Survival prediction of patients with sepsis from age, sex, and septic episode number alone

Identifying prognostic factors for survival in intensive care unit patients with SIRS or sepsis by machine learning analysis on electronic health records

BACKGROUND

Systemic inflammatory response syndrome (SIRS) and sepsis are the most common causes of in-hospital death. However, the characteristics associated with the improvement in the patient conditions during the ICU stay were not fully elucidated for each population as well as the possible differences between the two.

GOAL

The aim of this study is to highlight the differences between the prognostic clinical features for the survival of patients diagnosed with SIRS and those of patients diagnosed with sepsis by using a multi-variable predictive modeling approach with a reduced set of easily available measurements collected at the admission to the intensive care unit (ICU).

METHODS

Data were collected from 1,257 patients (816 non-sepsis SIRS and 441 sepsis) admitted to the ICU. We compared the performance of five machine learning models in predicting patient survival. Matthews correlation coefficient (MCC) was used to evaluate model performances and feature importance, and by applying Monte Carlo stratified Cross-Validation.

RESULTS

Extreme Gradient Boosting (MCC = 0.489) and Logistic Regression (MCC = 0.533) achieved the highest results for SIRS and sepsis cohorts, respectively. In order of importance, APACHE II, mean platelet volume (MPV), eosinophil counts (EoC), and C-reactive protein (CRP) showed higher importance for predicting sepsis patient survival, whereas, SOFA, APACHE II, platelet counts (PLTC), and CRP obtained higher importance in the SIRS cohort.

CONCLUSION

By using complete blood count parameters as predictors of ICU patient survival, machine learning models can accurately predict the survival of SIRS and sepsis ICU patients. Interestingly, feature importance highlights the role of CRP and APACHE II in both SIRS and sepsis populations. In addition, MPV and EoC are shown to be important features for the sepsis population only, whereas SOFA and PLTC have higher importance for SIRS patients.

Clinical Feature Ranking Based on Ensemble Machine Learning Reveals Top Survival Factors for Glioblastoma Multiforme

Glioblastoma multiforme (GM) is a malignant tumor of the central nervous system considered to be highly aggressive and often carrying a terrible survival prognosis. An accurate prognosis is therefore pivotal for deciding a good treatment plan for patients. In this context, computational intelligence applied to data of electronic health records (EHRs) of patients diagnosed with this disease can be useful to predict the patients' survival time. In this study, we evaluated different machine learning models to predict survival time in patients suffering from glioblastoma and further investigated which features were the most predictive for survival time. We applied our computational methods to three different independent open datasets of EHRs of patients with glioblastoma: the Shieh dataset of 84 patients, the Berendsen dataset of 647 patients, and the Lammer dataset of 60 patients. Our survival time prediction techniques obtained concordance index (C-index) = 0.583 in the Shieh dataset, C-index = 0.776 in the Berendsen dataset, and C-index = 0.64 in the Lammer dataset, as best results in each dataset. Since the original studies regarding the three datasets analyzed here did not provide insights about the most predictive clinical features for survival time, we investigated the feature importance among these datasets. To this end, we then utilized Random Survival Forests, which is a decision tree-based algorithm able to model non-linear interaction between different features and might be able to better capture the highly complex clinical and genetic status of these patients. Our discoveries can impact clinical practice, aiding clinicians and patients alike to decide which therapy plan is best suited for their unique clinical status.

A novel estimator for the two-way partial AUC

BACKGROUND

The two-way partial AUC has been recently proposed as a way to directly quantify partial area under the ROC curve with simultaneous restrictions on the sensitivity and specificity ranges of diagnostic tests or classifiers. The metric, as originally implemented in the tpAUC R package, is estimated using a nonparametric estimator based on a trimmed Mann-Whitney U-statistic, which becomes computationally expensive in large sample sizes. (Its computational complexity is of order [Formula: see text], where [Formula: see text] and [Formula: see text] represent the number of positive and negative cases, respectively). This is problematic since the statistical methodology for comparing estimates generated from alternative diagnostic tests/classifiers relies on bootstrapping resampling and requires repeated computations of the estimator on a large number of bootstrap samples.

METHODS

By leveraging the graphical and probabilistic representations of the AUC, partial AUCs, and two-way partial AUC, we derive a novel estimator for the two-way partial AUC, which can be directly computed from the output of any software able to compute AUC and partial AUCs. We implemented our estimator using the computationally efficient pROC R package, which leverages a nonparametric approach using the trapezoidal rule for the computation of AUC and partial AUC scores. (Its computational complexity is of order [Formula: see text], where [Formula: see text].). We compare the empirical bias and computation time of the proposed estimator against the original estimator provided in the tpAUC package in a series of simulation studies and on two real datasets.

RESULTS

Our estimator tended to be less biased than the original estimator based on the trimmed Mann-Whitney U-statistic across all experiments (and showed considerably less bias in the experiments based on small sample sizes). But, most importantly, because the computational complexity of the proposed estimator is of order [Formula: see text], rather than [Formula: see text], it is much faster to compute when sample sizes are large.

CONCLUSIONS

The proposed estimator provides an improvement for the computation of two-way partial AUC, and allows the comparison of diagnostic tests/machine learning classifiers in large datasets where repeated computations of the original estimator on bootstrap samples become too expensive to compute.

The paracrine isthmin1 transcriptionally regulated by C/EBPβ exacerbates pulmonary vascular leakage in murine sepsis

It is known that pulmonary vascular leakage, a key pathological feature of sepsis-induced lung injury, is largely regulated by perivascular cells. However, the underlying mechanisms have not been fully uncovered. In the present study, we aimed to evaluate the role of isthmin1, a secretory protein originating from alveolar epithelium, in the pulmonary vascular leakage during sepsis and to investigate the regulatory mechanisms of isthmin1 gene transcription. We observed an elevated isthmin1 gene expression in the pulmonary tissue of septic mice induced by cecal ligation and puncture (CLP), as well as in primary murine alveolar type II epithelial cells (ATII) exposed to lipopolysaccharide (LPS). Furthermore, we confirmed that isthmin1 derived from ATII contributes to pulmonary vascular leakage during sepsis. Specifically, adenovirus-mediated isthmin1 disruption in ATII led to a significant attenuation of the increased pulmonary microvascular endothelial cell (PMVEC) hyperpermeability in a PMVEC/ATII coculture system when exposed to LPS. In addition, adeno-associated virus 9 (AAV9)-mediated knockdown of isthmin1 in the alveolar epithelium of septic mice significantly attenuated pulmonary vascular leakage. Finally, mechanistic studies unveiled that nuclear transcription factor CCAAT/enhancer binding protein (C/EBP)β participates in isthmin1 gene activation by binding directly to the cis-regulatory element of isthmin1 locus and may contribute to isthmin1 upregulation during sepsis. Collectively, the present study highlighted the impact of the paracrine protein isthmin1, derived from ATII, on the exacerbation of pulmonary vascular permeability in sepsis and revealed a new regulatory mechanism for isthmin1 gene transcription.NEW & NOTEWORTHY This article addresses the role of the alveolar epithelial-secreted protein isthmin1 on the exacerbation of pulmonary vascular permeability in sepsis and identified nuclear factor CCAAT/enhancer binding protein (C/EBP)β as a new regulator of isthmin1 gene transcription. Targeting the C/EBPβ-isthmin1 regulatory axis on the alveolar side would be of great value in the treatment of pulmonary vascular leakage and lung injury induced by sepsis.

Quick Sequential Organ Failure Assessment Score, Lactate, and Neutrophil-Lymphocyte Ratio Help in Diagnosis and Mortality Prediction during Golden Hour of Sepsis in Emergency Department

Introduction

Sepsis is a life-threatening condition with a very narrow golden period in which confirmatory diagnosis may change the outcome dramatically. No confirmatory biomarker is available till date for early diagnosis of sepsis. This study aimed to evaluate the combined and independent role of quick sequential organ failure assessment (qSOFA) score, lactate, and neutrophil-lymphocyte ratio (NLR) in diagnosis and mortality prediction in early sepsis.

Methods

This was a hospital-based, single-center, prospective cohort study conducted in a tertiary care institute, Karnataka, India. Three hundred adult sepsis patients were recruited during 10-month period, and demographic data, qSOFA score, lactate, NLR, and culture samples were collected in ED within 1 h of admission. Outcome groups (survivor and nonsurvivor) were statistically analyzed with relative frequencies (%), median, mean ± standard deviation with 95% confidence interval (CI), univariate, bivariate, and multivariate logistic regression analysis, and Receiver -operating characteristic curve (ROC) curve to test the predictive ability of initial levels of three biomarkers.

Results

Sepsis was more prevalent among middle-aged male patients. Male gender (odds ratio [OR], 6.9; 95% CI: 1.61-30.1), qSOFA (OR, 154; 95% CI: 15-1565), and lactate (OR, 1.36; 95% CI: 22-833) show 97% (area under the curve) predictive accuracy of the model for sepsis on bivariate and multivariate logistic regression analysis. A significant rise in NLR was a poor outcome indicator on univariate analysis (P = 0.773).

Conclusion

All three biomarkers are good outcome predictors whereas qSOFA and lactate have diagnostic significance in early sepsis. These markers can be used for patient triaging, minimizing culture report dependence for treatment and ultimately the outcome.

A statistical comparison between Matthews correlation coefficient (MCC), prevalence threshold, and Fowlkes-Mallows index

Even if assessing binary classifications is a common task in scientific research, no consensus on a single statistic summarizing the confusion matrix has been reached so far. In recent studies, we demonstrated the advantages of the Matthews correlation coefficient (MCC) over other popular rates such as cross-entropy error, F1 score, accuracy, balanced accuracy, bookmaker informedness, diagnostic odds ratio, Brier score, and Cohen's kappa. In this study, we compared the MCC to other two statistics: prevalence threshold (PT), frequently used in obstetrics and gynecology, and Fowlkes-Mallows index, a metric employed in fuzzy logic and drug discovery. Through the investigation of the mutual relations among three metrics and the study of some relevant use cases, we show that, when positive data elements and negative data elements have the same importance, the Matthews correlation coefficient can be more informative than its two competitors, even this time.

Rapid Escherichia coli Cloned DNA Detection in Serum Using an Electrical Double Layer-Gated Field-Effect Transistor-Based DNA Sensor

In this study, a rapid diagnosis platform was developed for the detection of Escherichia coli O157:H7. An electrical double layer (EDL)-gated field-effect transistor-based biosensor (BioFET) as a point-of-care testing device is demonstrated with its high sensitivity, portability, high selectivity, quick response, and ease of use. The specially designed ssDNA probe was immobilized on the extended gate electrode to bind the target complementary DNA segment of E. coli, resulting in a sharp drain current change within minutes. The limit of detection for target DNA is validated to a concentration of 1 fM in buffer solution and serum. Meanwhile, the results of a Kelvin probe force microscope were shown to have reduced surface potential of the DNA immobilized sensors before and after the cDNA detection, which is consistent with the decreased drain current of the BioFET. A 1.2 kb E. coli duplex DNA synthesized in plasmid was sonicated and detected in serum samples with the sensor array. Gel electrophoresis was used to confirm the efficiency of sonication by elucidating the length of DNA. Those results show that the EDL-gated BioFET system is a promising platform for rapid identification of pathogens for future clinical needs.

Eleven quick tips for data cleaning and feature engineering

Applying computational statistics or machine learning methods to data is a key component of many scientific studies, in any field, but alone might not be sufficient to generate robust and reliable outcomes and results. Before applying any discovery method, preprocessing steps are necessary to prepare the data to the computational analysis. In this framework, data cleaning and feature engineering are key pillars of any scientific study involving data analysis and that should be adequately designed and performed since the first phases of the project. We call "feature" a variable describing a particular trait of a person or an observation, recorded usually as a column in a dataset. Even if pivotal, these data cleaning and feature engineering steps sometimes are done poorly or inefficiently, especially by beginners and unexperienced researchers. For this reason, we propose here our quick tips for data cleaning and feature engineering on how to carry out these important preprocessing steps correctly avoiding common mistakes and pitfalls. Although we designed these guidelines with bioinformatics and health informatics scenarios in mind, we believe they can more in general be applied to any scientific area. We therefore target these guidelines to any researcher or practitioners wanting to perform data cleaning or feature engineering. We believe our simple recommendations can help researchers and scholars perform better computational analyses that can lead, in turn, to more solid outcomes and more reliable discoveries.

Ten simple rules for organizing a special session at a scientific conference

Special sessions are important parts of scientific meetings and conferences: They gather together researchers and students interested in a specific topic and can strongly contribute to the success of the conference itself. Moreover, they can be the first step for trainees and students to the organization of a scientific event. Organizing a special session, however, can be uneasy for beginners and students. Here, we provide ten simple rules to follow to organize a special session at a scientific conference.

Social Determinants of Health Associated With the Development of Sepsis in Adults: A Scoping Review

Evaluating risk for sepsis is complicated due to limited understanding of how social determinants of health (SDoH) influence the occurence of the disease. This scoping review aims to identify gaps and summarize the existing literature on SDoH and the development of sepsis in adults.

DATA SOURCES

A literature search using key terms related to sepsis and SDoH was conducted using Medline and PubMed.

STUDY SELECTION

Studies were screened by title and abstract and then full text in duplicate. Articles were eligible for inclusion if they: 1) evaluated at least one SDoH on the development of sepsis, 2) participants were 18 years or older, and 3) the studies were written in English between January 1970 and January 2022. Systematic reviews, meta-analyses, editorials, letters, commentaries, and studies with nonhuman participants were excluded.

DATA EXTRACTION

Data were extracted in duplicate using a standardized data extraction form. Studies were grouped into five categories according to the SDoH they evaluated (race, socioeconomic status [SES], old age and frailty, health behaviors, and social support). The study characteristics, key outcomes related to incidence of sepsis, mortality, and summary statements were included in tables.

DATA SYNTHESIS

The search identified 637 abstracts, 20 of which were included after full-text screening. Studies evaluating SES, old age, frailty, and gender demonstrated an association between sepsis incidence and the SDoH. Studies that examined race demonstrated conflicting conclusions as to whether Black or White patients were at increased risk of sepsis. Overall, a major limitation of this analysis was the methodological heterogeneity between studies.

CONCLUSIONS

There is evidence to suggest that SDoH impacts sepsis incidence, particularly SES, gender, old age, and frailty. Future prospective cohort studies that use standardized methods to collect SDoH data, particularly race-based data, are needed to inform public health efforts to reduce the incidence of sepsis and help clinicians identify the populations most at risk.

Clinical applications of machine learning in the survival prediction and classification of sepsis: coagulation and heparin usage matter

BACKGROUND

Sepsis is a life-threatening syndrome eliciting highly heterogeneous host responses. Current prognostic evaluation methods used in clinical practice are characterized by an inadequate effectiveness in predicting sepsis mortality. Rapid identification of patients with high mortality risk is urgently needed. The phenotyping of patients will assistant invaluably in tailoring treatments.

METHODS

Machine learning and deep learning technology are used to characterize the patients' phenotype and determine the sepsis severity. The database used in this study is MIMIC-III and MIMIC-IV ('Medical information Mart for intensive care') which is a large, public, and freely available database. The K-means clustering is used to classify the sepsis phenotype. Convolutional neural network (CNN) was used to predict the 28-day survival rate based on 35 blood test variables of the sepsis patients, whereas a double coefficient quadratic multivariate fitting function (DCQMFF) is utilized to predict the 28-day survival rate with only 11 features of sepsis patients.

RESULTS

The patients were grouped into four clusters with a clear survival nomogram. The first cluster (C_1) was characterized by low white blood cell count, low neutrophil, and the highest lymphocyte proportion. C_2 obtained the lowest Sequential Organ Failure Assessment (SOFA) score and the highest survival rate. C_3 was characterized by significantly prolonged PTT, high SIC, and a higher proportion of patients using heparin than the patients in other clusters. The early mortality rate of patients in C_3 was high but with a better long-term survival rate than that in C_4. C_4 contained septic coagulation patients with the worst prognosis, characterized by slightly prolonged partial thromboplastin time (PTT), significantly prolonged prothrombin time (PT), and high septic coagulation disease score (SIC). The survival rate prediction accuracy of CNN and DCQMFF models reached 92% and 82%, respectively. The models were tested on an external dataset (MIMIC-IV) and achieved good performance. A DCQMFF-based application platform was established for fast prediction of the 28-day survival rate.

CONCLUSION

CNN and DCQMFF accurately predicted the sepsis patients' survival, while K-means successfully identified the phenotype groups. The distinct phenotypes associated with survival, and significant features correlated with mortality were identified. The findings suggest that sepsis patients with abnormal coagulation had poor outcomes, abnormal coagulation increase mortality during sepsis. The anticoagulation effects of appropriate heparin sodium treatment may improve extensive micro thrombosis-caused organ failure.

Pentraxin-3 as a predictive marker of mortality in sepsis: an updated systematic review and meta-analysis

BACKGROUND

The purpose of this study was to clarify the prognostic value of Pentraxin-3 (PTX3) on the mortality of patients with sepsis.

METHODS

Publications published up to January 2021 were retrieved from PubMed, EMBASE, and the Cochrane library. Data from eligible cohort and case-control studies were extracted for the meta-analysis. Multivariate regression analysis was used to evaluate the correlation of the outcomes with sample size and male proportion.

RESULTS

A total of 17 studies covering 3658 sepsis patients were included. PTX3 level was significantly higher in non-survivor compared to survivor patients (SMD (95% CI): -1.06 (-1.43, -0.69), P < 0.001). Increased PTX3 level was significantly associated with mortality (HR (95% CI): 2.09 (1.55, 2.81), P < 0.001). PTX3 showed good predictive capability for mortality (AUC:ES (95% CI): 0.73 (0.70, 0.77), P < 0.001). The outcome comparing PTX3 level in non-survivors vs. survivors and the outcome of the association between PTX3 and mortality were associated with sample size but not male proportion. AUC was associated with both sample size and male proportion.

CONCLUSIONS

PTX3 level was significantly higher in non-survivor compared to survivor patients with sepsis. Elevated PTX3 level was significantly associated with mortality. Furthermore, the level of PTX3 might predict patient mortality.

Prior Distribution Estimation of Monitored Information in the Intensive Care Unit with the Hidden Markov Model and Decision Tree Methods

In the intensive care unit, the monitored variables collected from sensors may have different behaviors among patients with different clinical basic information. Giving prior information of the monitored variables based on their specific basic information as soon as the patient is admitted will support the clinicians with better decisions during the surgery. Instead of black box models, the explainable hidden Markov model is proposed, which can estimate the possible distribution parameters of the monitored variables under different clinical basic information. A Student's t-test or correlation test is conducted further to test whether the parameters have a significant relationship with the basic variables. The specific relationship is explored by using a conditional inference tree, which is an explainable model giving deciding rules. Instead of point estimation, interval forecast is chosen as the performance metrics including coverage rate and relative interval width, which provide more reliable results. By applying the methods to an intensive care unit data set with more than 20 thousand patients, the model has good performance with an area under the ROC Curve value of 0.75, which means the hidden states can generally be correctly labelled. The significant test shows that only a few combinations of the basic and monitored variables are not significant under the 0.01 significant level. The tree model based on different quantile intervals provides different coverage and width combination choices. A coverage rate around 0.8 is suggested, which has a relative interval width of 0.77.

Assessment of Metabolic Dysfunction in Sepsis in a Retrospective Single-Centre Cohort

Objective

Our primary aim was to assess selected metabolic dysfunction parameters, both independently and as a complement to the SOFA score, as predictors of short-term mortality in patients with infection admitted to the intensive care unit (ICU).

Methods

We retrospectively enrolled all consecutive adult patients admitted to the eight ICUs of Lille University Hospital, between January 2015 and September 2016, with suspected or confirmed infection. We selected seven routinely measured biological and clinical parameters of metabolic dysfunction (maximal arterial lactatemia, minimal and maximal temperature, minimal and maximal glycaemia, cholesterolemia, and triglyceridemia), in addition to age and the Charlson's comorbidity score. All parameters and SOFA scores were recorded within 24 h of admission.

Results

We included 956 patients with infection, among which 295 (30.9%) died within 90 days. Among the seven metabolic parameters investigated, only maximal lactatemia was associated with higher risk of 90-day hospital mortality in SOFA-adjusted analyses (SOFA-adjusted OR, 1.17; 95%CI, 1.10 to 1.25; p < 0.001). Age and the Charlson's comorbidity score were also statistically associated with a poor prognosis in SOFA-adjusted analyses. We were thus able to develop a metabolic failure, age, and comorbidity assessment (MACA) score based on scales of lactatemia, age, and the Charlson's score, intended for use in combination with the SOFA score.

Conclusions

The maximal lactatemia level within 24 h of ICU admission is the best predictor of short-term mortality among seven measures of metabolic dysfunction. Our combined "SOFA + MACA" score could facilitate early detection of patients likely to develop severe infections. Its accuracy requires further evaluation.

Sepsis and Cognitive Assessment

Sepsis disproportionally affects people over the age of 65, and with an exponentially increasing older population, sepsis poses additional risks for cognitive decline. This review summarizes published literature for (1) authorship qualification; (2) the type of cognitive domains most often assessed; (3) timelines for cognitive assessment; (4) the control group and analysis approach, and (5) sociodemographic reporting. Using key terms, a PubMed database review from January 2000 to January 2021 identified 3050 articles, and 234 qualified as full text reviews with 18 ultimately retained as summaries. More than half (61%) included an author with an expert in cognitive assessment. Seven (39%) relied on cognitive screening tools for assessment with the remaining using a combination of standard neuropsychological measures. Cognitive domains typically assessed were declarative memory, attention and working memory, processing speed, and executive function. Analytically, 35% reported on education, and 17% included baseline (pre-sepsis) data. Eight (44%) included a non-sepsis peer group. No study considered sex or race/diversity in the statistical model, and only five studies reported on race/ethnicity, with Caucasians making up the majority (74%). Of the articles with neuropsychological measures, researchers report acute with cognitive improvement over time for sepsis survivors. The findings suggest avenues for future study designs.

Prognostic tools for elderly patients with sepsis: in search of new predictive models

As a tool to support clinical decision-making, Mortality Prediction Models (MPM) can help clinicians stratify and predict patient risk. There are numerous scoring systems for patients with sepsis that predict sepsis-related mortality and the severity of sepsis. But there are currently no MPMs for adults with sepsis who meet the criteria of "good." Clinicians are unlikely to use complex MPMs that require extensive or expensive data collection to impede workflow. Machine learning applied to minimal medical records of patients diagnosed with sepsis can be a useful tool. Progress is needed in the development and validation of clinical decision support tools that can assist in patient risk stratification, prognosis, discussion of patient outcomes, and shared decision making.

Combining Blood-Based Biomarkers to Predict Mortality of Sepsis at Arrival at the Emergency Department

Background

Our aim was to determine a useful combination of blood biomarkers that can predict 28-day mortality of sepsis upon arrival at the Emergency Department (ED).

Material/Methods

Based on Sepsis-3.0, 90 sepsis patients were enrolled and divided into survivor and nonsurvivor groups with day 28 as the study end point. After comparing the demographic data and clinical characteristics of patients, we evaluated the predictive validity of a combination of markers including interleukin-6 (IL-6), procalcitonin (PCT), and lactate at arrival at the ED. Independent risk factors were found by using univariate and multivariate logistic regression analyses, and the prognostic value of markers was determined by the area under the curve (AUC) of the receiver operating characteristic (ROC) curve.

Results

There were 67 (74.4%) survivors and 23 (25.6%) nonsurvivors. The levels of IL-6 (survivors vs nonsurvivors: median 205.30 vs 3499.00 pg/mL, P=0.012) and lactate (survivors vs nonsurvivors: median 2.37 vs 5.77 mmol/L, P=0.003) were significantly lower in survivor group compared with the nonsurvivor group. Markers including IL-6, PCT, lactate, and neutrophil-to-white blood cell ratio (NWR) were independent risk factors in predicting 28-day mortality due to sepsis. The combination of these 4 markers provided the best predictive performance for 28-day mortality of patients with sepsis, on arrival at the ED (AUC of 0.823, 95% confidence interval [CI] 0.723–0.924), and its accuracy, specificity, and sensitivity were 74.4% (95% CI 64.0–82.8%), 91% (95% CI 80.9–96.3%), and 65% (95% CI 42.8–82.8%), respectively.

Conclusions

The combination of IL-6, PCT, lactate, and NWR measurements is a potential predictor of 28-day mortality for patients with sepsis, at arrival at the ED. Further research is needed to confirm our findings.

Data analytics and clinical feature ranking of medical records of patients with sepsis

Background

Sepsis is a life-threatening clinical condition that happens when the patient’s body has an excessive reaction to an infection, and should be treated in one hour. Due to the urgency of sepsis, doctors and physicians often do not have enough time to perform laboratory tests and analyses to help them forecast the consequences of the sepsis episode. In this context, machine learning can provide a fast computational prediction of sepsis severity, patient survival, and sequential organ failure by just analyzing the electronic health records of the patients. Also, machine learning can be employed to understand which features in the medical records are more predictive of sepsis severity, of patient survival, and of sequential organ failure in a fast and non-invasive way.

Dataset and methods

In this study, we analyzed a dataset of electronic health records of 364 patients collected between 2014 and 2016. The medical record of each patient has 29 clinical features, and includes a binary value for survival, a binary value for septic shock, and a numerical value for the sequential organ failure assessment (SOFA) score. We disjointly utilized each of these three factors as an independent target, and employed several machine learning methods to predict it (binary classifiers for survival and septic shock, and regression analysis for the SOFA score). Afterwards, we used a data mining approach to identify the most important dataset features in relation to each of the three targets separately, and compared these results with the results achieved through a standard biostatistics approach.

Results and conclusions

Our results showed that machine learning can be employed efficiently to predict septic shock, SOFA score, and survival of patients diagnoses with sepsis, from their electronic health records data. And regarding clinical feature ranking, our results showed that Random Forests feature selection identified several unexpected symptoms and clinical components as relevant for septic shock, SOFA score, and survival. These discoveries can help doctors and physicians in understanding and predicting septic shock. We made the analyzed dataset and our developed software code publicly available online.

Supplementary Information

The online version contains supplementary material available at (10.1186/s13040-021-00235-0).