scholarly journals Artificial intelligence sepsis prediction algorithm learns to say “I don’t know”

2021 ◽  
Vol 4 (1) ◽  
Author(s):  
Supreeth P. Shashikumar ◽  
Gabriel Wardi ◽  
Atul Malhotra ◽  
Shamim Nemati
Keyword(s):  
Area Under The Curve ◽  
The United States ◽  
Prediction Algorithm ◽  
False Alarms ◽  
High Area ◽  
Life Saving ◽  
Sequential Prediction ◽  
Sepsis Risk ◽  
Deep Learning Model ◽  
Timely Administration

AbstractSepsis is a leading cause of morbidity and mortality worldwide. Early identification of sepsis is important as it allows timely administration of potentially life-saving resuscitation and antimicrobial therapy. We present COMPOSER (COnformal Multidimensional Prediction Of SEpsis Risk), a deep learning model for the early prediction of sepsis, specifically designed to reduce false alarms by detecting unfamiliar patients/situations arising from erroneous data, missingness, distributional shift and data drifts. COMPOSER flags these unfamiliar cases as indeterminate rather than making spurious predictions. Six patient cohorts (515,720 patients) curated from two healthcare systems in the United States across intensive care units (ICU) and emergency departments (ED) were used to train and externally and temporally validate this model. In a sequential prediction setting, COMPOSER achieved a consistently high area under the curve (AUC) (ICU: 0.925–0.953; ED: 0.938–0.945). Out of over 6 million prediction windows roughly 20% and 8% were identified as indeterminate amongst non-septic and septic patients, respectively. COMPOSER provided early warning within a clinically actionable timeframe (ICU: 12.2 [3.2 22.8] and ED: 2.1 [0.8 4.5] hours prior to first antibiotics order) across all six cohorts, thus allowing for identification and prioritization of patients at high risk for sepsis.

scholarly journals Artificial Intelligence Sepsis Prediction Algorithm Learns to Say "I don't know"

2021 ◽  
Author(s):  
Supreeth P. Shashikumar ◽  
Gabriel Wardi ◽  
Atul Malhotra ◽  
Shamim Nemati
Keyword(s):  
Area Under The Curve ◽  
The United States ◽  
Prediction Algorithm ◽  
False Alarms ◽  
High Area ◽  
Life Saving ◽  
Sequential Prediction ◽  
Sepsis Risk ◽  
Deep Learning Model ◽  
Timely Administration

Sepsis is a leading cause of morbidity and mortality worldwide. Early identification of sepsis is important as it allows timely administration of potentially life-saving resuscitation and antimicrobial therapy. We present COMPOSER (COnformal Multidimensional Prediction Of SEpsis Risk), a deep learning model for the early prediction of sepsis, specifically designed to reduce false alarms by detecting unfamiliar patients/situations arising from erroneous data, missingness, distributional shift and data drifts. COMPOSER flags these unfamiliar cases as 'indeterminate' rather than making spurious predictions. Six patient cohorts (515,720 patients) curated from two healthcare systems in the United States across intensive care units (ICU) and emergency departments (ED) were used to train and externally and temporally validate this model. In a sequential prediction setting, COMPOSER achieved a consistently high area under the curve (AUC) (ICU: 0.925-0.953; ED: 0.938-0.945). Out of over 6 million prediction windows roughly 20% and 8% were identified as 'indeterminate' amongst non-septic and septic patients, respectively. COMPOSER provided early warning within a clinically actionable timeframe (ICU: 12.2 [3.2 22.8] and ED: 2.1 [0.8 4.5] hours prior to first antibiotics order) across all six cohorts, thus allowing for identification and prioritization of patients at high risk for sepsis.

New Drugs of 2012: A Concise Overview of the NMEs and Trends for Innovative Brand Market in the United States

2013 ◽  
Vol 6 (2) ◽  
pp. 2009-2013
Author(s):  
D Samba Reddy
Keyword(s):  
Drug Approval ◽  
The United States ◽  
New Drugs ◽  
New Combination ◽  
Drug Status ◽  
Combination Drug ◽  
Human Cord Blood ◽  
Life Saving ◽  
Genomics And Proteomics ◽  
The Face

Thirty-nine (39) new drugs have been approved by the U.S. FDA in 2012, a record highest number of approvals since 1996. The record is a sign that pharma companies are poised to tap recent advances from genomics and proteomics. This list includes novel new drugs, known as new molecular entities (NMEs), biologics and new products. Many life-saving drugs are approved for marketing. The list includes a total of 10 drugs for cancer treatment, and nearly a quarter of those approved in 2012 had orphan drug status.  Among the breakthrough drugs approved in 2012 were ivacaftor (cystic fibrosis), vasmodegib (skin cancer), HPC-C (human cord blood product), ruxolitinib (myelofibrosis) and a new combination drug to treat HIV. In addition,  several unique products were approved for the treatment of macular degeneration, chronic weight management, overactive bladder, actinic keratosis, erectile dysfunction, glaucoma, respiratory distress syndrome, and COPD. The approval of 39 drugs in 2012 underscores a robust success rate and confirms that innovation is once again beginning to pay off. In the existing climate of reduced revenues in the face of generic competitions, the future and survival of big companies rests heavily on their unique niche products. It is apparent that big Pharma and a growing number of emerging Biotechs alike have focused their attention on developing new NMEs for rare diseases. In 2012, the length of the FDA’s review is shorter than agencies in other countries. Innovative models adopted for R&D strategies, communications, and new regulatory changes appear to shorten development timelines. Despite record drug approvals, there is bleak scope for blockbusters because most of these drugs have a limited market. The pipeline for blockbusters appears very low. However, there is unmet medical need for new drugs in autism, Alzheimer’s disease and epilepsy. Overall, the new drug approval list unveils unique and reemerging trends indicating that the pharma companies are poised for big growth from new brands approved for marketing for narrow-spectrum indications.    

scholarly journals Machine learning applied on chest x-ray can aid in the diagnosis of COVID-19: a first experience from Lombardy, Italy

2021 ◽  
Vol 5 (1) ◽  
Author(s):  
Isabella Castiglioni ◽  
Davide Ippolito ◽  
Matteo Interlenghi ◽  
Caterina Beatrice Monti ◽  
Christian Salvatore ◽  
...  
Keyword(s):  
Deep Learning ◽  
Area Under The Curve ◽  
Clinical Information ◽  
Training Dataset ◽  
Independent Dataset ◽  
X Ray ◽  
Learning Classifier ◽  
Chest X Ray ◽  
Polymerase Chain ◽  
Deep Learning Model

Abstract Background We aimed to train and test a deep learning classifier to support the diagnosis of coronavirus disease 2019 (COVID-19) using chest x-ray (CXR) on a cohort of subjects from two hospitals in Lombardy, Italy. Methods We used for training and validation an ensemble of ten convolutional neural networks (CNNs) with mainly bedside CXRs of 250 COVID-19 and 250 non-COVID-19 subjects from two hospitals (Centres 1 and 2). We then tested such system on bedside CXRs of an independent group of 110 patients (74 COVID-19, 36 non-COVID-19) from one of the two hospitals. A retrospective reading was performed by two radiologists in the absence of any clinical information, with the aim to differentiate COVID-19 from non-COVID-19 patients. Real-time polymerase chain reaction served as the reference standard. Results At 10-fold cross-validation, our deep learning model classified COVID-19 and non-COVID-19 patients with 0.78 sensitivity (95% confidence interval [CI] 0.74–0.81), 0.82 specificity (95% CI 0.78–0.85), and 0.89 area under the curve (AUC) (95% CI 0.86–0.91). For the independent dataset, deep learning showed 0.80 sensitivity (95% CI 0.72–0.86) (59/74), 0.81 specificity (29/36) (95% CI 0.73–0.87), and 0.81 AUC (95% CI 0.73–0.87). Radiologists’ reading obtained 0.63 sensitivity (95% CI 0.52–0.74) and 0.78 specificity (95% CI 0.61–0.90) in Centre 1 and 0.64 sensitivity (95% CI 0.52–0.74) and 0.86 specificity (95% CI 0.71–0.95) in Centre 2. Conclusions This preliminary experience based on ten CNNs trained on a limited training dataset shows an interesting potential of deep learning for COVID-19 diagnosis. Such tool is in training with new CXRs to further increase its performance.

Pure and Hybrid Deep Learning Models can Predict Pathologic Tumor Response to Neoadjuvant Therapy in Pancreatic Adenocarcinoma: A Pilot Study

2020 ◽  
pp. 000313482098255
Author(s):  
Michael D. Watson ◽  
Maria R. Baimas-George ◽  
Keith J. Murphy ◽  
Ryan C. Pickens ◽  
David A. Iannitti ◽  
...  
Keyword(s):  
Deep Learning ◽  
Neoadjuvant Therapy ◽  
Pancreatic Adenocarcinoma ◽  
Hybrid Model ◽  
Tumor Response ◽  
Tumor Regression ◽  
Area Under The Curve ◽  
Learning Model ◽  
Response To Therapy ◽  
Deep Learning Model

Background Neoadjuvant therapy may improve survival of patients with pancreatic adenocarcinoma; however, determining response to therapy is difficult. Artificial intelligence allows for novel analysis of images. We hypothesized that a deep learning model can predict tumor response to NAC. Methods Patients with pancreatic cancer receiving neoadjuvant therapy prior to pancreatoduodenectomy were identified between November 2009 and January 2018. The College of American Pathologists Tumor Regression Grades 0-2 were defined as pathologic response (PR) and grade 3 as no response (NR). Axial images from preoperative computed tomography scans were used to create a 5-layer convolutional neural network and LeNet deep learning model to predict PRs. The hybrid model incorporated decrease in carbohydrate antigen 19-9 (CA19-9) of 10%. Accuracy was determined by area under the curve. Results A total of 81 patients were included in the study. Patients were divided between PR (333 images) and NR (443 images). The pure model had an area under the curve (AUC) of .738 ( P < .001), whereas the hybrid model had an AUC of .785 ( P < .001). CA19-9 decrease alone was a poor predictor of response with an AUC of .564 ( P = .096). Conclusions A deep learning model can predict pathologic tumor response to neoadjuvant therapy for patients with pancreatic adenocarcinoma and the model is improved with the incorporation of decreases in serum CA19-9. Further model development is needed before clinical application.

External Validation of the Cardiac Surgery Score in a Quaternary Hospital in the United States of America

2021 ◽  
pp. 088506662110668
Author(s):  
Asha Singh ◽  
Chen Liang ◽  
Stephanie L. Mick ◽  
Chiedozie Udeh
Keyword(s):  
United States ◽  
Cardiac Surgery ◽  
United States Of America ◽  
Predictive Accuracy ◽  
External Validation ◽  
Cleveland Clinic ◽  
Area Under The Curve ◽  
The United States ◽  
Mortality Data ◽  
The Usa

Background The Cardiac Surgery Score (CASUS) was developed to assist in predicting post-cardiac surgery mortality using parameters measured in the intensive care unit. It is calculated by assigning points to ten physiologic variables and adding them to obtain a score (additive CASUS), or by logistic regression to weight the variables and estimate the probability of mortality (logistic CASUS). Both additive and logistic CASUS have been externally validated elsewhere, but not yet in the United States of America (USA). This study aims to validate CASUS in a quaternary hospital in the USA and compare the predictive performance of additive to logistic CASUS in this setting. Methods Additive and logistic CASUS (postoperative days 1-5) were calculated for 7098 patients at Cleveland Clinic from January 2015 to February 2017. 30-day mortality data were abstracted from institutional records and the Death Registries for Ohio State and the Centers for Disease Control. Given a low event rate, model discrimination was assessed by area under the curve (AUROC), partial AUROC (pAUC), and average precision (AP). Calibration was assessed by curves and quantified using Harrell's Emax, and Integrated Calibration Index (ICI). Results 30-day mortality rate was 1.37%. For additive CASUS, odds ratio for mortality was 1.41 (1.35-1.46, P <0.001). Additive and logistic CASUS had comparable pAUC and AUROC (all >0.83). However, additive CASUS had greater AP, especially on postoperative day 1 (0.22 vs. 0.11). Additive CASUS had better calibration curves, and lower Emax, and ICI on all days. Conclusions Additive and logistic CASUS discriminated well for postoperative 30-day mortality in our quaternary center in the USA, however logistic CASUS under-predicted mortality in our cohort. Given its ease of calculation, and better predictive accuracy, additive CASUS may be the preferred model for postoperative use. Validation in more typical cardiac surgery centers in the USA is recommended.

Compounding Racialized Vulnerability: COVID-19 in Prisons, Jails, and Migrant Detention Centers

2021 ◽  
Author(s):  
Matthew G.T. Denney ◽  
Ramon Garibaldo Valdez
Keyword(s):  
United States ◽  
Harm Reduction ◽  
Public Support ◽  
The United States ◽  
Correctional Facilities ◽  
Racial Inequalities ◽  
Detention Centers ◽  
Life Saving ◽  
Improved Sanitation ◽  
Immigrant Detention

Abstract Context: Carceral institutions are among the largest clusters of COVID-19 in the United States. As outbreaks have spread throughout prisons and detention centers, detainees have organized collectively to demand life-saving measures. Chief among these demands has been the call for decarceration: the release of detainees and inmates to prevent exposure to COVID-19. This paper theorizes the compounding racial vulnerability that has led to such a marked spread behind bars, mainly among race-class subjugated communities. Methods: We use journalistic sources and administrative data to provide an in-depth account of the spread of COVID-19 in American correctional facilities and of the mobilization to reduce contagions. We also use two survey experiments to describe public support for harm reduction and decarceration demands and measure the effects of information about (a) racial inequalities in prison, and (b) poor conditions inside migrant detention centers. Findings: We find that only one-third to one-half of respondents believe that response to COVID-19 in prisons and immigrant detention centers should be a high priority. We also find that Americans are much more supportive of harm reduction measures like improved sanitation than of releasing people from prisons and detention centers. Information about racial disparities increases support for releasing more people from prison. We do not find any significant effect of information about poor conditions in migrant detention centers. Conclusions: The conditions in prisons and migrant detention centers during the pandemic—and public opinion about them—highlight the realities of compounding racialized vulnerability in the United States.

scholarly journals Using Data Analytics to Predict Hospital Mortality in Sepsis Patients

2019 ◽  
Vol 14 (3) ◽  
pp. 40-57
Author(s):  
Yazan Alnsour ◽  
Rassule Hadidi ◽  
Neetu Singh
Keyword(s):  
Hospital Mortality ◽  
Predictive Model ◽  
Health Care Providers ◽  
Prediction Models ◽  
Predictive Analytics ◽  
Area Under The Curve ◽  
The United States ◽  
Hospital Death ◽  
Care Providers ◽  
Chi Square

Predictive analytics can be used to anticipate the risks associated with some patients, and prediction models can be employed to alert physicians and allow timely proactive interventions. Recently, health care providers have been using different types of tools with prediction capabilities. Sepsis is one of the leading causes of in-hospital death in the United States and worldwide. In this study, the authors used a large medical dataset to develop and present a model that predicts in-hospital mortality among Sepsis patients. The predictive model was developed using a dataset of more than one million records of hospitalized patients. The independent predictors of in-hospital mortality were identified using the chi-square automatic interaction detector. The authors found that adding hospital attributes to the predictive model increased the accuracy from 82.08% to 85.3% and the area under the curve from 0.69 to 0.84, which is favorable compared to using only patients' attributes. The authors discuss the practical and research contributions of using a predictive model that incorporates both patient and hospital attributes in identifying high-risk patients.

scholarly journals Quantified Kinematics to Evaluate Patient Chemotherapy Risks in Clinic

2020 ◽  
pp. 583-601
Author(s):  
Zaki Hasnain ◽  
Tanachat Nilanon ◽  
Ming Li ◽  
Aaron Mejia ◽  
Anand Kolatkar ◽  
...  
Keyword(s):  
Physical Activity ◽  
Health Care ◽  
Motion Capture ◽  
Performance Status ◽  
Area Under The Curve ◽  
The United States ◽  
Clinic Visit ◽  
Activity Levels ◽  
Health Care Encounters ◽  
The Right

PURPOSE Performance status (PS) is a key factor in oncologic decision making, but conventional scales used to measure PS vary among observers. Consumer-grade biometric sensors have previously been identified as objective alternatives to the assessment of PS. Here, we investigate how one such biometric sensor can be used during a clinic visit to identify patients who are at risk for complications, particularly unexpected hospitalizations that may delay treatment or result in low physical activity. We aim to provide a novel and objective means of predicting tolerability to chemotherapy. METHODS Thirty-eight patients across three centers in the United States who were diagnosed with a solid tumor with plans for treatment with two cycles of highly emetogenic chemotherapy were included in this single-arm, observational prospective study. A noninvasive motion-capture system quantified patient movement from chair to table and during the get-up-and-walk test. Activity levels were recorded using a wearable sensor over a 2-month period. Changes in kinematics from two motion-capture data points pre- and post-treatment were tested for correlation with unexpected hospitalizations and physical activity levels as measured by a wearable activity sensor. RESULTS Among 38 patients (mean age, 48.3 years; 53% female), kinematic features from chair to table were the best predictors for unexpected health care encounters (area under the curve, 0.775 ± 0.029) and physical activity (area under the curve, 0.830 ± 0.080). Chair-to-table acceleration of the nonpivoting knee ( t = 3.39; P = .002) was most correlated with unexpected health care encounters. Get-up-and-walk kinematics were most correlated with physical activity, particularly the right knee acceleration ( t = −2.95; P = .006) and left arm angular velocity ( t = −2.4; P = .025). CONCLUSION Chair-to-table kinematics are good predictors of unexpected hospitalizations, whereas the get-up-and-walk kinematics are good predictors of low physical activity.

Out-Of-Hospital Research in the Urban Environment

1993 ◽  
Vol 8 (S1) ◽  
pp. S20-S24
Author(s):  
Paul E. Pepe
Keyword(s):  
Prehospital Care ◽  
Large Population ◽  
Hospital Setting ◽  
The United States ◽  
Trauma Resuscitation ◽  
Urban Centers ◽  
Care Research ◽  
Life Saving ◽  
Artery Disease ◽  
Patient Studies

The urban prehospital setting is one of the best venues in which to examine life-saving resuscitation interventions. When the entire catchment of the urban emergency medical services (EMS) system is used, large-population patient studies can be generated. Certain unique features give several urban centers the ability to conduct clinical trials in the out-of-hospital setting. Without resuscitation at the scene, it is rare for cardiac arrest patients to survive. In the case of trauma resuscitation, prehospital care can impact outcome significantly. Since coronary artery disease and trauma kill nearly one-million persons annually in the United States, prehospital care research is a worthwhile endeavor. This rationale for prehospital care research is strengthened by the relatively high potential for full recovery.

scholarly journals Cry Wolf Effect? Evaluating the Impact of False Alarms on Public Responses to Tornado Alerts in the Southeastern United States

2019 ◽  
Vol 11 (3) ◽  
pp. 549-563 ◽  
Author(s):  
JungKyu Rhys Lim ◽  
Brooke Fisher Liu ◽  
Michael Egnoto
Keyword(s):  
United States ◽  
False Alarm ◽  
Protective Action ◽  
Southeastern United States ◽  
The United States ◽  
False Alarms ◽  
Protective Behavior ◽  
Tornado Warnings ◽  
Research Findings ◽  
The Impact

Abstract On average, 75% of tornado warnings in the United States are false alarms. Although forecasters have been concerned that false alarms may generate a complacent public, only a few research studies have examined how the public responds to tornado false alarms. Through four surveys (N = 4162), this study examines how residents in the southeastern United States understand, process, and respond to tornado false alarms. The study then compares social science research findings on perceptions of false alarms to actual county false alarm ratios and the number of tornado warnings issued by counties. Contrary to prior research, findings indicate that concerns about false alarm ratios generating a complacent public may be overblown. Results show that southeastern U.S. residents estimate tornado warnings to be more accurate than they are. Participants’ perceived false alarm ratios are not correlated with actual county false alarm ratios. Counterintuitively, the higher individuals perceive false alarm ratios and tornado alert accuracy to be, the more likely they are to take protective behavior such as sheltering in place in response to tornado warnings. Actual country false alarm ratios and the number of tornado warnings issued did not predict taking protective action.

Sign in / Sign up

Export Citation Format

Share Document