scholarly journals Quantifying the impact of addressing data challenges in prediction of length of stay

2021 ◽  
Vol 21 (1) ◽  
Author(s):  
Amin Naemi ◽  
Thomas Schmidt ◽  
Marjan Mansourvar ◽  
Ali Ebrahimi ◽  
Uffe Kock Wiil
Keyword(s):  
Length Of Stay ◽  
Predictive Models ◽  
Missing Values ◽  
Illness Severity ◽  
Support Vector ◽  
Admission Time ◽  
Address Data ◽  
Data Skewness ◽  
Classification And Regression ◽  
The Impact

Abstract Background Prediction of length of stay (LOS) at admission time can provide physicians and nurses insight into the illness severity of patients and aid them in avoiding adverse events and clinical deterioration. It also assists hospitals with more effectively managing their resources and manpower. Methods In this field of research, there are some important challenges, such as missing values and LOS data skewness. Moreover, various studies use a binary classification which puts a wide range of patients with different conditions into one category. To address these shortcomings, first multivariate imputation techniques are applied to fill incomplete records, then two proper resampling techniques, namely Borderline-SMOTE and SMOGN, are applied to address data skewness in the classification and regression domains, respectively. Finally, machine learning (ML) techniques including neural networks, extreme gradient boosting, random forest, support vector machine, and decision tree are implemented for both approaches to predict LOS of patients admitted to the Emergency Department of Odense University Hospital between June 2018 and April 2019. The ML models are developed based on data obtained from patients at admission time, including pulse rate, arterial blood oxygen saturation, respiratory rate, systolic blood pressure, triage category, arrival ICD-10 codes, age, and gender. Results The performance of predictive models before and after addressing missing values and data skewness is evaluated using four evaluation metrics namely receiver operating characteristic, area under the curve (AUC), R-squared score (R2), and normalized root mean square error (NRMSE). Results show that the performance of predictive models is improved on average by 15.75% for AUC, 32.19% for R2 score, and 11.32% for NRMSE after addressing the mentioned challenges. Moreover, our results indicate that there is a relationship between the missing values rate, data skewness, and illness severity of patients, so it is clinically essential to take incomplete records of patients into account and apply proper solutions for interpolation of missing values. Conclusion We propose a new method comprised of three stages: missing values imputation, data skewness handling, and building predictive models based on classification and regression approaches. Our results indicated that addressing these challenges in a proper way enhanced the performance of models significantly, which led to a more valid prediction of LOS.

scholarly journals Studying the Effect of Taking Statins before Infection in the Severity Reduction of COVID-19 with Machine Learning

2021 ◽  
Vol 2021 ◽  
pp. 1-12
Author(s):  
Alireza Davoudi ◽  
Mohsen Ahmadi ◽  
Abbas Sharifi ◽  
Roshina Hassantabar ◽  
Narges Najafi ◽  
...  
Keyword(s):  
Machine Learning ◽  
Diastolic Pressure ◽  
Systolic Pressure ◽  
Illness Severity ◽  
Support Vector ◽  
Learning Approaches ◽  
Angiotensin Converting Enzyme 2 ◽  
Decision Tree Method ◽  
The Impact

Statins can help COVID-19 patients’ treatment because of their involvement in angiotensin-converting enzyme-2. The main objective of this study is to evaluate the impact of statins on COVID-19 severity for people who have been taking statins before COVID-19 infection. The examined research patients include people that had taken three types of statins consisting of Atorvastatin, Simvastatin, and Rosuvastatin. The case study includes 561 patients admitted to the Razi Hospital in Ghaemshahr, Iran, during February and March 2020. The illness severity was encoded based on the respiratory rate, oxygen saturation, systolic pressure, and diastolic pressure in five categories: mild, medium, severe, critical, and death. Since 69.23% of participants were in mild severity condition, the results showed the positive effect of Simvastatin on COVID-19 severity for people that take Simvastatin before being infected by the COVID-19 virus. Also, systolic pressure for this case study is 137.31, which is higher than that of the total patients. Another result of this study is that Simvastatin takers have an average of 95.77 mmHg O2Sat; however, the O2Sat is 92.42, which is medium severity for evaluating the entire case study. In the rest of this paper, we used machine learning approaches to diagnose COVID-19 patients’ severity based on clinical features. Results indicated that the decision tree method could predict patients’ illness severity with 87.9% accuracy. Other methods, including the K -nearest neighbors (KNN) algorithm, support vector machine (SVM), Naïve Bayes classifier, and discriminant analysis, showed accuracy levels of 80%, 68.8%, 61.1%, and 85.1%, respectively.

Well Completion Optimization in Unconventional Reservoirs Using Machine Learning Methods

2021 ◽  
Author(s):  
Sohrat Baki ◽  
Cenk Temizel ◽  
Serkan Dursun
Keyword(s):  
Machine Learning ◽  
Predictive Models ◽  
Mapping Function ◽  
Volume Effect ◽  
Production Performance ◽  
Unconventional Reservoirs ◽  
Support Vector ◽  
Feature Engineering ◽  
Well Completion ◽  
The Impact

Abstract Unconventional reservoirs, mainly shale oil and natural gas, will continue to significantly help meet the ever-growing energy demands of global markets. Being complex in nature and having ultra-tight producing zones, unconventionals depends on effective well completion and stimulation treatments in order to be successful and economical. Within the last decade, thousands of unconventional wells have been drilled, completed and produced in North America. The scope of this work is exploring the primary impact of completion parameters such as lateral length, frac type, number of stages, proppant and fluid volume effect on the production performance of the wells in unconventional fields. The key attributes in completion, stimulation, and production for the wells were considered in machine learning workflow for building predictive models. Predictive models based on Neural Networks, Support Vector Machines or Decision Tree Based ensemble models, serves as mapping function from completion parameters to production in each well in the field. The completion parameters were analyzed in the workflow with respect to feature engineering and interpretation. This analysis resulted in key performance indicators for the region. Then the optimum values for the best production performing completions were identified for each well. Predictive models in the workflow were analyzed in accuracy and best model is used to understand the impact of completion parameters on the production rates. This study outlines an overall machine learning workflow, from feature engineering to interpretation of the machine learning models to quantify the effects of completion parameters on the production rate of the wells in unconventional fields

Admission time to hospital: a varying standard for a critical definition for admissions to an intensive care unit from the emergency department

2014 ◽  
Vol 38 (5) ◽  
pp. 575
Author(s):  
Shane Nanayakkara ◽  
Heike Weiss ◽  
Michael Bailey ◽  
Allison van Lint ◽  
Peter Cameron ◽  
...  
Keyword(s):  
Intensive Care Unit ◽  
Intensive Care ◽  
New Zealand ◽  
Length Of Stay ◽  
Performance Indicators ◽  
Decision Method ◽  
Admission Time ◽  
Definition Of ◽  
Time Of Admission ◽  
The Impact

Objective Time spent in the emergency department (ED) before admission to hospital is often considered an important key performance indicator (KPI). Throughout Australia and New Zealand, there is no standard definition of ‘time of admission’ for patients admitted through the ED. By using data submitted to the Australian and New Zealand Intensive Care Society Adult Patient Database, the aim was to determine the differing methods used to define hospital admission time and assess how these impact on the calculation of time spent in the ED before admission to an intensive care unit (ICU). Methods Between March and December of 2010, 61 hospitals were contacted directly. Decision methods for determining time of admission to the ED were matched to 67787 patient records. Univariate and multivariate analyses were conducted to assess the relationship between decision method and the reported time spent in the ED. Results Four mechanisms of recording time of admission were identified, with time of triage being the most common (28/61 hospitals). Reported median time spent in the ED varied from 2.5 (IQR 0.83–5.35) to 5.1 h (2.82–8.68), depending on the decision method. After adjusting for illness severity, hospital type and location, decision method remained a significant factor in determining measurement of ED length of stay. Conclusions Different methods are used in Australia and New Zealand to define admission time to hospital. Professional bodies, hospitals and jurisdictions should ensure standardisation of definitions for appropriate interpretation of KPIs as well as for the interpretation of studies assessing the impact of admission time to ICU from the ED. What is known about the topic? There are standards for the maximum time spent in the ED internationally, but these standards vary greatly across Australia. The definition of such a standard is critically important not only to patient care, but also in the assessment of hospital outcomes. Key performance indicators rely on quality data to improve decision-making. What does this paper add? This paper quantifies the variability of times measured and analyses why the variability exists. It also discusses the impact of this variability on assessment of outcomes and provides suggestions to improve standardisation. What are the implications for practitioners? This paper provides a clearer view on standards regarding length of stay in the ICU, highlighting the importance of key performance indicators, as well as the quality of data that underlies them. This will lead to significant changes in the way we standardise and interpret data regarding length of stay.

scholarly journals The Impact of Missing Value Imputation on the Interpretations of Predictive Models: A Case Study on One-year Mortality Prediction in ICU Patients with Acute Myocardial Infarction

2020 ◽  
Author(s):  
Seyedeh Neelufar Payrovnaziri ◽  
Aiwen Xing ◽  
Salman Shaeke ◽  
Xiuwen Liu ◽  
Jiang Bian ◽  
...  
Keyword(s):  
Myocardial Infarction ◽  
Acute Myocardial Infarction ◽  
Standard Deviation ◽  
Predictive Models ◽  
Missing Values ◽  
Prediction Models ◽  
Mean Squared Error ◽  
Mortality Prediction ◽  
One Year ◽  
The Impact

AbstractAcute Myocardial Infarction (AMI) is responsible for the death of millions of people annually around the world, which makes predictive analyses of AMI mortality risk necessary. Rich clinical data in electronic health records (EHR) makes such predictive modeling possible. However, missing values in EHR data is a major issue. Also, the interpretability of predictive models in medicine and healthcare is vital for medical professionals. Therefore, this study examines the impact of imputing missing values in EHR data on the performance and interpretations of predictive models. Our experiments showed a small standard deviation in root mean squared error of different runs of imputation under similar method does not necessarily imply small standard deviation in prediction models’ performance and interpretation. Our findings reveal that the imputation method and the level of missingness impact not only the predictive models’ performance but also the interpretation of the models in terms of feature importance.

Research on behavior recognition based on feature fusion of automatic coder and recurrent neural network

2020 ◽  
Vol 39 (6) ◽  
pp. 8927-8935
Author(s):  
Bing Zheng ◽  
Dawei Yun ◽  
Yan Liang
Keyword(s):  
Neural Network ◽  
Recurrent Neural Network ◽  
Behavior Pattern ◽  
Rapid Development ◽  
Video Data ◽  
Support Vector ◽  
Behavior Recognition ◽  
Learning Methods ◽  
The Impact ◽  
Internet Of Things Technology

Under the impact of COVID-19, research on behavior recognition are highly needed. In this paper, we combine the algorithm of self-adaptive coder and recurrent neural network to realize the research of behavior pattern recognition. At present, most of the research of human behavior recognition is focused on the video data, which is based on the video number. At the same time, due to the complexity of video image data, it is easy to violate personal privacy. With the rapid development of Internet of things technology, it has attracted the attention of a large number of experts and scholars. Researchers have tried to use many machine learning methods, such as random forest, support vector machine and other shallow learning methods, which perform well in the laboratory environment, but there is still a long way to go from practical application. In this paper, a recursive neural network algorithm based on long and short term memory (LSTM) is proposed to realize the recognition of behavior patterns, so as to improve the accuracy of human activity behavior recognition.

The Institute for Healthcare Improvement–NeuroPoint Alliance collaboration to decrease length of stay and readmission after lumbar spine fusion: using national registries to design quality improvement protocols

2020 ◽  
Vol 33 (6) ◽  
pp. 812-821
Author(s):  
Scott L. Zuckerman ◽  
Clinton J. Devin ◽  
Vincent Rossi ◽  
Silky Chotai ◽  
E. Hunter Dyer ◽  
...  
Keyword(s):  
Quality Improvement ◽  
Length Of Stay ◽  
Discharge Planning ◽  
Lumbar Fusion ◽  
Nonparametric Tests ◽  
Learning System ◽  
Design Quality ◽  
Implementation Phase ◽  
Healthcare Improvement ◽  
The Impact

OBJECTIVENational databases collect large amounts of clinical information, yet application of these data can be challenging. The authors present the NeuroPoint Alliance and Institute for Healthcare Improvement (NPA-IHI) program as a novel attempt to create a quality improvement (QI) tool informed through registry data to improve the quality of care delivered. Reducing the length of stay (LOS) and readmission after elective lumbar fusion was chosen as the pilot module.METHODSThe NPA-IHI program prospectively enrolled patients undergoing elective 1- to 3-level lumbar fusions across 8 institutions. A three-pronged approach was taken that included the following phases: 1) Research Phase, 2) Development Phase, and 3) Implementation Phase. Primary outcomes were LOS and readmission. From January to June 2017, a learning system was created utilizing monthly conference calls, weekly data submission, and continuous refinement of the proposed QI tool. Nonparametric tests were used to assess the impact of the QI intervention.RESULTSThe novel QI tool included the following three areas of intervention: 1) preoperative discharge assessment (location, date, and instructions), 2) inpatient changes (LOS rounding checklist, daily huddle, and pain assessments), and 3) postdischarge calls (pain, primary care follow-up, and satisfaction). A total of 209 patients were enrolled, and the most common procedure was a posterior laminectomy/fusion (60.2%). Seven patients (3.3%) were readmitted during the study period. Preoperative discharge planning was completed for 129 patients (61.7%). A shorter median LOS was seen in those with a known preoperative discharge date (67 vs 80 hours, p = 0.018) and clear discharge instructions (71 vs 81 hours, p = 0.030). Patients with a known preoperative discharge plan also reported significantly increased satisfaction (8.0 vs 7.0, p = 0.028), and patients with increased discharge readiness (scale 0–10) also reported higher satisfaction (r = 0.474, p < 0.001). Those receiving postdischarge calls (76%) had a significantly shorter LOS than those without postdischarge calls (75 vs 99 hours, p = 0.020), although no significant relationship was seen between postdischarge calls and readmission (p = 0.342).CONCLUSIONSThe NPA-IHI program showed that preoperative discharge planning and postdischarge calls have the potential to reduce LOS and improve satisfaction after elective lumbar fusion. It is our hope that neurosurgical providers can recognize how registries can be used to both develop and implement a QI tool and appreciate the importance of QI implementation as a separate process from data collection/analysis.

Big Data to Knowledge: Application of Machine Learning to Predictive Modeling of Therapeutic Response in Cancer.

2020 ◽  
Vol 21 ◽  
Author(s):  
Sukanya Panja ◽  
Sarra Rahem ◽  
Cassandra J. Chu ◽  
Antonina Mitrofanova
Keyword(s):  
Machine Learning ◽  
Missing Values ◽  
Therapeutic Response ◽  
Patient Data ◽  
Machine Learning Techniques ◽  
Support Vector ◽  
Complex Data ◽  
Human Machine Interaction ◽  
Data Repositories ◽  
Response Modeling

Background: In recent years, the availability of high throughput technologies, establishment of large molecular patient data repositories, and advancement in computing power and storage have allowed elucidation of complex mechanisms implicated in therapeutic response in cancer patients. The breadth and depth of such data, alongside experimental noise and missing values, requires a sophisticated human-machine interaction that would allow effective learning from complex data and accurate forecasting of future outcomes, ideally embedded in the core of machine learning design. Objective: In this review, we will discuss machine learning techniques utilized for modeling of treatment response in cancer, including Random Forests, support vector machines, neural networks, and linear and logistic regression. We will overview their mathematical foundations and discuss their limitations and alternative approaches all in light of their application to therapeutic response modeling in cancer. Conclusion: We hypothesize that the increase in the number of patient profiles and potential temporal monitoring of patient data will define even more complex techniques, such as deep learning and causal analysis, as central players in therapeutic response modeling.

11 Vitamin D Deficiency on Admission and the Association with Length of Stay in Patients with Thermal Injury: A Multi-center Analysis

2021 ◽  
Vol 42 (Supplement_1) ◽  
pp. S13-S14
Author(s):  
Sarah Zavala ◽  
Kate Pape ◽  
Todd A Walroth ◽  
Melissa A Reger ◽  
Katelyn Garner ◽  
...  
Keyword(s):  
Vitamin D ◽  
Length Of Stay ◽  
Vitamin D Deficiency ◽  
Burn Injury ◽  
Total Body Surface Area ◽  
Inhalation Injury ◽  
Hospital Length Of Stay ◽  
Burn Patients ◽  
Comfort Care ◽  
The Impact

Abstract Introduction In burn patients, vitamin D deficiency has been associated with increased incidence of sepsis. The objective of this study was to assess the impact of vitamin D deficiency in adult burn patients on hospital length of stay (LOS). Methods This was a multi-center retrospective study of adult patients at 7 burn centers admitted between January 1, 2016 and July 25, 2019 who had a 25-hydroxyvitamin D (25OHD) concentration drawn within the first 7 days of injury. Patients were excluded if admitted for a non-burn injury, total body surface area (TBSA) burn less than 5%, pregnant, incarcerated, or made comfort care or expired within 48 hours of admission. The primary endpoint was to compare hospital LOS between burn patients with vitamin D deficiency (defined as 25OHD &lt; 20 ng/mL) and sufficiency (25OHD ≥ 20 ng/mL). Secondary endpoints include in-hospital mortality, ventilator-free days of the first 28, renal replacement therapy (RRT), length of ICU stay, and days requiring vasopressors. Additional data collected included demographics, Charlson Comorbidity Index, injury characteristics, form of vitamin D received (ergocalciferol or cholecalciferol) and dosing during admission, timing of vitamin D initiation, and form of nutrition provided. Dichotomous variables were compared via Chi-square test. Continuous data were compared via student t-test or Mann-Whitney U test. Univariable linear regression was utilized to identify variables associated with LOS (p &lt; 0.05) to analyze further. Cox Proportional Hazard Model was utilized to analyze association with LOS, while censoring for death, and controlling for TBSA, age, presence of inhalation injury, and potential for a center effect. Results Of 1,147 patients screened, 412 were included. Fifty-seven percent were vitamin D deficient. Patients with vitamin D deficiency had longer LOS (18.0 vs 12.0 days, p &lt; 0.001), acute kidney injury (AKI) requiring RRT (7.3 vs 1.7%, p = 0.009), more days requiring vasopressors (mean 1.24 vs 0.58 days, p = 0.008), and fewer ventilator free days of the first 28 days (mean 22.9 vs 25.1, p &lt; 0.001). Univariable analysis identified burn center, AKI, TBSA, inhalation injury, admission concentration, days until concentration drawn, days until initiating supplementation, and dose as significantly associated with LOS. After controlling for center, TBSA, age, and inhalation injury, the best fit model included only deficiency and days until vitamin D initiation. Conclusions Patients with thermal injuries and vitamin D deficiency on admission have increased length of stay and worsened clinical outcomes as compared to patients with sufficient vitamin D concentrations.

scholarly journals A novel GPU based intrusion detection system using deep autoencoder with Fruitfly optimization

2021 ◽  
Vol 3 (6) ◽  
Author(s):  
R. Sekhar ◽  
K. Sasirekha ◽  
P. S. Raja ◽  
K. Thangavel
Keyword(s):  
Intrusion Detection ◽  
Intrusion Detection System ◽  
Missing Values ◽  
Detection System ◽  
Radial Basis Function Network ◽  
Detection Methods ◽  
Support Vector ◽  
Parameter Method ◽  
List Type ◽  
Fuzzy C Means

Abstract Intrusion Detection Systems (IDSs) have received more attention to safeguarding the vital information in a network system of an organization. Generally, the hackers are easily entering into a secured network through loopholes and smart attacks. In such situation, predicting attacks from normal packets is tedious, much challenging, time consuming and highly technical. As a result, different algorithms with varying learning and training capacity have been explored in the literature. However, the existing Intrusion Detection methods could not meet the desired performance requirements. Hence, this work proposes a new Intrusion Detection technique using Deep Autoencoder with Fruitfly Optimization. Initially, missing values in the dataset have been imputed with the Fuzzy C-Means Rough Parameter (FCMRP) algorithm which handles the imprecision in datasets with the exploit of fuzzy and rough sets while preserving crucial information. Then, robust features are extracted from Autoencoder with multiple hidden layers. Finally, the obtained features are fed to Back Propagation Neural Network (BPN) to classify the attacks. Furthermore, the neurons in the hidden layers of Deep Autoencoder are optimized with population based Fruitfly Optimization algorithm. Experiments have been conducted on NSL_KDD and UNSW-NB15 dataset. The computational results of the proposed intrusion detection system using deep autoencoder with BPN are compared with Naive Bayes, Support Vector Machine (SVM), Radial Basis Function Network (RBFN), BPN, and Autoencoder with Softmax. Article Highlights A hybridized model using Deep Autoencoder with Fruitfly Optimization is introduced to classify the attacks. Missing values have been imputed with the Fuzzy C-Means Rough Parameter method. The discriminate features are extracted using Deep Autoencoder with more hidden layers.

scholarly journals Association of Emergency Department Waiting Times With Patient Experience in Admitted and Discharged Patients

2021 ◽  
Vol 8 ◽  
pp. 237437352110114
Author(s):  
Andrew Nyce ◽  
Snehal Gandhi ◽  
Brian Freeze ◽  
Joshua Bosire ◽  
Terry Ricca ◽  
...  
Keyword(s):  
Emergency Department ◽  
Length Of Stay ◽  
Patient Experience ◽  
Waiting Times ◽  
Hospital Length ◽  
Short Length ◽  
Hospital Length Of Stay ◽  
Patient’S Experience ◽  
Discharged Patients ◽  
The Impact

Prolonged waiting times are associated with worse patient experience in patients discharged from the emergency department (ED). However, it is unclear which component of the waiting times is most impactful to the patient experience and the impact on hospitalized patients. We performed a retrospective analysis of ED patients between July 2018 and March 30, 2020. In all, 3278 patients were included: 1477 patients were discharged from the ED, and 1680 were admitted. Discharged patients had a longer door-to-first provider and door-to-doctor time, but a shorter doctor-to-disposition, disposition-to-departure, and total ED time when compared to admitted patients. Some, but not all, components of waiting times were significantly higher in patients with suboptimal experience (<100th percentile). Prolonged door-to-doctor time was significantly associated with worse patient experience in discharged patients and in patients with hospital length of stay ≤4 days. Prolonged ED waiting times were significantly associated with worse patient experience in patients who were discharged from the ED and in inpatients with short length of stay. Door-to-doctor time seems to have the highest impact on the patient’s experience of these 2 groups.

Sign in / Sign up

Export Citation Format

Share Document