Prediction Of Sepsis Research Articles

Machine Learning for Early Prediction of Sepsis in Intensive Care Unit (ICU) Patients.

Background and Objectives: Early detection of sepsis is crucial and can save lives. However, identifying sepsis early and accurately remains a difficult task in the medical field. This study aims to investigate a new machine-learning approach. By analyzing the clinical laboratory results and vital signs of adult patients in the ICU, this approach can predict and detect the initial signs of sepsis. Materials and Methods: To examine survival rates and predict outcomes, the study utilized several models, including the proportional hazards model and data mining algorithms. We analyzed data from the BESTCare database at KAMC, with a focus on patients aged 14 and older who were admitted to the ICU between April and October 2018. We conducted a thorough analysis of the medical records of a total of 1182 patients who were diagnosed with sepsis. Results: We studied two approaches to predict sepsis in ICU patients. The regression model utilizing survival analysis showed moderate predictive ability, emphasizing the importance of only three factors-time (from sepsis to an outcome; discharge or death), lactic acid, and temperature-had a significant p-value (p = 0.000568, p = 0.01, p = 0.02, respectively). Other data mining algorithms may have limitations due to their assumptions of variable independence and linear classification nature. Conclusions: To achieve progress and accuracy in the field of sepsis prediction, it is important to continuously strive for improvement. By meticulously cleaning and selecting data attributes, we can create a strong foundation for future advancements in this area.

Development and validation of a nomogram for predicting sepsis in patients with pyogenic liver abscess

Pyogenic liver abscess (PLA) is a severe condition that significantly increases the risk of sepsis. However, there is a notable dearth of research regarding the prediction of sepsis in PLA patients. The objective of this study was to develop and validate a prognostic nomogram for predicting sepsis in PLA patients. A total of 206 PLA patients were enrolled in our study, out of which 60 individuals (29.1%) met the Sepsis-3 criteria. Independent risk factors for sepsis were identified through univariate and multivariate logistic regression analyses. Subsequently, a nomogram was developed based on age, positive blood culture, procalcitonin, alanine aminotransferase, blood urea nitrogen, and d-dimer. The nomogram demonstrated excellent calibration and discrimination, as evidenced by the area under the receiver operating characteristic curve (AUC) values of 0.946 (95% confidence interval [CI], 0.912–0.979) and 0.980 (95%CI 0.951–1.000) in the derivation and validation cohorts, respectively. Furthermore, decision-curve analysis confirmed the clinical utility of the nomogram. This study provides valuable insights for the prevention of sepsis in PLA patients and underscores the potential application of the prognostic nomogram in clinical practice.

PoEMS: Policy Network-Based Early Warning Monitoring System for Sepsis in Intensive Care Units.

Sepsis is among the leading causes of morbidity and mortality in modern intensive care units (ICU). Due to accurate and early warning, the in-time antibiotic treatment of sepsis is critical for improving sepsis outcomes, contributing to saving lives, and reducing medical costs. However, the earlier prediction of sepsis onset is made, the fewer monitoring measurements can be processed, causing a lower prediction accuracy. In contrast, a more accurate prediction can be expected by analyzing more data but leading to the delayed warning associated with life-threatening events. In this study, we propose a novel deep reinforcement learning framework for solving early prediction of sepsis, called the Policy Network-based Early Warning Monitoring System (PoEMS). The proposed PoEMS provides accurate and early prediction results for sepsis onset based on analyzing varied-length electronic medical records (EMR). Furthermore, the system serves by monitoring the patient's health status consistently and provides an early warning only when a high risk of sepsis is detected. Additionally, a controlling parameter is designed for users to adjust the trade-off between earliness and accuracy, providing the adaptability of the model to meet various medical requirements in practical scenarios. Through a series of experiments on real-world medical data, the results demonstrate that our proposed PoEMS achieves a high AUROC result of more than 91% for early prediction, and predicts sepsis onset earlier and more accurately compared to other state-of-the-art competing methods.

Predicting the Onset of Sepsis Using Vital Signs Data: A Machine Learning Approach.

Sepsis is a major cause of mortality among hospitalized patients. Existing sepsis prediction methods face limitations due to their reliance on laboratory results and Electronic Medical Records (EMRs). This work aimed to develop a sepsis prediction model utilizing continuous vital signs monitoring, offering an innovative approach to sepsis prediction. Data from 48,886 Intensive Care Unit (ICU) patient stays were extracted from the Medical Information Mart for Intensive Care -IV dataset. A machine learning model was developed to predict sepsis onset based solely on vital signs. The model's efficacy was compared with the existing scoring systems of SIRS, qSOFA, and a Logistic Regression model. The machine learning model demonstrated superior performance at 6 hrs prior to sepsis onset, achieving 88.1% sensitivity and 81.3% specificity, surpassing existing scoring systems. This novel approach offers clinicians a timely assessment of patients' likelihood of developing sepsis.

Development and clinical impact assessment of a machine-learning model for early prediction of late-onset sepsis

Background and aimPreterm infants are prone to neonatal infections such as late-onset sepsis (LOS). The consequences of LOS can be severe and potentially life-threatening. Unfortunately, LOS often presents with unspecific symptoms, and early screening laboratory tests have limited diagnostic value and are often late. This study aimed to build a predictive algorithm to aid doctors in the early detection of LOS in very preterm infants. MethodsIn a retrospective cohort study, all consecutively admitted preterm infants (GA ≤ 32 weeks) from 2008 until 2019 were included. They were classified as LOS or control according to blood culture results, currently the gold standard. To generate features, routine and continuously measured oxygen saturation and heart rate data with a minute-by-minute sampling rate were extracted from electronic medical records. Care was taken not to include variables indicative of existing LOS suspicion. The timing of a positive blood culture served as a proxy for LOS-onset. An equivalent timestamp was generated in gestational-age-matched control patients without a positive blood culture. Three machine learning (ML) techniques (generalized additive models, logistic regression, and XGBoost) were used to build a classification algorithm. To simulate the performance of the algorithm in clinical practice, a simulation using multiple alarm thresholds was performed on hourly predictions for the total hospitalization period. Results292 infants with LOS were matched to 1497 controls. The median gestational age before matching was 28.1 and 30.3 weeks, respectively. Evaluation of the overall discriminative power of the LR algorithm yielded an AUC of 0.73 (p < 0.05) at the moment of clinical suspicion (t = 0). In the longitudinal simulation, our algorithm detects LOS in at least 47% of the patients before clinical suspicion without exceeding the alarm fatigue threshold of 3 alarms per day. Furthermore, medical experts evaluated the algorithm as clinically relevant regarding the feature contributions in the model explanations. ConclusionsAn ML algorithm was trained for the early detection of LOS. Performance was evaluated on both prediction horizons and in a clinical impact simulation. To the best of our knowledge, our assessment of clinical impact with a retrospective simulation on longitudinal data is the most extensive in the literature on LOS prediction to date. The clinically relevant algorithm, based on routinely collected data, can potentially accelerate clinical decisions in the early detection of LOS, even with limited inputs.

Incidence, outcomes and costs of severe sepsis and septic shock in geriatric trauma patients: Analysis of 2,563,463 hospitalizations at 3,284 hospitals.

Severe sepsis/septic shock (sepsis) is a leading cause of death in hospitalized trauma patients. Geriatric trauma patients are an increasing proportion of trauma care but little recent, large-scale, research exists in this high-risk demographic. The objectives of this study are to identify incidence, outcomes and costs of sepsis in geriatric trauma patients. Patients at short-term, nonfederal hospitals 65 years or older with ≥1 injury International Classification of Diseases, Tenth Revision, Clinical Modification code were selected from 2016 to 2019 Centers for Medicare & Medicaid Services Medicare Inpatient Standard Analytical Files. Sepsis was defined as International Classification of Diseases, Tenth Revision, Clinical Modification diagnosis codes R6520 and R6521. A log-linear model was used to examine the association of Sepsis with mortality, adjusting for age, sex, race, Elixhauser score, and Injury Severity Score. Dominance analysis using logistic regression was used to determine the relative importance of individual variables in predicting Sepsis. Institutional review board exemption was granted for this study. There were 2,563,436 hospitalizations from 3,284 hospitals (62.8% female; 90.4% White; 72.7% falls; median ISS, 6.0). Incidence of Sepsis was 2.1%. Sepsis patients had significantly worse outcomes. Mortality risk was significantly higher in septic patients (adjusted risk ratio, 3.98, 95% confidence interval, 3.92-4.04). Elixhauser score contributed the most to the prediction of Sepsis, followed by ISS (McFadden's R2 = 9.7% and 5.8%, respectively). Severe sepsis/septic shock occurs infrequently among geriatric trauma patients but is associated with increased mortality and resource utilization. Pre-existing comorbidities influence Sepsis occurrence more than Injury Severity Score or age in this group, identifying a population at high risk. Clinical management of geriatric trauma patients should focus on rapid identification and prompt aggressive action in high-risk patients to minimize the occurrence of sepsis and maximize survival. Therapeutic/Care Management; Level III.

Sepsis Prediction by Using a Hybrid Metaheuristic Algorithm: A Novel Approach for Optimizing Deep Neural Networks.

The early diagnosis of sepsis reduces the risk of the patient's death. Gradient-based algorithms are applied to the neural network models used in the estimation of sepsis in the literature. However, these algorithms become stuck at the local minimum in solution space. In recent years, swarm intelligence and an evolutionary approach have shown proper results. In this study, a novel hybrid metaheuristic algorithm was proposed for optimization with regard to the weights of the deep neural network and applied for the early diagnosis of sepsis. The proposed algorithm aims to reach the global minimum with a local search strategy capable of exploring and exploiting particles in Particle Swarm Optimization (PSO) and using the mental search operator of the Human Mental Search algorithm (HMS). The benchmark functions utilized to compare the performance of HMS, PSO, and HMS-PSO revealed that the proposed approach is more reliable, durable, and adjustable than other applied algorithms. HMS-PSO is integrated with a deep neural network (HMS-PSO-DNN). The study focused on predicting sepsis with HMS-PSO-DNN, utilizing a dataset of 640 patients aged 18 to 60. The HMS-PSO-DNN model gave a better mean squared error (MSE) result than other algorithms in terms of accuracy, robustness, and performance. We obtained the MSE value of 0.22 with 30 independent runs.

Value of Algorithm-Enabled Process Innovation: The Case of Sepsis

Problem definition: Algorithm-enabled decision support has an increasingly important role in supporting the day-to-day operations of healthcare organizations. Yet, fully realizing the value of algorithmic decision support lies critically in the opportunity to re-engineer the related processes and redefine roles in ways that make organizations more effective. We study how and when algorithm-enabled process innovation (AEPI) creates value in light of dynamic operational environments (i.e., workload) and behavioral responses to algorithmic predictions (i.e., algorithmic accuracy). Our context is an AEPI effort around a rule-based decision-support algorithm for early detection of sepsis—a costly condition that is the leading cause of death for hospitalized patients. We collaborated with a large U.S.-based hospital system and examined whether AEPI developed for sepsis care (sepsis AEPI) impacts patient mortality and when this impact is stronger or weaker. Methodology/results: We utilize a rich set of clinical and nonclinical data in empirically examining the impact of sepsis AEPI on patient mortality. We leverage the staggered implementation of sepsis AEPI across hospital units and conduct our estimation on a carefully matched sample. The matching utilizes data on patient vitals and the logic behind the algorithm to create a robust comparison group consisting of patient visits for which sepsis AEPI would have triggered an alert if it had been in place. Our empirical analysis shows that sepsis AEPI reduces the likelihood of death from sepsis (45% relative reduction in mortality risk due to sepsis). A higher-than-usual workload and an increase in the average number of inaccurate alert experience at a hospital unit (e.g., an oncology unit, which provides care for cancer patients), in general, reduces the effectiveness of AEPI. We also identify diminishing mortality benefits over prolonged periods of adoption; evaluation of the moderators over time helps explain this diminishing impact. Managerial implications: Our findings suggest that streamlining sepsis-care processes through a predictive algorithm (i.e., algorithm-based monitoring of real-time patient data and providing predictions, streamlined communication channels for coordinating care for a patient with sepsis prediction, and a more standardized process for sepsis diagnosis and treatment) can reduce the loss of life from sepsis. For the 3,739 sepsis patients in our study period, AEPI’s benefits would translate to 181 lives saved. We show that such value, however, is sensitive to operational and behavioral factors as the algorithm becomes a routine part of the day-to-day operations of the hospital. Funding: Financial support from University Hospitals is gratefully acknowledged. Supplemental Material: The online appendix is available at https://doi.org/10.1287/msom.2023.1226 .

A machine-learning approach for prediction of hospital mortality in cancer-related sepsis

ObjectiveTo develop a machine learning model to predict hospital mortality and identify risk factors in cancer-related sepsis patients. MethodWe obtained data from the Medical Information Mart for Intensive Care (MIMIC)-IV critical care data set, which included patients who diagnosed with cancer and fulfilled the definition of sepsis between 2008 and 2019. The data set was randomly split into a training set and a validation set. The dataset was imputed using the K-Nearest Neighbor (KNN) imputation model. An advanced machine learning model called CatBoost was established and then assessed by SHAP value. ResultsA total of 5081 patients were included in the final analysis. The cancer-related sepsis patients had a lower hospital survival (13.8% vs. 25.3%, P < 0.001) than non-cancer-related patients.For cancer-related sepsis patients, ensemble learning algorithms were superior to others with better accuracy and larger AUC, such as CatBoost (AUC: 0.828), LightGBM (AUC: 0.818), and Random Forest Classifier (AUC: 0.803). An evaluation of the performance suggested that the CatBoost model with the most powerful discrimination to predict hospital mortality, outperformed other models with a sensitivity of 76% and a specificity of 74%. The best cutoff was 0.223 for the CatBoost model. In addition, CatBoost also outperformed other severity scores such as SAPS-II (AUC: 0.725) and SOFA (AUC: 0.682). Urine output and the minimum BUN level on admission were the most important features for the hospital mortality prediction of cancer-related sepsis, while the patients’ age and the urine output on admission for non-cancer-related patients. ConclusionFor cancer-related sepsis patients, CatBoost model was a better prediction model. It was easy for clinicians to access by use of common clinical vital signs or laboratory examination parameters, which provides convenience for them to evaluate patient’s condition and make follow-up treatments.

FRI-539 - Prediction of sepsis in patient of acute on chronic liver failure (ACLF)-a machine learning approach

FRI-539 - Prediction of sepsis in patient of acute on chronic liver failure (ACLF)-a machine learning approach

Morphological changes in white blood cells in systemic inflammatory response syndrome (SIRS) with and without sepsis: An observational study.

This is an observational study which aims to research morphological changes of white blood cells in patients with Systemic Inflammatory Response Syndrome (SIRS) with and without sepsis and evaluate morphological changes in white blood cells as predictors of sepsis. Patients aged 18 years or more with SIRS with sepsis and SIRS without sepsis were included and those with haematological disorders or pregnant patients were excluded. A total of 52 patients with SIRS with sepsis and 32 patients of SIRS without sepsis were included. Peripheral blood smear was prepared from the venous blood sample drawn. The presence of toxic granules, cytoplasmic vacuoles, and Dohle bodies in both cases of SIRS with sepsis without sepsis were assessed and it was compared with culture-positive sepsis and shock. The difference in the presence of toxic granules (55.8% vs. 12.5%; p <0.001), cytoplasmic vacuoles (30.8% vs. 6.3%; p -0.012), and Dohle bodies (17.3% vs. 0%; p = 0.012) was significantly higher in the SIRS with sepsis group, compared to the SIRS without sepsis group. In the subgroup analysis of patients in the sepsis group, it was observed that patients with positive blood culture (9%) had a significantly higher proportion of toxic granules (100% vs. 51.1%; p=0.059), cytoplasmic vacuoles (40% vs. 29.8%; p=0.637) and Dohle bodies (40% vs. 14.9%; p=0.202). However, these differences were not statistically significant. Toxic granules and cytoplasmic vacuoles in the neutrophils of patients with SIRS with sepsis were found more frequently, compared to patients of SIRS without sepsis. Dohle bodies were found only in patients with sepsis and not in those with SIRS without sepsis.

Bioinformatics Analysis of Gene Expression Profiles for Diagnosing Sepsis and Risk Prediction in Patients with Sepsis.

: Although early recognition of sepsis is essential for timely treatment and can improve sepsis outcomes, no marker has demonstrated sufficient discriminatory power to diagnose sepsis. This study aimed to compare gene expression profiles between patients with sepsis and healthy volunteers to determine the accuracy of these profiles in diagnosing sepsis and to predict sepsis outcomes by combining bioinformatics data with molecular experiments and clinical information. We identified 422 differentially expressed genes (DEGs) between the sepsis and control groups, of which 93 immune-related DEGs were considered for further studies due to immune-related pathways being the most highly enriched. Key genes upregulated during sepsis, including S100A8, S100A9, and CR1, are responsible for cell cycle regulation and immune responses. Key downregulated genes, including CD79A, HLA-DQB2, PLD4, and CCR7, are responsible for immune responses. Furthermore, the key upregulated genes showed excellent to fair accuracy in diagnosing sepsis (area under the curve 0.747-0.931) and predicting in-hospital mortality (0.863-0.966) of patients with sepsis. In contrast, the key downregulated genes showed excellent accuracy in predicting mortality of patients with sepsis (0.918-0.961) but failed to effectively diagnosis sepsis.In conclusion, bioinformatics analysis identified key genes that may serve as biomarkers for diagnosing sepsis and predicting outcomes among patients with sepsis.

Thromboelastography in the Perioperative Period: A Literature Review.

Assessing coagulation status is essential for prompt intervention to reduce morbidity and mortality related to bleeding and thrombotic complications during the perioperative period. Traditional coagulation tests such as platelet count,activated partial thromboplastin time (aPTT), prothrombin time (PT), international normalized ratio (INR), and activated clotting time (ACT) provide only static evaluation. These tests are not designed for assessment of dynamically changing coagulation conditions during the perioperative time. However, viscoelastic coagulation testing such as thromboelastography (TEG) produces a rapid numerical and graphical representation that helps to detect and direct targeted hemostatic therapy. Searching the literature through PubMed, Medline, Ovid, CINAHL, and ClinicalTrials.gov we retrieved 210 studies, which represent the use of TEG in the perioperative period. The included studies were categorized under various settings such as trauma, obstetrics, orthopedics, intensive care unit (ICU), cardiovascular, transplant, and miscellaneous scenarios. TEG showed promising results in trauma surgeries in predicting mortality, hypercoagulability, and bleeding even when it was compared to conventional methods. TEG was also useful in monitoring anticoagulant therapy in orthopedic and obstetric surgeries; however, its role in predicting thrombotic events, hypercoagulability, or complications was questionable. In ICU patients, it showed promising results, especially in the prediction or improvement of sepsis, coagulopathy, thrombotic events, ICU duration, hospital stay, and ventilator duration. TEG parameters effectively predicted hypercoagulation in transplant surgeries. Regarding cardiovascular surgeries, they were effective in the prediction of the need for blood products, coagulopathy, thrombotic events, and monitoring anticoagulation therapy. More randomized clinical trials comparing TEG parameters with standardized tools are needed to produce robust results to standardize its use in different perioperative settings.

A Temperature- and Age-Adjusted Shock Index for Emergency Department Identification of Pediatric Sepsis

To empirically derive a novel temperature- and age-adjusted mean shock index (TAMSI) for early identification of sepsis and septic shock in children with suspected infection. We performed a retrospective cohort study of children aged 1 month to <18 years presenting to a single emergency department with suspected infection over a 10-year period. TAMSI was defined as (pulse rate - 10× [temperature - 37])/(mean arterial pressure). The primary outcome was sepsis, and the secondary outcome was septic shock. In the two-thirds training set, we determined TAMSI cutoffs for each age group using a minimum sensitivity of 85% and Youden Index. In the one-third validation data set, we calculated test characteristics for the TAMSI cutoffs and compared them with those for the Pediatric Advanced Life Support (PALS) tachycardia or systolic hypotension cutoffs. In the sepsis validation data set, the sensitivity-targeting TAMSI cutoff yielded a sensitivity of 83.5% (95% confidence interval [CI] 81.7% to 85.4%) and specificity of 42.8% (95% CI 42.4% to 43.3%) versus a sensitivity of 77.7% (95% CI 75.7% to 79.8%) and specificity of 60.0% (95% CI 59.5% to 60.4%) for PALS. For septic shock, the sensitivity-targeting TAMSI cutoff achieved a sensitivity of 81.3% (95% CI 75.2% to 87.4%) and a specificity of 83.5% (95% CI 83.2% to 83.8%) versus a sensitivity of 91.0% (95% CI 86.5% to 95.5%) and a specificity of 58.8% (95% CI 58.4% to 59.3%) for PALS. TAMSI yielded an increased positive likelihood ratio and similar negative likelihood ratio versus PALS. TAMSI achieved a similar negative likelihood ratio and improved positive likelihood ratio compared with PALS vital sign cutoffs for the prediction of septic shock, but it did not improve on PALS for sepsis prediction, among children with suspected infection.

Presepsin: gelsolin ratio, as a promising marker of sepsis-related organ dysfunction: a prospective observational study.

We aimed to facilitate the diagnosis and prognosis of sepsis-related organ dysfunction through analyzing presepsin (PSEP) and gelsolin (GSN) levels along with a novel marker, the presepsin:gelsolin (PSEP:GSN) ratio. Blood samples were collected from septic patients at the intensive care unit (ICU) at three time points (T1-3): T1: within 12 h after admission; T2: second day morning; T3: third day morning. Sampling points for non-septic ICU patients were T1 and T3. PSEP was measured by a chemiluminescence-based POCT method while GSN was determined by an automated immune turbidimetric assay. Data were compared with routine lab and clinical parameters. Patients were categorized by the Sepsis-3 definitions. PSEP:GSN ratio was evaluated in major sepsis-related organ dysfunctions including hemodynamic instability, respiratory insufficiency and acute kidney injury (AKI). In our single center prospective observational study, 126 patients were enrolled (23 control, 38 non-septic and 65 septic patients). In contrast to controls, significantly elevated (p < 0.001) admission PSEP:GSN ratios were found in non-septic and septic patients. Regarding 10-day mortality prediction, PSEP:GSN ratios were lower (p < 0.05) in survivors than in non-survivors during follow-up, while the prognostic performance of PSEP:GSN ratio was similar to widely used clinical scores (APACHE II, SAPS II, SOFA). PSEP:GSN ratios were also higher (p < 0.001) in patients with sepsis-related AKI than septic non-AKI patients during follow-up, especially in sepsis-related AKI patients needing renal replacement therapy. Furthermore, increasing PSEP:GSN ratios were in good agreement (p < 0.001) with the dosage and the duration of vasopressor requirement in septic patients. Moreover, PSEP:GSN ratios were markedly greater (p < 0.001) in patients with septic shock than in septic patients without shock. Compared to septic patients requiring oxygen supplementation, substantially elevated (p < 0.001) PSEP:GSN ratios were observed in septic patients with demand for mechanical ventilation, while higher PSEP:GSN ratios (p < 0.001) were also associated with extended periods of mechanical ventilation requirement in septic patients. PSEP:GSN ratio could be a useful complementary marker besides the routinely used SOFA score regarding the diagnosis and short term mortality prediction of sepsis. Furthermore, the significant increase of this biomarker may also indicate the need for prolonged vasopressor or mechanical ventilation requirement of septic patients. PSEP:GSN ratio could yield valuable information regarding the extent of inflammation and the simultaneous depletion of the patient's scavenger capacity during sepsis. NIH U.S. National Library of Medicine, ClinicalTrails.gov. Trial identifier: NCT05060679, (https://clinicaltrials.gov/ct2/show/NCT05060679) 23.03.2022, Retrospectively registered.

Weight a minute: The smaller and more immature, the more predictable the autonomic regulation?

To investigate the relation between autonomic regulation, measured using heart rate variability (HRV), body weight and degree of prematurity in infants. Further to assess utility to include body weight in a machine learning-based sepsis prediction algorithm. Longitudinal cohort study including 378 infants hospitalised in two neonatal intensive care units. Continuous vital sign data collection was performed prospectively from the time of NICU admission to discharge. Clinically relevant events were annotated retrospectively. HRV described using sample entropy of inter-beat intervals and assessed for its correlation with body weight measurements and age. Weight values were then added to a machine learning-based algorithm for neonatal sepsis detection. Sample entropy showed a positive correlation with increasing body weight and postconceptual age. Very low birth weight infants exhibited significantly lower HRV compared to infants with a birth weight >1500 g. This persisted when reaching similar weight and at the same postconceptual age. Adding body weight measures improved the algorithm's ability to predict sepsis in the overall population. We revealed a positive correlation of HRV with increasing body weight and maturation in infants. Restricted HRV, proven helpful in detecting acute events such as neonatal sepsis, might reflect prolonged impaired development of autonomic control.

An interpretable machine learning model for real-time sepsis prediction based on basic physiological indicators.

In view of the important role of risk prediction models in the clinical diagnosis and treatment of sepsis, and the limitations of existing models in terms of timeliness and interpretability, we intend to develop a real-time prediction model of sepsis with high timeliness and clinical interpretability. We used eight real-time basic physiological monitoring indicators of patients, including heart rate, respiratory rate, oxygen saturation, mean arterial pressure, systolic blood pressure, diastolic blood pressure, temperature and blood glucose, extracted three-hour dynamic feature sequences, and calculated 3 linear parameters (mean, standard deviation, and endpoint value), a 24-dimensional feature vector was constructed, and finally a real-time sepsis prediction model was constructed based on the Local Interpretable Model-Agnostic Explanation (LIME) interpretability method. The area under the receiver operating characteristic curve (AUROC), Accuracy and F1 scores of Extremely Randomized Trees we built were higher than those of other models, with AUROC above 0.76, showing the best performance. The Imbalance XGBoost has a high specificity (0.86) in predicting sepsis. The LIME local interpretable model we built can display a large amount of valid model prediction details for clinical workers' reference, including the prediction probability and the influence of each feature on the prediction result, thus effectively assisting the work of clinical workers and improving diagnostic efficiency. This model can provide real-time dynamic early warning of sepsis for critically ill patients under supervision and provide a reference for clinical decision support. At the same time, interpretive analysis of sepsis prediction models can improve the credibility of the models.

Improving Outcomes of Same-sitting Bilateral Flexible Ureteroscopy for Renal Stones in Real-world Practice—Lessons Learnt from Global Multicenter Experience of 1250 Patients

BackgroundBilateral kidney stones are commonly treated in staged procedures. ObjectiveTo evaluate outcomes after same-sitting bilateral retrograde intrarenal surgery (SSB-RIRS) for renal stones. Design, setting, and participantsData from adults who underwent bilateral RIRS in 21 centers were retrospectively reviewed (from January 2015 to June 2022). The inclusion criteria were unilateral/bilateral symptomatic bilateral stone(s) of any size/location in both kidneys and bilateral stones on follow-up with symptom/stone progression. Stone-free rate (SFR) was defined as absence of any fragment >3 mm at 3 mo. Outcome measurements and statistical analysisContinuous variables are presented as medians and 25–75th percentiles. A multivariable logistic regression analysis was performed to evaluate independent predictors of sepsis and bilateral SFR. Results and limitationsA total of 1250 patients were included. The median age was 48.0 (36–61) yr. Of the patients, 58.2% were prestented. The median stone diameter was 10 mm on both sides. Multiple stones were present in 45.3% and 47.9% of the left and right kidneys, respectively. Surgery was stopped in 6.8% of cases. The median surgical time was 75.0 (55–90) min. Complications were transient fever (10.7%), fever/infection needing prolonged stay (5.5%), sepsis (2%), and blood transfusion (1.3%). Bilateral and unilateral SFRs were 73.0% and 17.4%, respectively. Female (odds ratio [OR] 2.97, 95% confidence interval [CI] 1.18–7.49, p = 0.02), no antibiotic prophylaxis (OR 5.99, 95% CI 2.28–15.73, p < 0.001), kidney anomalies (OR 5.91, 95% CI 1.96–17.94, p < 0.001), surgical time ≥100 min (OR 2.86, 95% CI 1.12–7.31, p = 0.03) were factors associated with sepsis. Female (OR 1.88, 95% CI 1.35–2.62, p < 0.001), bilateral prestenting (OR 2.16, 95% CI 1.16–7.66, p = 0.04), and the use of high-power holmium:YAG laser (OR 1.63, 95% CI 1.14–2.34, p < 0.01) and thulium fiber laser (OR 2.50, 95% CI 1.32–4.74, p < 0.01) were predictors of bilateral SFR. Limitations were retrospective study and no cost analysis. ConclusionsSSB-RIRS is an effective treatment with an acceptable complication rate in selected patients with kidney stones. Patient summaryIn this large multicenter study, we looked at outcomes after same-sitting bilateral retrograde intrarenal surgery (SSB-RIRS) for renal stones in a large cohort. We found that SSB-RIRS was associated with acceptable morbidity and good stone clearance after a single session.

Prehospital qSOFA, mSOFA, and NEWS2 performance for sepsis prediction: A prospective, multi-center, cohort study.

Nowadays, there is no gold standard score for prehospital sepsis and sepsis-related mortality identification. The aim of the present study was to analyze the performance of qSOFA, NEWS2 and mSOFA as sepsis predictors in patients with infection-suspected in prehospital care. The second objective is to study the predictive ability of the aforementioned scores in septic-shock and in-hospital mortality. Prospective, ambulance-based, and multicenter cohort study, developed by the emergency medical services, among patients (n = 535) with suspected infection transferred by ambulance with high-priority to the emergency department (ED). The study enrolled 40 ambulances and 4 ED in Spain between 1 January 2020, and 30 September 2021. All the variables used in the scores, in addition to socio-demographic data, standard vital signs, prehospital analytical parameters (glucose, lactate, and creatinine) were collected. For the evaluation of the scores, the discriminative power, calibration curve and decision curve analysis (DCA) were used. The mSOFA outperformed the other two scores for mortality, presenting the following AUCs: 0.877 (95%CI 0.841-0.913), 0.761 (95%CI 0.706-0.816), 0.731 (95%CI 0.674-0.788), for mSOFA, NEWS, and qSOFA, respectively. No differences were found for sepsis nor septic shock, but mSOFA's AUCs was higher than the one of the other two scores. The calibration curve and DCA presented similar results. The use of mSOFA could provide and extra insight regarding the short-term mortality and sepsis diagnostic, backing its recommendation in the prehospital scenario.

New approaches to missing biomedical data recovery for machine learning

Missing data is a common problem for medical data sets, especially large ones. This issue is of major importance since it can influence the analysis and further use of the data, e.g., for machine learning purposes. There are various methods for recovering missing data.One such method is to remove observations with missing values, but this is not very usefulgiven the limited amount of data available. Another commonly used approach is the LastObservation Carried Forward (LOCF). But most such methods are not universal and may needadjustments to the data set at hand. This article describes the possibility of solving thisproblem in the case of multimodal time series of biomedical data coming from patients withsepsis. It describes and compares three approaches tailored to a sepsis dataset, which isanalyzed and finally used to build a sepsis prediction system based on clinical data routinelyrecorded in an intensive care unit.