scholarly journals Predictive modeling of hospital Length of Stay in COVID-19 patients using Artificial Neural Networks  

2021 ◽  
Author(s):  
Azam Orooji ◽  
Mostafa Shanbehzadeh ◽  
Hadi Kazemi-Arpanahi ◽  
Mohsen Shafiee
Keyword(s):  
Neural Network ◽  
Length Of Stay ◽  
Correlation Coefficient ◽  
Prediction Models ◽  
Hospital Length ◽  
Hospital Length Of Stay ◽  
P Value ◽  
Bed Management ◽  
Hospital Bed ◽  
Artificial Neural

Abstract BackgroundThe current pandemic of coronavirus disease (COVID-19) causes unexpected economic burdens to worldwide health organizations with severe shortages in hospital bed capacity and other related medical resources. Therefore, predicting the length of stay (LOS) is essential to ensure optimal allocating scarce hospital resources and inform evidence-based decision-making. Thus, the purpose of this research is to construct a model for predicting COVID-19 patients' hospital LOS by multiple multilayer perceptron-artificial neural network (MLP-ANN) algorithms. Material and MethodsUsing a single-center registry, the records of 1225 laboratory-confirmed COVID-19 hospitalized cases from February 9, 2020, to December 20, 2020, were analyzed. The correlation coefficient technique was developed to determine the most significant variables as the input of the ANN models. Only variables with a correlation coefficient at the P-value< 0.2 were used in model construction. Ultimately the prediction models were developed based on 12 ANN techniques according to selected variables. ResultsAfter implementing feature selection, a total of 20 variables was determined as the most relevant predictors to build the models. The results indicated that the best performance belongs to a neural network with 20 and 10 neurons in the hidden layer of the Bayesian Regularization classifier for whole and selected features with RMSE of 1.6213 and 2.2332, respectively. ConclusionThe developed model in this study can help in the better calculation of LOS in COVID-19 patients. This model also can be leveraged in hospital bed management and optimized resource utilization.

scholarly journals Multimodal Analgesia in Lumbar Fusion Surgery: Impact Amidst the Opioid Epidemic

Neurosurgery ◽  
2019 ◽  
Vol 66 (Supplement_1) ◽  
Author(s):  
Yi Lu ◽  
Erica Bertoncini
Keyword(s):  
Length Of Stay ◽  
Lumbar Fusion ◽  
Multimodal Analgesia ◽  
Hospital Length ◽  
Hospital Length Of Stay ◽  
Control Group ◽  
P Value ◽  
Standard Group ◽  
Fusion Surgery ◽  
Hospital Stays

Abstract INTRODUCTION Spine surgery traditionally relies on opioid analgesics for postoperative pain management. Opioids are associated with prolonged hospital stays and opioid use disorders. Opioid-focused prescribing habits in surgery have partially contributed to the opioid epidemic. METHODS A retrospective analysis was performed comparing patients receiving a multimodal analgesia regimen after lumbar fusion surgery vs control group receiving standard analgesia regimen. The multimodal regimen consisted of Acetaminophen 975 mg TID, Toradol 7.5 mg Q6 hours for 24-ho followed by Celebrex 100 mg BID for 7-d, Robaxin 500 mg Q6 hours prn for muscle spasms, Gabapentin 300 mg/100 mg TID for 4-wk, and prn narcotic. The standard regimen consisted of Acetaminophen 975 mg TID, narcotic prn, and muscle relaxant prn. There were 12 patients in the multimodal group and 26 patients in the control group evaluated over 3-mo and 6-mo time periods respectively. Primary outcomes included hospital length-of-stay, total and IV narcotic requirements in Morphine Milligram Equivalent (MME), and VASS pain scores. RESULTS Study results demonstrate differences between patient populations when focusing on the opioid-naïve participants. Opioid-naïve patients in the multimodal group were found to have significantly lower IV narcotic requirement than the control (0.22+/−0.67 mg/d for multimodal vs 5.36+/−5.56 mg/d for standard group, P-value = .001). These patients also had shorter hospital stays than the control (2.78+/−0.83 d for multimodal vs 3.53+/−1.17 d for standard group) but the difference was just below our threshold for significance (P-value = .066). Including both opioid-naïve and opioid-tolerant patients, no significant differences were found in hospital length-of-stay, MME, IV narcotic requirement nor VASS score between the multimodal group and the control groups (P-values of .46, .81, .36, and .91, respectively). CONCLUSION Overall, the study favors using multimodal analgesia in those undergoing lumbar spinal fusion surgeries as evident by considerably reduced IV narcotic requirement and nearly significant shortened hospital length-of-stay in opioid-naïve patients compared to control.

scholarly journals Neural networks and hospital length of stay: an application to support healthcare management with national benchmarks and thresholds

2021 ◽  
Vol 19 (1) ◽  
Author(s):  
Roberto Ippoliti ◽  
Greta Falavigna ◽  
Cristian Zanelli ◽  
Roberta Bellini ◽  
Gianmauro Numico
Keyword(s):  
Neural Networks ◽  
Artificial Neural Networks ◽  
Length Of Stay ◽  
Prediction Models ◽  
Average Length ◽  
Logistic Models ◽  
Healthcare Management ◽  
Hospital Length Of Stay ◽  
North West ◽  
Artificial Neural

Abstract Background The problem of correct inpatient scheduling is extremely significant for healthcare management. Extended length of stay can have negative effects on the supply of healthcare treatments, reducing patient accessibility and creating missed opportunities to increase hospital revenues by means of other treatments and additional hospitalizations. Methods Adopting available national reference values and focusing on a Department of Internal and Emergency Medicine located in the North-West of Italy, this work assesses prediction models of hospitalizations with length of stay longer than the selected benchmarks and thresholds. The prediction models investigated in this case study are based on Artificial Neural Networks and examine risk factors for prolonged hospitalizations in 2018. With respect current alternative approaches (e.g., logistic models), Artificial Neural Networks give the opportunity to identify whether the model will maximize specificity or sensitivity. Results Our sample includes administrative data extracted from the hospital database, collecting information on more than 16,000 hospitalizations between January 2018 and December 2019. Considering the overall department in 2018, 40% of the hospitalizations lasted more than the national average, and almost 3.74% were outliers (i.e., they lasted more than the threshold). According to our results, the adoption of the prediction models in 2019 could reduce the average length of stay by up to 2 days, guaranteeing more than 2000 additional hospitalizations in a year. Conclusions The proposed models might represent an effective tool for administrators and medical professionals to predict the outcome of hospital admission and design interventions to improve hospital efficiency and effectiveness.

scholarly journals HOSPITAL LENGTH OF STAY PREDICTION BASED ON PATIENT EXAMINATION USING NEURAL NETWORK

2021 ◽  
Vol 8 (1) ◽  
pp. 28
Author(s):  
Rabiatul Adawiyah
Keyword(s):  
Neural Network ◽  
Machine Learning ◽  
Length Of Stay ◽  
Health Services ◽  
Parameter Optimization ◽  
Hospital Length ◽  
Learning Rate ◽  
Hospital Length Of Stay ◽  
Grid Search ◽  
Accuracy Rate

<p>Length of stay (LOS) or length of stay is the main indicator in improving health services, which is expected to continue to increase along with population growth. The population of Indonesia is included in the largest category in the world. This was followed by an increase in the number of inpatient and emergency unit visits which had a high burden of care costs. This study aims to provide a solution by predicting LOS using Neural Network (NN). Predictions can be used as a consideration in improving effective and efficient health services. To improve the performance of the NN algorithm, we implement parameter optimization using the Grid Search to find the combination of the number of epochs, learning rate and momentum that can produce the best accuracy value. The results showed that NN could predict LOS with an accuracy rate of 89.22% when using the default parameter. Meanwhile, by performing parameter optimization using the Grid Search to find the ideal combination of parameters for learning rate, momentum and epoch, the accuracy rate is increased to 92.20%.</p><p><strong>Keywords</strong>: <em>length of stay, neural network, patient examination, machine learning</em></p><p><em>Length of st</em>ay (LOS) atau lama rawat inap merupakan indikator utama dalam peningkatan pelayanan kesehatan yang diperkirakan akan terus meningkat bersamaan dengan jumlah pertumbuhan penduduk. Jumlah penduduk Indonesia termasuk dalam kategori terbanyak di dunia. Hal ini diikuti dengan pertambahan jumlah kunjungan pasien rawat inap dan unit gawat darurat yang memiliki beban biaya perawatan yang tinggi. Penelitian ini bertujuan untuk memberikan solusi dengan melakukan prediksi LOS menggunakan Neural Network (NN). Prediksi dapat digunakan sebagai pertimbangan dalam peningkatan pelayanan kesehatan yang efektif dan efisien. Untuk meningkatkan performansi dari algoritma NN, kami menerapkan optimasi parameter menggunakan Grid Search untuk menemukan kombinasi jumlah epoch, learning rate dan momentum yang dapat menghasilkan nilai akurasi terbaik. Hasil penelitian menunjukkan bahwa NN dapat memprediksi LOS dengan tingkat akurasi 89,22% jika menggunakan default parameter. Sedangkan dengan melakukan optimasi parameter menggunakan Grid Search untuk menemukan kombinasi ideal parameter learning rate, momentum dan epoch, maka tingkat akurasi meningkat menjadi 92,20%.</p><p><strong>Kata kunci</strong>:<em> length of stay, neural network, patient examination, machine learning</em></p>

Abstract 15580: The Impact of Sex and Gender-related Factors on Length-of-stay Following NSTEMI: A Multicountry Analysis

Circulation ◽  
2020 ◽  
Vol 142 (Suppl_3) ◽  
Author(s):  
Uri Bender ◽  
Colleen M Norris ◽  
Valeria Raparelli ◽  
Tadiri Christina ◽  
Louise Pilote
Keyword(s):  
Length Of Stay ◽  
Cardiac Risk ◽  
Hospital Length ◽  
Hospital Length Of Stay ◽  
Bivariate Analysis ◽  
P Value ◽  
Sex And Gender ◽  
Related Factors ◽  
And Gender ◽  
The Impact

Introduction: Gender refers to psycho-socio-cultural characteristics typically ascribed to men, women and gender-diverse individuals and has been shown to be associated with adverse clinical outcomes in AMI independent of sex. Substantial heterogeneity in hospital length of stay exists among patients admitted with NSTEMI. Whether sex and gender-based differences contribute to length-of-stay (LOS) among patients with NSTEMI remains unknown. Methods: To examine the relationship between sex, gender-related factors and LOS in adults hospitalized for NSTEMI, data from the GENESIS-PRAXY (n=1,210, Canada, U.S. and Switzerland), EVA (n=430, Italy) and VIRGO (n=3,572, U.S., Spain and Australia) studies of adults hospitalized for AMI were combined and analyzed. A best-fit linear regression model was selected through incremental analysis by stepwise addition of gender-related variables thought to be different in either impact or distribution between men and women. Results: Among the overall cohort (n=5,212), 2,218 participants with a diagnosis of NSTEMI were included in the final cohort (66% women, mean age 48.5 years, 67.8% U.S.). Half of the patients had a LOS of longer than 4 days (n=1,124) and were more likely to be white and have a clustering of cardiac risk factors in comparison to those with shorter LOS. No association between sex and LOS was observed in the bivariate analysis (p=0.87). In the multivariable model adjusted for sex, age, country of hospitalization, level of education, marital status, employment status, income, and social support, age (0.062 days/year, p=0.0002), being employed (-0.63 days in workers, p=0.01) and the treatment country relative to Canada (Italy=4.1 days; Spain=1.7 days; and the U.S.=-1.0 days, all p-value<0.001) were significant predictors of LOS. Conclusions: Employed individuals are more likely to experience a shorter LOS following NSTEMI. Variation in LOS exists across different countries and is likely due to institutional policy, resource allocation, and differences in cultural and psychosocial influences.

Abstract P400: Racial Disparities in Subarachnoid Hemorrhage Outcomes-A Single Center Experience

Stroke ◽  
2021 ◽  
Vol 52 (Suppl_1) ◽  
Author(s):  
Jerome Deas ◽  
Eyad Almallouhi ◽  
Chirantan Banerjee
Keyword(s):  
Subarachnoid Hemorrhage ◽  
Length Of Stay ◽  
Hospital Length ◽  
University Hospital ◽  
Hospital Length Of Stay ◽  
P Value ◽  
High Morbidity ◽  
Chi Square ◽  
Single Center Experience ◽  
Discharge Outcomes

Introduction: Subarachnoid hemorrhage (SAH) has high morbidity and mortality, and prior studies have reported outcome disparities between African American (AA) and Caucasian patients. We compared demographics, risk factors, and discharge outcomes among different ethnicities treated at our comprehensive stroke center. Methods: We used data on all SAH patients admitted between July 2014 and March 2020 to our university hospital in the Southeast United States. Race was categorized into AA, Caucasian, and “other.” Pearson chi-square test and analysis of variance were used to compare these variables between the different groups. Results: A total of 578 SAH patients were identified (39% AA patients, 54% Caucasian, and 7% other). Admission Glascow Coma Score (GCS) and Hunt & Hess scores were comparable between the 3 groups. AA patients were significantly younger (51 vs 59 in Caucasian group vs 56 years in Other, p-value <0.001) and had higher BP at admission (systolic BP 152 vs 144 vs 145, p=0.002, diastolic BP 86 vs 80 vs 81, p<0.001). AA patients were more likely to have a history of hypertension (p<0.001) and had higher BMI (30 vs 28.1 vs 26, p=0.003) and Hemoglobin A1c (5.8 vs 5.6 vs 6.1, p=0.013). Modified Rankin scale (mRS) at discharge, in-hospital mortality, and discharge destination were similar between the groups, but AA patients had a longer mean hospital length of stay (19 vs 14 vs 17 days, p=0.035). Conclusion: In our cohort, AA SAH patients were significantly younger and had more comorbidities at admission. Although they had a higher length of stay, discharge outcomes were comparable to other races.

Abstract P245: Increased Hospital Length of Stay as a Predictor of Heart Failure Readmission

Circulation ◽  
2012 ◽  
Vol 125 (suppl_10) ◽  
Author(s):  
Yao Djilan ◽  
Robert G Zoble ◽  
Philip Foulis ◽  
Adam Zoble ◽  
Britta I Neugaard
Keyword(s):  
Heart Failure ◽  
Length Of Stay ◽  
Medical Center ◽  
Hospital Length ◽  
Hospital Length Of Stay ◽  
P Value ◽  
Proportional Hazard ◽  
Hazard Regression ◽  
Cox Proportional Hazard ◽  
Cox Proportional Hazard Regression

Background: Heart Failure (HF) is a major cause of morbidity and mortality with a global prevalence estimated at 16 million. Understanding the risk factors and predictors assists in identifying patients at high risk of subsequent re-hospitalization and death. Consequently, we hypothesized that hospital Length of Stay (LOS) can predict the risk of HF patients’ readmission within a year from an indexed HF hospitalization. Methods: A total of 1,255 HF patients admitted to a VA medical center between January 2004 and December 2009, were identified with IDC-9 code for HF. In order to reduce selection bias, patients were required to have at least on HF readmission from the index HF hospitalization. LOS was categorized in two groups (less than 7 days and more than 7days). The study sample was analyzed using Kaplan Meier (KM) estimate and Cox proportional hazard regression in order to assess the risk of patients’ readmission within a year from their indexed HF hospitalization at a significance level of 0.05. Results: Among the HF patients identified, 511 met the inclusion criteria. The KM curves below show that patients with LOS greater than 7days are more likely to be readmitted compared to those with LOS less than 7 days. Cox proportional hazard regression confirmed this finding with a hazard ratio of 1.39 (CI: 1.08 – 1.79) and a p-value = 0.012 while adjusting for age, tobacco use, diabetes and medications at first HF discharge. Conclusion: From this study sample, we can predict that patients with LOS greater than 7 days are 1.39 times more likely to be readmitted within a year compared to those with LOS less than 7 days.

Abstract 17103: Elevated NT-proBNP Level is Not a Prognostic Marker for Patients Hospitalized With SARS-CoV-2 Infection

Circulation ◽  
2020 ◽  
Vol 142 (Suppl_3) ◽  
Author(s):  
Farukh G Ikram ◽  
Priyanka Acharya ◽  
Nicholas Nguyen ◽  
Manavjot S Sidhu
Keyword(s):  
Length Of Stay ◽  
Surrogate Marker ◽  
Hospital Length ◽  
Hospital Length Of Stay ◽  
Major Adverse Cardiovascular Events ◽  
P Value ◽  
Chi Square ◽  
Exact Test ◽  
Laboratory Marker ◽  
Icu Admission

Introduction: NT-proBNP is a widely available and frequently utilized laboratory marker in patients with congestive heart failure. Its availability and use as a surrogate marker for elevated ventricular filling pressures and ventricular strain that frequently manifests as shortness of breath makes it an attractive laboratory marker to triage patients with SARS-CoV-2 infection. Hypothesis: Elevated NT-proBNP is correlated with higher mortality and/or rate of major adverse cardiovascular events (MACE) in patients admitted with SARS-CoV-2 infection. Methods: A retrospective analysis was performed on 225 patients admitted for SARS-CoV-2 infection at a major metropolitan hospital located in the Southwestern United States from the period of March 2020 to May 2020. NT-proBNP levels were recorded in 117 patients (52.7%) on admission. Elevated NT-proBNP was defined as: above 450 pg/ml in patients less than 50 years old, above 900 pg/ml if 50 to 75 years old, and above 1,800 pg/ml if above 75 years old. The primary endpoint was a composite of MACE and mortality during hospitalization. MACE was defined as stroke, myocardial infarction, DVT or PE and shock requiring vasopressor support. Two-sample Wilcoxon rank-sum (Mann-Whitney) test, Pearson’s chi square test and Fisher’s exact test were utilized for data analysis. Results: Of the 117 SARS-CoV-2 positive patients that had admission NT-proBNP levels available for analysis, 23 (19.66%) met age-adjusted criteria for elevated NT-proBNP. There was no significant correlation between elevated NT-proBNP levels and MACE (p = 0.482) or mortality (p = 0.737) in patients with SARS-CoV-2 infections. There was no statistical difference in total length of stay (p-value = 0.6384) or ICU admission (p-value = 0.354) between those with elevated admission NT-proBNP and those without. Conclusions: An elevated NT-proBNP at time of admission does not significantly correlate with higher rates of MACE, hospital length of stay, ICU admission, and mortality in patients admitted with SARS-CoV-2 infection.

scholarly journals 182. Missed Opportunities to Discontinue Unnecessary Vancomycin During Pharmacist Therapeutic Monitoring

2020 ◽  
Vol 7 (Supplement_1) ◽  
pp. S98-S98
Author(s):  
Corey J Medler ◽  
Mary Whitney ◽  
Juan Galvan-Cruz ◽  
Ron Kendall ◽  
Rachel Kenney ◽  
...  
Keyword(s):  
Length Of Stay ◽  
Adverse Drug Events ◽  
Academic Medical Center ◽  
Total Duration ◽  
Hospital Length ◽  
Hospital Length Of Stay ◽  
Therapeutic Monitoring ◽  
Resistant Bacteria ◽  
Therapeutic Drug ◽  
Missed Opportunities

Abstract Background Unnecessary and prolonged IV vancomycin exposure increases risk of adverse drug events, notably nephrotoxicity, which may result in prolonged hospital length of stay. The purpose of this study is to identify areas of improvement in antimicrobial stewardship for vancomycin appropriateness by clinical pharmacists at the time of therapeutic drug monitoring (TDM). Methods Retrospective, observational cohort study at an academic medical center and a community hospital. Inclusion: patient over 18 years, received at least three days of IV vancomycin where the clinical pharmacy TDM service assessed for appropriate continuation for hospital admission between June 19, 2019 and June 30, 2019. Exclusion: vancomycin prophylaxis or administered by routes other than IV. Primary outcome was to determine the frequency and clinical components of inappropriate vancomycin continuation at the time of TDM. Inappropriate vancomycin continuation was defined as cultures positive for methicillin-susceptible Staphylococcus aureus (MRSA), vancomycin-resistant bacteria, and non-purulent skin and soft tissue infection (SSTI) in the absence of vasopressors. Data was reported using descriptive statistics and measures of central tendency. Results 167 patients met inclusion criteria with 38.3% from the ICU. SSTIs were most common indication 39 (23.4%) cases, followed by pneumonia and blood with 34 (20.4%) cases each. At time of vancomycin TDM assessment, vancomycin continuation was appropriate 59.3% of the time. Mean of 4.22 ± 2.69 days of appropriate vancomycin use, 2.18 ± 2.47 days of inappropriate use, and total duration 5.42 ± 2.94. 16.4% patients developed an AKI. Majority of missed opportunities were attributed to non-purulent SSTI (28.2%) and missed MRSA nares swabs in 21% pneumonia cases (table 1). Conclusion Vancomycin is used extensively for empiric treatment of presumed infections. Appropriate de-escalation of vancomycin therapy is important to decrease the incidence of adverse effects, decreasing hospital length of stay, and reduce development of resistance. According to the mean duration of inappropriate therapy, there are opportunities for pharmacy and antibiotic stewardship involvement at the time of TDM to optimize patient care (table 1). Missed opportunities for vancomycin de-escalation Disclosures All Authors: No reported disclosures

Sign in / Sign up

Export Citation Format

Share Document