Predictive Scoring Model Research Articles

A Novel Prediction Model for Risk of Acute-on-Chronic Liver Failure in Chronic Hepatitis B Virus-Infected Patients With Flare of Hepatitis

Background: Chronic patients with flare of hepatitis B virus (HBV) may progress to acute-on-chronic liver failure (ACLF), which is characterized by high mortality. Uniform models to predict risk of ACLF are lacking. We aimed to present an accurate risk prediction model that incorporates clinical manifestations and laboratory results in chronic HBV-infected patients with flare of hepatitis. Methods: We selected the best variable combination using recursive feature elimination (RFE) from a perspective cohort including 360 patients with hepatitis flare at the Third Affiliated Hospital of Sun Yat-sen University to develop a risk prediction model and nomogram scoring system, then validated the model using two external independent sets consisting of 96 and 65 individuals, and compared its performance with End-Stage Liver Disease (MELD) score model. Findings: An excellent prediction model called the PAT model using the five final identified predictors including prothrombin time (PT), total bilirubin (TBil), ascites, pneumonia and hemoglobin (HGB) was constructed. The model could achieve an Area under the Receiver Operating Characteristic curve (AUROC) of 0.964 (95% confidence interval [CI]: 0.949-0.980) in the development set and performed well in external validation cohorts (validation cohort 1, AUROC = 0.909, 95% CI 0.840-0.977; validation cohort 2, AUROC = 0.898, 95% CI 0.813-0.983, both of which were significantly higher than the MELD score model (P <0.001). Interpretation: The easy-to-use PAT model can accurately predict the risk of ACLF risk in chronic hepatitis B patients with flare of hepatitis, which is expected to become a widely accepted prediction model and guide therapeutic options. Funding Statement: This work was supported by a grant for National Key RD National Natural Science Foundation of China (81773176); Science and Technology Program of Guangzhou (201804010474); the 5010 Project of Clinical Research in Sun Yat-sen University (2016009); Medical Scientific Research Foundation of Guangdong Province (A2017118); Major Science and Technology Projects in 13th Five-Year (2018ZX10715004-001-09); Natural Science Foundation of Guangdong (2018A030310272); Guangdong basic and Applied Basic Research Foundation (2019A1515110166) ; Guangdong Medical Science and Technology Research Fund (A2018366); 3rd Affiliated Hospital of Sun Yat-Sen University, Clinical Research Program (P000-277) and Meizhou Science and Technology Project (2019B0203003). Declaration of Interests: The authors declare that they have no competing interests. Ethics Approval Statement: The study was approved by the institutional ethics committee of the Third Affiliated Hospital of Sun Yat-sen University.

Predicting 30-days mortality for MIMIC-III patients with sepsis-3: a machine learning approach using XGboost

BackgroundSepsis is a significant cause of mortality in-hospital, especially in ICU patients. Early prediction of sepsis is essential, as prompt and appropriate treatment can improve survival outcomes. Machine learning methods are flexible prediction algorithms with potential advantages over conventional regression and scoring system. The aims of this study were to develop a machine learning approach using XGboost to predict the 30-days mortality for MIMIC-III Patients with sepsis-3 and to determine whether such model performs better than traditional prediction models.MethodsUsing the MIMIC-III v1.4, we identified patients with sepsis-3. The data was split into two groups based on death or survival within 30 days and variables, selected based on clinical significance and availability by stepwise analysis, were displayed and compared between groups. Three predictive models including conventional logistic regression model, SAPS-II score prediction model and XGBoost algorithm model were constructed by R software. Then, the performances of the three models were tested and compared by AUCs of the receiver operating characteristic curves and decision curve analysis. At last, nomogram and clinical impact curve were used to validate the model.ResultsA total of 4559 sepsis-3 patients are included in the study, in which, 889 patients were death and 3670 survival within 30 days, respectively. According to the results of AUCs (0.819 [95% CI 0.800–0.838], 0.797 [95% CI 0.781–0.813] and 0.857 [95% CI 0.839–0.876]) and decision curve analysis for the three models, the XGboost model performs best. The risk nomogram and clinical impact curve verify that the XGboost model possesses significant predictive value.ConclusionsUsing machine learning technique by XGboost, more significant prediction model can be built. This XGboost model may prove clinically useful and assist clinicians in tailoring precise management and therapy for the patients with sepsis-3.

Estimation of Major Adverse Cardiovascular Events in Patients With Myocardial Infarction Undergoing Primary Percutaneous Coronary Intervention: A Risk Prediction Score Model From a Derivation and Validation Study.

Background: The population with myocardial infarction (MI) undergoing primary percutaneous coronary intervention (PPCI) is growing, but validated models to guide their clinical management are lacking. This study aimed to develop and validate prognostic models to predict major adverse cardiovascular events (MACEs) in patients with MI undergoing PPCI.Methods and Results: Models were developed in 4,151 patients with MI who underwent PPCI in Fuwai Hospital between January 2010 and June 2017, with a median follow-up of 698 days during which 544 MACEs occurred. The predictors included in the models were age, a history of diabetes mellitus, atrial fibrillation, chronic kidney disease, coronary artery bypass grafting, the Killip classification, ejection fraction at admission, the high-sensitivity C-reactive protein (hs-CRP) level, the estimated glomerular filtration rate, the d-dimer level, multivessel lesions, and the culprit vessel. The models had good calibration and discrimination in the derivation and internal validation with C-indexes of 0.74 and 0.60, respectively, for predicting MACEs. The new prediction model and Thrombolysis in Myocardial Infarction (TIMI) risk score model were compared using the receiver operating characteristic curve. The areas under the curve of the new prediction model and TIMI risk score model were 0.806 and 0.782, respectively (difference between areas = 0.024 < 0.05; z statistic, 1.718).Conclusion: The new prediction model could be used in clinical practice to support risk stratification as recommended in clinical guidelines.

Development and Validation of a Risk Scoring Model for Early Prediction of Severe Colon Ischemia.

Colon ischemia (CI) is injury to the intestines secondary to insufficient blood flow. Its clinical severity can range from mild to life-threatening. To investigate predictive risk factors for CI and propose a scoring model for severe outcomes. We retrospectively analyzed the medical records of patients admitted to Chungnam National University Hospital from January 2010 to December 2018. CI was defined as severe when patients required surgery immediately or after initial conservative management, death occurred after hospitalization, or symptoms persisted after 2weeks. By controlling for possible confounders from the logistic regression analysis, we obtained a new risk scoring model for the early prediction of severe CI. Furthermore, using the area under the receiver operating characteristics curve (AUROC), we assessed the accuracy of the model. A total of 274 patients endoscopically diagnosed with CI were included, of whom 181 had severe CI. In the multivariate analysis, tachycardia, elevated C-reactive protein, Favier endoscopic classification stage ≥ 2, and history of hypertension were independently and significantly associated with severe CI. The AUROC of the model was 0.749. This risk scoring model based on the presence of tachycardia, elevated C-reactive protein level, unfavorable endoscopic findings by Favier's classification, and the history of hypertension could be used to predict severe CI outcomes at an early stage.

Development and validation of a tool for predicting type 2 diabetes in Mexican women of reproductive age

IntroductionDiabetes is a worldwide problem with a greater impact in developing countries, where many people are unaware of their risk. In Mexico, women show the greatest risk for T2D. Current risk scores have been developed and validated in predominantly older European cohorts. They are not the best option in Mexican women. The development of a risk model/score in this population would be useful. ObjectiveTo develop and validate a risk model and score that incorporates the most relevant risk factors for T2D in Mexican women of reproductive age. MethodsThe study was carried out in two phases, with the first phase being the development of the predictive model and the second phase the validation of the model in a separate independent population. A cohort of Mexican patients of reproductive age (“Derivation Cohort”) was used to create the predictive model. It included data on 3161 women. Risk factors for identification were assessed using Cox proportional hazards regression. Finally a score with a range of 0 to 19 points was developed to identify the 2.4 year probability of developing DM2 in Mexican women of reproductive age. Results147 new cases of T2D (4.6%) were identified in the Derivation Cohort model, 97 of 925 participants (10.48%) in the validation cohort. The risk factor predictors of T2D were: history of gestational diabetes (HR 2.69, 95% CI 1.10–6.58), BMI (HR 1.03, 95% CI 1.01–1.06), hypertriglyceridemia (HR 1.54, 95% CI 1.11–2.14) and fasting blood glucose (HR 1.06, 95% CI 1.05–1.08), with an AUC of 0.75. The AUC in the validation cohort was 0.91 (95% CI 0.87–0.94). The score had a sensitivity of 73% and specificity of 67% at a cutoff of ≥15. ConclusionsA predictive model and risk score was developed to detect cases at risk for incident T2D. It was generated using the characteristics of Mexican women of reproductive age. This risk score is a step forward in attempting to address the generational legacy that diabetes in pregnancy could have on women and their children.

Comparison of the Performance of the NCDR CathPCI Score and Mehran Score to Predict Acute Kidney Injury Post-Percutaneous Coronary Intervention.

A comparison of acute kindney injury (AKI) post-percutaneous coronary intervention (PCI) prediction models is lacking. In this study, we aim to compare the National Cardiovascular Data Registry (NCDR) CathPCI score to the Mehran score in acute coronary syndrome (ACS) vs non-ACS patients. We included patients who received PCI at our facility between July 2015 and December 2017. We excluded patients without a pre- and/or post-PCI serum creatinine, patients on dialysis at the time of PCI and patients with missing variables required to calculate the predictive scoring model. The primary outcome of this study was AKI post-PCI. Performance of the NCDR CathPCI score and the Mehran score were evaluated by comparing the area under the receiver-operating characteristic curve (AUROC) for both scores. The analysis included 1,507 patients. In non-ACS patients, the Mehran score performed better than the NCDR CathPCI score with AUROC 0.75 and 0.68 respectively (p=0.014). When categorized into 4 risk groups, a Mehran score ≥ 2 had a sensitivity of 86% and a Mehran score of ≥ 3 had a specificity of 83% in non-ACS patients. In contrast, when the NCDR CathPCI score was categorized into risk groups, it was not able to predict the risk of AKI (p=0.78) with sensitivity of 0% for the intermediate and high risk group. In ACS patients, the NCDR CathPCI score was superior in predicting the risk for AKI with AUROC 0.79 versus 0.74 (p=.019). In predicting AKI post-PCI, the NCDR CathPCI score performed better in ACS populations, and the Mehran score performed better in the non-ACS population.

Preeclampsia Risk Prediction Model for Chinese Pregnant Patients With Systemic Lupus Erythematosus.

To screen for a high risk of preeclampsia in women with systemic lupus erythematosus (SLE). A total of 513 antenatal care records of pregnant patients with SLE were obtained, and the data were randomly assigned to either a development set (n = 342) or a validation set (n = 171). Preeclampsia predictors were identified with stepwise regression, and a coefficient B of each variable was used to establish a prediction model and risk scoring system. Goodness-of-fit was assessed by the Hosmer-Lemeshow and Omnibus tests, and the area under the receiver operating characteristic curve (area under the curve) was used to assess discrimination. Validation was performed using the validation set. The preeclampsia incidence was 14.4% in the pregnant patients with SLE. A mean arterial pressure (MAP) ≥96.5 mm Hg (odds ratio [OR] 213.15 [95% confidence interval (95% CI) 24.39-999.99]), prepregnancy hypertension (OR 18.19 [95% CI 2.67-125.01]), a hematologic disorder (OR 4.13 [95% CI 1.03-16.67]), positive IgM anticardiolipin antibodies (aCLs) (OR 19.85 [95% CI 1.11-333.33]), serum albumin <31.5 grams/liter (OR 9.88 [95% CI 2.07-47.62]), serum uric acid ≥303 μmoles/liter (OR 5.58 [95% CI 1.40-22.22]), and 24-hour urinary protein ≥0.286 grams (OR 14.39 [95% CI 2.43-83.33]) were selected for the preeclampsia prediction model. The area under the curve was 0.975. Preeclampsia prediction model scores >4 indicated a high risk of preeclampsia. For the validation set, the preeclampsia prediction accuracy was 93.6% (sensitivity 88.5%, specificity 94.5%). A model for predicting the risk of preeclampsia in pregnant patients with SLE was established on the basis of MAP, prepregnancy hypertension, hematologic disorders, IgM aCLs, albumin, uric acid, and 24-hour urinary protein. The model had good predictive efficiency and can help clinicians improve pregnancy outcomes in high-risk women with early interventions.

GES: A validated simple score to predict the risk of HCC in patients with HCV-GT4-associated advanced liver fibrosis after oral antivirals.

Hepatocellular carcinoma (HCC) risk persists after hepatitis C virus (HCV) eradication with direct-acting antivirals (DAAs), particularly in patients with cirrhosis. Identifying those who are likely to develop HCC is a critical unmet medical need. Our aim is to develop a score that offers individualized patient HCC risk prediction. This two-centre prospective study included 4400 patients, with cirrhosis and advanced fibrosis who achieved a sustained virologic response (SVR), including 2372 patients (derivation cohort). HCC-associated factors were identified by multivariable Cox regression analysis to develop a scoring model for prediction of HCC risk; and subsequently internally and externally validated in two independent cohorts of 687 and 1341 patients. In the derivation cohort, the median follow-up was 23.51±8.21months, during which 109 patients (4.7%) developed HCC. Age, sex, serum albumin, α fetoprotein and pretreatment fibrosis stage were identified as risk factors for HCC. A simple predictive model (GES) score was constructed. The 2-year cumulative HCC incidence using Kaplan-Meier method was 1.2%, 3.3% and 7.1% in the low-risk, medium-risk and high-risk groups respectively. Internal and external validation showed highly significant difference among the three risk groups (P<.001) with regard to cumulative HCC risk. GES score has high predictive ability value (Harrell's C statistic 0.801), that remained robustly consistent across two independent validation cohorts (Harrell's C statistic 0.812 and 0.816). GES score is simple with validated good predictive ability for the development of HCC after eradication of HCV and may be useful for HCC risk stratification in those patients.

AutoScore: A Machine Learning-Based Automatic Clinical Score Generator and Its Application to Mortality Prediction Using Electronic Health Records.

BackgroundRisk scores can be useful in clinical risk stratification and accurate allocations of medical resources, helping health providers improve patient care. Point-based scores are more understandable and explainable than other complex models and are now widely used in clinical decision making. However, the development of the risk scoring model is nontrivial and has not yet been systematically presented, with few studies investigating methods of clinical score generation using electronic health records.ObjectiveThis study aims to propose AutoScore, a machine learning–based automatic clinical score generator consisting of 6 modules for developing interpretable point-based scores. Future users can employ the AutoScore framework to create clinical scores effortlessly in various clinical applications.MethodsWe proposed the AutoScore framework comprising 6 modules that included variable ranking, variable transformation, score derivation, model selection, score fine-tuning, and model evaluation. To demonstrate the performance of AutoScore, we used data from the Beth Israel Deaconess Medical Center to build a scoring model for mortality prediction and then compared the data with other baseline models using the receiver operating characteristic analysis. A software package in R 3.5.3 (R Foundation) was also developed to demonstrate the implementation of AutoScore.ResultsImplemented on the data set with 44,918 individual admission episodes of intensive care, the AutoScore-created scoring models performed comparably well as other standard methods (ie, logistic regression, stepwise regression, least absolute shrinkage and selection operator, and random forest) in terms of predictive accuracy and model calibration but required fewer predictors and presented high interpretability and accessibility. The nine-variable, AutoScore-created, point-based scoring model achieved an area under the curve (AUC) of 0.780 (95% CI 0.764-0.798), whereas the model of logistic regression with 24 variables had an AUC of 0.778 (95% CI 0.760-0.795). Moreover, the AutoScore framework also drives the clinical research continuum and automation with its integration of all necessary modules.ConclusionsWe developed an easy-to-use, machine learning–based automatic clinical score generator, AutoScore; systematically presented its structure; and demonstrated its superiority (predictive performance and interpretability) over other conventional methods using a benchmark database. AutoScore will emerge as a potential scoring tool in various medical applications.

Establishment of an antepartum predictive scoring model to identify candidates for vaginal birth after cesarean

BackgroundEvidence-based medicine has shown that successful vaginal birth after cesarean (VBAC) is associated with fewer complications than an elective repeat cesarean. Although spontaneous vaginal births and reductions in cesarean delivery (CD) rates have been advocated, the risk factors for VBAC complications remain unclear and failed trials of labor (TOL) can lead to adverse pregnancy outcomes.MethodsTo construct an antepartum predictive scoring model for VBAC. Retrospective analysis of charts from 1062 women who underwent TOL at no less than 28 gestational weeks with vertex singletons and no more than one prior CD.ResultsWe constructed our scoring model based on the following variables: maternal age, previous vaginal delivery, interdelivery interval (time between prior cesarean and the following delivery), presence of prior cesarean TOL, dystocia as prior CD indication, intertuberous diameter, maternal predelivery body mass index, gestational age at delivery, estimated fetal weight, and hypertensive disorders. Previous vaginal delivery was the most influential variable. The nomogram showed an area under the curve of 77.7% (95% confidence interval, 73.8–81.5%; sensitivity, 78%; specificity, 70%; cut-off, 13 points). The Kappa value to judge the consistency of the results between the predictive model and the actual results was 0.71(95% confidence interval 0.65–0.77) indicating strong consistency. We used the cut-off to divide the VBAC women into two groups according to the success of the TOL. The maternal and neonatal outcomes such as labor time, number of deliveries by midwives, postpartum hemorrhage, uterine rupture, neonatal asphyxia, puerperal infection were significantly different between the two groups.ConclusionsOur predictive scoring model incorporates easily ascertainable variables and can be used to personalize antepartum counselling for successful TOLs after cesareans.

Early detection of non-invasive ventilation failure among acute respiratory failure patients in the emergency department

BackgroundNon-invasive mechanical ventilation (NIV) has become an alternative to an invasive artificial airway for the management of acute respiratory failure (ARF). NIV failure causes delayed intubation, which eventually has been associated with increased morbidity and mortality. This study aimed to develop the clinical scoring system of NIV failure in ARF patients.MethodsThis study was a diagnostic, retrospectively cross-sectional, and exploratory model at the Emergency Medicine Department in Ramathibodi Hospital between February 2017 and December 2017. We included all of the acute respiratory failure patients aged > 18 years and received non-invasive ventilation (NIV). Clinical factors associated with NIV failure were recorded. The predictive model and prediction score for NIV failure were developed by multivariable logistic regression analysis.ResultA total of 329 acute respiratory failure patients have received NIV success (N = 237) and failure (N = 92). This study showed that NIV failure was associated with heart rate > 110 bpm, systolic BP < 110 mmHg, SpO2 < 90%, arterial pH < 7.30 and serum lactate. The clinical scores were classified into three groups: low, moderate, and high.ConclusionWe suggested that the novel clinical scoring of the NIV failure in this study may use as a good predictor for NIV failure in the emergency room.

Adopting the Triple Aim Framework in the Saudi Healthcare System: A Delphi Study.

PurposeHelping healthcare systems to optimize performance using particular metrics through objective monitoring will positively impact an organization’s progress toward strategic goals and objectives. The Institute of Healthcare Improvement introduced the Triple Aim framework for guiding new or transforming health systems, concurrently improving population health and patients’ care experience, and reducing per capita cost. Consensus to determine applicable and appropriate measures to monitor this transformation within the scope of the three objectives is required. Thus, the study aimed to adopt the Triple Aim measures and reach a consensus among experts in healthcare systems on the applicability of the measures in the health system of Saudi Arabia.MethodsA two-round Delphi study was conducted with 17 invited experts who were knowledgeable and experienced in healthcare systems and administration, quality improvement, and strategic planning. These rounds were based on the measures of the three objectives, where they were introduced as statements and grouped under each respective aim. The experts were instructed to score each measure using a 5-point Likert scale and were invited to formulate new measures related to the same aim. Statements reaching a consensus level of 80% were considered applicable measures for the Saudi health system.ResultsA total of 17 measures were circulated among experts; out of which, 16 measures reached a consensus. The 16 measures represent the three main domains of the Triple Aim model, ie, population health, experience of care, and cost per capita. The measure that failed to reach a consensus was the predictive model scores because it requires medical knowledge, where the majority of the experts were non-physicians.ConclusionA Delphi study was used to reach consensus among experts on the Triple Aim measures as a first step to building a solid foundation for the population management required to implement these initiatives in the future.

EARLY PALLIATIVE CARE INTERVENTION REDUCES ICU READMISSIONS IN HIGH-RISK PATIENTS

SESSION TITLE: Palliative Care and End-of-Life Issues Posters SESSION TYPE: Original Investigation Posters PRESENTED ON: October 18-21, 2020 PURPOSE: To determine whether a process of early palliative care intervention prompted by a patient-specific estimate of poor prognosis leads to improved patient outcomes in the medical intensive care unit (ICU). METHODS: We conducted a prospective two-phase study including usual care (UC) (N=53) and usual care plus targeted pro-active palliative care intervention (PC) (N=51) (either by consulting supportive care services or shared decision making discussion by the intensivists), each lasting four months on consecutive patients admitted to the ICU who survived ≥48 hours. For both phases, patients predicted to be at high risk (≥40% 6-month mortality) were identified using our validated integrated prognostic model1 [Surprise question “No” response-Would you be surprised if this patient died in the next six months?-, APACHE IV score, and Charlson Comorbidity Index (CCI) score]. RESULTS: Compared to UC, the pro-active PC intervention phase showed significant reduction in ICU readmissions (35.8% vs 16%, p=0.022) and increased identification of the patient’s legal decision maker (98% vs 83%, p=0.009). There were no differences in means between groups (UC vs PC) for age (67.9 vs 68.1 years, p=.938), CCI score (6.6 vs 6.1, p=.489), and mortality prediction model score (51.6 vs 54.3, p=.141). Intervention-arm patients had higher BMI (34.0 vs 28.3 kg/m2, p=0.005) and APACHE IV scores (65.3 vs 55.8, p=0.019), suggestive of a sicker critically ill patient population. There was no significant difference between groups (UC vs PC) in mean length of stay in ICU (5.5 vs 6.2 days, p=.454) or in hospital (13.0 vs 14.7 days, p=.466). There was no difference between groups (UC vs PC) in death at hospital discharge (60% vs 48%, p=.208). CONCLUSIONS: This pilot study demonstrates feasibility and improved patient-specific outcomes of end-of-life care. CLINICAL IMPLICATIONS: Systematic identification of high-risk ICU patients using the integrated mortality prediction model and pro-active primary palliative care intervention improves shared-decision making with the appropriate legal decision-maker and ICU readmission rates. This study promotes the role of family engagement and empowerment which is in compliance with Society of Critical Care Medicine (SCCM) ICU liberation bundle (A-F). These findings provide support for a large, multisite trial to see if they can be confirmed and if this approach represents a way to improve ICU patient care. DISCLOSURES: No relevant relationships by Lori Constantine, source=Web Response No relevant relationships by Stacey Culp, source=Web Response No relevant relationships by Sarah Hadique, source=Web Response No relevant relationships by Erica Mokaya, source=Web Response No relevant relationships by Alvin Moss, source=Web Response No relevant relationships by Hassan Mujahid, source=Web Response No relevant relationships by Rahul Sangani, source=Web Response

A study on credit scoring modeling with different feature selection and machine learning approaches

A bit hurdle for financial institutions is to decide potential candidates to give a line of credit identifying the right people without any credit risk. For such a crucial decision, past demographic and financial data of debtors is important to build an automated artificial intelligence credit score prediction model based on machine learning classifier. In addition, for building robust and accurate machine learning models, important input predictors (debtor's information) must be selected. The present computational work focuses on building a credit scoring prediction model. A publicly available German credit data is incorporated in this study. An improvement in the credit scoring prediction has been shown with the use of different feature selection techniques (such as Information-gain, Gain-Ratio and Chi-Square) and machine learning classifiers (Bayesian, Naïve Bayes, Random Forest, Decision Tree (C5.0) and SVM (support Vector Machine)). Further, a comparative analysis is performed between different machine learning classifiers and between different feature selection techniques. Different evaluation metrics are considered for analyzing performance of the models (such as accuracy, F-measure, false positive rate, false negative rate and training time). After analysis, a best combination of machine learning classifier and feature selection technique are identified. In this study, a combination of random forest (RF) and Chi-Square (CS) is found good, among other combinations, with respect to good performance accuracy, F-measure and low false positive and false negative rates. However, training time for this particular combination was found to be slightly higher. Result of C5.0 with chi-square was comparable with the best one. This study provides an opportunity to financial institutions to build an automated model for better credit scoring.

Predictive scoring models for persistent gram-negative bacteremia that reduce the need for follow-up blood cultures: a retrospective observational cohort study

BackgroundAlthough the risk factors for positive follow-up blood cultures (FUBCs) in gram-negative bacteremia (GNB) have not been investigated extensively, FUBC has been routinely carried out in many acute care hospitals. We attempted to identify the risk factors and develop a predictive scoring model for positive FUBC in GNB cases.MethodsAll adults with GNB in a tertiary care hospital were retrospectively identified during a 2-year period, and GNB cases were assigned to eradicable and non-eradicable groups based on whether removal of the source of infection was possible. We performed multivariate logistic analyses to identify risk factors for positive FUBC and built predictive scoring models accordingly.ResultsOut of 1473 GNB cases, FUBCs were carried out in 1268 cases, and the results were positive in 122 cases. In case of eradicable source of infection, we assigned points according to the coefficients from the multivariate logistic regression analysis: Extended spectrum beta-lactamase-producing microorganism (+ 1 point), catheter-related bloodstream infection (+ 1), unfavorable treatment response (+ 1), quick sequential organ failure assessment score of 2 points or more (+ 1), administration of effective antibiotics (− 1), and adequate source control (− 2). In case of non-eradicable source of infection, the assigned points were end-stage renal disease on hemodialysis (+ 1), unfavorable treatment response (+ 1), and the administration of effective antibiotics (− 2). The areas under the curves were 0.861 (95% confidence interval [95CI] 0.806–0.916) and 0.792 (95CI, 0.724–0.861), respectively. When we applied a cut-off of 0, the specificities and negative predictive values (NPVs) in the eradicable and non-eradicable sources of infection groups were 95.6/92.6% and 95.5/95.0%, respectively.ConclusionsFUBC is commonly carried out in GNB cases, but the rate of positive results is less than 10%. In our simple predictive scoring model, zero scores—which were easily achieved following the administration of effective antibiotics and/or adequate source control in both groups—had high NPVs. We expect that the model reported herein will reduce the necessity for FUBCs in GNB cases.

Bile cholesterol and viscosity, the keys to discriminating adenomatous polyps from cholesterol polyps by a novel predictive scoring model

BackgroundAdenomatous gallbladder polyps, premalignant lesions of the gallbladder, have fatal outcomes, whereas cholesterol polyps have benign features. Herein, we proposed a novel, predictive scoring model of adenomatous polyps to distinguish them from cholesterol polyps, by analyzing bile components and bile viscosity.MethodsPatients with gallbladder polyp pathologically confirmed after cholecystectomies were analyzed. After dividing patients into two groups (adenomatous or cholesterol polyps), the clinicopathologic profiles and bile nature, including components and viscosity were compared and a predictive scoring model for adenomatous polyps was assessed.ResultsEleven adenomatous polyps and 96 cholesterol polyps were analyzed. The variables significantly associated with adenomatous polyps were age > 55 years (OR = 23.550, p = 0.020), bile viscosity< 7.5 s− 1 (OR = 22.539, p = 0.012), and bile cholesterol< 414.5 mg/dl (OR = 10.004, p = 0.023) and the points for each variable in the predictive scoring model were allocated as 3, 3, and 2, respectively. Final scores ranged from 0 to 8 points and the best performance of model at a cutoff of ≥6 points had 90.9% of sensitivity and 80.2% of specificity.ConclusionsBile viscosity and bile cholesterol accompanied by age were revealed as significant predictors of adenomatous polyps, distinguishing them from cholesterol polyps of gallbladder. It can be the cornerstone for creating accurate guidelines for preoperatively determining treatment strategies of gallbladder polyps.

Establishment of risk prediction model and risk score for in-hospital mortality in patients with AECOPD.

Risk stratification for patients with acute exacerbation of chronic obstructive pulmonary disease (AECOPD) may help clinicians choose appropriate treatments and improve the quality of care. A total of 695 patients hospitalized with AECOPD from January 2015 to December 2017 were considered. They were assigned to a death and a survival cohort. The independent prognostic factors were determined by multivariate logistic regression analysis. Meanwhile, we also compared the new scale with three other scores and tested the new scale internally and externally. A new risk score was created, made up of six independent variables: age, D-dimer, albumin, cardiac troponin I, partial pressure of carbon dioxide and oxygenation index. The area under the receiver operator characteristic curve (AUROC) for the model was 0.929, and the other three CURB-65, DECAF and BAP-65 models were 0.718, 0.922 and 0.708. The Cohen's kappa coefficient between the new scale and DECAF was calculated to be 0.648, suggesting that there is a substantial consistency between the two. In the internal and external validation cohorts, 490 and 500 patients were recruited with a total mortality rate of 5.15%. The AUROC for in-hospital mortality was 0.937 in the internal cohort and 0.914 in external cohort, which was significantly better than the scores for CURB-65 and BAP-65, but it was not significantly different from the DECAF. The new scale may help to stratify the risk of in-hospital mortality of AECOPD. The DECAF performed as well as the new instrument, and it appears to be valid in Chinese patients.

Predictive risk score model for severe fever with thrombocytopenia syndrome mortality based on qSOFA and SIRS scoring system

BackgroundSevere fever with thrombocytopenia syndrome (SFTS) is a severe systemic virus infectious disease usually having multi-organ dysfunction which resembles sepsis.MethodsData of 321 patients with laboratory-confirmed SFTS from May 2013 to July 2017 were retrospectively analyzed. Demographic and clinical characteristics, calculated quick sequential organ failure assessment (qSOFA) score and systemic inflammatory response syndrome (SIRS) criteria for survivors and nonsurvivors were compared. Independent risk factors associated with in-hospital mortality were obtained using multivariable logistic regression analysis. Risk score models containing different risk factors for mortality in stratified patients were established whose predictive values were evaluated using the area under ROC curve (AUC).ResultsOf 321 patients, 87 died (27.1%). Age (p < 0.001) and percentage numbers of patients with qSOFA≥2 and SIRS≥2 (p < 0.0001) were profoundly greater in nonsurvivors than in survivors. Age, qSOFA score, SIRS score and aspartate aminotransferase (AST) were independent risk factors for mortality for all patients. qSOFA score was the only common risk factor in all patients, those age ≥ 60 years and those enrolled in the intensive care unit (ICU). A risk score model containing all these risk factors (Model1) has high predictive value for in-hospital mortality in these three groups with AUCs (95% CI): 0.919 (0.883–0.946), 0.929 (0.862–0.944) and 0.815 (0.710–0.894), respectively. A model only including age and qSOFA also has high predictive value for mortality in these groups with AUCs (95% CI): 0.872 (0.830–0.906), 0.885(0.801–0.900) and 0.865 (0.767–0.932), respectively.ConclusionsRisk models containing qSOFA have high predictive validity for SFTS mortality.

Predictive Model for Postoperative Ambulatory Function after Lower Extremity Bypass in Chronic Limb-Threatening Ischemia

In chronic limb-threatening ischemia, maintenance or recovery of ambulatory function is an important goal of treatment. This study aimed to develop a predictive model for ambulatory ability 1year after bypass based on preoperative risk factors, including the Wound, Ischemia, and foot Infection (WIfI) classification. We analyzed 146 patients with chronic limb-threatening ischemia (154 limbs) who underwent bypass to below the knee arteries. The patients were classified into 2 groups based on ambulatory status 1year postoperatively: postoperative ambulation (99 patients, 104 limbs) and postoperative nonambulation (47 patients, 50 limbs). Various factors associated with postoperative ambulation were analyzed and a predictive model of postoperative ambulation was developed. Multivariate logistic regression analysis detected preoperative nonambulatory status, functional nonindependence in daily living, older age, WIfI wound grade 3, chronic obstructive pulmonary disease, and hemodialysis as independent risk factors for postoperative nonambulation. The predictive scoring model (scores ranging from -5.0 to 4.4) comprising these risk factors discriminated the postoperative ambulatory status well: the probabilities of postoperative ambulatory ability were ≥85% in those with a score ≤-2, 50% in those with a score of zero, and ≤15% in those with a score ≥2. The area under the receiver operating characteristic curve was 0.898, indicating good performance of the model. Preoperative nonambulatory status, functional nonindependence, advanced age, high WIfI wound grade, chronic obstructive pulmonary disease, and hemodialysis were important predictors of postoperative nonambulatory status. The predictive model will help us identify patients who will benefit from bypass surgery.

Prediction model and risk scores of ICU admission and mortality in COVID-19.

This study aimed to develop risk scores based on clinical characteristics at presentation to predict intensive care unit (ICU) admission and mortality in COVID-19 patients. 641 hospitalized patients with laboratory-confirmed COVID-19 were selected from 4997 persons under investigation. We performed a retrospective review of medical records of demographics, comorbidities and laboratory tests at the initial presentation. Primary outcomes were ICU admission and death. Logistic regression was used to identify independent clinical variables predicting the two outcomes. The model was validated by splitting the data into 70% for training and 30% for testing. Performance accuracy was evaluated using area under the curve (AUC) of the receiver operating characteristic analysis (ROC). Five significant variables predicting ICU admission were lactate dehydrogenase, procalcitonin, pulse oxygen saturation, smoking history, and lymphocyte count. Seven significant variables predicting mortality were heart failure, procalcitonin, lactate dehydrogenase, chronic obstructive pulmonary disease, pulse oxygen saturation, heart rate, and age. The mortality group uniquely contained cardiopulmonary variables. The risk score model yielded good accuracy with an AUC of 0.74 ([95% CI, 0.63-0.85], p = 0.001) for predicting ICU admission and 0.83 ([95% CI, 0.73-0.92], p<0.001) for predicting mortality for the testing dataset. This study identified key independent clinical variables that predicted ICU admission and mortality associated with COVID-19. This risk score system may prove useful for frontline physicians in clinical decision-making under time-sensitive and resource-constrained environment.