scholarly journals Neural network analysis of clinical variables predicts escalated care in COVID-19 patients: a retrospective study

PeerJ ◽  
2021 ◽  
Vol 9 ◽  
pp. e11205
Author(s):  
Joyce Q. Lu ◽  
Benjamin Musheyev ◽  
Qi Peng ◽  
Tim Q. Duong
Keyword(s):  
Operating Characteristic ◽  
Vital Signs ◽  
Area Under The Curve ◽  
Blood Gases ◽  
Characteristic Analysis ◽  
Test Dataset ◽  
Clinical Variables ◽  
Icu Admission ◽  
Reactive Protein ◽  
Laboratory Variables

This study sought to identify the most important clinical variables that can be used to determine which COVID-19 patients hospitalized in the general floor will need escalated care early on using neural networks (NNs). Analysis was performed on hospitalized COVID-19 patients between 7 February 2020 and 4 May 2020 in Stony Brook Hospital. Demographics, comorbidities, laboratory tests, vital signs and blood gases were collected. We compared those data obtained at the time in emergency department and the time of intensive care unit (ICU) upgrade of: (i) COVID-19 patients admitted to the general floor (N = 1203) vs. those directly admitted to ICU (N = 104), and (ii) patients not upgraded to ICU (N = 979) vs. those upgraded to the ICU (N = 224) from the general floor. A NN algorithm was used to predict ICU admission, with 80% training and 20% testing. Prediction performance used area under the curve (AUC) of the receiver operating characteristic analysis (ROC). We found that C-reactive protein, lactate dehydrogenase, creatinine, white-blood cell count, D-dimer and lymphocyte count showed temporal divergence between COVID-19 patients hospitalized in the general floor that were upgraded to ICU compared to those that were not. The NN predictive model essentially ranked the same laboratory variables to be important predictors of needing ICU care. The AUC for predicting ICU admission was 0.782 ± 0.013 for the test dataset. Adding vital sign and blood-gas data improved AUC (0.822 ± 0.018). This work could help frontline physicians to anticipate downstream ICU need to more effectively allocate healthcare resources.

scholarly journals Deep Neural Network Analysis of Clinical Variables Predicts Escalated Care in COVID-19 Patients

2020 ◽  
Author(s):  
Joyce Lu ◽  
Benjamin Musheyev ◽  
Qi Peng ◽  
Tim Duong
Keyword(s):  
Neural Network ◽  
Deep Learning ◽  
Laboratory Tests ◽  
Deep Neural Network ◽  
Vital Signs ◽  
Area Under The Curve ◽  
Blood Gases ◽  
Characteristic Analysis ◽  
Clinical Variables ◽  
Icu Admission

Abstract This study sought to identify the most important clinical variables that can be used to determine which COVID-19 patients will need escalated care early on using deep-learning neural networks. Analysis was performed on hospitalized COVID-19 patients between February 7, 2020 and May 4, 2020 in Stony Brook Hospital. Demographics, comorbidities, laboratory tests, vital signs, and blood gases were collected. We compared data obtained at the time in emergency department and the time of intensive care unit (ICU) upgrade of: i) COVID-19 patients admitted to the general floor (N=1203) versus those directly admitted to ICU (N=104), and ii) patients not upgraded to ICU (N=979) versus those upgraded to the ICU (N=224) from the general floor. A deep neural network algorithm was used to predict ICU admission, with 80% training and 20% testing. Prediction performance used area under the curve (AUC) of the receiver operating characteristic analysis (ROC). We found that C-reactive protein, lactate dehydrogenase, creatinine, white-blood cell count, D-dimer, and lymphocyte count showed temporal divergence between patients were upgraded to ICU compared to those were not. The deep learning predictive model ranked essentially the same set of laboratory variables to be important predictors of needing ICU care. The AUC for predicting ICU admission was 0.782±0.013 for the test dataset. Adding vital sign and blood-gas data improved AUC (0.861±0.018). This study identified a few laboratory tests that were predictive of escalated care. This work could help frontline physicians to anticipate downstream ICU needs to more effectively allocate healthcare resources.

Soluble CD14 Subtype in Peripheral Blood is a Biomarker for Early Diagnosis of Sepsis

2020 ◽  
Vol 51 (6) ◽  
pp. 614-619
Author(s):  
Wuqiong Zhou ◽  
Heping Rao ◽  
Qiuming Ding ◽  
Xiang Lou ◽  
Jianjiang Shen ◽  
...  
Keyword(s):  
Septic Shock ◽  
Early Diagnosis ◽  
Peripheral Blood ◽  
Operating Characteristic ◽  
White Blood Cells ◽  
Area Under The Curve ◽  
The Other ◽  
Characteristic Analysis ◽  
Soluble Cd14 ◽  
Reactive Protein

Abstract Objective To study the value of serum soluble CD14 subtype (sCD14-ST) in early diagnosis of sepsis. Methods Seventy-two patients were diagnosed with systemic inflammatory response syndrome, sepsis, or septic shock. Peripheral blood was collected at 0, 12, 24, and 48 hours after admission to the hospital. Levels of sCD14-ST, procalcitonin (PCT), hypersensitive C-reactive protein (CRP), and white blood cells (WBC) were determined. Results Levels of sCD14-ST in the patients with septic shock were higher than those in the other patients (P < .01) and peaked at 48 h. PCT and CRP levels were similar in the patients at admission but increased by 5 times to 10 times in the next 48 h, especially in the patients with septic shock. WBC levels remained high and did not change dramatically. Receiver operating characteristic analysis revealed that the area under the curve, sensitivity, and specificity values of sCD14-ST to diagnose sepsis were much higher than those of the other markers. Conclusion Compared with PCT, CRP, and WBC, sCD14-ST is a better biomarker for the early diagnosis of sepsis.

scholarly journals Biomarker Significance of Interleukin-18 in Juvenile Idiopathic Arthritis

2020 ◽  
pp. 3200-3207
Author(s):  
Wasan W. Al-bassam ◽  
Ali H. Ad'hiah ◽  
Khadier Z. Mayouf
Keyword(s):  
Juvenile Idiopathic Arthritis ◽  
Inflammatory Cytokines ◽  
Operating Characteristic ◽  
Area Under The Curve ◽  
Interleukin 18 ◽  
Characteristic Analysis ◽  
Reactive Protein ◽  
Significant Area ◽  
Pro Inflammatory Cytokines

Juvenile idiopathic arthritis (JIA) represents a group of multifactorial autoinflammatory arthritis diseases. A dysregulated production of pro-inflammatory cytokines is proposed to have a role in the pathogenesis of the disease. Interleukin-18 (IL-18) is one of these pro-inflammatory cytokines. Therefore, this study aimed to define the role of IL-18 in the pathogenesis of JIA. Accordingly, the serum level of IL-18 was determined in 59 Iraqi JIA patients and 58 matched controls. The results revealed a significantly increased median of IL-18 in the patients as compared to the control. A similar increased level was observed in subgroups of patients characterized according to gender, seropositivity for C-reactive protein and rheumatoid factors, juvenile arthritis disease activity score 27 (JADAS27), type of medication, and JIA subtypes. However, JADAS27 showed a significant positive correlation with IL-18 level. Receiver operating characteristic analysis revealed that IL-18 occupied a significant area under the curve, and therefore its significance as a biomarker was suggested. In conclusion, IL-18 is an important biomarker for JIA and may have a role in pathogenesis of disease.

scholarly journals Identification of an optimal prolactin threshold to determine prolactinoma size using receiver operating characteristic analysis

2021 ◽  
Vol 11 (1) ◽  
Author(s):  
Bianca M. Leca ◽  
Maria Mytilinaiou ◽  
Marina Tsoli ◽  
Andreea Epure ◽  
Simon J. B. Aylwin ◽  
...  
Keyword(s):  
Receiver Operating Characteristic ◽  
Operating Characteristic ◽  
Tumour Size ◽  
Area Under The Curve ◽  
Serum Prolactin ◽  
Pituitary Neoplasms ◽  
Characteristic Analysis ◽  
Roc Curve Analysis ◽  
Baseline Serum ◽  
Receiver Operating

AbstractProlactinomas represent the most common type of secretory pituitary neoplasms, with a therapeutic management that varies considerably based on tumour size and degree of hyperprolactinemia. The aim of the current study was to evaluate the relationship between serum prolactin (PRL) concentrations and prolactinoma size, and to determine a cut-off PRL value that could differentiate micro- from macro-prolactinomas. A retrospective cohort study of 114 patients diagnosed with prolactinomas between 2007 and 2017 was conducted. All patients underwent gadolinium enhanced pituitary MRI and receiver operating characteristic (ROC) analyses were performed. 51.8% of patients in this study were men, with a mean age at the time of diagnosis of 42.32 ± 15.04 years. 48.2% of the total cohort were found to have microadenomas. Baseline serum PRL concentrations were strongly correlated to tumour dimension (r = 0.750, p = 0.001). When performing the ROC curve analysis, the area under the curve was 0.976, indicating an excellent accuracy of the diagnostic method. For a value of 204 μg/L (4338 mU/L), sensitivity and specificity were calculated at 0.932 and 0.891, respectively. When a cut off value of 204 μg/L (4338 mU/L) was used, specificity was 93.2%, and sensitivity 89.1%, acceptable to reliably differentiate between micro- and macro- adenomas.

scholarly journals Early Warning Scores at time of ICU admission to Predict Mortality in critically ill COVID-19 Patients

2021 ◽  
pp. 1-15
Author(s):  
Asha Tyagi ◽  
Surbhi Tyagi ◽  
Ananya Agrawal ◽  
Aparna Mohan ◽  
Devansh Garg ◽  
...  
Keyword(s):  
Developing Country ◽  
Cox Regression ◽  
Vital Signs ◽  
Developed Countries ◽  
Operating Characteristics ◽  
Icu Mortality ◽  
Prospective Data ◽  
Icu Admission ◽  
Reactive Protein ◽  
Time Of Admission

Abstract Objective: To assess ability of NEWS2, SIRS, qSOFA and CRB-65 calculated at the time of Intensive Care Unit (ICU) admission for predicting ICU-mortality in patients of laboratory confirmed COVID-19 infection. Methods: This prospective data analysis was based on chart reviews for laboratory confirmed COVID-19 patients admitted to ICUs over a 1month period. The NEWS2, CURB-65, qSOFA and SIRS were calculated from the first recorded vital signs upon admission to ICU and assessed for predicting mortality. Results: Total of 140 patients aged between 18 to 95 years were included in the analysis of whom majority were >60 years (47.8%), with evidence of pre-existing comorbidities (67.1%). The commonest symptom at presentation was dyspnea (86.4%). Based upon the Receiver Operating Characteristics-Area Under Curve (AUC), the best discriminatory power to predict ICU mortality was for the CRB65 (AUC: 0.720 [95% CI: 0.630 – 0.811]) followed closely by NEWS2 (AUC: 0.712 [95% CI: 0.622 – 0.803]). Additionally, a multivariate cox regression model showed Glasgow Coma Score at time of admission [P < 0.001; adjusted Hazard Ratio = 0.808 (95% CI: 0.715-0.911)] to be the only significant predictor of ICU mortality. Conclusion: CRB65 and NEWS2 scores assessed at the time of ICU admission offer only a fair discriminatory value for predicting mortality. Further evaluation after adding laboratory markers such as C-reactive protein and D-dimer may yield a more useful prediction model. Much of the earlier data is from developed countries and uses scoring at time of hospital admission. This study was from a developing country, with the scores assessed at time of ICU admission, rather than the emergency department as with existing data from developed countries, for patients with moderate/severe COVID disease. Since the scores showed some utility for predicting ICU mortality even when measured at time of ICU admission, their use in allocation of limited ICU resources in a developing country merits further research.

Procalcitonin and Conventional Markers of Inflammation in Peritoneal Dialysis Patients and Peritonitis

2006 ◽  
Vol 26 (2) ◽  
pp. 240-248 ◽  
Author(s):  
Galip Guz ◽  
Bulent Colak ◽  
Kenan Hizel ◽  
Kadriye A. Reis ◽  
Yasemin Erten ◽  
...  
Keyword(s):  
Peritoneal Dialysis ◽  
Operating Characteristic ◽  
Characteristic Curve ◽  
Area Under The Curve ◽  
Hemodialysis Patients ◽  
Biochemical Tests ◽  
Reactive Protein ◽  
Markers Of Inflammation ◽  
Median Serum ◽  
Significant Difference

Objectives To determine the significance of a newly described marker of inflammation procalcitonin (PCT), and to investigate its relationship to conventional markers of inflammation, such as C-reactive protein (CRP), fibrinogen, and erythrocyte sedimentation rate (ESR), in patients on peritoneal dialysis (PD) and with peritonitis. Design A prospective, observational clinical study. Setting The Nephrology Division of a University-affiliated teaching hospital. Patients and Methods 51 consecutive patients on PD were included in the study. Of this number, 16 developed peritonitis during the observational period. Baseline PCT, CRP, and fibrinogen concentrations and ESR of 51 PD patients were determined at a time point (TB) prior to any evidence of infection. These results were compared with laboratory values from 74 hemodialysis patients and 34 nonuremic control subjects. All PD patients then were followed prospectively for evidence of peritonitis. In addition to routine blood tests, including hemoglobin and leukocyte count, and routine biochemical tests, blood samples were taken to measure PCT, CRP, and fibrinogen concentrations and ESR at the time (T0) when patients first were diagnosed with PD peritonitis and also on the 4th (T4) and the 14th (T14) days after treatment for peritonitis was initiated. PCT was assayed by immunoluminometry. Results No significant difference was observed between baseline median serum PCT concentrations in PD and hemodialysis patients; however, in both groups, baseline median PCT concentrations were significantly higher than those of nonuremic controls ( p < 0.05). The 16 patients on PD who developed peritonitis had 21 PD peritonitis episodes during the study period. The increased PCT concentration observed at T0 in PD peritonitis episodes decreased with therapy, and this change was statistically significant ( p < 0.05). In a receiver operating characteristic curve analysis for peritonitis, the area under the curve (AUC) for PCT was 0.80, which was significantly lower than the AUC for CRP and greater than the AUCs for fibrinogen and ESR. The sensitivity of PCT for peritonitis was lower than the sensitivity of conventional markers of inflammation; however, the specificity of PCT was higher. Conclusions Median serum PCT concentration in PD patients was significantly higher than in nonuremic controls but not hemodialysis patients. Serum PCT concentrations may serve as a useful adjunct to traditional markers of inflammation in detecting and monitoring inflammation and peritonitis in PD patients.

scholarly journals Measuring the palpable pulsatility length as a physical examination test in defining the severity of inflow stenosis for hemodialysis fistulas

2019 ◽  
Vol 20 (1) ◽  
Author(s):  
Matt Chiung-Yu Chen ◽  
Mei-Jui Weng ◽  
Misoso Yi-Wen Wu ◽  
Yi-Chun Liu ◽  
Wen-Che Chi
Keyword(s):  
Operating Characteristic ◽  
Area Under The Curve ◽  
High Sensitivity ◽  
Characteristic Analysis ◽  
Main Trunk ◽  
Arteriovenous Fistulas ◽  
Cannulation Site ◽  
Before And After ◽  
Physical Examinations ◽  
Hemodialysis Fistulas

Abstract Background Pulsatility is an important property of hemodialysis arteriovenous fistulas (AVF) and can be perceived by the fingers as a gradual decrease in strength downstream from the anastomosis along the main trunk of the fistula. The distance from the point at which the pulse becomes imperceptible to the anastomosis is termed the palpable pulsatility length (PPL); we considered this length may play a role in assessing the severity of inflow stenosis for hemodialysis fistulas. Methods This study was performed by retrospective analysis of routinely collected data. Physical examinations and fistula measurements were performed in a selected population of 76 hemodialysis patients with mature fistulas during half a year. Fistula measurements included the PPL before and after treatment and the distance between the anastomosis and the arterial cannulation site (aPump length). The aPump index (API) was calculated by dividing the PPL by the aPump length. Angiograms were reviewed to determine the location and severity of stenosis. PPL and API were used to detect the critical inflow stenosis, which indicates severe inflow stenosis of an AVF. Results Receiver operating characteristic analysis showed that the area under the curve was 0.895 for API and 0.878 for PPL. A cutoff value of API < 1.29 and PPL < 11.0 cm were selected to detect the critical inflow stenosis. The sensitivity was 96.0% versus 80.0% and specificity was 84.31% versus 84.31% for API and PPL, respectively. Conclusions PPL and API are useful tools in defining the severity of pure inflow stenosis for mature AVFs in the hands of trained examiners with high sensitivity and specificity.

scholarly journals Usefulness of Lymphocyte ratio and CRP Level for Early Detection of Surgical Site Infection Following spinal surgery

2019 ◽  
Author(s):  
Wenbo Wei ◽  
Shajie Dang ◽  
Dapeng Duan ◽  
Liqun Gong ◽  
Jue Wang ◽  
...  
Keyword(s):  
Surgical Site Infection ◽  
Spinal Surgery ◽  
Operating Characteristic ◽  
Characteristic Curve ◽  
Area Under The Curve ◽  
Site Infection ◽  
C Reactive Protein ◽  
Lymphocyte Ratio ◽  
Reactive Protein ◽  
Operating Characteristic Curve

Abstract Background: To investigate the significant laboratory markers for early diagnosis of surgical site infection after spinal surgery. And determine the diagnostic cut-off values of these markers Methods: A total of 67 patients participated in the study: 11 patients who developed surgical site infection after spinal surgery (SSI Group) and 56 patients were compared with the infected group in terms of age，gender, operating time and intraoperative blood loss (Non-SSI Group). The white blood cell (WBC) count , WBC differential , C-reactive protein (CRP) and erythrocyte sedimentation rate(ESR) were determined before and 1, 3 and 7 days postoperatively . Then, we determine the diagnostic cutoff for these markers by using the receiver operating characteristic curve. Results: The CRP, ESR and WBC were significantly higher in the SSI group at 3 and 7 days postoperatively. The lymphocyte ratio at 3 days postoperatively was significantly lower in the SSI Group. Using the receiver operating characteristic curve，lymphocyte ratio <11.5% at 3 days postoperatively (sensitivity 90.9%, specificity 75.4%, area under the curve [AUC] 0.919), and C-reactive protein level >26 mg/dL at 7 days postoperatively (sensitivity 90.9%, specificity 87.7%, area under the curve [AUC] 0.954) were the significant laboratory marker for early detection of SSI Conclusion: Lymphocyte ratio<11.5% at 3 days and C-reactive protein levels>26.5mg/dl at 7 days after spinal surgery are reliable markers of SSI.

scholarly journals 694. Prediction Tool for Infective Endocarditis in Beta-hemolytic Streptococcal Bacteremia

2021 ◽  
Vol 8 (Supplement_1) ◽  
pp. S448-S449
Author(s):  
Ryo Hasegawa ◽  
Takahiro Matsuo ◽  
Osamu Takahashi ◽  
Nobuyoshi Mori
Keyword(s):  
Infective Endocarditis ◽  
Predictive Value ◽  
Area Under The Curve ◽  
Tertiary Hospital ◽  
Predictive Score ◽  
Characteristic Analysis ◽  
Group B Streptococci ◽  
Reactive Protein ◽  
Multivariable Analyses ◽  
Group B

Abstract Background Although beta-hemolytic streptococci (BHS) is a rare causative pathogen of infective endocarditis (IE), IE is a serious condition and it is important to predict IE in BHS bacteremia (BHS-IE). The purpose of this study was to develop a predictive score for BHS-IE. Methods We conducted a retrospective study comparing the clinical features of BHS-IE and BHS-non infective endocarditis (BHS-nIE) in adult patients with BHS bacteremia at a 520-bed tertiary hospital in Tokyo, Japan from 2004 to 2020. IE was diagnosed according to modified Duke's criteria, and both “Definite” and “Possible” were included. Univariate and multivariable analyses were conducted using logistic regression. Results Among 250 patients with BHS bacteremia, 47 (19%) were diagnosed with BHS-IE. The median (IQR) patient age was 71 (59, 84) years and 121 (68%) were male. The proportions of A, B, C/G groups were 14%, 38.4%, and 47.6%, respectively. Five predictors, either independently associated with BHS-IE or clinically relevant, were used to develop the prediction score: C-reactive protein ≥ 10 mg/dl (2 points); Group B Streptococci (1 point); Auscultation of heart murmur (1 point); Platelet count &lt; 150 /µl (1 point); and Hypotension (systolic blood pressure &lt; 90 mmHg or on vasopressor) (1 point). In a receiver operating characteristic analysis, the area under the curve was 0.74 (95% confidence interval [CI]: 0.66 - 0.82). The cut-point was 2. A score ≥2 had a sensitivity of 87% (95%CI: 0.743 - 0.952), a specificity of 37% (95%CI: 0.308 - 0.445), a positive predictive value of 24%, and a negative predictive value of 93%, respectively. Conclusion We developed the score to help clinicians rule out IE in BHS bacteremia. Further research is warranted for validation. Disclosures All Authors: No reported disclosures

Ambulant erworbene Pneumonie: Die Mortalität nachhaltig senken

2020 ◽  
Vol 8 (5) ◽  
pp. 252-253
Author(s):  
Stefan Krüger
Keyword(s):  
Operating Characteristic ◽  
Distress Syndrome ◽  
Area Under The Curve ◽  
Scoring Systems ◽  
Roc Curves ◽  
Community Acquired Pneumonia ◽  
Ambulant Erworbene Pneumonie ◽  
Icu Admission ◽  
Failure Assessment

Background: The study aimed to investigate the predictive value of the quick sequential organ failure assessment (qSOFA) for clinical outcomes in emergency patients with community-acquired pneumonia (CAP). Methods: A total of 742 CAP cases from the emergency department (ED) were enrolled in this study. The scoring systems including the qSOFA, SOFA and CURB-65 (confusion, urea, respiratory rate, blood pressure and age) were used to predict the prognostic outcomes of CAP in ICU-admission, acute respiratory distress syndrome (ARDS) and 28-day mortality. According to the area under the curve (AUC) of the receiver operating characteristic (ROC) curves, the accuracies of prediction of the scoring systems were analyzed among CAP patients. Results: The AUC values of the qSOFA, SOFA and CURB-65 scores for ICU-admission among CAP patients were 0.712 (95%CI: 0.678–0.745, P < 0.001), 0.744 (95%CI: 0.711–0.775, P < 0.001) and 0.705 (95%CI: 0.671–0.738, P < 0.001), respectively. For ARDS, the AUC values of the qSOFA, SOFA and CURB-65 scores were 0.730 (95%CI: 0.697–0.762, P < 0.001), 0.724 (95%CI: 0.690–0.756, P < 0.001) and 0.749 (95%CI: 0.716–0.780, P < 0.001), respectively. After 28 days of follow-up, the AUC values of the qSOFA, SOFA and CURB-65 scores for 28-day mortality were 0.602 (95%CI: 0.566–0.638, P < 0.001), 0.587 (95%CI: 0.551–0.623, P < 0.001) and 0.614 (95%CI: 0.577–0.649, P < 0.001) in turn. There were no statistical differences between qSOFA and SOFA scores for predicting ICU-admission (Z = 1.482, P = 0.138), ARDS (Z = 0.321, P = 0.748) and 28-day mortality (Z = 0.573, P = 0.567). Moreover, we found no differences to predict the ICU-admission (Z = 0.370, P = 0.712), ARDS (Z = 0.900, P = 0.368) and 28-day mortality (Z = 0.768, P = 0.442) using qSOFA or CURB-65 scores. Conclusion: qSOFA was not inferior to SOFA or CURB-65 scores in predicting the ICU-admission, ARDS and 28-day mortality of patients presenting in the ED with CAP.

Sign in / Sign up

Export Citation Format

Share Document