Risk Score For Early Prediction Research Articles

Development and external validation of a dynamic risk score for early prediction of cardiogenic shock in cardiac intensive care units using machine learning.

Myocardial infarction and heart failure are major cardiovascular diseases that affect millions of people in the US with the morbidity and mortality being highest among patients who develop cardiogenic shock. Early recognition of cardiogenic shock allows prompt implementation of treatment measures. Our objective is to develop a new dynamic risk score, called CShock, to improve early detection of cardiogenic shock in cardiac intensive care unit (ICU). We developed and externally validated a deep learning-based risk stratification tool, called CShock, for patients admitted into the cardiac ICU with acute decompensated heart failure and/or myocardial infarction to predict onset of cardiogenic shock. We prepared a cardiac ICU dataset using MIMIC-III database by annotating with physician adjudicated outcomes. This dataset that consisted of 1500 patients with 204 having cardiogenic/mixed shock was then used to train CShock. The features used to train the model for CShock included patient demographics, cardiac ICU admission diagnoses, routinely measured laboratory values and vital signs, and relevant features manually extracted from echocardiogram and left heart catheterization reports. We externally validated the risk model on the New York University (NYU) Langone Health cardiac ICU database that was also annotated with physician adjudicated outcomes. The external validation cohort consisted of 131 patients with 25 patients experiencing cardiogenic/mixed shock. CShock achieved an area under the receiver operator characteristic curve (AUROC) of 0.821 (95% CI 0.792-0.850). CShock was externally validated in the more contemporary NYU cohort and achieved an AUROC of 0.800 (95% CI 0.717-0.884), demonstrating its generalizability in other cardiac ICUs. Having an elevated heart rate is most predictive of cardiogenic shock development based on Shapley values. The other top ten predictors are having an admission diagnosis of myocardial infarction with ST-segment elevation, having an admission diagnosis of acute decompensated heart failure, Braden Scale, Glasgow Coma Scale, Blood urea nitrogen, Systolic blood pressure, Serum chloride, Serum sodium, and Arterial blood pH. The novel CShock score has the potential to provide automated detection and early warning for cardiogenic shock and improve the outcomes for the millions of patients who suffer from myocardial infarction and heart failure.

A novel risk score for venous thromboembolism in lung cancer patients: a retrospective cohort study

Background: Venous thromboembolism (VTE) is a common and potentially fatal complication in patients with lung cancer. This study aimed to develop and validate a risk score for early prediction of VTE in these patients. Methods: Four hundred and one patients with lung cancer from three pulmonology departments hospitalized between January 2011 and December 2021 were retrospectively assessed. The population was divided into two groups: a Development Group (182 patients) and a validation group (199 patients). In the development group, the risk score system was developed, via univariate and multivariate analyses, based on demographic and clinicopathological variables; it was then validated in the validation group. Results: The incidence of VTE was 26.8% in the development group. It was 25.8%, and 27.6% in the internal and external validation groups, respectively. Hemoglobin level <10g/l, metastasis, histological type poorly or undifferentiated non-small cell carcinoma, and active smoking were the items of the risk score system. This score allowed proper stratification of patients with either high or low risk of VTE in the development group (c statistic =0.703). The patients in the development group were classified into 3 risk groups: low risk (scores 0-1), moderate risk (scores 2-3), and high risk (scores 4-5). When validated in the validation group, there was a moderate loss of predictive power of the score (c statistic=0.641), but the categorization of the patients by the score remained clinically useful. Conclusions: This risk score requires prospective validation studies on a nationwide scale in order to use it as a valid tool for the prevention of VTE in lung cancer.

Clinical risk score for early prediction of recurring SARS-CoV-2 positivity in non-critical patients.

Recurrent positive results in quantitative reverse transcriptase-PCR (qRT-PCR) tests have been commonly observed in COVID-19 patients. We aimed to construct and validate a reliable risk stratification tool for early predictions of non-critical COVID-19 survivors' risk of getting tested re-positive within 30 days. We enrolled and retrospectively analyzed the demographic data and clinical characters of 23,145 laboratory-confirmed cases with non-critical COVID-19. Participants were followed for 30 days and randomly allocated to either a training (60%) or a validation (40%) cohort. Multivariate logistic regression models were employed to identify possible risk factors with the SARS-CoV-2 recurrent positivity and then incorporated into the nomogram. The study showed that the overall proportion of re-positive cases within 30 days of the last negative test was 24.1%. In the training cohort, significantly contributing variables associated with the 30-day re-positivity were clinical type, COVID-19 vaccination status, myalgia, headache, admission time, and first negative conversion, which were integrated to build a nomogram and subsequently translate these scores into an online publicly available risk calculator (https://anananan1.shinyapps.io/DynNomapp2/). The AUC in the training cohort was 0.719 [95% confidence interval (CI), 0.712-0.727] with a sensitivity of 66.52% (95% CI, 65.73-67.30) and a specificity of 67.74% (95% CI, 66.97-68.52). A significant AUC of 0.716 (95% CI, 0.706-0.725) was obtained for the validation cohort with a sensitivity of 62.29% (95% CI, 61.30-63.28) and a specificity of 71.26% (95% CI, 70.34-72.18). The calibration curve exhibited a good coherence between the actual observation and predicted outcomes. The risk model can help identify and take proper management in high-risk individuals toward the containment of the pandemic in the community.

The Caprini Risk Score for Early Prediction of Mortality in Patients With Acute Coronary Syndrome.

The Caprini Risk Score (CRS) is a validated predictive instrument for venous thrombosis. Previous investigators have shown that a high CRS is associated with a higher risk of mortality from thrombotic diseases. The aim of this study was to assess the association between the CRS and prognosis of patients with acute coronary syndrome (ACS). Secondary analysis of data from a retrospective cohort study was conducted. Patients were classified into 3 CRS-based categories (CRS ≤ 2, CRS = 3-4, and CRS ≥ 5, indicating low, medium, and high, respectively). Kaplan-Meier curves and Cox regression models were used to assess the prognosis of patients with ACS. All-cause mortality and cardiac mortality were the end points. Two hundred fifty-four patients (12.8%) died during follow-up. Multivariate Cox regression models identified CRS as an independent risk factor for all-cause mortality among patients with ACS (CRS = 3-4 vs CRS ≤ 2, hazard ratio: 3.268, 95% confidence interval: 1.396-7.647, P = .006; CRS ≥ 5 vs CRS ≤ 2, hazard ratio: 4.099, 95% confidence interval: 1.708-9.841, P = .002). Pearson correlation analysis showed a positive correlation between CRS and fibrinogen level ( r = 0.486, R2 = 0.765, P < .001) as well as D-dimer level ( r = 0.480, R2 = 0.465, P < .001). The CRS is a useful prognostic assessment instrument for patients with ACS, and the risk stratification of patients with ACS can be achieved based on their CRS at admission.

A risk score for early predicting bloodstream infections in febrile obstetric patients: a pilot study.

Early prediction of bloodstream infections (BSI) among obstetric patients remains to be a challenge for clinicians. The objective of this study was to develop a risk score and assess its discriminative ability in febrile obstetric patients in a maternal intensive care unit (ICU). Between May 2015 and August 2020, a total of 497 febrile obstetric patients were categorized into BSI group (n = 276) and Non-BSI group (n = 221) based on the result of blood cultures. White blood cell count, C-reactive protein (CRP), procalcitonin (PCT), time of interval from amniorrhea to fever (IFAF) and maximum body temperature (Tmax) were compared between the two groups. All patients were divided into training set (n = 298) and validation set (n = 199). The risk score was established using univariate and multivariate logistic regression from patients in the training set, and its discriminative ability was tested among patients in the validation set. The levels of neutrophil, CRP, PCT, IFAF and Tmax were significantly higher in BSI group than those in Non-BSI group. PROM, Tmax, neutrophil and CRP acted as independent predictive factors for BSI in the training set. The area under the receiver operating characteristic curve of risk score for early prediction of BSI in the training, validation set and the whole population was 0.829 (95% CI 0.783-0.876), 0.848 (95% CI 0.792-0.903) and 0.838 (95% CI 0.803-0.873), respectively. The risk score has a feasible discriminatory ability in early prediction of BSI in febrile obstetric patients.

Polygenic Risk Score for Early Prediction of Sepsis Risk in the Polytrauma Screening Cohort.

BackgroundIncreasing genetic variants associated with sepsis have been identified by candidate-gene and genome-wide association studies, but single variants conferred minimal alterations in risk prediction. Our aim is to evaluate whether a weighted genetic risk score (wGRS) that aggregates information from multiple variants could improve risk discrimination of traumatic sepsis.MethodsSixty-four genetic variants potential relating to sepsis were genotyped in Chinese trauma cohort. Genetic variants with mean decrease accuracy (MDA) > 1.0 by random forest algorithms were selected to construct the multilocus wGRS. The area under the curve (AUC) and net reclassification improvement (NRI) were adopted to evaluate the discriminatory and reclassification ability of weighted genetic risk score (wGRS).ResultsSeventeen variants were extracted to construct the wGRS in 883 trauma patients. The wGRS was significantly associated with sepsis after trauma (OR = 2.19, 95% CI = 1.53–3.15, P = 2.01 × 10–5) after being adjusted by age, sex, and ISS. Patients with higher wGRS have an increasing incidence of traumatic sepsis (Ptrend = 6.81 × 10–8), higher SOFA (Ptrend = 5.00 × 10–3), and APACHE II score (Ptrend = 1.00 × 10–3). The AUC of the risk prediction model incorporating wGRS into the clinical variables was 0.768 (95% CI = 0.739–0.796), with an increase of 3.40% (P = 8.00 × 10–4) vs. clinical factor-only model. Furthermore, the NRI increased 25.18% (95% CI = 17.84–32.51%) (P = 6.00 × 10–5).ConclusionOur finding indicated that genetic variants could enhance the predictive power of the risk model for sepsis and highlighted the application among trauma patients, suggesting that the sepsis risk assessment model will be a promising screening and prediction tool for the high-risk population.

A practical risk score for early prediction of neurological outcome after out-of-hospital cardiac arrest: MIRACLE2.

The purpose of this study was to develop a practical risk score to predict poor neurological outcome after out-of-hospital cardiac arrest (OOHCA) for use on arrival to a Heart Attack Centre. From May 2012 to December 2017, 1055 patients had OOHCA in our region, of whom 373 patients were included in the King's Out of Hospital Cardiac Arrest Registry (KOCAR). We performed prediction modelling with multivariable logistic regression to identify predictors of the primary outcome to derive a risk score. This was externally validated in two independent cohorts comprising 473 patients. The primary endpoint was poor neurological outcome at 6-month follow-up (Cerebral Performance Category 3-5). Seven independent predictors of outcome were identified: missed (unwitnessed) arrest, initial non-shockable rhythm, non-reactivity of pupils, age (60-80 years-1 point; >80 years-3 points), changing intra-arrest rhythms, low pH <7.20, and epinephrine administration (2 points). The MIRACLE2 score had an area under the curve (AUC) of 0.90 in the development and 0.84/0.91 in the validation cohorts. Three risk groups were defined-low risk (MIRACLE2 ≤2-5.6% risk of poor outcome); intermediate risk (MIRACLE2 of 3-4-55.4% of poor outcome); and high risk (MIRACLE2 ≥5-92.3% risk of poor outcome). The MIRACLE2 score had superior discrimination than the OHCA [median AUC 0.83 (0.818-0.840); P < 0.001] and Cardiac Arrest Hospital Prognosis models [median AUC 0.87 (0.860-0.870; P = 0.001] and equivalent performance with the Target Temperature Management score [median AUC 0.88 (0.876-0.887); P = 0.092]. The MIRACLE2 is a practical risk score for early accurate prediction of poor neurological outcome after OOHCA, which has been developed for simplicity of use on admission.

Development of a new risk score for invasive cancer in branch-duct intraductal papillary mucinous neoplasms according to morphological characterization by EUS.

Background and Objective:The management of branch-duct intraductal papillary mucinous neoplasms (BD-IPMNs) is determined by a number of guidelines. The current weight of risk factors by EUS predicting invasive cancer is unknown. The aim of this study is to develop a risk score for early prediction of invasive cancer according to morphological characterization by EUS in a surgical cohort.Materials and Methods:This is an observational, multicenter retrospective study. All consecutive patients with a histologically proven BD-IPMN who underwent previous EUS between 2005 and 2017 were included. Morphological features by EUS were evaluated. A score using a logistic regression model was performed to assess the risk of invasive cancer.Results:Of 335 patients who underwent pancreatic surgery, 131 (median age: 66 years, 50.4% – male) were included. By multivariable analysis, lymph nodes (odds ratio [OR]: 17.7 [confidence interval (CI) 95%: 2.8–112.6], P = 0.002, 4 points), main pancreatic duct ≥10 mm (OR: 8.6 [CI 95%: 1.9–39.5], P = 0.006, 2 points), abrupt change of pancreatic duct (OR: 5.5 [CI 95%: 1.4–22.2], P = 0.016, 1.5 points), and solid component (OR: 4.2 [CI 95%: 1.3–13.6], P = 0.017, 1 point) were independent factors associated with invasive cancer and included in the model. The following categories of the score (0–8.5 points) – A (0–1), B (1.5–3), C (3.5–5), and D (5.5–8.5 points) – presented a positive predictive value of 8.5%, 38.9%, 62.5%, and 100%, respectively. The area under the curve was 0.857 (P < 0.001), with an overall sensitivity and specificity of 84% and 70% in the internal validation of the score.Conclusion:This EUS predictive score for invasive cancer in BD-IPMN has a high accuracy and could be an additional tool to consider in patient management.

Development and Validation of a Risk Score for Prediction of Venous Thromboembolism in Patients With Lung Cancer.

This study aimed to develop and validate a risk score for early prediction of venous thromboembolism (VTE) in patients with lung cancer. A total of 827 patients with lung cancer from February 2013 to February 2018 in our hospital were retrospectively analyzed. Demographic and clinicopathological variables independently correlated to VTE were applied to develop the risk score in the development group while examined in the validation group. The regression coefficients of multivariable logistic regression test were applied to assign a risk score system. The incidence of VTE was 12.3%, 12.7%, and 11.8% in all patients, in the development and validation groups, respectively. The 496 patients in the development group were classified into 3 groups: low risk (scores ≤3), moderate risk (scores 4-5), and high risk (scores ≥6). The risk of VTE was significantly and positively related to the risk scores in both development and validation groups. The risk score system aided proper stratification of patients with either high or low risk of VTE in the development and validation groups (c statistic = 0.819 and 0.827, respectively). This risk score system based on the factors with most significant correlation showed good predictive ability and is potentially useful for predicting VTE in patients with lung cancer. However, it was developed and validated by a retrospective analysis and has significant limitations, and a prospective validation with all the classic variables assessing the thrombotic risk is needed for a solid conclusion.

Risk score for early prediction of invasive cancer of BD-IPMN according to morphological characterization in EUS in patients who underwent to pancreatic surgery

Risk score for early prediction of invasive cancer of BD-IPMN according to morphological characterization in EUS in patients who underwent to pancreatic surgery

Development of a Simple Clinical Risk Score for Early Prediction of Severe Dengue in Adult Patients.

We aimed to develop and validate a risk score to aid in the early identification of laboratory-confirmed dengue patients at high risk of severe dengue (SD) (i.e. severe plasma leakage with shock or respiratory distress, or severe bleeding or organ impairment). We retrospectively analyzed data of 1184 non-SD patients at hospital presentation and 69 SD patients before SD onset. We fit a logistic regression model using 85% of the population and converted the model coefficients to a numeric risk score. Subsequently, we validated the score using the remaining 15% of patients. Using the derivation cohort, two scoring algorithms for predicting SD were developed: models 1 (dengue illness ≤4 days) and 2 (dengue illness >4 days). In model 1, we identified four variables: age ≥65 years, minor gastrointestinal bleeding, leukocytosis, and platelet count ≥100×109 cells/L. Model 1 (ranging from −2 to +6 points) showed good discrimination between SD and non-SD, with an area under the receiver operating characteristic curve (AUC) of 0.848 (95% confidence interval [CI], 0.771–0.924). The optimal cutoff value for model 1 was 1 point, with a sensitivity and specificity for predicting SD of 70.3% and 90.6%, respectively. In model 2 (ranging from 0 to +3 points), significant predictors were age ≥65 years and leukocytosis. Model 2 showed an AUC of 0.859 (95% CI, 0.756–0.963), with an optimal cutoff value of 1 point (sensitivity, 80.3%; specificity, 85.8%). The median interval from hospital presentation to SD was 1 day. This finding underscores the importance of close monitoring, timely resuscitation of shock including intravenous fluid adjustment and early correction of dengue-related complications to prevent the progressive dengue severity. In the validation data, AUCs of 0.904 (95% CI, 0.825–0.983) and 0.917 (95% CI, 0.833–1.0) in models 1 and 2, respectively, were achieved. The observed SD rates (in both cohorts) were <3% for patients with a score <1 point, but >50% for those with a score of ≥2 points, irrespective of the day of illness onset, suggesting that our simple risk score can be easily implemented in resource-limited countries for early prediction of dengue patients at risk of SD provided that they have rapid dengue confirmed tests. For patients with other acute febrile illnesses or bacterial infections usually have SD risk score of >1. Thus, these scoring algorithms cannot totally replace good clinical judgement of the physician, and most importantly, early differentiating dengue from other febrile illnesses is critical for appropriate monitoring and management.

Middle age self-report risk score predicts cognitive functioning and dementia in 20–40 years

IntroductionOn the basis of the proxy measures of cognitive reserve, we created a middle age self-report risk score for early prediction of dementia. MethodsWe used a longitudinal population-based study of 2602 individuals with a replication sample (N = 1011). Risk score at a mean age of 47 years was based on questions on educational and occupational attainments. Cognitive status at a mean age of 74 was determined via two validated telephone instruments. ResultsThe prevalence of dementia was 10% after a mean follow-up of 28 years. Risk score was a good predictor of dementia: area under the curve = 0.77 (95% confidence interval, 0.74–0.80). The risk of dementia decreased as a function of risk score from 36% to 0%. The risk score was significantly associated with cognition after a mean follow-up of 39 years in the replication sample. DiscussionSelf-report risk score predicted cognitive functioning and dementia risk 20–40 years later.

Risk Factors for Acute Kidney Injury in Severe Rhabdomyolysis

BackgroundAcute kidney injury (AKI) is a life-threatening complication of severe rhabdomyolysis. This study was conducted to assess risk factors for AKI and to develop a risk score for early prediction.MethodsRetrospective observational cohort study with a 9-year follow-up, carried out in an acute-care teaching-affiliated hospital. A total of 126 patients with severe rhabdomyolysis defined as serum creatine kinase (CK) > 5,000 IU/L fulfilled the inclusion criteria. Univariate and logistic regression analyses were performed to determine risk factors for AKI. Based on the values obtained for each variable, a risk score and prognostic probabilities were estimated to establish the risk for developing AKI.ResultsThe incidence of AKI was 58%. Death during hospitalization was significantly higher among patients with AKI, compared to patients without AKI (19.2% vs 3.6%, p = 0.008). The following variables were independently associated with AKI: peak CK (odds ratio [OR] 4.9, 95%CI 1.4-16.8), hypoalbuminemia (< 33 mg/dL, [OR 5.1, 95%CI 1.4-17-7]), metabolic acidosis (OR 5.3, 95%CI 1.4-20.3), and decreased prothrombin time (OR 4.4, 95% CI 1.3-14.5). A risk score for AKI was calculated for each patient, with an OR of 1.72 (95%CI 1.45-2.04). The discrimination value of the predictive model was established by means of a ROC curve, with the area under the curve of 0.871 (p<0.001).ConclusionsThe identification of independent factors associated with AKI and a risk score for early prediction of this complication in patients with severe rhabdomyolysis may be useful in clinical practice, particularly to implement early preventive measures.