Mortality Risk Calculator Research Articles

Biomarkers Predictive of Extubation and Survival of COVID-19 Patients.

PurposeMany patients with COVID-19 who develop acute respiratory distress syndrome (ARDS) require prolonged periods of mechanical ventilation. Mechanical ventilation may amplify ventilator-associated complications and extend resource utilization. A better understanding of prognostic indicators could help in the planning and distribution of resources, particularly in resource-limited areas. We analyzed laboratory studies of intubated COVID-19 patients with the goal of identifying biomarkers that may predict extubation success and survival to discharge.MethodsA retrospective chart review was performed on all COVID-19 patients requiring mechanical ventilation between January 3, 2020, and January 7, 2020, in a single academic tertiary care center in Northeastern New York State. The electronic medical record was used to collect 14 laboratory variables at three time points: admission, intubation, and extubation (including terminal extubation) for all intubated intensive care unit (ICU) patients treated for COVID-19. Mean laboratory values were analyzed with the Mann-Whitney U test. Categorical variables were analyzed with the two-sample Wilcoxon rank-sum test.ResultsSeventy-two patients met the inclusion criteria. Forty-three patients were male. The mean age was 61 years. The overall mortality was 50%. On admission, intubated patients who survived had significantly higher platelet counts (p=0.024), and absolute lymphocyte counts (ALC; p=0.047). Notably, ferritin (p=0.018) and aspartate transaminase (AST; p=0.0045) levels were lower in survivors.At the time of intubation, survivors again had a higher platelet count (p=0.024) and ALC (p=0.037) levels. They had a lower D-dimer (p=0.0014), ferritin (p=0.0015), lactate dehydrogenase (LDH; p=0.0145), and AST (p=0.018) compared to intubated patients who died.At extubation, survivors had higher platelet count (p=0.0002), ALC (p=0.0013), and neutrophil/lymphocyte ratio (NLR; p=0.0024). Survivors had lower d-dimer (p=0.035), ferritin (p=0.0012), CRP (p=0.045), LDH (p=0.002), AST (p<0.001), and ALK (p=0.0048).ConclusionsBiomarkers associated with increased risk of mortality include platelet count, ALC, lymphocyte percentage, NLR, D-dimer, ferritin, C-reactive protein (CRP), AST, alanine transaminase (ALT), and alkaline phosphatase (ALK). This study provides additional evidence that these biomarkers have prognostic value in patients with severe COVID-19. The goal is to find objective surrogate markers of disease improvement or success of extubation. When considered within the larger body of data, it is our hope that a mortality risk calculator can be generated for intubated COVID-19 patients.

Machine learning prediction of COVID-19 mortality in cancer patients.

1558 Background: COVID-19 is a challenge for clinical decision-making in cancer patients and the allocation of healthcare resources. An accurate prognosis prediction to effectively triage patients is needed, especially in the community oncology practice. Methods:Nationwide cohort from Oncoclínicas Brazil was used to validate previously developed multivariable logistic regression (mLR) model (Ferrari et al, JCO GO 2021) and to construct a machine learning Random Forest (RF) algorithm as predictor of 30-day mortality after SARS-CoV-2 detection by RT-PCR in cancer patients diagnosed in an outpatient setting. To find the most important baseline clinical determinants of early COVID-19-related death via Gini index, a RF with 100,000 trees was trained in 75% of the dataset, and the performance was assessed in the remaining 25%. We then compared the accuracy of different models in terms of sensitivity, specificity and area under the receiver operating characteristics curves (AUC). Results:From March to December 2020, 533 patients with COVID-19 were prospectively registered in the database. Median age was 60 years (19-93) and 67% were female. Most frequent cancers were breast in 34%, hematological in 16%, and gastrointestinal in 15%. Comorbidities were common (52%), as was current/former smoking history (17%). Most patients were on active systemic therapy or radiotherapy (84%) in the advanced or metastatic disease setting (55%). The overall mortality rate was 15% (CI95% 12%-18%). We validated the original mLR model trained in the first 198 patients: management in a non-curative setting (odds ratio [OR] 3.7), age ≥ 60 years (OR 2.3), and current/former smoking (OR 1.9) were significant predictors of death in the expanded cohort. Presence of comorbidities (OR 1.9) also defined poor outcome in the updated mLR model, which yielded low sensitivity (74%), specificity (68%) and AUC (0.78). With RF modeling, the most significant predictors of 30-day death after COVID-19 (in decreasing order) were older age, treatment of advanced or metastatic disease, tumor type (respiratory tract, brain and unknown primary cancers had higher mortality), COVID-related symptom burden at baseline evaluation and treatment regimen (immunotherapy combinations had higher mortality). The RF model demonstrated high sensitivity (89%), specificity (88%) and AUC (0.96). Conclusions:The results highlight the possibility that machine learning algorithms are able to predict early mortality after COVID-19 in cancer patients with high accuracy. The proposed prediction model may be helpful in the prompt identification of high-risk patients based on clinical features alone, without having to wait for the results of additional tests such as laboratory or radiologic studies. It can also help prioritize medical resources and redefine vaccination strategies. A web-based mortality risk calculator will be created for clinical decision support.

668 Utilizing Risk Score Assessment to Maximize Sleep Research Participant Safety During the COVID-19 Pandemic

IntroductionResearch study recruitment has been profoundly affected by the COVID-19 pandemic, demonstrated by significant delays or pauses. Various guidelines pertaining to in-person visits have applied to research. Some call for exclusion of participants that the CDC has labeled “at increased risk”.1 For obstructive sleep apnea (OSA) studies, these guidelines have caused a sharp decrease in the number of new participants. This decrease is due to high rates of OSA comorbidities including obesity and diabetes. New evidence-based risk scores have been developed using individual- and community-level factors. The use of more refined COVID-19 risk scores can help protect patient safety while allowing research to continue.MethodsThe risk score assessment used for this study (COVID-19 Mortality Risk Calculator; Johns Hopkins University, Baltimore, MD)2 is evidence-based and uses a set of risk factors and community-level pandemic dynamics in the state of residence.3,4 It was compared to the list of CDC medical conditions that are considered to put an individual “at increased risk.” Both measures were calculated retrospectively on current participants to determine how many could safely attend in-person visits based on each risk assessment method.ResultsSample characteristics of the 110 participants were: mean age: 49.5±13.7(24–76); mean BMI: 32.3±5.3(20.9–46.1); mean AHI: 24.3±21.4(5.1–110). Mortality Risk Calculator scores were: 91(82.7%) close to/lower than average [Level 1]; 12(10.9%) moderately elevated; 6(5.5%) substantially elevated; 1(0.9%) high; and 0(0%) very high [Level 5]. Using CDC guidance, 63 (57.3%) had at least one at-risk condition and 47 (42.7%) had 0. Using only Level 1 of the Risk Calculator would allow an additional 28 (25%) participants to attend in-person visits; using Levels 1 and 2 would allow an additional 40 (37%) participants.ConclusionPolicies based on CDC at-risk conditions resulted in higher levels of participant exclusion in research during the COVID-19 pandemic than use of an evidence-based Mortality Risk Calculator. This analysis shows that researchers can use risk-adjusted scores to make informed decisions about study participation that balances both participant safety and research study progress.Support (if any)This project was supported in part by Department of Veteran Affairs VA HSRD IIR 16–277, VA RRD D2651-R, and VA San Diego Healthcare System Research Service.

Real-time mortality risk calculator following pancreatoduodenectomy: quantifying the impact of perioperative events

Presenter: Timothy M Pawlik MD, PhD, MPH | The Ohio State University Background: Despite advances in surgical techniques, pancreatoduodenectomy (PD) still carries a risk of post-operative mortality. Traditionally, mortality risk prediction has been based solely on pre-operative patient-level risk factors. In reality, risk of in-hospital mortality following PD is more dynamic and also depends on intra- and post-operative events. We sought to develop a mortality risk calculator based on “real-time” peri-operative characteristics occurring during the entire PD hospital course. Methods: Patients who underwent PD between 2014-2018 were identified in the American College of Surgeons National Surgical Quality Improvement Program (NSQIP) database. The study cohort was randomly divided into training and validation cohorts. Three models (pre-, intra-, and post-operative) to predict 30-day mortality were developed based on peri-operative variables selected by stepwise cox regression analyses. Patients were classified into low-, intermediate-, or high-risk relative to 30-day mortality. Model performance was assessed using area under the curve (AUC) and model fit was assessed using the corrected Akaike information criterion (AICc) and Kaplan-Meier plots. Results: Among 17,683 patients who underwent PD, 284 (1.6%) individuals died within 30-days. A number of patient-level factors in the pre-operative (age, sex, hypertension, currently on dialysis, ASA class, laboratory data of albumin, total bilirubin, and white blood cells), intra-operative (operative time, blood transfusion, and arterial resection), and post-operative (pneumonia, pulmonary embolism, acute renal failure, cerebrovascular accident/stroke, myocardial infarction, pancreatic fistula, and reoperation) periods were associated with 30-day mortality and were therefore incorporated into a risk calculator (Table). Risk of 30-day mortality incrementally increased in the different models with each higher risk category (pre-operative model: low, 0.8% vs. intermediate, 1.8% vs. high, 3.7%; log-rank p<0.001, AICc = 5459)(intra-operative model: low, 0.7% vs. intermediate, 1.8% vs. high, 4.0%; log-rank p<0.001, AICc = 5429)(post-operative model: low, 0.2% vs. intermediate, 1.1%, vs. high, 7.1%; log-rank p<0.001, AICc = 5084). Of note, the predictive ability of the real-time mortality risk calculator was good based on only preoperative factors (AUC; training cohort: 0.696, validation cohort: 0.681). Incorporating intra-operative variables into the model increased the AUC (training: 0.724, validation: 0.702), while the model that incorporated post-operative factors was most accurate in predicting 30-day mortality (training: AUC of 0.871, validation: 0.850). Importantly, roughly 1 in 4 patients (n = 4,270, 24.1%; kappa = 0.60) had a discordant calculated risk of 30-day mortality comparing the pre- versus intra-operative risk models (increased risk: n = 2,298, 53.8%; decreased risk: n = 1,972, 46.2%). The discordance in risk assessment was even more pronounced (n = 5,673, 32.1%; kappa = 0.47) among patients comparing the pre- versus post-operative risk models (increased risk: n = 2,870, 50.6%; decreased risk: n = 2,803, 49.4%). The calculator is available: https://ktsahara.shinyapps.io/Real-timePD/. Conclusion: Risk of 30-day mortality was not static and changed over the hospital course following PD. Pre-operative risk assessment was frequently discordant with risk assessed at other points during the hospital course. The proposed “real-time” calculator may help stratify postoperative mortality risk better in the real-world clinical setting.

LUCAS: Development and Validation of a Simple Mortality Risk Calculator that Predicts 60-Day Survival of Adult SARS-CoV-2 Positive Patients at Hospital Admission Using Rapid Routine Tests

Background: There is an urgent need to develop a simplified risk tool that enables rapid triaging of SARS CoV-2 positive patients during hospital admission, which complements current practice. Many predictive tools developed to date are complex, rely on multiple blood results and past medical history, do not include chest X ray results and rely on Artificial Intelligence rather than simplified algorithms. Our aim was to develop a simplified risk-tool based on five parameters and CXR image data that predicts the 60-day survival of adult SARS CoV-2 positive patients at hospital admission.Methods: We analysed the NCCID database of patient blood variables and CXR images from 19 hospitals across the UK contributed clinical data on SARS CoV-2 positive patients using multivariable logistic regression. The initial dataset was non-randomly split between development and internal validation dataset with 1434 and 310 SARS CoV-2 positive patients, respectively. External validation of final model conducted on 741 Accident and Emergency admissions with suspected SARS CoV-2 infection from a separate NHS Trust which was not part of the initial NCCID data set.Findings: The LUCAS mortality score included five strongest predictors (lymphocyte count, urea, CRP, age, sex), which are available at any point of care with rapid turnaround of results. Our simple multivariable logistic model showed high discrimination for fatal outcome with the AUC-ROC in development cohort 0.765 (95% confidence interval (CI): 0.738 - 0.790), in internal validation cohort 0.744 (CI: 0.673 - 0.808), and in external validation cohort 0.752 (CI: 0.713 - 0.787). The discriminatory power of LUCAS mortality score was increased slightly when including the CXR image data (for normal versus abnormal): internal validation AUC-ROC 0.770 (CI: 0.695 - 0.836) and external validation AUC-ROC 0.791 (CI: 0.746 - 0.833). The discriminatory power of LUCAS and LUCAS + CXR performed in the upper quartile of pre-existing risk stratification scores with the added advantage of using only 5 predictors.Interpretation: This simplified prognostic tool derived from objective parameters can be used to obtain valid predictions of mortality in patients within 60 days SARS CoV-2 RT-PCR results. This free-to-use simplified tool can be used to assist the triage of patients into low, moderate, high or very high risk of fatality and is available at https://mdscore.net/.Funding Statement: EPSRC Impact Acceleration account fund EP/R511705/1; AS is funded by a Wellcome Trust fellowship 205188/Z/16/Z; University of Brighton COVID-19 Research Urgency Fund.Declaration of Interests: All researchers declare no competing interests.Ethics Approval Statement: De-identified and pseudo-anonymised patient data were obtained from data sets which were approved by the ethics committee as part of the existing Cardiac MRI Database NHS REC IRAS Ref: 222349 and University of Brighton REC (8011).

Individual and community-level risk for COVID-19 mortality in the United States.

Reducing COVID-19 burden for populations will require equitable and effective risk-based allocations of scarce preventive resources, including vaccinations1. To aid in this effort, we developed a general population risk calculator for COVID-19 mortality based on various sociodemographic factors and pre-existing conditions for the US population, combining information from the UK-based OpenSAFELY study with mortality rates by age and ethnicity across US states. We tailored the tool to produce absolute risk estimates in future time frames by incorporating information on pandemic dynamics at the community level. We applied the model to data on risk factor distribution from a variety of sources to project risk for the general adult population across 477 US cities and for the Medicare population aged 65 years and older across 3,113 US counties, respectively. Validation analyses using 54,444 deaths from 7 June to 1 October 2020 show that the model is well calibrated for the US population. Projections show that the model can identify relatively small fractions of the population (for example 4.3%) that might experience a disproportionately large number of deaths (for example 48.7%), but there is wide variation in risk across communities. We provide a web-based risk calculator and interactive maps for viewing community-level risks.

In-Hospital Mortality and Prediction in an Urban U.S. Population With COVID-19.

Coronavirus disease 2019 (COVID-19) has touched every aspect of society, and as the pandemic continues around the globe, many of the clinical factors that influence the disease course remain unclear. A useful clinical decision-making tool is a risk stratification model to determine in-hospital mortality as defined in this study. The study was performed at Robert Wood Johnson University Hospital (RWJUH) in New Brunswick, New Jersey, USA. Data was extracted from our electronic medical records on 44 variables that included demographic, clinical, laboratory tests, treatments, and mortality information. We used the least absolute shrinkage and selection operator regression with corrected Akaike’s information criterion to identify a subset of variables that yielded the smallest estimated prediction error for the risk of in-hospital mortality. During the study period, 808 COVID-19 patients were admitted to RWJUH. The sample size was limited to patients with at least one confirmed in-house positive nasopharyngeal swab COVID-19 test. Pregnant patients or those who were transferred to our facility were excluded. Patients who were in observation and were discharged from the emergency room were also excluded. A total of 403 patients had complete values for all variables and were eligible for the study. We identified significant clinical, laboratory, and radiologic variables determining severe outcomes and mortality. An in-hospital mortality risk calculator was created after the identification of significant factors for the specific cohort, which were abnormal CT scan or chest X-ray, chronic kidney disease, age, white blood cell count, platelet count, alanine aminotransferase, and aspartate transaminase with a sensitivity, specificity, and negative predictive value of 82%, 72%, and 93%, respectively. While numerous reports from around the globe have helped outline the pandemic, demographic factors vary widely. This study is more applicable to an urban, highly diverse population in the United States.

The Emergency Surgery Score accurately predicts the need for postdischarge respiratory and renal support after emergent laparotomies: A prospective EAST multicenter study.

The Emergency Surgery Score (ESS) was recently validated as an accurate mortality risk calculator for emergency general surgery. We sought to prospectively evaluate whether ESS can predict the need for respiratory and/or renal support (RRS) at discharge after emergent laparotomies (EL). This is a post hoc analysis of a 19-center prospective observational study. Between April 2018 and June 2019, all adult patients undergoing EL were enrolled. Preoperative, intraoperative, and postoperative variables were systematically collected. In this analysis, patients were excluded if they died during the index hospitalization, were discharged to hospice, or transferred to other hospitals. A composite variable, the need for RRS, was defined as the need for one or more of the following at hospital discharge: tracheostomy, ventilator dependence, or dialysis. Emergency Surgery Score was calculated for all patients, and the correlation between ESS and RRS was examined using the c-statistics method. From a total of 1,649 patients, 1,347 were included. Median age was 60 years, 49.4% were men, and 70.9% were White. The most common diagnoses were hollow viscus organ perforation (28.1%) and small bowel obstruction (24.5%); 87 patients (6.5%) had a need for RRS (4.7% tracheostomy, 2.7% dialysis, and 1.3% ventilator dependence). Emergency Surgery Score predicted the need for RRS in a stepwise fashion; for example, 0.7%, 26.2%, and 85.7% of patients required RRS at an ESS of 2, 12, and 16, respectively. The c-statistics for the need for RRS, the need for tracheostomy, ventilator dependence, or dialysis at discharge were 0.84, 0.82, 0.79, and 0.88, respectively. Emergency Surgery Score accurately predicts the need for RRS at discharge in EL patients and could be used for preoperative patient counseling and for quality of care benchmarking. Prognostic and epidemiological, level III.

A comparison of the new, parsimonious tool Surgical Risk Preoperative Assessment System (SURPAS) to the American College of Surgeons (ACS) risk calculator in emergency surgery

BackgroundThe Surgical Risk Preoperative Assessment System accurately predicts postoperative complications in elective surgery using only 8 preoperative variables, but its performance in emergency surgery has not been evaluated. We hypothesized that the Surgical Risk Preoperative Assessment System accurately predicts postoperative complications in emergency surgery and compared its performance to that of the American College of Surgeons Surgical Risk Calculator. MethodsWe calculated the Surgical Risk Preoperative Assessment System and the American College of Surgeons Surgical Risk Calculator risk estimates in a random sample of 1,010 emergency surgery cases from the American College of Surgeons National Surgical Quality Improvement Program 2014 to 2017 database. Risk estimates were compared with known outcomes. Analyses included the Hosmer-Lemeshow goodness of fit graphs and P values, c-indexes, and Brier scores. ResultsResults were similar between the Surgical Risk Preoperative Assessment System and the American College of Surgeons Surgical Risk Calculator for mortality, urinary tract infection, cardiac, venous thromboembolism, and renal complications. The American College of Surgeons Surgical Risk Calculator underestimated morbidity compared with the Surgical Risk Preoperative Assessment System (16.04% American College of Surgeons Surgical Risk Calculator vs 24.88% Surgical Risk Preoperative Assessment System vs 24.3% observed). Both calculators overestimated readmission (7.67% American College of Surgeons Surgical Risk Calculator vs 5.18% Surgical Risk Preoperative Assessment System vs 4.1% observed). ConclusionBoth calculators predicted mortality, urinary tract infection, cardiac, venous thromboembolism, and renal complications well, but readmissions relatively poorly. The Surgical Risk Preoperative Assessment System estimated overall morbidity accurately, while the American College of Surgeons Surgical Risk Calculator underestimated this risk.

Prospective validation of the Emergency Surgery Score in emergency general surgery: An Eastern Association for the Surgery of Trauma multicenter study.

The Emergency Surgery Score (ESS) was recently developed and retrospectively validated as an accurate mortality risk calculator for emergency general surgery. We sought to prospectively validate ESS, specifically in the high-risk nontrauma emergency laparotomy (EL) patient. This is an Eastern Association for the Surgery of Trauma multicenter prospective observational study. Between April 2018 and June 2019, 19 centers enrolled all adults (aged >18 years) undergoing EL. Preoperative, intraoperative, and postoperative variables were prospectively and systematically collected. Emergency Surgery Score was calculated for each patient and validated using c-statistic methodology by correlating it with three postoperative outcomes: (1) 30-day mortality, (2) 30-day complications (e.g., respiratory/renal failure, infection), and (3) postoperative intensive care unit (ICU) admission. A total of 1,649 patients were included. The mean age was 60.5 years, 50.3% were female, and 71.4% were white. The mean ESS was 6, and the most common indication for EL was hollow viscus perforation. The 30-day mortality and complication rates were 14.8% and 53.3%; 57.0% of patients required ICU admission. Emergency Surgery Score gradually and accurately predicted 30-day mortality; 3.5%, 50.0%, and 85.7% of patients with ESS of 3, 12, and 17 died after surgery, respectively, with a c-statistic of 0.84. Similarly, ESS gradually and accurately predicted complications; 21.0%, 57.1%, and 88.9% of patients with ESS of 1, 6, and 13 developed postoperative complications, with a c-statistic of 0.74. Emergency Surgery Score also accurately predicted which patients required intensive care unit admission (c-statistic, 0.80). This is the first prospective multicenter study to validate ESS as an accurate predictor of outcome in the EL patient. Emergency Surgery Score can prove useful for (1) perioperative patient and family counseling, (2) triaging patients to the intensive care unit, and (3) benchmarking the quality of emergency general surgery care. Prognostic study, level III.

Validation and improvement of a highly predictive bariatric surgery mortality risk calculator to include sleeve gastrectomy using MBSAQIP 2015-2017 data.

Initial development of a prominent bariatric surgery mortality risk calculator comprising cases that now account for <10% of commonly performed operations. Whether the previously highly predictive model is valid with more recent data is unknown. To validate and improve a bariatric-surgery-specific mortality calculator with updated case mix and outcomes data. Metabolic and Bariatric Surgery Accreditation and Quality Improvement Program-accredited bariatric surgery programs. The Metabolic and Bariatric Accreditation and Quality Improvement Program Participant Use File from years 2015 to 2017 was used for the analysis. C-statistics were calculated with observed death as the outcome and estimated 30-day mortality risk as the only predictor and receiver operating characteristic curve was plotted. Similar analyses were repeated for each body mass index (BMI) subgroup. Backward selection logistic regression was used to investigate the potential of improving the robustness of the model. Patients were predominantly female (n = 446,149, 80.4%) and white (n = 409,350, 73.7%) with a mean (standard deviation) age of 45.4 (12.0) years and BMI of 44.5 (8.4) kg/m2, and the most commonly performed operation was sleeve gastrectomy (n = 338,061, 60.9%). Assessing previous model using present data, area under the curve was .7412. By BMI subgroup, area under the curve for BMI <45 kg/m2 was .7645, for BMI 45 to 60 kg/m2 was .7586, and for BMI >60 kg/m2 was .6576. The present study found that the model previously developed maintains discrimination with changing surgical procedures. Though variables in the initial calculator are helpful, additional factors should be considered when weighing risk, such as sex, previous surgery, and renal function. Future studies are needed to determine whether changes in modifiable risk factors will impact mortality rates.

Can the Emergency Surgery Score (ESS) predict outcomes in emergency general surgery patients with missing data elements? A nationwide analysis

BackgroundThe Emergency Surgery Score (ESS) is an accurate mortality risk calculator for emergency general surgery (EGS). We sought to assess whether ESS can accurately predict 30-day morbidity, mortality, and requirement for postoperative Intensive Care Unit (ICU) care in patients with missing data variables. MethodsAll EGS patients with one or more missing ESS variables in the 2007–2015 ACS-NSQIP database were included. ESS was calculated assuming that a missing variable is normal (i.e. no additional ESS points). The correlation between ESS and morbidity, mortality, and postoperative ICU level of care was assessed using the c-statistics methodology. ResultsOut of a total of 4,456,809 patients, 359,849 were EGS, and of those 256,278 (71.2%) patients had at least one ESS variable missing. ESS correlated extremely well with mortality (c-statistic = 0.94) and postoperative requirement of ICU care (c-statistic = 0.91) and well with morbidity (c-statistic = 0.77). ConclusionESS performs well in predicting outcomes in EGS patients even when one or more data elements are missing and remains a useful bedside tool for counseling EGS patients and for benchmarking the quality of EGS care.

COVID-19 Mortality Risk Assessment: An International Multi-Center Study

Background: Timely identification of COVID-19 patients at high risk of mortality can significantly improve patient management and resource allocation within hospitals. This study seeks to develop and validate a data-driven personalized mortality risk calculator for hospitalized COVID-19 patients. Methods: De-identified data was obtained for 3,927 COVID-19 positive patients from six independent centers, comprising 33 different hospitals. Demographic, clinical, and laboratory variables were collected at hospital admission. The COVID-19 Mortality Risk (CMR) tool was developed using the XGBoost algorithm to predict mortality. Its discrimination performance was subsequently evaluated on three validation cohorts. Findings: The derivation cohort of 3,062 patients has an observed mortality rate of 26.84%. Increased age, decreased oxygen saturation (≤ 93%), elevated levels of C-reactive protein (≥ 130 mg/L), blood urea nitrogen (≥ 18 mg/dL), and blood creatinine (≥ 1.2 mg/dL) were identified as primary risk factors, validating clinical findings. The model obtains out-of-sample AUCs of 0.90 (95% CI, 0.87-0.94) on the derivation cohort. In the validation cohorts, the model obtains AUCs of 0.92 (95% CI, 0.88-0.95) on Seville patients, 0.87 (95% CI, 0.84-0.91) on Hellenic COVID-19 Study Group patients, and 0.81 (95% CI, 0.76-0.85) on Hartford Hospital patients. The CMR tool is available as an online application at covidanalytics.io/mortality_calculator and is currently in clinical use. Interpretation: The CMR model leverages machine learning to generate accurate mortality predictions using commonly available clinical features. This is the first risk score trained and validated on a cohort of COVID-19 patients from Europe and the United States. Funding: No funding was provided.Declaration of Interests: All authors have no competing interests to declare that may be relevant to the submitted work.Ethics Approval Statement: All independent organizations and the Massachusetts Institute of Technology institutional review boards approved this protocol as minimal-risk research using data collected for standard clinical practice and waived the requirement for informed consent. The survey was anonymous and confidentiality of information was assured.

Development and validation of a real-time mortality risk calculator before, during and after hepatectomy: an analysis of the ACS NSQIP database

BackgroundAlthough most conventional risk prediction models have been based on preoperative information, intra- and post-operative events may be more relevant to mortality after surgery. We sought to develop a mortality risk calculator based on real time characteristics associated with hepatectomy. MethodsPatients who underwent hepatectomy between 2014 and 2017 were identified in the ACS-NSQIP dataset. Three prediction models (pre-, intra-, post-operative) were developed and validated using perioperative data. ResultsAmong 14,720 patients, 197 (1.3%) experienced 30-day mortality. The predictive ability of the real-time mortality risk calculator was very good based on only preoperative factors (AUC; training cohort: 0.813, validation cohort: 0.731). Incorporating intra-operative variables into the model increased the AUC (training: 0.838, validation: 0.777), while the post-operative model achieved an AUC of 0.922 in the training and 0.885 in the validation cohorts, respectively. While patients with low preoperative risk had only very small fluctuations in the estimated 30-day mortality risk during the intraoperative (Δ0.4%) and postoperative (Δ0.6%) phases, patients who were already deemed high risk preoperatively had additional increased mortality risk based on factors that occurred in the intraoperative (Δ5.4%) and postoperative (Δ9.3%) periods. ConclusionA real-time mortality risk calculator may better help clinicians identify patients at risk of death at the different stages of the surgical episode.

The competitive endogenous RNA regulatory network reveals potential prognostic biomarkers for overall survival in hepatocellular carcinoma.

The aim of the present study is to construct a competitive endogenous RNA (ceRNA) regulatory network by using differentially expressed long noncoding RNAs (lncRNAs), microRNAs (miRNAs), and mRNAs in patients with hepatocellular carcinoma (HCC), and to construct a prognostic model for predicting overall survival (OS) of HCC patients. Differentially expressed lncRNAs, miRNAs, and mRNAs were explored between HCC tissues and normal liver tissues. A prognostic model was built for predicting OS of HCC patients and receiver operating characteristic curves were used to evaluate the performance of the prognostic model. There were 455 differentially expressed lncRNAs, 181 differentially expressed miRNAs, and 5035 differentially expressed mRNAs. A ceRNA regulatory network was constructed based on 43 lncRNAs, 37 miRNAs, and 105 mRNAs. Eight mRNA biomarkers (H2AFX, SQSTM1, ITM2A, PFKP, TPD52L1, ACSL4, STRN3, and CPEB3) were identified as independent risk factors by multivariate Cox regression and were used to develop a prognostic model for OS. The C‐indexes in the model group were 0.776 (95% confidence interval [CI], 0.730‐0.822), 0.745 (95% CI, 0.699‐0.791), and 0.789 (95% CI, 0.743‐0.835) for 1‐, 3‐, and 5‐year OS, respectively. The current study revealed potential molecular biological regulation pathways and prognostic biomarkers by the ceRNA regulatory network. A prognostic model based on prognostic mRNAs in the ceRNA network might be helpful to predict the individual mortality risk for HCC patients. The individual mortality risk calculator can be used by visiting the following URL: https://zhangzhiqiao.shinyapps.io/Smart_cancer_predictive_system_HCC/.

National risk prediction model for perioperative mortality in non-cardiac surgery.

Many multivariable models to calculate mortality risk after surgery are limited by insufficient sample size at development or by application to cohorts distinct from derivation populations. The aims of this study were to validate the Surgical Outcome Risk Tool (SORT) for a New Zealand population and to develop an extended NZRISK model to calculate 1-month, 1-year and 2-year mortality after non-cardiac surgery. Data from the New Zealand National Minimum Data Set for patients having surgery between January 2013 and December 2014 were used to validate SORT. A random 75 per cent split of the data was used to develop the NZRISK model, which was validated in the other 25 per cent of the data set. External validation of SORT in the 360 140 patients who underwent surgery in the study period showed good discrimination (area under the receiver operating characteristic curve (AUROC) value of 0·906) but poor calibration (McFadden's pseudo-R2 0·137, calibration slope 5·32), indicating it was invalid in this national surgical population. Internal validation of the NZRISK model, which incorporates sex and ethnicity in addition to the variables used in SORT for 1-month, 1-year and 2-year outcomes, demonstrated excellent discrimination with AUROC values of 0·921, 0·904 and 0·895 respectively, and excellent calibration (McFadden's pseudo-R2 0·275, 0·308 and 0·312 respectively). Calibration slopes were 1·12, 1·02 and 1·02 respectively. The SORT performed poorly in this national population. However, inclusion of sex and ethnicity in the NZRISK model improved performance. Calculation of mortality risk beyond 30 days after surgery adds to the utility of this tool for shared decision-making.

Development and validation of a clinical risk calculator for mortality after colectomy for fulminant Clostridium difficile colitis.

Clostridium difficile colitis is an increasingly important cause of morbidity and mortality. Fulminant C. difficile colitis (FCDC) is a severe form of the colitis driven by a significant systemic inflammatory response, and managed with a total abdominal colectomy. Despite surgery, postoperative mortality rates remain high. The aim of this study was to develop a bedside calculator to predict the risk of 30-day postoperative mortality for patients with FCDC. After institutional review board approval, the American College of Surgeons National Surgical Quality Improvement Program database (2005-2015) was used to include adult patients who underwent emergency surgery for FCDC. A priori preoperative predictors of mortality were selected from the literature: age, immunosuppression, preoperative shock, intubation, and laboratory values. The predictive accuracy of different logistic regression models was measured by calculating the area under the receiver-operating characteristic curve. A cohort of 124 patients from Québec was used to validate the developed mortality calculator. A total of 557 patients met the inclusion criteria, and the overall mortality was 44%. After developing the calculator, no statistically significant differences were found in comparison with the American College of Surgeons National Surgical Quality Improvement Program probability of mortality available in the database (area under the receiver operating curve, 75.61 vs. 75.14; p = 0.79). External validation with the cohort of patients from Quebec showed an area under the curve of 74.0% (95% confidence interval, 65.0-82.9). A clinically applicable calculator using preoperative variables to predict postoperative mortality for patients with FCDC was developed and externally validated. This calculator may help guide preoperative decision making. Prognostic and epidemiological study, level III.

Life after 90: Predictors of mortality and performance of the ACS-NSQIP risk calculator in 4,724 nonagenarian patients undergoing emergency general surgery

The decision to emergently operate on nonagenarian patients (NONAs) can be complex due to the uncertainty about outcomes and goals of care at this advanced age. We sought to study: (1) the outcomes and predictors of mortality for NONAs undergoing emergency general surgery (EGS) and (2) the accuracy of ACS-NSQIP mortality risk calculator in this special population. Using the 2007 to 2015 ACS-NSQIP database, we included all patients older than 90 years of age who underwent an emergent operation with a Current Procedural Terminology (CPT) code for "digestive system." Multivariable logistic regression analyses were performed to identify independent predictors of 30-day mortality. NONAs' mortality rates for different combinations of risk factors were also studied and compared to the ACS-NSQIP calculator-predicted mortality rates. Out of a total of 4,456,809 patients, 4,724 NONAs were included. The overall 30-day patient mortality and morbidity rates were 21% and 45%, respectively. In multivariable analyses, several independent predictors of 30-day mortality were identified, including recent history of weight loss, history of steroid use, smoking, functional dependence, hypoalbuminemia and sepsis or septic shock. The mortality among NONAs with a history of steroid use and a recent history of weight loss was 100%. Similarly, the mortality of NONAs with recent history of weight loss who presented with preoperative septic shock was 93%. The ACS-NSQIP calculator significantly and consistently underestimated the risk of mortality in all NONAs undergoing EGS. Most NONAs undergoing EGS survive the hospital stay and the first 30 postoperative days, even in the presence of significant preexisting comorbidities. However, the combination of recent weight loss with either steroid use or septic shock nearly ensures mortality and should be used in the discussions with patients and families before a decision to operate is made. The ACS-NSQIP surgical risk calculator should be used with caution in these high-risk patients. Prognostic study, level III.

Abstract P079: Prediction of All-Cause Mortality in Diabetic Patients

Introduction: Diabetics people have an increased risk of developing health problems, including macro-vascular complications, and have more than double risk of death for any causes. Early identification of high-risk diabetic patients is therefore important so that appropriate interventions can be provided to help reduce the risk of premature mortality. Hypothesis: The purpose of this study was to investigate factors associated with all-cause mortality in diabetic patients and create a mortality risk calculator to aid in treatment decisions. Methods: Longitudinal analysis on 2,286 patients with type 2 diabetes individuated in the general population of the Moli-sani study, a population-based cohort study that recruited 24,325 men and women aged≥35 yr of Southern Italy, 2005-2010. Diabetes was defined as use of antidiabetic treatment or fasting blood glucose ≥126 mg/dl at baseline. The cohort was followed up until December 31 st 2015. Overall mortality was assessed by the Italian mortality registry (ReNCaM registry) and validated by Italian death certificates. Association of risk factors with mortality was assessed by multivariable Cox-regression survival analysis. A global score for mortality was obtained summing up scores (σ) obtained as rescaled beta coefficients of the Cox regression for each conditions. Results: Over a median follow-up of 7.8 years (17,652 person-years), 318 all-cause deaths were ascertained. Starting from a large panel of potential risk factors, we settled a final multivariable model including conditions independently associated with all-cause mortality, specifically: age (σ=0, 2 or 7 for age&lt;55 yr, 55-70 yr or ≥70 yr, respectively), sex (σ=0 or 2 for women or men), diabetic drugs in use (σ=0, 1 or 2 for none, yes but not insulin, insulin, respectively), history of cardiovascular disease (σ=0 or 2 for no or yes), cigarette smoking (σ=0, 2 or 3 for no, former or current), C-reactive protein (CRP) (σ=0, 1 or 3 for CRP≤1 mg/L, 1-2.99 mg/L or 3-10 mg/L), estimated glomerular filtration rate (eGFR, ml/min/1.73m 2 ) (σ=0, 1 or 3 for eGFR≥90, 60-90 or &lt;60) and adhesion to Mediterranean Diet (σ=0, -1 or -2 for low (0-3 points), intermediate or high adhesion (≥6 points)). C-statistic for the final model was 0.74. Diabetic patients stratified according to global score clearly were discriminated as mortality risk; specifically hazard ratio and predicted risk of death at 10 years were 1.0 (referent) and 4.5%, 2.80 (95%CI: 2.00 to 3.93) and 12.1% and 8.47 (6.14 to 11.68) and 32.2% for values of the global score 0 to 5 (prevalence 15.7%), 8 to10 (19.9%) and 14 to 24 (16.8%), respectively (P for difference in Kaplan-Meier curves &lt;0.0001). Conclusions: A global score of risk for all-cause mortality in diabetic patients has been developed, easy to use in clinical practice which include classical CVD risk factors and indicators of nephropathy, low grade inflammation and adherence to Mediterranean Diet.

Performance of CT-based low skeletal muscle index, low mean muscle attenuation, and bioelectric impedance derived low phase angle in the detection of an increased risk of nutrition related mortality

The aim of this study was to evaluate in a head to head comparison the performance of bioimpedance derived low phase angle (PA), CT-based low skeletal muscle index (SMI) and low mean muscle attenuation (MA), and Pandora Score (PS) to detect an increased nutrition related mortality in hospitalized patients. A total of 7736 patients were hospitalized in Dessau community hospital (Nov 11, 2016-Feb 05, 2017). In 227/7736 patients an abdominal CT scan was obtained and low values (below threshold) of skeletal muscle index (SMI) and mean muscle attenuation (MA) were obtained at L3 level using Slice-O-Matic. In 180/227 patients, malnutrition screening (NRS-2002≥3), determination of low PA (<5th percentile) and calculation of 30-day mortality risk (PS≥5%) were done within 6 days after the CT scan. PA was low in 116 (64%), SMI was low in 95 (53%), and MA was below threshold in 137 (77%) patients. Using hospital mortality (8.3%) as a gold standard, low SMI, low MA, or low PA showed high sensitivity (SMI 93.3%; MA 86.7%; PA 86.7%) and high negative predictive (NPV) values (SMI 98.9%; MA 95.3%; PA 96.9%) and improved specificity (NRS+SMI 65.5%; NRS+MA 50.3%; NRS+PA 50.3%) when combined with the criterion NRS≥3. Using hospital mortality as a gold standard, a PS≥36 (predicted mortality≥5%) showed a sensitivity of 73.3%, a specificity of 62.2% and an NPV of 96.4%. PS predicted mortality was ≥5% in 70 (39%) patients and was higher (median 5.2%; IQR: 2.8; 11.8 vs 1.7%; IQR: 1.0; 2.8; p<0.001) in patients with a low PA and this was also observed for low vs normal SMI (median 5.7%; IQR: 2.6; 11.1 vs 2.7%; IQR:1.4;5.5; p<0.001) but not for low vs normal MA (median 3.9%; IQR: 2.1; 8.8 vs 3.4%; IQR: 2.0; 8.1; ns). Determination of phase angle<5th percentile is a convenient non-invasive and inexpensive bedside method with high sensitivity and a high negative predictive value to detect patients at risk of nutrition related mortality with a performance comparable to CT derived skeletal muscle index or mean muscle attenuation.