scholarly journals Predicting bloodstream infection outcome using machine learning

2021 ◽  
Vol 11 (1) ◽  
Author(s):  
Yazeed Zoabi ◽  
Orli Kehat ◽  
Dan Lahav ◽  
Ahuva Weiss-Meilik ◽  
Amos Adler ◽  
...  
Keyword(s):  
Machine Learning ◽  
Patient Outcomes ◽  
Medical Records ◽  
Bloodstream Infections ◽  
Operating Characteristics ◽  
Is Implementation ◽  
Adverse Health Outcomes ◽  
Receiver Operating Characteristics Curve ◽  
History Of ◽  
Poor Outcomes

AbstractBloodstream infections (BSI) are a main cause of infectious disease morbidity and mortality worldwide. Early prediction of BSI patients at high risk of poor outcomes is important for earlier decision making and effective patient stratification. We developed electronic medical record-based machine learning models that predict patient outcomes of BSI. The area under the receiver-operating characteristics curve was 0.82 for a full featured inclusive model, and 0.81 for a compact model using only 25 features. Our models were trained using electronic medical records that include demographics, blood tests, and the medical and diagnosis history of 7889 hospitalized patients diagnosed with BSI. Among the implications of this work is implementation of the models as a basis for selective rapid microbiological identification, toward earlier administration of appropriate antibiotic therapy. Additionally, our models may help reduce the development of BSI and its associated adverse health outcomes and complications.

scholarly journals Predicting bloodstream infection outcome using machine learning

2021 ◽  
Author(s):  
Yazeed Zoabi ◽  
Orli Kehat ◽  
Dan Lahav ◽  
Ahuva Weiss-Meilik ◽  
Amos Adler ◽  
...  
Keyword(s):  
Electronic Medical Records ◽  
Patient Outcomes ◽  
Medical Records ◽  
Bloodstream Infections ◽  
Operating Characteristics ◽  
Is Implementation ◽  
Adverse Health Outcomes ◽  
Receiver Operating Characteristics Curve ◽  
History Of ◽  
Poor Outcomes

Bloodstream infections (BSI) are a main cause of infectious disease morbidity and mortality worldwide. Early prediction of patients at high risk of poor outcomes of BSI is important for earlier decision making and effective patient stratification. We developed electronic medical record-based ma-chine learning models that predict patient outcomes of BSI. The area under the receiver-operating characteristics curve was 0.82 for a full featured inclusive model, and 0.81 for a compact model using only 25 features. Our models were trained, using electronic medical records that include demographics, blood tests, and the medical and diagnosis history of 7,889 hospitalized patients diagnosed with BSI. Among the implications of this work is implementation of the models as a basis for selective rapid microbiological identification, toward earlier administration of appropriate antibiotic therapy. Additionally, our models may help reduce the development of BSI and its associated adverse health outcomes and complications.

scholarly journals Towards the Prediction of Rearrest during Out-of-Hospital Cardiac Arrest

Entropy ◽  
2020 ◽  
Vol 22 (7) ◽  
pp. 758
Author(s):  
Andoni Elola ◽  
Elisabete Aramendi ◽  
Enrique Rueda ◽  
Unai Irusta ◽  
Henry Wang ◽  
...  
Keyword(s):  
Machine Learning ◽  
Cardiac Arrest ◽  
Patient Outcomes ◽  
Learning Model ◽  
Spontaneous Circulation ◽  
Operating Characteristics ◽  
Machine Learning Model ◽  
Validation Procedure ◽  
Electrocardiogram Ecg ◽  
Hospital Cardiac Arrest

A secondary arrest is frequent in patients that recover spontaneous circulation after an out-of-hospital cardiac arrest (OHCA). Rearrest events are associated to worse patient outcomes, but little is known on the heart dynamics that lead to rearrest. The prediction of rearrest could help improve OHCA patient outcomes. The aim of this study was to develop a machine learning model to predict rearrest. A random forest classifier based on 21 heart rate variability (HRV) and electrocardiogram (ECG) features was designed. An analysis interval of 2 min after recovery of spontaneous circulation was used to compute the features. The model was trained and tested using a repeated cross-validation procedure, on a cohort of 162 OHCA patients (55 with rearrest). The median (interquartile range) sensitivity (rearrest) and specificity (no-rearrest) of the model were 67.3% (9.1%) and 67.3% (10.3%), respectively, with median areas under the receiver operating characteristics and the precision–recall curves of 0.69 and 0.53, respectively. This is the first machine learning model to predict rearrest, and would provide clinically valuable information to the clinician in an automated way.

scholarly journals Early measurement of blood sST2 is a good predictor of death and poor outcomes in patients admitted for COVID-19 infection

2021 ◽  
Author(s):  
M Sánchez-Marteles ◽  
J Rubio-Gracia ◽  
N. Peña-Fresneda ◽  
V Garcés-Horna ◽  
B Gracia-Tello ◽  
...  
Keyword(s):  
Prognostic Markers ◽  
Multivariable Analysis ◽  
Operating Characteristics ◽  
Clinical Value ◽  
End Points ◽  
Receiver Operating Characteristics Curve ◽  
Tertiary Teaching ◽  
Patients At Risk ◽  
Poor Outcomes

AbstractImportanceAlthough several biomarkers have shown correlation to prognosis in COVID-19 patients, their clinical value is limited because of lack of specificity, suboptimal sensibility, or poor dynamic behavior.ObjectiveIn search of better prognostic markers in COVID-19, we hypothesized that circulating soluble ST2 (sST2) could be associated to a worse outcome, prompted by our previous knowledge of sST2 involvement in heart failure-associated lung deterioration, and by mounting evidence favoring a role of IL-33/ST2 axis in the disease.Design, Setting and participantsOne hundred and fifty-two patients admitted for confirmed COVID-19 infection were included in a prospective non-interventional, observational study carried out in a tertiary teaching center. Blood samples were drawn at admission, 48-72 hours later and at discharge. sST2 concentrations, and routine blood laboratory were analyzed.Main outcomesPrimary end-points were admission at intensive care unit (ICU) and, mortality. Other outcomes were a need for high oxygen flow therapy (HOF) or increasing treatment at 48/72 hours.ResultsMedian age was 57.5 years (SD: 12.8), 60.4% males. Ten per cent of patients (n=15) were derived to ICU and/or died during admission. The rest stayed hospitalized 8(IQR:6) days on average. About 34% (n=47), 38% (n=53) and 48.5% (n=66) needed HOF, up-titrate therapy or both, respectively.Median (IQR) sST2 serum concentration (ng/mL) rose to 53.1(30.9) at admission, peaked at 48-72h (79.5[64]) and returned to admission levels at discharge (44.9[36.7]), remaining significantly elevated above healthy donor values (18.6[15.1]).A concentration of sST2 above 58.9 ng/mL identified patients progressing to ICU admission or death. These results remained significant after multivariable analysis. The area under the receiver operating characteristics curve (AUC) of sST2 for the occurrence of end-points was 0.776 (p=0.001). Admission sST2 was higher in patients who needed up-tritate therapy.Conclusions and relevanceIn patients admitted for COVID-19 infection, measurement of sST2 measurement early within 24h after at admission was able to identify patients at risk of severe complications or death.

Diagnostic re-evaluation and predictors of congenital hypothyroidism with eutopic thyroid gland in Jiangxi, China

2021 ◽  
Vol 0 (0) ◽  
Author(s):  
Shao-Hong Chen ◽  
Bi-Cheng Yang ◽  
Jiang-Ying Li ◽  
Ping Xu ◽  
Feng Wang
Keyword(s):  
Congenital Hypothyroidism ◽  
Operating Characteristics ◽  
Dose Requirement ◽  
Bone Maturation ◽  
Screening Programs ◽  
Receiver Operating Characteristics Curve ◽  
Thyroid Glands ◽  
History Of ◽  
Predictive Role

Abstract Objectives An increase in the incidence of congenital hypothyroidism (CH) with eutopic gland has been reported worldwide due to neonatal screening programs. In this study, we aimed to determine the prevalence of transient CH (TCH) and to investigate predictive factors that could distinguish between permanent and transient CH in patients with eutopic thyroid glands. Methods We retrospectively reviewed 508 children treated for CH with eutopic thyroid glands between June 1998 and June 2020 in Jiangxi Newborn Screening Center. All patients were treated with levothyroxine and underwent Diagnostic re-evaluation after 2–3 years of age. Patients were classified as having TCH or permanent CH (PCH) during follow-up. Results Of the 508 patients initially treated for CH with a normally located gland, 335 patients (65.9%) were classified in the TCH group and 173 (34.1%) in the PCH group based on the defined criteria. Multivariate analysis revealed that TCH was associated with a lower levothyroxine dose at 24 months of age (p<0.001) and a lower likelihood of having a first-degree family history of CH (p=0.026) than PCH. Gender, prematurity, low birth weight, initial CH severity such as serum TSH and FT4 levels, or bone maturation delay at diagnosis had no effect. Receiver operating characteristics curve analysis showed that a cutoff of 2.3 μg/kg/day for levothyroxine dose requirement at 24 months of age had a sensitivity of 71% and a specificity of 70% for predicting transient CH, with values below this threshold considered predictive of transient CH. Conclusions TCH presents a significant portion of patients with CH. The levothyroxine dose requirement at 24 months of age has a predictive role in differentiating TCH from PCH in CH patients with eutopic thyroid glands.

scholarly journals Machine learning to predict mortality after rehabilitation among patients with severe stroke

2020 ◽  
Vol 10 (1) ◽  
Author(s):  
Domenico Scrutinio ◽  
Carlo Ricciardi ◽  
Leandro Donisi ◽  
Ernesto Losavio ◽  
Petronilla Battista ◽  
...  
Keyword(s):  
Machine Learning ◽  
Logistic Regression ◽  
Regression Model ◽  
Predictive Value ◽  
Logistic Regression Model ◽  
Clinical Decision Making ◽  
Clinical Decision ◽  
Severe Disability ◽  
Operating Characteristics ◽  
Receiver Operating Characteristics Curve

AbstractStroke is among the leading causes of death and disability worldwide. Approximately 20–25% of stroke survivors present severe disability, which is associated with increased mortality risk. Prognostication is inherent in the process of clinical decision-making. Machine learning (ML) methods have gained increasing popularity in the setting of biomedical research. The aim of this study was twofold: assessing the performance of ML tree-based algorithms for predicting three-year mortality model in 1207 stroke patients with severe disability who completed rehabilitation and comparing the performance of ML algorithms to that of a standard logistic regression. The logistic regression model achieved an area under the Receiver Operating Characteristics curve (AUC) of 0.745 and was well calibrated. At the optimal risk threshold, the model had an accuracy of 75.7%, a positive predictive value (PPV) of 33.9%, and a negative predictive value (NPV) of 91.0%. The ML algorithm outperformed the logistic regression model through the implementation of synthetic minority oversampling technique and the Random Forests, achieving an AUC of 0.928 and an accuracy of 86.3%. The PPV was 84.6% and the NPV 87.5%. This study introduced a step forward in the creation of standardisable tools for predicting health outcomes in individuals affected by stroke.

scholarly journals Epidemiology and outcomes of bloodstream infections in patients discharged from the emergency department

CJEM ◽  
2015 ◽  
Vol 17 (1) ◽  
pp. 27-37 ◽  
Author(s):  
Justine Chan ◽  
Jenna Wong ◽  
Raphael Saginur ◽  
Alan J. Forster ◽  
Carl van Walraven
Keyword(s):  
Emergency Department ◽  
Elderly Patients ◽  
Antibiotic Treatment ◽  
Patient Outcomes ◽  
Medical Records ◽  
Bloodstream Infections ◽  
Population Based ◽  
Specific Patient ◽  
Type Method ◽  
Increased Risk

AbstractObjectiveTo determine the outcomes of patients discharged from the emergency department (ED) with a bloodstream infection (BSI) and how these outcomes are influenced by antibiotic treatment.MethodWe identified every BSI in adult patients discharged from our ED to the community between July 1, 2002, and March 31, 2011. The medical records of all cases were reviewed to determine antibiotic treatment in the ED and at discharge. Microorganism sensitivities were used to determine whether antibiotics were appropriate. These data were linked to population-based administrative data to determine specific patient outcomes within the subsequent 2-week period: death, urgent hospitalization, or an unplanned return to the ED.ResultsA total of 480 adults with BSI were identified (1.49 cases per 1,000 adults discharged from the department). Compared to controls (321,048 patients), BSI patients had a significantly higher risk of urgent hospitalization (adjusted OR 2.1 [95% CI 1.6–2.8]) and unplanned return to the ED (adjusted OR 4.1 [95% CI 3.3–4.9]). Outcome risk was significantly lowered in BSI patients who received appropriate antibiotics in the ED and at discharge. In elderly patients, the risk of urgent hospitalization increased significantly as the time to appropriate antibiotics was delayed.ConclusionsBSI patients discharged from the ED have a significantly increased risk of urgent hospitalization and unplanned return to the ED in the subsequent 2 weeks. These risks decrease significantly with the timely provision of appropriate antibiotics. Our results support the aggressive use of measures ensuring that such patients receive appropriate antibiotics as soon as possible.

scholarly journals A Study on Singapore’s Ageing Population in the Context of Eldercare Initiatives Using Machine Learning Algorithms

2021 ◽  
Vol 5 (4) ◽  
pp. 51
Author(s):  
Easwaramoorthy Rangaswamy ◽  
Girija Periyasamy ◽  
Nishad Nawaz
Keyword(s):  
Machine Learning ◽  
Learning Algorithms ◽  
Healthcare Systems ◽  
Machine Learning Algorithms ◽  
Ageing Population ◽  
Operating Characteristics ◽  
Industry Standard ◽  
Receiver Operating Characteristics Curve ◽  
The Government ◽  
The Impact

Ageing has always directly impacted the healthcare systems and, more specifically, the eldercare costs, as initiatives related to eldercare need to be addressed beyond the regular healthcare costs. This study aims to examine the general issues of eldercare in the Singapore context, as the population of the country is ageing rapidly. The main objective of the study is to examine the eldercare initiatives of the government and their likely impact on the ageing population. The methodology adopted in this study is Cross-Industry Standard Process for Data Mining (CRISP-DM). Reviews related to the impact of an ageing population on healthcare systems in the context of eldercare initiatives were studied. Analysis methods include correlation and machine learning algorithms, such as Decision Tree, Logistic Regression and Receiver Operating Characteristics curve analysis. Suggestions have been provided for various healthcare and eldercare systems’ initiatives and needs that are required to transform to cope with the ageing population.

scholarly journals Early Measurement of Blood sST2 Is a Good Predictor of Death and Poor Outcomes in Patients Admitted for COVID-19 Infection

2021 ◽  
Vol 10 (16) ◽  
pp. 3534
Author(s):  
Marta Sánchez-Marteles ◽  
Jorge Rubio-Gracia ◽  
Natacha Peña-Fresneda ◽  
Vanesa Garcés-Horna ◽  
Borja Gracia-Tello ◽  
...  
Keyword(s):  
Multivariable Analysis ◽  
Receiver Operating Characteristics ◽  
Early Measurement ◽  
Operating Characteristics ◽  
Clinical Value ◽  
Blood Samples ◽  
Primary Endpoints ◽  
Receiver Operating Characteristics Curve ◽  
Patients At Risk ◽  
Poor Outcomes

Although several biomarkers have shown correlation to prognosis in COVID-19 patients, their clinical value is limited because of lack of specificity, suboptimal sensibility or poor dynamic behavior. We hypothesized that circulating soluble ST2 (sST2) could be associated to a worse outcome in COVID-19. In total, 152 patients admitted for confirmed COVID-19 were included in a prospective non-interventional, observational study. Blood samples were drawn at admission, 48–72 h later and at discharge. sST2 concentrations and routine blood laboratory were analyzed. Primary endpoints were admission at intensive care unit (ICU) and mortality. Median age was 57.5 years [Standard Deviation (SD: 12.8)], 60.4% males. 10% of patients (n = 15) were derived to ICU and/or died during admission. Median (IQR) sST2 serum concentration (ng/mL) rose to 53.1 (30.9) at admission, peaked at 48–72 h (79.5(64)) and returned to admission levels at discharge (44.9[36.7]). A concentration of sST2 above 58.9 ng/mL was identified patients progressing to ICU admission or death. Results remained significant after multivariable analysis. The area under the receiver operating characteristics curve (AUC) of sST2 for endpoints was 0.776 (p = 0.001). In patients admitted for COVID-19 infection, early measurement of sST2 was able to identify patients at risk of severe complications or death.

Glucose Levels and Outcome After Primary Intraventricular Hemorrhage

2019 ◽  
Vol 16 (1) ◽  
pp. 40-46
Author(s):  
Rui Guo ◽  
Ruiqi Chen ◽  
Chao You ◽  
Lu Ma ◽  
Hao Li ◽  
...  
Keyword(s):  
Diabetes Mellitus ◽  
Blood Glucose ◽  
Blood Glucose Level ◽  
Incidence Rate ◽  
Glucose Level ◽  
Intraventricular Hemorrhage ◽  
Laboratory Data ◽  
History Of ◽  
Poor Outcomes ◽  
The Impact

Background and Purpose: Hyperglycemia is reported to be associated with poor outcome in patients with spontaneous Intracerebral Hemorrhage (ICH), but the association between blood glucose level and outcomes in Primary Intraventricular Hemorrhage (PIVH) remains unclear. We sought to identify the parameters associated with admission hyperglycemia and analyze the impact of hyperglycemia on clinical outcome in patients with PIVH. Methods: Patients admitted to Department of Neurosurgery, West China Hospital with PIVH between 2010 and 2016 were retrospectively included in our study. Clinical, radiographic, and laboratory data were collected. Univariate and multivariate logistic regression analyses were used to identify independent predictors of poor outcomes. Results: One hundred and seventy patients were included in the analysis. Mean admission blood glucose level was 7.78±2.73 mmol/L and 10 patients (5.9%) had a history of diabetes mellitus. History of diabetes mellitus (P = 0.01; Odds Ratio [OR], 9.10; 95% Confidence Interval [CI], 1.64 to 50.54) was independent predictor of admission critical hyperglycemia defined at 8.17 mmol/L. Patients with admission critical hyperglycemia poorer outcome at discharge (P < 0.001) and 90 days (P < 0.001). After adjustment, admission blood glucose was significantly associated with discharge (P = 0.01; OR, 1.30; 95% CI, 1.06 to 1.59) and 90-day poor outcomes (P = 0.03; OR, 1.27; 95% CI, 1.03 to 1.58), as well as mortality at 90 days (P = 0.005; OR, 1.41; 95% CI, 1.11 to 1.78). In addition, admission critical hyperglycemia showed significantly increased the incidence rate of pneumonia in PIVH (P = 0.02; OR, 6.04; 95% CI 1.27 to 28.80) even after adjusting for the confounders. Conclusion: Admission blood glucose after PIVH is associated with discharge and 90-day poor outcomes, as well as mortality at 90 days. Admission hyperglycemia significantly increases the incidence rate of pneumonia in PIVH.

The use of a Domain Ontology for the Management of Essential Hypertension (Preprint)

2020 ◽  
Author(s):  
Emma Chavez ◽  
Vanessa Perez ◽  
Angélica Urrutia
Keyword(s):  
Decision Making ◽  
Essential Hypertension ◽  
Medical History ◽  
Medical Records ◽  
Medical Training ◽  
Relevant Information ◽  
Domain Ontology ◽  
Free Text ◽  
Snomed Ct ◽  
History Of

BACKGROUND : Currently, hypertension is one of the diseases with greater risk of mortality in the world. Particularly in Chile, 90% of the population with this disease has idiopathic or essential hypertension. Essential hypertension is characterized by high blood pressure rates and it´s cause is unknown, which means that every patient might requires a different treatment, depending on their history and symptoms. Different data, such as history, symptoms, exams, etc., are generated for each patient suffering from the disease. This data is presented in the patient’s medical record, in no order, making it difficult to search for relevant information. Therefore, there is a need for a common, unified vocabulary of the terms that adequately represent the diseased, making searching within the domain more effective. OBJECTIVE The objective of this study is to develop a domain ontology for essential hypertension , therefore arranging the more significant data within the domain as tool for medical training or to support physicians’ decision making will be provided. METHODS The terms used for the ontology were extracted from the medical history of de-identified medical records, of patients with essential hypertension. The Snomed-CT’ collection of medical terms, and clinical guidelines to control the disease were also used. Methontology was used for the design, classes definition and their hierarchy, as well as relationships between concepts and instances. Three criteria were used to validate the ontology, which also helped to measure its quality. Tests were run with a dataset to verify that the tool was created according to the requirements. RESULTS An ontology of 310 instances classified into 37 classes was developed. From these, 4 super classes and 30 relationships were obtained. In the dataset tests, 100% correct and coherent answers were obtained for quality tests (3). CONCLUSIONS The development of this ontology provides a tool for physicians, specialists, and students, among others, that can be incorporated into clinical systems to support decision making regarding essential hypertension. Nevertheless, more instances should be incorporated into the ontology by carrying out further searched in the medical history or free text sections of the medical records of patients with this disease.

Sign in / Sign up

Export Citation Format

Share Document