \u003cp\u003e84 (12.2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e64 (13.2)\u003c/p\u003e \u003cp\u003e234 (48.3)\u003c/p\u003e \u003cp\u003e124 (25.6)\u003c/p\u003e \u003cp\u003e62 (12.8)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"1\" nameend=\"c6\" namest=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSleep quality, n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"1\" nameend=\"c6\" namest=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003enormal\u003c/p\u003e \u003cp\u003ebad\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e399 (34.1)\u003c/p\u003e \u003cp\u003e772 (65.9)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e321 (46.7)\u003c/p\u003e \u003cp\u003e366 (53.3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e78 (16.1)\u003c/p\u003e \u003cp\u003e406 (83.9)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"1\" nameend=\"c6\" namest=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eActivities of daily living, n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"1\" nameend=\"c6\" namest=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003enormal\u003c/p\u003e \u003cp\u003edisorder\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1090 (93.1)\u003c/p\u003e \u003cp\u003e81 (6.9)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e665 (96.8)\u003c/p\u003e \u003cp\u003e22 (3.2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e425 (87.8)\u003c/p\u003e \u003cp\u003e59 (12.2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"1\" nameend=\"c6\" namest=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHemoglobin, n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"1\" nameend=\"c6\" namest=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u0026lt;90\u003c/p\u003e \u003cp\u003e90ཞ160\u003c/p\u003e \u003cp\u003e\u0026gt;160\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e80 (6.8)\u003c/p\u003e \u003cp\u003e905 (77.3)\u003c/p\u003e \u003cp\u003e186 (15.9)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e36 (5.2)\u003c/p\u003e \u003cp\u003e521 (75.8)\u003c/p\u003e \u003cp\u003e130 (18.9)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e44 (9.1)\u003c/p\u003e \u003cp\u003e384 (79.3)\u003c/p\u003e \u003cp\u003e56 (11.6)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"1\" nameend=\"c6\" namest=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHematocrit (median [IQR],vol%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e42 (37,47)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e43 (38,48)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e41 (36,45)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"1\" nameend=\"c6\" namest=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBeta-blockers, n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.254\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"1\" nameend=\"c6\" namest=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003cp\u003eno\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e741 (63.3)\u003c/p\u003e \u003cp\u003e430 (36.7)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e444 (64.6)\u003c/p\u003e \u003cp\u003e243 (35.4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e297 (61.4)\u003c/p\u003e \u003cp\u003e187 (38.6)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"1\" nameend=\"c6\" namest=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eStatins, n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.911\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"1\" nameend=\"c6\" namest=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003cp\u003eno\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e641 (54.7)\u003c/p\u003e \u003cp\u003e530 (45.3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e377 (54.9)\u003c/p\u003e \u003cp\u003e310 (45.1)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e264 (54.5)\u003c/p\u003e \u003cp\u003e220 (45.5)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"1\" nameend=\"c6\" namest=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDiuretics, n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.865\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"1\" nameend=\"c6\" namest=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003cp\u003eno\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e980 (83.7)\u003c/p\u003e \u003cp\u003e191 (16.3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e576 (83.8)\u003c/p\u003e \u003cp\u003e111 (16.2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e404 (83.5)\u003c/p\u003e \u003cp\u003e80 (16.5)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"1\" nameend=\"c6\" namest=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNitrates, n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.146\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"1\" nameend=\"c6\" namest=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003cp\u003eno\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e67 (5.7)\u003c/p\u003e \u003cp\u003e1104 (94.3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e45 (6.6)\u003c/p\u003e \u003cp\u003e642 (93.4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e22 (4.5)\u003c/p\u003e \u003cp\u003e462 (95.5)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"1\" nameend=\"c6\" namest=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCalcium channel blockers, n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.617\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"1\" nameend=\"c6\" namest=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003cp\u003eno\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e122 (10.4)\u003c/p\u003e \u003cp\u003e1049 (89.6)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e69 (10)\u003c/p\u003e \u003cp\u003e618 (90)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e53 (11)\u003c/p\u003e \u003cp\u003e431 (89)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"1\" nameend=\"c6\" namest=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePsychotropics, n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.048\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"1\" nameend=\"c6\" namest=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003cp\u003eno\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e103 (8.8)\u003c/p\u003e \u003cp\u003e1068 (91.2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e51 (7.4)\u003c/p\u003e \u003cp\u003e636 (92.6)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e52 (10.7)\u003c/p\u003e \u003cp\u003e432 (89.3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"1\" nameend=\"c6\" namest=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAnxiety, n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"1\" nameend=\"c6\" namest=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNone\u003c/p\u003e \u003cp\u003eMild\u003c/p\u003e \u003cp\u003esevere\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e961 (82.1)\u003c/p\u003e \u003cp\u003e150 (12.8)\u003c/p\u003e \u003cp\u003e60 (5.1)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e622 (90.5)\u003c/p\u003e \u003cp\u003e47 (6.8)\u003c/p\u003e \u003cp\u003e18 (2.6)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e339 (70)\u003c/p\u003e \u003cp\u003e103 (21.3)\u003c/p\u003e \u003cp\u003e42 (8.7)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"1\" nameend=\"c6\" namest=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDepression, n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"1\" nameend=\"c6\" namest=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNone\u003c/p\u003e \u003cp\u003eMild\u003c/p\u003e \u003cp\u003esevere\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e955 (81.6)\u003c/p\u003e \u003cp\u003e156 (13.3)\u003c/p\u003e \u003cp\u003e60 (5.1)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e616 (89.7)\u003c/p\u003e \u003cp\u003e56 (8.2)\u003c/p\u003e \u003cp\u003e15 (2.2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e339 (70)\u003c/p\u003e \u003cp\u003e100 (20.7)\u003c/p\u003e \u003cp\u003e45 (9.3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"1\" nameend=\"c6\" namest=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSocial support, n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.027\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"1\" nameend=\"c6\" namest=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003elow level\u003c/p\u003e \u003cp\u003emiddle level\u003c/p\u003e \u003cp\u003ehihg level\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e61 (5.2)\u003c/p\u003e \u003cp\u003e1082 (92.4)\u003c/p\u003e \u003cp\u003e28 (2.4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e26 (3.8)\u003c/p\u003e \u003cp\u003e646 (94)\u003c/p\u003e \u003cp\u003e15 (2.2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e35 (7.2)\u003c/p\u003e \u003cp\u003e436 (90.1)\u003c/p\u003e \u003cp\u003e13 (2.7)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"1\" nameend=\"c6\" namest=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec16\" class=\"Section2\"\u003e \u003ch2\u003e3.2 Feature selection\u003c/h2\u003e \u003cp\u003eLASSO was used to perform dimensionality reduction analysis on 28 potential predictive variables. Some predictors with weak correlations to the outcome variable were automatically excluded as their coefficients were compressed to zero (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eA). The model was further optimized using 10-fold cross-validation. To ensure the model\u0026rsquo;s fitting performance and simplicity, we selected the lambda value corresponding to the smallest standard error. Under this lambda value, the original 28 variables were simplified to 5 key variables (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eB). The five variables identified by LASSO were incorporated into the multivariate logistic regression analysis. The results are presented in Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e2\u003c/span\u003e. When five predictive variables, NYHA classification (OR\u0026thinsp;=\u0026thinsp;1.688, 95% CI: 1.351\u0026ndash;2.116, P\u0026thinsp;\u0026lt;\u0026thinsp;0.001), sleep quality (OR\u0026thinsp;=\u0026thinsp;3.929, 95% CI: 2.745\u0026ndash;5.701, P\u0026thinsp;\u0026lt;\u0026thinsp;0.001), ADL (OR\u0026thinsp;=\u0026thinsp;2.508, 95% CI: 1.292\u0026ndash;5.089, P\u0026thinsp;=\u0026thinsp;0.008), anxiety (OR\u0026thinsp;=\u0026thinsp;1.899, 95% CI: 1.196\u0026ndash;3.060, P\u0026thinsp;=\u0026thinsp;0.007), and depression (OR\u0026thinsp;=\u0026thinsp;1.218, 95% CI: 0.786\u0026ndash;1.896, P\u0026thinsp;=\u0026thinsp;0.377), were combined, the corresponding AIC value was minimal, indicating optimal model performance.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab3\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eResults of Logistic Regression\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"7\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eVariables\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eB\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eSE\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eOR\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eCI ( 95%)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eWald χ\u003csup\u003e2\u003c/sup\u003e\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u003cem\u003eP\u003c/em\u003e value\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNYHA class\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.524\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.11426\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e1.688\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e1.688 (1.351\u0026ndash;2.116)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e4.583\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e\u0026lt;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSleep quality\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e1.368\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.18608\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e3.929\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e3.929 (2.745\u0026ndash;5.701)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e7.354\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e\u0026lt;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eActivities of daily living\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.919\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.34783\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e2.508\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e2.508(1.292\u0026ndash;5.089)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e2.643\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.008\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAnxiety\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.641\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.23849\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e1.899\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e1.898 (1.196\u0026ndash;3.060)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e2.689\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.007\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDepression\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.198\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.22349\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e1.218\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e1.218 (0.786\u0026ndash;1.896)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.884\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.377\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec17\" class=\"Section2\"\u003e \u003ch2\u003e3.3 Model performance\u003c/h2\u003e \u003cp\u003eWe trained and internally validated the model based on these five features, as shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e. Table\u0026nbsp;\u003cspan refid=\"Tab5\" class=\"InternalRef\"\u003e3\u003c/span\u003e summarizes the performance of the five ML models. RF, XGBoost, and LASSO-LR all exhibited favorable predictive performance in the training set and validation set. The AUC for RF was slightly higher compared to XGBoost and LASSO-LR, with AUC values of 0.761 (95% CI: 0.729\u0026ndash;0.794) in the training set, and 0.721 (95% CI: 0.669\u0026ndash;0.774) in the validation set, indicating its moderate to good classification accuracy and predictive performance. In addition, it also has better overall performance metrics, such as accuracy (reflecting the correctness of predictions), sensitivity (identifying positive cases), specificity (identifying negative cases), and F1 Score (balancing recall and precision). This suggested that RF has a slightly superior predictive ability for fatigue risk in CHF patients compared to other models.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab4\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003e\u003cb\u003eA\u003c/b\u003e The performance of the model in the training set\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"7\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eModel\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAUC\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eAccuracy\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eSensitivity\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eSpecificity\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eF1 Score\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003eBrier\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDT\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.736\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.707\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.662\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.739\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.65\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.198\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRF\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.761\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.713\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.647\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.759\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.65\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.234\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eXGBoost\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.753\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.705\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.653\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.741\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.645\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.197\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eKNN\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.669\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.680\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.490\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.813\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.557\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.285\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLR\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.759\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.707\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.659\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.741\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.649\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.194\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab5\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003e\u003cb\u003eB\u003c/b\u003e The performance of the model in the validation set\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"7\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eModel\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAUC\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eAccuracy\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eSensitivity\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eSpecificity\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eF1 Score\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003eBrier\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDT\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.684\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.673\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.626\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.707\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.615\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.216\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRF\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.721\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.676\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.612\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.722\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.612\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.263\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eXGBoost\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.706\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.676\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.619\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.717\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.615\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.210\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eKNN\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.633\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.642\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.442\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.785\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.508\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.198\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLR\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.718\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.676\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.626\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.712\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.617\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.208\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec18\" class=\"Section2\"\u003e \u003ch2\u003e3.4 Interpretation of the final model\u003c/h2\u003e \u003cp\u003eIn this study, the contribution of each feature to the RF model was quantified using SHAP values. The features are ranked in descending order of their contribution as follows: NYHA classification, anxiety, sleep quality, depression and ADL, as depicted in Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e. Each point represents a sample, with the horizontal axis displaying the SHAP values for each feature. Colors indicate the direction and relative magnitude of the feature within the classification, with red corresponding to high values, blue to low values, and purple to the mean. Taking the first row as an example, a high NYHA functional class (red) has a positive impact on the prediction, while a low NYHA functional class (blue) has a negative impact.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e"},{"header":"4 Discussion","content":"\u003cp\u003eAs one of the landmark symptoms of CHF, fatigue is prevalent among CHF patients and significantly impacts their quality of life. In this study, a total of 1171 patients with CHF were included. Among them, 484 patients developed fatigue, with an incidence of 41.3%, which was consistent with the results of the study by Walke et al.\u003csup\u003e28\u003c/sup\u003e. Despite the relatively high incidence of fatigue, medical staff and patients often overlook the experience of fatigue and its implications\u003csup\u003e\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e\u003c/sup\u003e, which suggest that healthcare professionals should strengthen the screening of CHF patients who are at high risk for fatigue to reduce the incidence.\u003c/p\u003e \u003cp\u003eCurrently, there is still a lack of risk prediction tools for fatigue in CHF patients. With the advent of the big data era, an increasing number of scholars are using ML to construct risk prediction models, which play an important role in disease diagnosis and management, as well as personalized treatment\u003csup\u003e\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e\u003c/sup\u003e. These models not only effectively reduce human intervention in the evaluation process, but also make the prediction results more accurate\u003csup\u003e\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e\u003c/sup\u003e. Therefore, this study introduced five ML algorithms to construct a fatigue risk prediction model for CHF patients, including LASSO-LR, RF, KNN, XGBoost, and DT. Multiple ML algorithms were selected to construct predictive models because they have different computational characteristics and have been widely used in other studies\u003csup\u003e\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e,\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e\u003c/sup\u003e. By integrating multiple decision trees, the RF model effectively processes high-dimensional, non-linear cardiovascular data that may contain missing values. Furthermore, the training process of RF can be parallelized, significantly expediting model construction and enhancing data fitting compared to traditional regression methods. Additionally, RF is characterized by high prediction accuracy and robust anti-noise capabilities, which contribute to improving the overall robustness of the model\u003csup\u003e\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e,\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eIn addition, the use of LASSO regression in ML for variable selection offers notable advantages. Through L1 regularization, it automatically shrinks the coefficients of insignificant variables to zero, thereby enabling efficient variable selection. Moreover, this approach effectively mitigates the risks of multicollinearity and overfitting\u003csup\u003e\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e\u003c/sup\u003e. As a result, the model can achieve a higher level of precision and sensitivity when predicting fatigue in patients with CHF. Another advantage of our research is the introduction of the SHAP method to explain the RF model. ML models are often referred to as black box models, lacking transparency in decision-making. While the SHAP algorithm can clearly explain the complex relationship between features and prediction results, and has significant advantages in model interpretability and visualization\u003csup\u003e\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e\u003c/sup\u003e. Through the output of SHAP, we found that the risk factors for fatigue of CHF patients were NYHA classification, anxiety, sleep quality, depression, and ADL.\u003c/p\u003e \u003cp\u003eNYHA classification is generally regarded as the cornerstone for evaluating the functional status and the severity of symptoms in patients with HF\u003csup\u003e36\u003c/sup\u003e. The results of our study showed that NYHA classification has the greatest contribution to fatigue, and the higher the NYHA classification, the greater the risk of fatigue in CHF patients. Multiple studies have also demonstrated that NYHA classification is significantly associated with fatigue in patients with CHF\u003csup\u003e\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e,\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e\u003c/sup\u003e. Therefore, it is necessary to promptly identify changes in patients' cardiac function and conduct in-depth research to strengthen the understanding of the relationship between NYHA classification and fatigue.\u003c/p\u003e \u003cp\u003eAnxiety is another strong predictor. Persistent mobility issues, palpitations, chest tightness, repeated hospitalizations, and uncertainty about the disease can all contribute to CHF patients\u0026rsquo; prolonged anxiety\u003csup\u003e\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e\u003c/sup\u003e. Long-term anxiety, in turn, can exacerbate the burden on the body, leading to excessive physical exertion, weakened immunity and increased fatigue\u003csup\u003e\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e\u003c/sup\u003e. However, few studies have paid attention to the anxiety of patients with CHF, which indicated that medical staff may need to devote more attention to the anxiety of CHF patients and provide them with mental health support as well as management of symptom distress.\u003c/p\u003e \u003cp\u003eInterestingly, in our study, the results of multivariate logistic regression indicated that depression was not significantly associated with fatigue in CHF patients, but it had the smallest AIC value when combined with the other four variables. This suggested that depression has a potential contribution to the prediction of fatigue and can improve the predictive performance of the model through the combined effect with other variables. Consequently, our research incorporated depression into the model as a predictive factor for fatigue. The relationship between depression and fatigue in CHF patients may be influenced by reduced cerebral blood flow, along with the effects of depression on executive dysfunction and individual independence\u003csup\u003e\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e,\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e\u003c/sup\u003e. Future studies could further analyze the association between fatigue and depression in CHF patients from the perspectives of variable interactions, dynamic relationships, and potential mechanisms.\u003c/p\u003e \u003cp\u003eSleep quality ranked third in the importance of risk factors. Research has found that the reduced cardiac output in patients with CHF leads to insufficient cerebral perfusion, which may cause changes in the function of the central nervous system related to sleep regulation, leading to alterations in the patients' sleep characteristics. Additionally, factors such as the use of diuretics disrupt the patients' circadian rhythms and negatively influence patients\u0026rsquo; sleep quality, ultimately resulting in physical fatigue\u003csup\u003e\u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e,\u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e42\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eADL is also a predictor of fatigue. On the one hand, impaired ADL indicates that the patient is in a poor state of bone marrow suppression and malnutrition, which leads to a reduction in muscle mass and the occurrence of physical fatigue. On the other hand, limited self-care ability further caused muscle atrophy, decreased cardiopulmonary function, and aggravated fatigue\u003csup\u003e\u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e\u003c/sup\u003e. At present, data on the effects of exercise suggests that helping patients increase their activity levels may be the best way to manage fatigue\u003csup\u003e\u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e44\u003c/span\u003e\u003c/sup\u003e. Therefore, healthcare professionals should encourage CHF patients to develop personalized exercise plans based on their individual circumstances, gradually increasing their activity levels to improve physical fitness and alleviate fatigue.\u003c/p\u003e \u003cp\u003e \u003cem\u003eLimitations\u003c/em\u003e \u003c/p\u003e \u003cp\u003eThis study has several limitations. Firstly, the data collection for this study was limited to two tertiary hospitals in Yunnan Province, and the sample coverage has certain geographical limitations. Secondly, our study is a cross-sectional study, which can effectively reveal the correlation between variables, but cannot infer causal relationships, and it is difficult to conduct in-depth analysis of longitudinal clinical data, which may result in selection bias. Thirdly, the large amount of missing data for NT-proBNP affected the integrity of the data and limited our ability to conduct a more comprehensive analysis of the NT-proBNP indicator. Finally, this study did not conduct external validation, which might have limited the generalization ability of the model. Therefore, we fully verified the reliability and stability of the model through the internal validation method of K-fold cross-validation. To further enhance the model's clinical application value, future studies should expand the sample size and conduct multi-center external validation, incorporating additional objective measurement indicators to enhance the model's generalizability and applicability.\u003c/p\u003e"},{"header":"Conclusion","content":"\u003cp\u003eIn summary, by comparing the performance of various ML algorithms in predicting fatigue risk, we ultimately selected RF as the optimal model. Through SHAP interpretation of the model, we ranked the contribution of five predictive variables, including NYHA classification, anxiety, sleep quality, depression, and ADL. This provides a basis for healthcare professionals to CHF patients at high risk for fatigue at an early stage, and provide personalized treatment.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e \u003ch2\u003eConflicts of Interest\u003c/h2\u003e \u003cp\u003eThe authors have no conflicts of interest in this study.\u003c/p\u003e \u003c/p\u003e \u003cp\u003e \u003cstrong\u003eEthics approval\u003c/strong\u003e \u003cp\u003e The procedures followed in this study comply with the Declaration of Helsinki and were approved by the Ethics Committee of the First Affiliated Hospital of Kunming Medical University (Approval No.: 2024-L-8). All participants signed written informed consent forms.\u003c/p\u003e \u003c/p\u003e\u003ch2\u003eFunding\u003c/h2\u003e \u003cp\u003eThis research was supported by the Reserve Talent Project of Academic and Technological Leaders for Young and Middle-aged People in Yunnan Province, China, No. 202205AC160017 and the Yunnan Fundamental Research Kunming Medical University, No 202301AY070001-153.\u003c/p\u003e\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\u003cp\u003eMin Zhou (first author) contributed to the study design, data collection and analysis, and writing the manuscript. Fang Ma (corresponding author) was responsible for designing the study and revising the manuscript. Jingran Yang, Yimei Zhang, Yu Wang were responsible for data collection. Ruijie Yanglan, Qinlan Li were responsible for the data curation. Yangjuan Bai, Wei Wei contributed to the review of the manuscript. All authors read and approved the final manuscript.\u003c/p\u003e\u003ch2\u003eData Availability\u003c/h2\u003e\u003cp\u003eThe data that support the findings of this study are available on request from the corresponding author.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eHunt SA, American College of Cardiology, American Heart Association Task Force on Practice Guidelines (Writing Committee to Update the 2001 Guidelines for the Evaluation and Management of Heart Failure). ACC/AHA 2005 guideline update for the diagnosis and management of chronic heart failure in the adult: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (Writing Committee to Update the 2001 Guidelines for the Evaluation and Management of Heart Failure). J Am Coll Cardiol. 2005;46:e1\u0026ndash;82.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKim JSR, Risbud R, Gray C, Banerjee D, Trivedi R. The Dyadic Experience of Managing Heart Failure A Qualitative Investigation. J Cardiovasc Nurs. 2020;35:12\u0026ndash;8.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAlpert CM, Smith MA, Hummel SL, Hummel EK. Symptom burden in heart failure: assessment, impact on outcomes, and management. Heart Fail Rev. 2017;22:25\u0026ndash;39.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMcDonagh TA, Metra M, Adamo M, Gardner RS, Baumbach A, B\u0026ouml;hm M, Burri H, Butler J, Čelutkienė J, Chioncel O, Cleland JGF, Coats AJS, Crespo-Leiro MG, Farmakis D, Gilard M, Heymans S, Hoes AW, Jaarsma T, Jankowska EA, Lainscak M, Lam CSP, Lyon AR, McMurray JJV, Mebazaa A, Mindham R, Muneretto C, Francesco Piepoli M, Price S, Rosano GMC, Ruschitzka F, Kathrine Skibelund A, ESC Scientific Document Group. 2021 ESC Guidelines for the diagnosis and treatment of acute and chronic heart failure. Eur Heart J. 2021;42:3599\u0026ndash;726.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHiga H, Lennie TA, Chung ML, Tsuchihashi-Makaya M. Associations of multidimensional fatigue with the physical, psychological, and situational factors in outpatients with heart failure: a cross-sectional study. Eur J Cardiovasc Nurs 2022:zvac117.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSchjoedt I, Sommer I, Bjerrum MB. Experiences and management of fatigue in everyday life among adult patients living with heart failure: A systematic review of qualitative evidence. JBI Database Syst Reviews Implement Rep. 2016;14:68\u0026ndash;115.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChen L-H, Li C-Y, Shieh S-M, Yin W-H, Chiou A-F. Predictors of fatigue in patients with heart failure: Fatigue in patients with heart failure. J Clin Nurs. 2009;19:1588\u0026ndash;96.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eStanek EJ, Oates MB, McGhan WF, Denofrio D, Loh E. Preferences for treatment outcomes in patients with heart failure: Symptoms versus survival. J Card Fail. 2000;6:225\u0026ndash;32.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWhitehead L. The Family Experience of Fatigue in Heart Failure. J Fam Nurs. 2017;23:138\u0026ndash;56.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFalk K, Swedberg K, Gaston-Johansson F, Ekman I. Fatigue is a prevalent and severe symptom associated with uncertainty and sense of coherence in patients with chronic heart failure. Eur J Cardiovasc Nurs 2007;6.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTang W-R, Yu C-Y, Yeh S-J. Fatigue and its related factors in patients with chronic heart failure. J Clin Nurs. 2010;19:69\u0026ndash;78.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eH\u0026auml;gglund L, Boman K, Stenlund H, Lundman B, Brulin C. Factors related to fatigue among older patients with heart failure in primary health care. Int J Older People Nurs. 2008;3:96\u0026ndash;103.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHiga H, Lennie TA, Chung ML, Tsuchihashi-Makaya M. Associations of multidimensional fatigue with the physical, psychological, and situational factors in outpatients with heart failure: a cross-sectional study. Eur J Cardiovasc Nurs. 2023;22:647\u0026ndash;54.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHuang S-T, Ke X, Huang Y-P, Wu Y-X, Yu X-Y, Liu H-K, Liu D. A prediction model for moderate to severe cancer-related fatigue in colorectal cancer after chemotherapy: A prospective case\u0026ndash;control study. Support Care Cancer. 2023;31:426.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCourtier N, Gambling T, Enright S, Barrett-Lee P, Abraham J, Mason MD. A prognostic tool to predict fatigue in women with early-stage breast cancer undergoing radiotherapy. Breast. 2013;22:504\u0026ndash;9.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJing L, Ulloa Cerna AE, Good CW, Sauers NM, Schneider G, Hartzel DN, Leader JB, Kirchner HL, Hu Y, Riviello DM, Stough JV, Gazes S, Haggerty A, Raghunath S, Carry BJ, Haggerty CM, Fornwalt BK. A Machine Learning Approach to Management of Heart Failure Populations. JACC Heart Fail. 2020;8:578\u0026ndash;87.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKokori E, Patel R, Olatunji G, Ukoaka BM, Abraham IC, Ajekiigbe VO, Kwape JM, Babalola AE, Udam NG, Aderinto N. Machine learning in predicting heart failure survival: a review of current models and future prospects. Heart Fail Rev. 2024;30:431\u0026ndash;42.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYang J, Xie Y, Wang T, Pu Y, Ye T, Huang Y, Song B, Cheng F, Yang Z, Zhang X. Machine learning-based risk prediction of mild cognitive impairment in patients with chronic heart failure: A model development and validation study. Geriatr Nurs. 2025;62:145\u0026ndash;56.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHeidenreich PA, Bozkurt B, Aguilar D, Allen LA, Byun JJ, Colvin MM, Deswal A, Drazner MH, Dunlay SM, Evers LR, Fang JC, Fedson SE, Fonarow GC, Hayek SS, Hernandez AF, Khazanie P, Kittleson MM, Lee CS, Link MS, Milano CA, Nnacheta LC, Sandhu AT, Stevenson LW, Vardeny O, Vest AR, Yancy CW. 2022 AHA/ACC/HFSA Guideline for the Management of Heart Failure: A Report of the American College of Cardiology/American Heart Association Joint Committee on Clinical Practice Guidelines. Circulation. 2022;145:e895\u0026ndash;1032.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSmets EM, Garssen B, Bonke B, De Haes JC. The Multidimensional Fatigue Inventory (MFI) psychometric qualities of an instrument to assess fatigue. J Psychosom Res. 1995;39:315\u0026ndash;25.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAluzaite K, Al-Mandhari R, Osborne H, Ho C, Williams M, Sullivan M-M, Hobbs CE, Schultz M. Detailed Multi-Dimensional Assessment of Fatigue in Inflammatory Bowel Disease. Inflamm Intest Dis. 2019;3:192\u0026ndash;201.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePurcell A, Fleming J, Bennett S, Burmeister B, Haines T. Determining the minimal clinically important difference criteria for the Multidimensional Fatigue Inventory in a radiotherapy population. Support Care Cancer. 2010;18:307\u0026ndash;15.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMiao Y, Liu X, Liu Z, Xie H, Deng G. The preliminary revision of the Chinese version of the Multidimensional Fatigue Scale in military primary medical staff. Chin J Mental Health. 2008:658\u0026ndash;60.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePavlovic NV, Gilotra NA, Lee CS, Ndumele C, Mammos D, Dennisonhimmelfarb C, AbshireSaylor M. Fatigue in Persons With Heart Failure: A Systematic Literature Review and Meta-Synthesis Using the Biopsychosocial Model of Health. J Card Fail. 2022;28:283\u0026ndash;315.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWilliams BA. The clinical epidemiology of fatigue in newly diagnosed heart failure. BMC Cardiovasc Disord. 2017;17:122.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiao T, Gan Y, Liu L, Du Y, Li G, Guo M. The Association of Sleep Quality With Anorexia in Patients With Heart Failure: Do Anxiety and Depressive Symptoms Mediate This Association. J Cardiovasc Nurs 2025.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLawton MP, Brody EM. Assessment of older people: self-maintaining and instrumental activities of daily living. Gerontologist. 1969;9:179\u0026ndash;86.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWalke LM, Gallo WT, Tinetti ME, Fried TR. The burden of symptoms among community-dwelling older persons with advanced chronic disease. Arch Intern Med. 2004;164:2321\u0026ndash;4.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eIwasaki Y, Takeuchi I, Stanev V, Kusne AG, Ishida M, Kirihara A, Ihara K, Sawada R, Terashima K, Someya H, Uchida K, Saitoh E, Yorozu S. Machine-learning guided discovery of a new thermoelectric material. Sci Rep. 2019;9:2751.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLorenzoni G, Sabato SS, Lanera C, Bottigliengo D, Minto C, Ocagli H, De Paolis P, Gregori D, Iliceto S, Pisan\u0026ograve; F. Comparison of machine learning techniques for prediction of hospitalization in heart failure patients. JCM. 2019;8:1298.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHu Y, Ma F, Hu M, Shi B, Pan D, Ren J. Development and validation of a machine learning model to predict the risk of readmission within one year in HFpEF patients. Int J Med Informatics. 2025;194:105703.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eShi G, Liu G, Gao Q, Zhang S, Wang Q, Wu L, He P, Yu Q. A random forest algorithm-based prediction model for moderate to severe acute postoperative pain after orthopedic surgery under general anesthesia. BMC Anesthesiol. 2023;23:361.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMansoor H, Elgendy IY, Segal R, Bavry AA, Bian J. Risk prediction model for in-hospital mortality in women with ST-elevation myocardial infarction: A machine learning approach. Heart Lung. 2017;46:405\u0026ndash;11.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMcNeish DM. Using Lasso for Predictor Selection and to Assuage Overfitting: A Method Long Overlooked in Behavioral Sciences. Multivar Behav Res. 2015;50:471\u0026ndash;84.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLin Q, Zhao W, Zhang H, Chen W, Lian S, Ruan Q, Qu Z, Lin Y, Chai D, Lin X. Predicting the risk of heart failure after acute myocardial infarction using an interpretable machine learning model. Front Cardiovasc Med. 2025;12:1444323.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDickstein K, Cohen-Solal A, Filippatos G, McMurray JJV, Ponikowski P, Poole-Wilson PA, Str\u0026ouml;mberg A, van Veldhuisen DJ, Atar D, Hoes AW, Keren A, Mebazaa A, Nieminen M, Priori SG, Swedberg K, ESC Committee for Practice Guidelines (CPG). failure 2008: the Task Force for the diagnosis and treatment of acute and chronic heart failure 2008 of the European Society of Cardiology. Developed in collaboration with the Heart Failure Association of the ESC (HFA) and endorsed by the European Society of Intensive Care Medicine (ESICM). Eur J Heart Fail. 2008;10:933\u0026ndash;89. ESC guidelines for the diagnosis and treatment of acute and chronic heart.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eInamdar A, Inamdar A. Heart failure: Diagnosis, management and utilization. JCM. 2016;5:62.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRashid S, Qureshi AG, Noor TA, Yaseen K, Sheikh MAA, Malik M, Malik J. Anxiety and Depression in Heart Failure: An Updated Review. Curr Probl Cardiol. 2023;48:101987.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFalk K, Patel H, Swedberg K, Ekman I. Fatigue in patients with chronic heart failure - a burden associated with emotional and symptom distress. Eur J Cardiovasc Nurs. 2009;8:91\u0026ndash;6.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFink AM, Gonzalez RC, Lisowski T, Pini M, Fantuzzi G, Levy WC, Piano MR. Fatigue, inflammation, and projected mortality in heart failure. J Card Fail. 2012;18:711\u0026ndash;6.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRiegel B, Ratcliffe SJ, Sayers SL, Potashnik S, Buck HG, Jurkovitz C, Fontana S, Weaver TE, Weintraub WS, Goldberg LR. Determinants of excessive daytime sleepiness and fatigue in adults with heart failure. Clin Nurs Res. 2012;21:271\u0026ndash;93.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRedeker NS, Stein S. Characteristics of sleep in patients with stable heart failure versus a comparison group. Heart lung: J Crit care 2006;35.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGomes-Neto M, Dur\u0026atilde;es AR, Concei\u0026ccedil;\u0026atilde;o LSR, Roever L, Silva CM, Alves IGN, Ellingsen \u0026Oslash;, Carvalho VO. Effect of combined aerobic and resistance training on peak oxygen consumption, muscle strength and health-related quality of life in patients with heart failure with reduced left ventricular ejection fraction: a systematic review and meta-analysis. Int J Cardiol. 2019;293:165\u0026ndash;75.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHsiung P, Lin P-C, Lin T-Y, Wu W-T, Sun J-L, Chou P-L. Effectiveness of Exercise on Fatigue for Patients With Heart Failure: A Systematic Review and Meta-Analysis. J Appl Gerontol. 2025;44:312\u0026ndash;26.\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Chronic heart failure, Fatigue, Risk factors, Machine learning","lastPublishedDoi":"10.21203/rs.3.rs-8543031/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8543031/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eBackground\u003c/h2\u003e \u003cp\u003eAccurately identifying high-risk individuals with fatigue among patients with chronic heart failure (CHF) is crucial for improving their quality of life. This study aimed to construct a risk prediction model for fatigue in patients with CHF based on machine learning (ML) algorithms.\u003c/p\u003e\u003ch2\u003eMethod\u003c/h2\u003e \u003cp\u003eThe study population consisted of patients diagnosed with CHF at two tertiary hospitals in Yunnan from May 10, 2024, to October 31, 2024. LASSO (Least Absolute Shrinkage and Selection Operator) and logistic regression were employed for variable selection. Prediction models were developed and validated using five ML algorithms, and the model\u0026rsquo;s performance was assessed using several metrics, including the area under the receiver operating characteristic curve (ROC AUC), accuracy, sensitivity, specificity, F1 score, and brier score. SHAP (SHapley Additive exPlanations) plots were utilized for model interpretation.\u003c/p\u003e\u003ch2\u003eResults\u003c/h2\u003e \u003cp\u003eA total of 1171 CHF patients were included. Among the five ML models, Random Forest (RF) had the best predictive performance and was the optimal prediction model for fatigue in CHF patients. The best predictors identified included New York Heart Association (NYHA) classification, anxiety, sleep quality, depression, and activities of daily living (ADL).\u003c/p\u003e\u003ch2\u003eConclusion\u003c/h2\u003e \u003cp\u003eThe RF model demonstrated robust performance in predicting fatigue risk in CHF patients, providing a valuable tool for healthcare professionals to identify high-risk individuals and implement timely interventions.\u003c/p\u003e\u003ch2\u003eTrial registration:\u003c/h2\u003e \u003cp\u003eNot applicable\u003c/p\u003e","manuscriptTitle":"Machine Learning-Based Risk Prediction Model for Fatigue in Chronic Heart Failure Patients","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-01-27 00:08:32","doi":"10.21203/rs.3.rs-8543031/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"ed7d577d-1bc0-4b4e-8b83-df9e2a178ef4","owner":[],"postedDate":"January 27th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2026-04-27T13:56:24+00:00","versionOfRecord":[],"versionCreatedAt":"2026-01-27 00:08:32","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8543031","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8543031","identity":"rs-8543031","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}