Dataset Information

Neural Network Prediction of ICU Length of Stay Following Cardiac Surgery Based on Pre-Incision Variables.

ABSTRACT:

Background

Advanced predictive analytical techniques are being increasingly applied to clinical risk assessment. This study compared a neural network model to several other models in predicting the length of stay (LOS) in the cardiac surgical intensive care unit (ICU) based on pre-incision patient characteristics.

Methods

Thirty six variables collected from 185 cardiac surgical patients were analyzed for contribution to ICU LOS. The Automatic Linear Modeling (ALM) module of IBM-SPSS software identified 8 factors with statistically significant associations with ICU LOS; these factors were also analyzed with the Artificial Neural Network (ANN) module of the same software. The weighted contributions of each factor ("trained" data) were then applied to data for a "new" patient to predict ICU LOS for that individual.

Results

Factors identified in the ALM model were: use of an intra-aortic balloon pump; O2 delivery index; age; use of positive cardiac inotropic agents; hematocrit; serum creatinine ? 1.3 mg/deciliter; gender; arterial pCO2. The r2 value for ALM prediction of ICU LOS in the initial (training) model was 0.356, p <0.0001. Cross validation in prediction of a "new" patient yielded r2 = 0.200, p <0.0001. The same 8 factors analyzed with ANN yielded a training prediction r2 of 0.535 (p <0.0001) and a cross validation prediction r2 of 0.410, p <0.0001. Two additional predictive algorithms were studied, but they had lower prediction accuracies. Our validated neural network model identified the upper quartile of ICU LOS with an odds ratio of 9.8(p <0.0001).

Conclusions

ANN demonstrated a 2-fold greater accuracy than ALM in prediction of observed ICU LOS. This greater accuracy would be presumed to result from the capacity of ANN to capture nonlinear effects and higher order interactions. Predictive modeling may be of value in early anticipation of risks of post-operative morbidity and utilization of ICU facilities.

SUBMITTER: LaFaro RJ

PROVIDER: S-EPMC4692524 | biostudies-literature | 2015

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Neural Network Prediction of ICU Length of Stay Following Cardiac Surgery Based on Pre-Incision Variables.

LaFaro Rocco J RJ Pothula Suryanarayana S Kubal Keshar Paul KP Inchiosa Mario Emil ME Pothula Venu M VM Yuan Stanley C SC Maerz David A DA Montes Lucresia L Oleszkiewicz Stephen M SM Yusupov Albert A Perline Richard R Inchiosa Mario Anthony MA

PloS one 20151228 12

<h4>Background</h4>Advanced predictive analytical techniques are being increasingly applied to clinical risk assessment. This study compared a neural network model to several other models in predicting the length of stay (LOS) in the cardiac surgical intensive care unit (ICU) based on pre-incision patient characteristics.<h4>Methods</h4>Thirty six variables collected from 185 cardiac surgical patients were analyzed for contribution to ICU LOS. The Automatic Linear Modeling (ALM) module of IBM-SP ...[more]

PMID: 26710254

Similar Datasets

Project description:BackgroundIn the context of increased attention afforded to hospital efficiency and improved but safe patient throughput, decreasing unnecessary hospital length of stay (LOS) is imperative. Given that lumbar spine procedures may be among a hospital's most profitable services, identifying patients at risk of increased healthcare resource utilization prior to surgery is a valuable opportunity to develop targeted pre- and peri-operative intervention and quality improvement initiatives. The purpose of the present investigation was to examine patient factors that predict prolonged LOS as well as discharge disposition following elective, posterior, lumbar spine surgery.MethodsWe employed a retrospective cohort analysis on 779 consecutive patients treated with lumbar surgery without fusion. Our primary outcome measures were extended LOS (three or more midnights) and discharge disposition. Patient sociodemographic, procedural, and discharge characteristics were adjusted for in our analysis. Sociodemographic variables included Area of Deprivation Index (ADI), a comprehensive metric of socioeconomic status, utilizing income, education, employment, and housing quality based on patient zip code. Multivariable logistic regression and ordinal logistic regression analyses were performed to assess whether covariates were independently predictive of extended LOS and discharge disposition, respectively.Results779 patients were studied, with a median age of 66 years (±15) and a median LOS of 1 midnight (range, 1-10 midnights). Patients in the most disadvantaged ADI quintile (adjusted odds ratio, aOR 2.48 95% CI 1.15-5.47), those who underwent a minimally-invasive or tubular retractor surgery (aOR 3.03 95% CI 1.02-8.56), those who had an intra-operative drain placed (aOR 4.46 95% CI 2.53-7.26), who had a cerebrospinal fluid leak (aOR 3.46 95% CI 1.55-7.58), who were discharged anywhere but home (aOR 17.11 95% CI 9.24-33.00), and those who were evaluated by physical therapy (aOR 7.23 95% CI 2.13-45.30) or OT (aOR 2.20 95% CI 1.13-4.22) had a significantly increased chance of an extended LOS. Preoperative opioid use was not associated with an increased LOS following surgery (aOR 1.12 95% CI 0.56-1.46). Extended LOS was not associated with post-discharge emergency department representation or unplanned readmission within 90 days following discharge (p=0.148). Patients who were older (aOR 1.99 95% CI 1.62-2.48), in higher quintiles on ADI (3rd quintile; aOR 1.90 95% CI 1.12-3.23, 4th quintile; aOR 1.79, 95% CI 1.05-3.05, 5th quintile; aOR 2.16 95% CI 1.26-3.75), who had a CSF leak (aOR 2.18 95% CI 1.22-3.86), or who had a longer procedure duration (aOR 1.38 95% CI 1.17-1.62) were more likely to require additional services or be sent to a subacute facility upon discharge.ConclusionsPatient sociodemographics, along with procedural factors, and discharge disposition were all associated with an increased likelihood of prolonged LOS and resource intensive discharges following elective lumbar spine surgery. Several of these factors could be reliably identified pre-operatively and may be amenable to targeted preoperative intervention. Improving discharge disposition planning in the peri-operative period may allow for more efficient use of hospitalization and inpatient and post-acute resources.

Project description:Machine learning can predict outcomes and determine variables contributing to precise prediction, and can thus classify patients with different risk factors of outcomes. This study aimed to investigate the predictive accuracy for mortality and length of stay in intensive care unit (ICU) patients using machine learning, and to identify the variables contributing to the precise prediction or classification of patients. Patients (n = 12,747) admitted to the ICU at Chiba University Hospital were randomly assigned to the training and test cohorts. After learning using the variables on admission in the training cohort, the area under the curve (AUC) was analyzed in the test cohort to evaluate the predictive accuracy of the supervised machine learning classifiers, including random forest (RF) for outcomes (primary outcome, mortality; secondary outcome, length of ICU stay). The rank of the variables that contributed to the machine learning prediction was confirmed, and cluster analysis of the patients with risk factors of mortality was performed to identify the important variables associated with patient outcomes. Machine learning using RF revealed a high predictive value for mortality, with an AUC of 0.945 (95% confidence interval [CI] 0.922-0.977). In addition, RF showed high predictive value for short and long ICU stays, with AUCs of 0.881 (95% CI 0.876-0.908) and 0.889 (95% CI 0.849-0.936), respectively. Lactate dehydrogenase (LDH) was identified as a variable contributing to the precise prediction in machine learning for both mortality and length of ICU stay. LDH was also identified as a contributing variable to classify patients into sub-populations based on different risk factors of mortality. The machine learning algorithm could predict mortality and length of stay in ICU patients with high accuracy. LDH was identified as a contributing variable in mortality and length of ICU stay prediction and could be used to classify patients based on mortality risk.

Project description:Background Postoperative infection contributes to the worsening of congenital cardiac surgery (CCS) outcomes. Surgical site infection (SSI), bloodstream infection (BSI) and ventilator associated pneumonia (VAP) are common. An additional bundle of preventive measures against central-line associated bloodstream infection (CLABSI) bundle was implemented in April 2019. Objectives To compare the incidence of major infections after pediatric CCS before and after the implementation of the CLABSI bundle and to identify risk factors for major infections. Methods We conducted a single-center, retrospective study to assess the incidence of major infections including bloodstream infection (BSI), surgical site infection (SSI), and ventilator-associated pneumonia (VAP) after pediatric CCS one year before and after implementation of the CLABSI bundle during April 2018–March 2020. The demographics and outcomes of the patients were explored, and risk factors for major infections were identified using multivariate analysis. Results A total of 548 children (53% male) underwent CCS with a median age of 1.9 years (range 0.01–17.5 years). The median Aristotle Basic Complexity score was 7.1 (range 3–14.5). The CLABSI bundle was applied in 262 patients. Overall mortality was 5.5%. 126 patients (23%) experienced major postoperative infections. During the year after the implementation of the CLABSI bundle, BSI was reduced from 8.4% to 3.1% (p = 0.01), with a smaller reduction in VAP (21% to 17.6%; p = 0.33). The incidence of SSI was unchanged (1.7% to 1.9%; p = 0.77). The independent risk factors for major infections were age at surgery <6 months (p = 0.04), postoperative ventilator usage >2 days (p < 0.01), central line usage >4 days (p = 0.04), and surgery during the pre-CLABSI bundle period (p = 0.01). Conclusion Following the implementation of the CLABSI prevention package in our pediatric CCS unit, the incidence of BSI was significantly reduced. The incidence of VAP tended to decrease, while the SSI was unchanged. Sustainability of the prevention package through nurse empowerment and compliance audits is an ongoing challenge.

Dataset Information

Neural Network Prediction of ICU Length of Stay Following Cardiac Surgery Based on Pre-Incision Variables.

Background

Methods

Results

Conclusions

Publications

Neural Network Prediction of ICU Length of Stay Following Cardiac Surgery Based on Pre-Incision Variables.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets