Prognostic Machine Learning Models for First-Year Mortality in Incident Hemodialysis Patients: Development and Validation Study.
Ontology highlight
ABSTRACT:
SUBMITTER: Sheng K
PROVIDER: S-EPMC7661257 | biostudies-literature | 2020 Oct
REPOSITORIES: biostudies-literature
Publications
Prognostic Machine Learning Models for First-Year Mortality in Incident Hemodialysis Patients: Development and Validation Study.
Sheng Kaixiang K Zhang Ping P Yao Xi X Li Jiawei J He Yongchun Y Chen Jianghua J
JMIR medical informatics 20201029 10
<h4>Background</h4>The first-year survival rate among patients undergoing hemodialysis remains poor. Current mortality risk scores for patients undergoing hemodialysis employ regression techniques and have limited applicability and robustness.<h4>Objective</h4>We aimed to develop a machine learning model utilizing clinical factors to predict first-year mortality in patients undergoing hemodialysis that could assist physicians in classifying high-risk patients.<h4>Methods</h4>Training and testing ...[more]
Similar Datasets
Project description:BackgroundIn health research, several chronic diseases are susceptible to competing risks (CRs). Initially, statistical models (SM) were developed to estimate the cumulative incidence of an event in the presence of CRs. As recently there is a growing interest in applying machine learning (ML) for clinical prediction, these techniques have also been extended to model CRs but literature is limited. Here, our aim is to investigate the potential role of ML versus SM for CRs within non-complex data (small/medium sample size, low dimensional setting).MethodsA dataset with 3826 retrospectively collected patients with extremity soft-tissue sarcoma (eSTS) and nine predictors is used to evaluate model-predictive performance in terms of discrimination and calibration. Two SM (cause-specific Cox, Fine-Gray) and three ML techniques are compared for CRs in a simple clinical setting. ML models include an original partial logistic artificial neural network for CRs (PLANNCR original), a PLANNCR with novel specifications in terms of architecture (PLANNCR extended), and a random survival forest for CRs (RSFCR). The clinical endpoint is the time in years between surgery and disease progression (event of interest) or death (competing event). Time points of interest are 2, 5, and 10 years.ResultsBased on the original eSTS data, 100 bootstrapped training datasets are drawn. Performance of the final models is assessed on validation data (left out samples) by employing as measures the Brier score and the Area Under the Curve (AUC) with CRs. Miscalibration (absolute accuracy error) is also estimated. Results show that the ML models are able to reach a comparable performance versus the SM at 2, 5, and 10 years regarding both Brier score and AUC (95% confidence intervals overlapped). However, the SM are frequently better calibrated.ConclusionsOverall, ML techniques are less practical as they require substantial implementation time (data preprocessing, hyperparameter tuning, computational intensity), whereas regression methods can perform well without the additional workload of model training. As such, for non-complex real life survival data, these techniques should only be applied complementary to SM as exploratory tools of model's performance. More attention to model calibration is urgently needed.
| S-EPMC9951458 | biostudies-literature
Project description:BackgroundEchocardiography (echo) based machine learning (ML) models may be useful in identifying patients at high-risk of all-cause mortality.MethodsWe developed ML models (ResNet deep learning using echo videos and CatBoost gradient boosting using echo measurements) to predict 1-year, 3-year, and 5-year mortality. Models were trained on the Mackay dataset, Taiwan (6083 echos, 3626 patients) and validated in the Alberta HEART dataset, Canada (997 echos, 595 patients). We examined the performance of the models overall, and in subgroups (healthy controls, at risk of heart failure (HF), HF with reduced ejection fraction (HFrEF) and HF with preserved ejection fraction (HFpEF)). We compared the models' performance to the MAGGIC risk score, and examined the correlation between the models' predicted probability of death and baseline quality of life as measured by the Kansas City Cardiomyopathy Questionnaire (KCCQ).FindingsMortality rates at 1-, 3- and 5-years were 14.9%, 28.6%, and 42.5% in the Mackay cohort, and 3.0%, 10.3%, and 18.7%, in the Alberta HEART cohort. The ResNet and CatBoost models achieved area under the receiver-operating curve (AUROC) between 85% and 92% in internal validation. In external validation, the AUROCs for the ResNet (82%, 82%, and 78%) were significantly better than CatBoost (78%, 73%, and 75%), for 1-, 3- and 5-year mortality prediction respectively, with better or comparable performance to the MAGGIC score. ResNet models predicted higher probability of death in the HFpEF and HFrEF (30%-50%) subgroups than in controls and at risk patients (5%-20%). The predicted probabilities of death correlated with KCCQ scores (all p < 0.05).InterpretationEcho-based ML models to predict mortality had good internal and external validity, were generalizable, correlated with patients' quality of life, and are comparable to an established HF risk score. These models can be leveraged for automated risk stratification at point-of-care.FundingFunding for Alberta HEART was provided by an Alberta Innovates - Health Solutions Interdisciplinary Team Grant no. AHFMRITG 200801018. P.K. holds a Canadian Institutes of Health Research (CIHR) Sex and Gender Science Chair and a Heart & Stroke Foundation Chair in Cardiovascular Research. A.V. and V.S. received funding from the Mitacs Globalink Research Internship.
| S-EPMC10006431 | biostudies-literature
Project description:BackgroundBesides the classic logistic regression analysis, non-parametric methods based on machine learning techniques such as random forest are presently used to generate predictive models. The aim of this study was to evaluate random forest mortality prediction models in haemodialysis patients.MethodsData were acquired from incident haemodialysis patients between 1995 and 2015. Prediction of mortality at 6 months, 1 year and 2 years of haemodialysis was calculated using random forest and the accuracy was compared with logistic regression. Baseline data were constructed with the information obtained during the initial period of regular haemodialysis. Aiming to increase accuracy concerning baseline information of each patient, the period of time used to collect data was set at 30, 60 and 90 days after the first haemodialysis session.ResultsThere were 1571 incident haemodialysis patients included. The mean age was 62.3 years and the average Charlson comorbidity index was 5.99. The mortality prediction models obtained by random forest appear to be adequate in terms of accuracy [area under the curve (AUC) 0.68-0.73] and superior to logistic regression models (ΔAUC 0.007-0.046). Results indicate that both random forest and logistic regression develop mortality prediction models using different variables.ConclusionsRandom forest is an adequate method, and superior to logistic regression, to generate mortality prediction models in haemodialysis patients.
| S-EPMC8247746 | biostudies-literature
Project description:ImportanceThere is an absence of mortality risk assessment tools in first-episode psychosis (FEP) that could enable personalized interventions.ObjectiveTo examine the feasibility of machine learning (ML) in discerning mortality risk in FEP and to assess whether such risk predictions can inform pharmacotherapy choices.Design, setting, and participantsIn this prognostic study, Swedish nationwide cohort data (from July 1, 2006, to December 31, 2021) were harnessed for model development and validation. Finnish cohort data (from January 1, 1998, to December 31, 2017) were used for external validation. Data analyses were completed between December 2022 and December 2023.Main outcomes and measuresFifty-one nationwide register variables, encompassing demographics and clinical and work-related histories, were subjected to ML to predict future mortality risk. The ML model's performance was evaluated by calculating the area under the receiver operating characteristic curve (AUROC). The comparative effectiveness of pharmacotherapies in patients was assessed and was stratified by the ML model to those with predicted high mortality risk (vs low risk), using the between-individual hazard ratio (HR). The 5 most important variables were then identified and a model was retrained using these variables in the discovery sample.ResultsThis study included 24 052 Swedish participants (20 000 in the discovery sample and 4052 in the validation sample) and 1490 Finnish participants (in the validation sample). Swedish participants had a mean (SD) age of 29.1 (8.1) years, 62.1% were men, and 418 died with 2 years. Finnish participants had a mean (SD) age of 29.7 (8.0) years, 61.7% were men, and 31 died within 2 years. The discovery sample achieved an AUROC of 0.71 (95% CI, 0.68-0.74) for 2-year mortality prediction. Using the 5 most important variables (ie, the top 10% [substance use comorbidities, first hospitalization duration due to FEP, male sex, prior somatic hospitalizations, and age]), the final model resulted in an AUROC of 0.70 (95% CI, 0.63-0.76) in the Swedish sample and 0.67 (95% CI, 0.56-0.78) in the Finnish sample. Individuals with predicted high mortality risk had an elevated 15-year risk in the Swedish sample (HR, 3.77 [95% CI, 2.92-4.88]) and an elevated 20-year risk in the Finnish sample (HR, 3.72 [95% CI, 2.67-5.18]). For those with predicted high mortality risk, long-acting injectable antipsychotics (HR, 0.45 [95% CI, 0.23-0.88]) and mood stabilizers (HR, 0.64 [95% CI, 0.46-0.90]) were associated with decreased mortality risk. Conversely, for those predicted to survive, only oral aripiprazole (HR, 0.38 [95% CI, 0.20-0.69]) and risperidone (HR, 0.38 [95% CI, 0.18-0.82]) were associated with decreased mortality risk.Conclusions and relevanceIn this prognostic study, an ML-based model was developed and validated to predict mortality risk in FEP. These findings may help to develop personalized interventions to mitigate mortality risk in FEP.
| S-EPMC10949098 | biostudies-literature
Project description:Since January 2002, Medicare has provided payment for medical nutrition therapy for patients with chronic kidney disease. Few patients receive dietary counseling before end-stage renal disease onset; whether such counseling is associated with improved outcomes is unknown.Retrospective cohort analysis.Patients who initiated hemodialysis therapy on June 1, 2005, to May 31, 2007, in the United States for whom predialysis dietitian care was reported on the Centers for Medicare & Medicaid Services Medical Evidence Report.Dietitian care before end-stage renal disease onset.Time to death.Propensity score for dietitian care calculated using logistic regression; Cox regression analysis used to compare time to death by predialysis dietitian care overall and stratified by tertiles of propensity score, adjusting for baseline characteristics.Most patients (88%) received no dietitian care; 9% received dietitian care for 12 months or less, and 3% received dietitian care for more than 12 months before dialysis therapy initiation (total N = 156,440). Predialysis dietitian care was associated independently with higher albumin and lower total cholesterol levels at dialysis therapy initiation. There was evidence of an independent association between predialysis dietitian care for longer than 12 months and decreased mortality during the first year on dialysis therapy for the second tertile of propensity score. Adjusted mortality HRs were 1.16 (95% CI, 0.44-3.09; P = 0.8), 0.81 (95% CI, 0.71-0.93; P = 0.002), and 0.93 (95% CI, 0.86-1.01; P = 0.1) in the first, second, and third tertiles of propensity score, respectively.Information for dietitian care was missing for 18.6% of Medical Evidence Reports and has low sensitivity; including only incident dialysis patients precluded evaluation of an association between dietitian care and chronic kidney disease progression; the observational design allowed the possibility of residual confounding.Our study suggests an independent association between predialysis dietitian care for more than 12 months and lower mortality during the first year on dialysis therapy.
| S-EPMC4882105 | biostudies-literature
Project description:The single leading cause of mortality on hemodialysis is sudden cardiac death. Whether measures of electrophysiologic substrate independently associate with mortality is unknown. We examined measures of electrophysiologic substrate in a prospective cohort of 571 patients on incident hemodialysis enrolled in the Predictors of Arrhythmic and Cardiovascular Risk in End Stage Renal Disease Study. A total of 358 participants completed both baseline 5-minute and 12-lead electrocardiogram recordings on a nondialysis day. Measures of electrophysiologic substrate included ventricular late potentials by the signal-averaged electrocardiogram and spatial mean QRS-T angle measured on the averaged beat recorded within a median of 106 days (interquartile range, 78-151 days) from dialysis initiation. The cohort was 59% men, and 73% were black, with a mean±SD age of 55±13 years. Transthoracic echocardiography revealed a mean±SD ejection fraction of 65.5%±12.0% and a mean±SD left ventricular mass index of 66.6±22.3 g/m2.7 During 864.6 person-years of follow-up, 77 patients died; 35 died from cardiovascular causes, of which 15 were sudden cardiac deaths. By Cox regression analysis, QRS-T angle ?75° significantly associated with increased risk of cardiovascular mortality (hazard ratio, 2.99; 95% confidence interval, 1.31 to 6.82) and sudden cardiac death (hazard ratio, 4.52; 95% confidence interval, 1.17 to 17.40) after multivariable adjustment for demographic, cardiovascular, and dialysis factors. Abnormal signal-averaged electrocardiogram measures did not associate with mortality. In conclusion, spatial QRS-T angle but not abnormal signal-averaged electrocardiogram significantly associates with cardiovascular mortality and sudden cardiac death independent of traditional risk factors in patients starting hemodialysis.
| S-EPMC5084888 | biostudies-other
Project description:ObjectiveEarly prediction of long-term outcomes in patients with sepsis-induced cardiorenal syndrome (CRS) remains a great challenge in clinical practice. Herein, we aimed to construct a nomogram and machine learning model for predicting the 1-year mortality risk in patients with sepsis-induced CRS.MethodsThis retrospective study enrolled 340 patients diagnosed with sepsis-induced CRS in Shanghai Tongji Hospital between January 2015 and May 2019, as a discovery cohort. Two predictive models, the nomogram and machine learning model, were used to predict 1-year mortality. The prognostic variables used to develop the nomogram were identified based on a forward stepwise binary logistic regression, and the predictive ability of the nomogram was evaluated by the areas under the receiver operating characteristic curve (AUC) and the calibration curve. Meanwhile, machine learning (ML) techniques, such as support vector machine, random forest (RF), and gradient boosted decision tree, were assessed mainly by accuracy and AUC. Feature ranking analysis was performed using the ML algorithm. Both nomogram and ML models were externally validated by an independent cohort of 103 patients diagnosed with sepsis-induced CRS between June 2019 and December 2020.ResultsAge, sequential sepsis-related organ failure score (SOFA), serum myoglobin (MYO), vasopressor use, and mechanical ventilation were identified as independent risk factors for 1-year mortality in the nomogram predictive model. In the discovery cohort, the nomogram yielded higher AUC for predicting mortality than did the SOFA score (0.855 [95% CI: 0.815-0.895] vs. 0.756 [95% CI: 0.705-0.808]). For ML, the model developed by RF showed the highest accuracy (0.765) and AUC (0.854). In feature ranking analysis, factors such as age, MYO, SOFA score, vasopressor use, and baseline serum creatinine were identified as important features affecting 1-year prognosis. Moreover, the nomogram and RF model both performed well in external validation, with an AUC of 0.877 and 0.863, respectively.ConclusionOur nomogram and ML models showed that age, SOFA score, serum MYO levels, and the use of vasopressors during hospitalization were the main factors influencing the risk of long-term mortality. Our models may serve as useful tools for assessing long-term prognosis in patients with sepsis-induced CRS.
| S-EPMC9099150 | biostudies-literature
Project description:In the general population, low serum magnesium levels are associated with poor outcomes and death. While limited data suggest that low baseline magnesium levels may be associated with higher mortality in hemodialysis (HD) patients, the impact of changes in magnesium levels over time is unknown.We examined the association of time-varying serum magnesium levels with all-cause mortality using multivariable time-varying survival models adjusted for clinical characteristics and other time-varying laboratory measures.9,359 maintenance HD patients treated in a large dialysis organization between 2007 and 2011.Time-varying serum magnesium levels across 5 magnesium increments (<1.8, 1.8-<2.0, 2.0-<2.2, 2.2-<2.4, and ?2.4mg/dL).All-cause mortality.2,636 individuals died over 5 years. Time-varying serum magnesium levels < 2.0mg/dL were associated with higher mortality after adjustment for demographics and comorbid conditions, including hypertension, diabetes, and malignancies (reference: magnesium, 2.2-<2.4mg/dL): adjusted HRs for serum magnesium level < 1.8 and 1.8 to <2.0mg/dL were 1.39 (95% CI, 1.23-1.58; P<0.001) and 1.20 (95% CI, 1.06-1.36; P=0.004), respectively. Some associations were attenuated to the null after incremental adjustment for laboratory test results, particularly serum albumin. However, among patients with serum albumin measurements, low albumin level (<3.5g/dL) and magnesium level < 2.0mg/dL were associated with an additional death risk (adjusted HR, 1.17; 95% CI, 1.05-1.31; P=0.004), whereas patients with high serum albumin levels (?3.5g/dL) exhibited low death risk (adjusted HRs of 0.53 and 0.53 [P?0.001] for magnesium < 2.0 and ?2.0mg/dL, respectively; reference: albumin < 3.5g/dL and magnesium ? 2.0mg/dL).Causality cannot be determined, and residual confounding cannot be excluded given the observational study design.Lower serum magnesium levels are associated with higher mortality in HD patients, including those with hypoalbuminemia. Interventional studies are warranted to examine whether correction of hypomagnesemia ameliorates adverse outcomes in this population.
| S-EPMC4658300 | biostudies-literature
Project description:Patients with mantle cell lymphoma (MCL), an incurable B-cell malignancy, benefit from accurate pretreatment disease stratification. We curated an extensive database of 862 patients diagnosed between 2014 and 2022. A machine learning (ML) gradient-boosted model incorporated baseline features from clinicopathologic, cytogenetic, and genomic data with high predictive power discriminating between patients with indolent or responsive MCL and those with aggressive disease (AUC ROC = 0.83). In addition, we utilized the gradient-boosted framework as a robust feature selection method for multivariate logistic and survival modeling. The best ML models incorporated features from clinical and genomic data types highlighting the need for correlative molecular studies in precision oncology. As proof of concept, we launched our most accurate and practical models using an application interface, which has potential for clinical implementation. We designated the 20-feature ML model-based index the "integrative MIPI" or iMIPI and a similar 10-feature ML index the "integrative simplified MIPI" or iMIPI-s. The top 10 baseline prognostic features represented in the iMIPI-s are: lactase dehydrogenase (LDH), Ki-67%, platelet count, bone marrow involvement percentage, hemoglobin levels, the total number of observed somatic mutations, TP53 mutational status, Eastern Cooperative Oncology Group performance level, beta-2 microglobulin, and morphology. Our findings emphasize that prognostic applications and indices should include molecular features, especially TP53 mutational status. This work demonstrates the clinical utility of complex ML models and provides further evidence for existing prognostic markers in MCL.SignificanceOur model is the first to integrate a dynamic algorithm with multiple clinical and molecular features, allowing for accurate predictions of MCL disease outcomes in a large patient cohort.
| S-EPMC10395375 | biostudies-literature
Project description:BackgroundControversy surrounds which factors are important for predicting early mortality after dialysis initiation (DI). We investigated associations of predialysis course and circumstances affecting planning and execution of DI with mortality following DI.MethodsAmong 1580 patients participating in the Peridialysis study, a study of causes and timing of DI, we registered features of predialysis course, clinical and biochemical data at DI, incidence of unplanned suboptimal DI, contraindications to peritoneal dialysis (PD) or hemodialysis (HD), and modality preference, actual choice, and cause of modality choice. Patients were followed for 12 months or until transplantation. A flexible parametric model was used to identify independent factors associated with all-cause mortality.ResultsFirst-year mortality was 19.33%. Independent factors predicting death were high age, comorbidity, clinical contraindications to PD or HD, suboptimal DI, high eGFR, low serum albumin, hyperphosphatemia, high C-reactive protein, signs of overhydration and cerebral symptoms at DI. Among 1061 (67.2%) patients who could select dialysis modality based on personal choice, 654 (61.6%) chose PD, 368 (34.7%) center HD and 39 (3.7%) home HD. The 12-months survival did not differ significantly between patients receiving PD and in-center HD.ConclusionsFirst-year mortality in incident dialysis patients was in addition to high age and comorbidity, associated with clinical contraindications to PD or HD, clinical symptoms, hyperphosphatemia, inflammation, and suboptimal DI. In patients with a "free" choice of dialysis modality based on their personal preferences, PD and in-center HD led to broadly similar short-term outcomes.
| S-EPMC9235232 | biostudies-literature