scholarly journals Ruling In and Ruling Out COVID-19: Computing SARS-CoV-2 Infection Risk From Symptoms, Imaging and Test Data.

2020 ◽  
Author(s):  
Chistopher D'Ambrosia ◽  
Henrik Christensen ◽  
Eliah Aronoff-Spencer
Keyword(s):  
Bayesian Inference ◽  
Test Data ◽  
Medical Center ◽  
Metric Learning ◽  
Evaluation Process ◽  
Infection Risk ◽  
Molecular Diagnostic ◽  
Diagnostic Challenge ◽  
Diagnostic Uncertainty ◽  
Inference Network

Background: Assigning meaningful probabilities of SARS CoV2 infection risk presents a diagnostic challenge across the continuum of care. Methods: We integrated patient symptom and test data using machine learning and Bayesian inference to quantify individual patient risk of SARS CoV 2 infection. We trained models with 100,000 simulated patient profiles based on thirteen symptoms, estimated local prevalence, imaging, and molecular diagnostic performance from published reports. We tested these models with consecutive patients who presented with a COVID 19 compatible illness at the University of California San Diego Medical Center over 14 days starting in March 2020. Results: We included 55 consecutive patients with fever (78%) or cough (77%) presenting for ambulatory (n=11) or hospital care (n=44). 51% (n=28) were female, 49% were age <60. Common comorbidities included diabetes (22%), hypertension (27%), cancer (16%) and cardiovascular disease (13%). 69% of these (n=38) were RT-PCR confirmed positive for SARS CoV2 infection, 11 had repeated negative nucleic acid testing and an alternate diagnosis. Bayesian inference network, distance metric learning, and ensemble models discriminated between patients with SARS CoV2 infection and alternate diagnoses with sensitivities of 81.6 to 84.2%, specificities of 58.8 to 70.6%, and accuracies of 61.4 to 71.8%. After integrating imaging and laboratory test statistics with the predictions of the Bayesian inference network, changes in diagnostic uncertainty at each step in the simulated clinical evaluation process were highly sensitive to location, symptom, and diagnostic test choices. Conclusions: Decision support models that incorporate symptoms and available test results can help providers diagnose SARS CoV2 infection in real world settings.

scholarly journals Computing SARS-CoV-2 Infection Risk From Symptoms, Imaging, and Test Data: Diagnostic Model Development (Preprint)

2020 ◽  
Author(s):  
Christopher D'Ambrosia ◽  
Henrik Christensen ◽  
Eliah Aronoff-Spencer
Keyword(s):  
Machine Learning ◽  
Bayesian Inference ◽  
Test Data ◽  
Diagnostic Performance ◽  
Medical Center ◽  
Infection Risk ◽  
Molecular Diagnostic ◽  
Diagnostic Model ◽  
Diagnostic Challenge ◽  
Inference Network

BACKGROUND Assigning meaningful probabilities of SARS-CoV-2 infection risk presents a diagnostic challenge across the continuum of care. OBJECTIVE The aim of this study was to develop and clinically validate an adaptable, personalized diagnostic model to assist clinicians in ruling in and ruling out COVID-19 in potential patients. We compared the diagnostic performance of probabilistic, graphical, and machine learning models against a previously published benchmark model. METHODS We integrated patient symptoms and test data using machine learning and Bayesian inference to quantify individual patient risk of SARS-CoV-2 infection. We trained models with 100,000 simulated patient profiles based on 13 symptoms and estimated local prevalence, imaging, and molecular diagnostic performance from published reports. We tested these models with consecutive patients who presented with a COVID-19–compatible illness at the University of California San Diego Medical Center over the course of 14 days starting in March 2020. RESULTS We included 55 consecutive patients with fever (n=43, 78%) or cough (n=42, 77%) presenting for ambulatory (n=11, 20%) or hospital care (n=44, 80%). In total, 51% (n=28) were female and 49% (n=27) were aged &lt;60 years. Common comorbidities included diabetes (n=12, 22%), hypertension (n=15, 27%), cancer (n=9, 16%), and cardiovascular disease (n=7, 13%). Of these, 69% (n=38) were confirmed via reverse transcription-polymerase chain reaction (RT-PCR) to be positive for SARS-CoV-2 infection, and 20% (n=11) had repeated negative nucleic acid testing and an alternate diagnosis. Bayesian inference network, distance metric learning, and ensemble models discriminated between patients with SARS-CoV-2 infection and alternate diagnoses with sensitivities of 81.6%-84.2%, specificities of 58.8%-70.6%, and accuracies of 61.4%-71.8%. After integrating imaging and laboratory test statistics with the predictions of the Bayesian inference network, changes in diagnostic uncertainty at each step in the simulated clinical evaluation process were highly sensitive to location, symptom, and diagnostic test choices. CONCLUSIONS Decision support models that incorporate symptoms and available test results can help providers diagnose SARS-CoV-2 infection in real-world settings.

scholarly journals Computing SARS-CoV-2 Infection Risk From Symptoms, Imaging, and Test Data: Diagnostic Model Development

10.2196/24478 ◽  
2020 ◽  
Vol 22 (12) ◽  
pp. e24478
Author(s):  
Christopher D'Ambrosia ◽  
Henrik Christensen ◽  
Eliah Aronoff-Spencer
Keyword(s):  
Machine Learning ◽  
Bayesian Inference ◽  
Test Data ◽  
Diagnostic Performance ◽  
Medical Center ◽  
Infection Risk ◽  
Molecular Diagnostic ◽  
Diagnostic Model ◽  
Diagnostic Challenge ◽  
Inference Network

Background Assigning meaningful probabilities of SARS-CoV-2 infection risk presents a diagnostic challenge across the continuum of care. Objective The aim of this study was to develop and clinically validate an adaptable, personalized diagnostic model to assist clinicians in ruling in and ruling out COVID-19 in potential patients. We compared the diagnostic performance of probabilistic, graphical, and machine learning models against a previously published benchmark model. Methods We integrated patient symptoms and test data using machine learning and Bayesian inference to quantify individual patient risk of SARS-CoV-2 infection. We trained models with 100,000 simulated patient profiles based on 13 symptoms and estimated local prevalence, imaging, and molecular diagnostic performance from published reports. We tested these models with consecutive patients who presented with a COVID-19–compatible illness at the University of California San Diego Medical Center over the course of 14 days starting in March 2020. Results We included 55 consecutive patients with fever (n=43, 78%) or cough (n=42, 77%) presenting for ambulatory (n=11, 20%) or hospital care (n=44, 80%). In total, 51% (n=28) were female and 49% (n=27) were aged <60 years. Common comorbidities included diabetes (n=12, 22%), hypertension (n=15, 27%), cancer (n=9, 16%), and cardiovascular disease (n=7, 13%). Of these, 69% (n=38) were confirmed via reverse transcription-polymerase chain reaction (RT-PCR) to be positive for SARS-CoV-2 infection, and 20% (n=11) had repeated negative nucleic acid testing and an alternate diagnosis. Bayesian inference network, distance metric learning, and ensemble models discriminated between patients with SARS-CoV-2 infection and alternate diagnoses with sensitivities of 81.6%-84.2%, specificities of 58.8%-70.6%, and accuracies of 61.4%-71.8%. After integrating imaging and laboratory test statistics with the predictions of the Bayesian inference network, changes in diagnostic uncertainty at each step in the simulated clinical evaluation process were highly sensitive to location, symptom, and diagnostic test choices. Conclusions Decision support models that incorporate symptoms and available test results can help providers diagnose SARS-CoV-2 infection in real-world settings.

Fitting fatigue test data with a novel S-N curve using frequentist and Bayesian inference

2017 ◽  
Vol 105 ◽  
pp. 128-143 ◽  
Author(s):  
Davide Leonetti ◽  
Johan Maljaars ◽  
H.H. (Bert) Snijder
Keyword(s):  
Bayesian Inference ◽  
Fatigue Test ◽  
Test Data

scholarly journals Intraocular Tuberculosis

2013 ◽  
Vol 47 (4) ◽  
pp. 193-201 ◽  
Author(s):  
Swapnil Parchand ◽  
Vishali Gupta
Keyword(s):  
Diagnostic Tests ◽  
Early Stage ◽  
Specific Treatment ◽  
Good Prognosis ◽  
Molecular Diagnostic ◽  
Interferon Gamma Release Assay ◽  
Diagnostic Challenge ◽  
Global Epidemiology ◽  
Tuberculosis Therapy ◽  
Intraocular Tuberculosis

ABSTRACT Intraocular tuberculosis remains a major diagnostic challenge and it is extremely important to establish the diagnosis as the specific treatment helps in reducing the recurrences, thus reducing ocular morbidity. The present review aims to describe the global epidemiology and pathogenesis of intraocular tuberculosis with clinical spectrum and different presentations. The challenges in establishing the diagnosis with role of conventional tests like PPD skin test as well as current diagnostic tests including interferon gamma release assay and molecular diagnostic tests are discussed. The treatment requires anti-tuberculosis therapy with the use of concomitant corticosteroids and carries good prognosis provided the treatment is started in the early stage. How to cite this article Parchand S, Gupta V, Gupta A, Sharma A. Intraocular Tuberculosis. J Postgrad Med Edu Res 2013;47(4):193-201.

scholarly journals 1005. The Antibiotic Optimization DOOR: Refining Assessment of Antibiotic Therapy with Desirability of Outcome Ranking

2019 ◽  
Vol 6 (Supplement_2) ◽  
pp. S354-S354
Author(s):  
Steven Smoke ◽  
Vishal Patel ◽  
Nicole Leonida ◽  
Maria DeVivo
Keyword(s):  
Retrospective Study ◽  
Antibiotic Therapy ◽  
Medical Center ◽  
Academic Medical Center ◽  
Treatment Success ◽  
Molecular Diagnostic ◽  
Single Measure ◽  
Optimal Therapy ◽  
Desirable Outcome ◽  
The Impact

Abstract Background Desirability of outcome ranking (DOOR) is a novel methodology for incorporating multiple outcomes into a single value to more comprehensively compare therapeutic strategies. Its primary application has been limited to antibiotic clinical trials, incorporating treatment success and antibiotic toxicity into a single measure. We describe the application of DOOR methodology to a retrospective study evaluating antibiotic optimization. Methods This was a single-center, retrospective quasi-experimental study conducted at an academic medical center evaluating the impact of prospective pharmacist review of rapid molecular diagnostic testing (RDT) of blood cultures on antibiotic optimization. Two 8-week time periods were evaluated, corresponding to RDT implementation prior to prospective pharmacist review (RDT-only) and RDT with prospective pharmacist review (RDT-PPR). Patients with a positive blood culture who were not on optimal therapy at the time of gram stain were included in the study. Outcomes included the percentage of patients who received optimal therapy, time to optimal antibiotic therapy, and percentage of patients who had therapy de-escalated. An antibiotic optimization DOOR was created with 3 ordinal ranks. The most desirable outcome, rank one, was patients receiving optimal therapy with no missed de-escalation opportunities. Rank two was patients receiving optimal therapy with a missed de-escalation opportunity. The least desirable outcome, rank three, consisted of patients not receiving optimal antibiotic therapy. Time to optimal therapy was used as a tiebreaker for patients in ranks one and two. Results A total of 19 and 29 patients were included in the pre and post-intervention periods, respectively. The percentage of patients reaching optimal therapy was 84% (16/19) and 97% ([28/29], P = 0.16). Median time to optimal therapy was 30:28:26 and 22:40:17 (P = 0.32), respectively. DOOR analysis indicated that the probability of a better outcome for the RDT-PPR group than the RDT-only group was 58% (95% CI 54–62). Conclusion In this small retrospective study, the use of a novel composite methodology identified the benefit of an intervention that was not detected by standard comparison of individual outcomes. Disclosures All authors: No reported disclosures.

Adaptive Damage Detection Using Tunable Piezoelectric Admittance Sensor and Intelligent Inference

2014 ◽  
Author(s):  
K. Zhou ◽  
Q. Shuai ◽  
J. Tang
Keyword(s):  
Bayesian Inference ◽  
Damage Detection ◽  
Structural Model ◽  
Experimental Studies ◽  
Frequency Measurement ◽  
Measurement Information ◽  
Accurate Identification ◽  
Measurement Capability ◽  
Uncertainty Measurement ◽  
Inference Network

The piezoelectric impedance/admittance-based damage detection has been recognized to be sensitive to small-sized damage due to its high frequency measurement capability. Recently, a new class of admittance-based damage detection schemes has been proposed, in which the piezoelectric transducer is integrated with a tunable inductive circuitry. The present research focuses on exploiting the tunable nature of the piezoelectric admittance sensor for the effective identification of damage. In particular, we incorporate the Bayesian inference network into the damage detection process which can intelligently guide the accurate identification of damage location and severity by taking full advantage of the baseline model and measurement as well as the online measurement. As the tunable sensor can provide greatly enriched measurement information, the Bayesian inference can adequately utilize such information and furthermore directly and continuously update the structural model until the model prediction matches with the measurement results. This new approach takes into account the model uncertainty, measurement error, and incompleteness of measurements. Extensive numerical analyses and experimental studies are carried out on a panel structure for methodology demonstration and validation.

scholarly journals Acute presentation of intestinal malrotation in adults: a report of two cases

2010 ◽  
Vol 92 (7) ◽  
pp. e15-e18 ◽  
Author(s):  
Thomas Hanna ◽  
Jacob A Akoh
Keyword(s):  
Case Reports ◽  
Delayed Diagnosis ◽  
Adult Life ◽  
Intestinal Malrotation ◽  
Unusual Complication ◽  
Fetal Life ◽  
Diagnostic Challenge ◽  
Multiple Fractures ◽  
Diagnostic Uncertainty ◽  
First Case

Introduction Intestinal malrotation is a rare developmental abnormality occurring as a result of incomplete rotation during fetal life. It usually presents in the first few weeks of life, but may persist unrecognised into adult life. We report two interesting cases in elderly patients both characterised by a significant diagnostic challenge due to atypical clinical and radiological signs and in one case an unusual complication following laparotomy. Case Reports The first case was a 64-year-old man initially treated for diverticulitis but at laparotomy was found to have malrotation of the midgut and a perforated left-sided appendicitis. The second case was a 76-year-old woman admitted with multiple fractures and increasing abdominal distension following a fall. Ten days after admission, she underwent right hemicolectomy to treat faecal peritonitis due to multiple caecal perforations complicating volvulus in the presence of midgut malrotation. Conclusions These cases illustrate challenges associated with managing patients with undiagnosed intestinal malrotation. Delayed diagnosis is a common feature in several case reports describing atypical presentation of appendicitis in patients with malrotation. While abdominal CT scan can remove much of the diagnostic uncertainty, the diagnosis of malrotation can be missed unless there is a high index of suspicion.

Infection risk in endovascular neurointerventions: a comparative analysis of 549 cases with and without prophylactic antibiotic use

2020 ◽  
Vol 132 (3) ◽  
pp. 797-801 ◽  
Author(s):  
Jan-Karl Burkhardt ◽  
Omar Tanweer ◽  
Miguel Litao ◽  
Pankaj Sharma ◽  
Eytan Raz ◽  
...  
Keyword(s):  
Antibiotic Prophylaxis ◽  
Medical Center ◽  
Prophylactic Antibiotic ◽  
Systemic Infection ◽  
Antibiotic Use ◽  
Infection Risk ◽  
Prophylactic Antibiotics ◽  
Oral Antibiotic ◽  
Systematic Analysis ◽  
Significant Difference

OBJECTIVEA systematic analysis on the utility of prophylactic antibiotics for neuroendovascular procedures has not been performed. At the authors’ institution there is a unique setup to address this question, with some attending physicians using prophylactic antibiotics (cefazolin or vancomycin) for all of their neurointerventions while others generally do not.METHODSThe authors performed a retrospective review of the last 549 neurointerventional procedures in 484 patients at Tisch Hospital, NYU Langone Medical Center. Clinical and radiological data were collected for analysis, including presence of prophylactic antibiotic use, local or systemic infection, infection laboratory values, and treatment. Overall, 306 aneurysms, 117 arteriovenous malformations/arteriovenous fistulas, 86 tumors, and 40 vessel stenosis/dissections were treated with coiling (n = 109), Pipeline embolization device (n = 197), embolization (n = 203), or stenting (n = 40).RESULTSAntibiotic prophylaxis was used in 265 of 549 cases (48%). There was no significant difference between patients with or without antibiotic prophylaxis in sex (p = 0.48), presence of multiple interventions (p = 0.67), diseases treated (p = 0.11), or intervention device placed (p = 0.55). The mean age of patients in the antibiotic prophylaxis group (53.4 years) was significantly lower than that of the patients without prophylaxis (57.1 years; p = 0.014). Two mild local groin infections (0.36%) and no systemic infections (0%) were identified in this cohort, with one case in each group (1/265 [0.38%] vs 1/284 [0.35%]). Both patients recovered completely with local drainage (n = 1) and oral antibiotic treatment (n = 1).CONCLUSIONSThe risk of infection associated with endovascular neurointerventions with or without prophylactic antibiotic use was very low in this cohort. The data suggest that the routine use of antibiotic prophylaxis seems unnecessary and that to prevent antibiotic resistance and reduce costs antibiotic prophylaxis should be reserved for selected patients deemed to be at increased infection risk.

Sign in / Sign up

Export Citation Format

Share Document