Predictors Of Length Of Stay Research Articles

Determinants of Hospital Stay Duration Post-Colorectal Surgery

Introduction: Hospital length of stay (LOS) remains a vital metric for assessing patient outcomes and healthcare resource utilization. Given the substantial financial impact of diagnosing and treating colorectal anomalies, coupled with an increased susceptibility to postoperative complications, it is crucial to understand the factors affecting LOS following colorectal surgery. Our primary objective was to investigate the preoperative, intraoperative, and postoperative risk factors that have substantial influence over LOS following a colorectal procedure.
 Methods: This study analyzed data from a retrospective study of adults who underwent various colorectal surgeries (colostomy, ileostomy, small bowel resection, etc.) at Cleveland Clinic Foundation (January 2005
 - December 2014). Predictor variables were categorized into preoperative (patient demographics, medical history, comorbidities, lifestyle factors), intraoperative, and postoperative factors. LOS was grouped into short-term (SLOS) (≤ 7 days), medium-term (MLOS) (8-30 days), and long-term (LLOS) (> 30 days) stays. Multinomial logistic regression models assessed predictor effects on LOS.
 Results: Among the 7874 patients, 50.7% were females, with a minimum age of 20 years. SLOS were observed in 61.1%, MLOS in 37.6%, and LLOS in 1.3% of patients. Advanced age correlated with prolonged LOS, possibly due to age-related health challenges like weak immune systems. Coagulopathy, and fluid and electrolyte disorders raised MLOS and LLOS risk, likely due to complications like significant bleeding and electrolyte imbalances. Surgery duration predicted longer LOS, elevating LLOS and MLOS by 52% and 42%. Postoperative infections were associated to extended stays, possibly due to subsequent interventions, monitoring and recovery delays.
 Conclusion: Our study revealed that key preoperative predictors of LOS included Age, coagulopathy, fluid and electrolyte disorders, severe weight loss, and drug abuse. Notably, intraoperative factors such as surgical approach (open vs laparoscopic) and surgery duration, alongside postoperative complications including superficial and serious infections, significantly influenced LOS. By incorporating these insights into the preoperative planning, clinicians could potentially develop tailored interventions to mitigate risk factors and enhance postoperative recovery, thus potentially reducing LOS and improving patient outcomes.

Investigating Length of Stay Patterns and Its Predictors in the South Wales Trauma Network.

Length of stay (LOS) is frequently employed as a performance metric for trauma care. Following the establishment of the trauma network worldwide, the assessment and prediction of LOS in different levels of trauma centres have been extensively studied. However, assessing the total patient length of stay from a whole trauma network perspective is unclear. The objective of this study was to systematically analyse the overall Length of Stay (LOS) pattern within the SWTN before its establishment and in the immediate time after its foundation and, secondly, to assess the association between relevant impact factors and LOS. A retrospective secondary analysis based on the trauma admission dataset from Trauma Audit and Research Network(TARN) dataset was conducted. The studied sample covered around 18000 patients admitted to trauma centres from South Wales Major trauma network between January 2012 and October 2021. The primary outcome is the total length of stay in the trauma network. Statistical tests were applied to examine the difference between normal and outlier LOS. Data visualisation was utilised to demonstrate the LOS patterns and potential association between LOS and relevant demographic and clinical predictors. The distribution of length of stay in SWTN follows a right-skewed distribution with a median of 10 (IQR, 5-18) and a mean of 15.92 days. There were 1520 patients with outliers for LOS. A significant difference (p¡ 0.05) was found between the normal and outlier groups of LOS based on demographic (age, gender and residential information) and clinical characteristics(ward type, maximum of anatomically-based injury severity score(AIS) and probability of survival). Age group, maximum AIS score on specific injured region, ward type and its interaction effect with the number of admissions may associated with the LOS. Specifically, patients admitted to the geriatric ward exhibited notably prolonged LOS, and individuals with more than 2 admissions to long-term care and recovery-related wards such as neurosurgical rehabilitation, spinal injuries and burns wards also displayed elevated LOS. Our finding supports prior evidence indicating elderly people are vulnerable to longer stays. Moreover, concerning the types of admission wards, patients admitted to rehabilitation wards who underwent more than 2 hospitalisations also faced an increased risk of prolonged stay. Based on these results, policymakers and healthcare providers should contemplate expanding the allocation of medical resources to this demographic to mitigate the length of stay and optimise associated healthcare costs.

Zero-Inflated Count Regression Models in Solving Challenges Posed by Outlier-Prone Data; an Application to Length of Hospital Stay.

Ignoring outliers in data may lead to misleading results. Length of stay (LOS) is often considered a count variable with a high frequency of outliers. This study exemplifies the potential of robust methodologies in enhancing the accuracy and reliability of analyses conducted on skewed and outlier-prone count data of LOS. The application of Zero-Inflated Poisson (ZIP) and robust Zero-Inflated Poisson (RZIP) models in solving challenges posed by outlier LOS data were evaluated. The ZIP model incorporates two components, tackling excess zeros with a zero-inflation component and modeling positive counts with a Poisson component. The RZIP model introduces the Robust Expectation-Solution (RES) algorithm to enhance parameter estimation and address the impact of outliers on the model's performance. Data from 254 intensive care unit patients were analyzed (62.2% male). Patients aged 65 or older accounted for 58.3% of the sample. Notably, 38.6% of patients exhibited zero LOS. The overall mean LOS was 5.89 (± 9.81) days, and 9.45% of cases displayed outliers. Our analysis using the RZIP model revealed significant predictors of LOS, including age, underlying comorbidities (p<0.001), and insurance status (p=0.013). Model comparison demonstrated the RZIP model's superiority over ZIP, as evidenced by lower Akaike information criteria (AIC) and Bayesians information criteria (BIC) values. The application of the RZIP model allowed us to uncover meaningful insights into the factors influencing LOS, paving the way for more informed decision-making in hospital management.

The association of traditional obesity parameters with the length of stay among patients with coronary artery disease: A cross-sectional study.

There is a strong association between obesity and coronary artery disease (CAD). Obesity is measured using traditional obesity parameters, such as body mass index, body adiposity index, waist circumference (WC), and hip circumference. The aim of this study is to explore the association between traditional obesity parameters and the length of stay (LOS) among hospitalized CAD patients. An original correlative descriptive study was carried out using secondary data analysis, in which 220 hospitalized Jordanian CAD patients were recruited from Jordan northern and middle regions. Age, WC, triglycerides, and high- sensitivity C-reactive protein were all positive predictors of the total hospital LOS among hospitalized patients with CAD. The WC, age, triglycerides, and high-sensitivity C-reactive protein levels were significantly positively associated with total LOS. Healthcare providers, including nurses, should take into account these significant positive predictors of LOS to achieve better health outcomes and improve patient satisfaction.

A Prospective Multicenter Comparison of Trauma and Injury Severity Score, American Society of Anesthesiologists Physical Status, and National Surgical Quality Improvement Program Calculator's Ability to Predict Operative Trauma Outcomes.

Trauma outcome prediction models have traditionally relied upon patient injury and physiologic data (eg, Trauma and Injury Severity Score [TRISS]) without accounting for comorbidities. We sought to prospectively evaluate the role of the American Society of Anesthesiologists physical status (ASA-PS) score and the National Surgical Quality Improvement Program Surgical Risk-Calculator (NSQIP-SRC), which are measurements of comorbidities, in the prediction of trauma outcomes, hypothesizing that they will improve the predictive ability for mortality, hospital length of stay (LOS), and complications compared to TRISS alone in trauma patients undergoing surgery within 24 hours. A prospective, observational multicenter study (9/2018-2/2020) of trauma patients ≥18 years undergoing operation within 24 hours of admission was performed. Multiple logistic regression was used to create models predicting mortality utilizing the variables within TRISS, ASA-PS, and NSQIP-SRC, respectively. Linear regression was used to create models predicting LOS and negative binomial regression to create models predicting complications. From 4 level I trauma centers, 1213 patients were included. The Brier Score for each model predicting mortality was found to improve accuracy in the following order: 0.0370 for ASA-PS, 0.0355 for NSQIP-SRC, 0.0301 for TRISS, 0.0291 for TRISS+ASA-PS, and 0.0234 for TRISS+NSQIP-SRC. However, when comparing TRISS alone to TRISS+ASA-PS ( P = .082) and TRISS+NSQIP-SRC ( P = .394), there was no significant improvement in mortality prediction. NSQIP-SRC more accurately predicted both LOS and complications compared to TRISS and ASA-PS. TRISS predicts mortality better than ASA-PS and NSQIP-SRC in trauma patients undergoing surgery within 24 hours. The TRISS mortality predictive ability is not improved when combined with ASA-PS or NSQIP-SRC. However, NSQIP-SRC was the most accurate predictor of LOS and complications.

Finding the best trade-off between performance and interpretability in predicting hospital length of stay using structured and unstructured data.

This study aims to develop high-performing Machine Learning and Deep Learning models in predicting hospital length of stay (LOS) while enhancing interpretability. We compare performance and interpretability of models trained only on structured tabular data with models trained only on unstructured clinical text data, and on mixed data. The structured data was used to train fourteen classical Machine Learning models including advanced ensemble trees, neural networks and k-nearest neighbors. The unstructured data was used to fine-tune a pre-trained Bio Clinical BERT Transformer Deep Learning model. The structured and unstructured data were then merged into a tabular dataset after vectorization of the clinical text and a dimensional reduction through Latent Dirichlet Allocation. The study used the free and publicly available Medical Information Mart for Intensive Care (MIMIC) III database, on the open AutoML Library AutoGluon. Performance is evaluated with respect to two types of random classifiers, used as baselines. The best model from structured data demonstrates high performance (ROC AUC = 0.944, PRC AUC = 0.655) with limited interpretability, where the most important predictors of prolonged LOS are the level of blood urea nitrogen and of platelets. The Transformer model displays a good but lower performance (ROC AUC = 0.842, PRC AUC = 0.375) with a richer array of interpretability by providing more specific in-hospital factors including procedures, conditions, and medical history. The best model trained on mixed data satisfies both a high level of performance (ROC AUC = 0.963, PRC AUC = 0.746) and a much larger scope in interpretability including pathologies of the intestine, the colon, and the blood; infectious diseases, respiratory problems, procedures involving sedation and intubation, and vascular surgery. Our results outperform most of the state-of-the-art models in LOS prediction both in terms of performance and of interpretability. Data fusion between structured and unstructured text data may significantly improve performance and interpretability.

2416. Outcomes of Patients with Ventilator-Associated Pneumonia in For-Profit Hospitals Compared to Non-Profit hospitals: A National Retrospective Cohort

Abstract Background Healthcare-associated infections (HAI) constitute a significant burden to patient outcomes and healthcare costs. Ventilator-associated pneumonia (VAP) carries a high mortality and represents one of the costliest HAI. There is uncertainty if treating this subgroup of patients in For-Profit (FP) hospitals portents better outcomes than Nonprofit (NP) hospitals. Methods A retrospective cohort of patients admitted from January 2016 through December 2020 was extracted from the National Inpatient Sample database. We compared two cohorts of patients with a primary diagnosis of VAP: patients admitted to FP hospitals and patients admitted to NP hospitals. The primary end-point was all-cause inpatient mortality. A multivariate logistic regression analysis was conducted. The secondary end-points included: comparing the requirement of vasopressor use and Acute Kidney Injury (AKI) requiring hemodialysis (HD); describing associated comorbidities; and comparing if there was a difference in length of stay (LoS) and hospital cost between the 2 groups. In order to compare LoS and hospital cost, a multivariate linear regression was conducted. Results A total of 6 155 hospitalizations under the primary diagnosis of VAP were included. No statistically significant difference in mortality was noted (FP: 4.24 vs. NP:8; p = 0.182), but age (aOR 1.01, p = 0.020) and Charlson Comorbidity Index(CCI) (aOR 1.20; p &amp;lt; 0.001) were modest predictors of mortality for both groups (FP vs. NP). There was no statistically significant difference in mean LoS between these two groups (FP:10.61 days vs. NP: 9.57, p=0.398). CCI (Coef 0.64, p= 0.01) and admission to large-size hospitals (Coef 3.03, p &amp;lt; 0.001) were predictors of LoS for both groups. Admission to for-profit hospitals predicted higher hospitalization costs (p = 0.020). Furthermore, admission to medium-size hospitals, large-size hospitals, and urban hospitals were also predictors of higher hospital cost (p &amp;lt; 0.05). No statistically significant difference was found in vasopressor use and AKI requiring HD among both groups (FP vs. NP) (p &amp;lt; 0.05). Conclusion Admission to For-Profit hospitals did not portend better outcomes compared to Nonprofit hospitals. Conversely, admission to For-Profit hospitals predicted higher total hospital costs. Disclosures All Authors: No reported disclosures

Implementation of an Enhanced Recovery after Surgery Pathway for Transgender and Gender-Diverse Individuals Undergoing Chest Reconstruction Surgery: An Observational Cohort Study.

Enhanced Recovery After Surgery (ERAS) protocols are associated with improved clinical outcomes in cisgender breast surgery patients. However, a paucity of research exists regarding transgender and gender-diverse individuals (TGD) in the ERAS framework. The primary objective of this observational cohort study is to describe the implementation of a gender-affirming ERAS protocol and its relationship to hospital length of stay (LOS) in TGD patients following chest reconstruction surgery. The secondary aim is to identify intraoperative predictors of LOS and define variables associated with adverse outcomes. We identified 362 patients in three epochs: a traditional group (n = 144), a partial ERAS implementation group (n = 92), and an ERAS group (n = 126). Exploratory multivariable median regression modeling was performed to identify independent predictors of LOS. We report that the traditional group's median hospital LOS was 1.1 days compared to 0.3 days in the ERAS group. Intraoperative tranexamic acid administration was associated with significantly shorter LOS (p < 0.001), reduced postoperative drainage (p < 0.001), and fewer returns to the operating room within 24 h (p = 0.047). Our data suggest that implementing a multimodal ERAS gender-affirming pathway was associated with improved patient-centered surgical outcomes such as decreased return to the operating room for hematoma evacuation, higher rates of discharge home, and reduced postoperative drainage output.

Predicting COVID-19 Hospital Stays with Kolmogorov–Gabor Polynomials: Charting the Future of Care

Optimal allocation of ward beds is crucial given the respiratory nature of COVID-19, which necessitates urgent hospitalization for certain patients. Several governments have leveraged technology to mitigate the pandemic’s adverse impacts. Based on clinical and demographic variables assessed upon admission, this study predicts the length of stay (LOS) for COVID-19 patients in hospitals. The Kolmogorov–Gabor polynomial (a.k.a., Volterra functional series) was trained using regularized least squares and validated on a dataset of 1600 COVID-19 patients admitted to Khorshid Hospital in the central province of Iran, and the five-fold internal cross-validated results were presented. The Volterra method provides flexibility, interactions among variables, and robustness. The most important features of the LOS prediction system were inflammatory markers, bicarbonate (HCO3), and fever—the adj. R2 and Concordance Correlation Coefficients were 0.81 [95% CI: 0.79–0.84] and 0.94 [0.93–0.95], respectively. The estimation bias was not statistically significant (p-value = 0.777; paired-sample t-test). The system was further analyzed to predict “normal” LOS ≤ 7 days versus “prolonged” LOS &gt; 7 days groups. It showed excellent balanced diagnostic accuracy and agreement rate. However, temporal and spatial validation must be considered to generalize the model. This contribution is hoped to pave the way for hospitals and healthcare providers to manage their resources better.

Developing a length of stay prediction model for newborns, achieving better accuracy with greater usability

BackgroundOne in ten newborn children is born prematurely. The elongated length of stay (LOS) of these children in the Neonatal Intensive Care Unit (NICU) has important implications on hospital occupancy figures, healthcare and management costs, as well as the psychology of parents. In order to allow accurate planning and resource allocation, this study aims to create a generalizable and robust model to predict the NICU LOS of preterm newborns. MethodsData were collected from a large tertiary center NICU between 2011 and 2018 and relates to 5,362 newborns. The selected model was externally validated using a data set of 8,768 newborns from another tertiary center NICU. This report compares several models, such as Random Forest (RF), quantile RF, and other feature selection methods, including LASSO and AIC step-forward selection. In addition, a novel step-forward selection based on False Discovery Rate (FDR) for quantile regression is presented and evaluated. ResultsA high-orderquantile regression model for predicting preterm newborns’ LOS that uses only four features available at birth had more attractive properties than other richer ones. The model achieved a Mean Absolute Error (MAE) of 6.26 days on the internal validation set (average LOS 27.04) and an MAE of 6.04 days on the external validation set (average LOS 29.32). The suggested model surpassed the accuracy obtained by models in the literature. It is shown empirically that the FDR-based selection has better properties than the AIC-based step-forward selection approach. ConclusionThis paper demonstrates a process to create a predictive model for NICU LOS in preterm newborns, where each step is reasoned. We obtain a simple and robust model for NICU LOS prediction, which achieves far better results than the current model used for financing NICUs. Utilizing this model, we have created an easy-to-use online web application to ease parents' worries and to assist NICU management: https://tzviel.shinyapps.io/calcuLOS.

A deep learning approach for inpatient length of stay and mortality prediction

PurposeAccurate prediction of the Length of Stay (LoS) and mortality in the Intensive Care Unit (ICU) is crucial for effective hospital management, and it can assist clinicians for real-time demand capacity (RTDC) administration, thereby improving healthcare quality and service levels. MethodsThis paper proposes a novel one-dimensional (1D) multi-scale convolutional neural network architecture, namely 1D-MSNet, to predict inpatients’ LoS and mortality in ICU. First, a 1D multi-scale convolution framework is proposed to enlarge the convolutional receptive fields and enhance the richness of the convolutional features. Following the convolutional layers, an atrous causal spatial pyramid pooling (SPP) module is incorporated into the networks to extract high-level features. The optimized Focal Loss (FL) function is combined with the synthetic minority over-sampling technique (SMOTE) to mitigate the imbalanced-class issue. ResultsOn the MIMIC-IV v1.0 benchmark dataset, the proposed approach achieves the optimum R-Square and RMSE values of 0.57 and 3.61 for the LoS prediction, and the highest test accuracy of 97.73% for the mortality prediction. ConclusionThe proposed approach presents a superior performance in comparison with other state-of-the-art, and it can effectively perform the LoS and mortality prediction tasks.

Health outcome predictive modelling in intensive care units

The literature in Intensive Care Units (ICUs) data analysis focuses on predictions of length-of-stay (LOS) and mortality based on patient acuity scores such as Acute Physiology and Chronic Health Evaluation (APACHE), Sequential Organ Failure Assessment (SOFA), to name a few. Unlike ICUs in other areas around the world, ICUs in Ontario, Canada, collect two primary intensive care scoring scales, a therapeutic acuity score called the “Multiple Organs Dysfunctional Score” (MODS) and a nursing workload score called the “Nine Equivalents Nursing Manpower Use Score” (NEMS). The dataset analyzed in this study contains patients’ NEMS and MODS scores measured upon patient admission into the ICU and other characteristics commonly found in the literature. Data were collected between January 1st, 2015 and May 31st, 2021, at two teaching hospital ICUs in Ontario, Canada. In this work, we developed logistic regression, random forests (RF) and neural networks (NN) models for mortality (discharged or deceased) and LOS (short or long stay) predictions. Considering the effect of mortality outcome on LOS, we also combined mortality and LOS to create a new categorical health outcome called LMClass (short stay & discharged, short stay & deceased, or long stay without specifying mortality outcomes), and then applied multinomial regression, RF and NN for its prediction. Among the models evaluated, logistic regression for mortality prediction results in the highest area under the curve (AUC) of 0.795 and also for LMClass prediction the highest accuracy of 0.630. In contrast, in LOS prediction, RF outperforms the other methods with the highest AUC of 0.689. This study also demonstrates that MODS and NEMS, as well as their components measured upon patient arrival, significantly contribute to health outcome prediction in ICUs.

S1379 Hyperbilirubinemia on Admission May Predict Length of Stay in Alcoholic Hepatitis With a Maddrey’s Discriminant Function Score Below 32

S1379 Hyperbilirubinemia on Admission May Predict Length of Stay in Alcoholic Hepatitis With a Maddrey’s Discriminant Function Score Below 32

The predictive value of the preoperative albumin-to-fibrinogen ratio for postoperative hospital length of stay in liver cancer patients.

Hepatocellular carcinoma (HCC) is a significant global health burden, with postoperative hospital length of stay (LOS) impacting patient outcomes and healthcare costs. Existing nutritional, inflammatory, and coagulation indices can predict LOS, with particular interest in albumin, fibrinogen, and D-dimer. This study investigates the predictive value of preoperative albumin-to-fibrinogen ratio (AFR) and albumin-to-D-dimer ratio (ADR) for postoperative LOS in HCC patients. This retrospective study involved 462 adult HCC patients who underwent partial hepatic lesion excision between February 2016 and August 2022. We analyzed demographic and clinical data, including preoperative blood samples, surgical approach, and LOS. The primary outcome measure was LOS, calculated from the date of surgery to the date of hospital discharge. Preoperative AFR and ADR were calculated. The ROC curves determined optimal cutoff points. The Cox proportional hazards model, Kaplan-Meier method, and the log-rank test were used for statistical analysis. The study established an optimal AFR cutoff value of 15.474, with a higher AUC value than ADR, indicating superior predictive potential for postoperative LOS. Participants with high-AFR (AFR > 15.474) had a shorter median LOS (13 vs. 15 days, p < 0.001) compared to those with low-AFR (AFR ≤15.474). Multivariate analysis revealed high-AFR (HR: 1.99; p < 0.001) as a positive influence on LOS reduction, whereas Child-Pugh rated as B (HR: 0.49; p < 0.001), laparotomy (HR: 0.37; p < 0.001) and total bilirubin >20.5 μmol/L (HR: 0.58; p < 0.001) negatively impacted LOS reduction. Subgroup analysis confirmed AFR's predictive ability for patients experiencing reduced or prolonged LOS due to Child-Pugh score, surgical methods, and total bilirubin concentrations. Even within normal albumin and fibrinogen levels, patients with high-AFR exhibited a shorter LOS (all p < 0.001). Our findings underscore the value of the AFR as a reliable predictor of LOS in HCC patients. An AFR greater than 15.474 consistently correlated with a shorter LOS, suggesting its potential clinical utility in guiding perioperative management and resource allocation in HCC patients.

In-hospital mortality and length of stay of patients with hypertensive crisis treated at public hospitals in Harari Regional State, Eastern Ethiopia.

Hypertensive crisis poses substantial cardiovascular morbidity and mortality. This study aimed to assess in-hospital mortality, length of stay (LOS), and their predictors among patients with hypertensive crisis treated at public hospitals in Harari Regional State, Eastern Ethiopia. An institutional-based retrospective cohort study was conducted from October 1 to 31, 2022. The medical records of 328 patients with hypertensive crisis treated at two public hospitals between September 1, 2017 and August 31, 2022 were reviewed. Cox proportional hazards regression and negative binomial regression were used to identify predictors of in-hospital mortality and LOS, respectively. The in-hospital mortality rate of patients with hypertensive crisis was 18.94 (95% confidence interval (CI): 12.08-29.70) per 1000 person-day observation. The median (interquartile range) LOS of these patients was 10 (4-120) hours. Age ≥65 years (adjusted hazard ratio (AHR): 3.30; 95% CI: 1.17- 9.33); increment in initial systolic blood pressure (AHR: 1.040; 95% CI: 1.014-1.066); and having acute brain-related damage (AHR: 4.02; 95% CI: 1.48-10.88) were predictors of in-hospital mortality. Rural residence (adjusted incident-rate ratio (IRR): 1.34; 95% CI: 1.03-1.75); having a history of medication discontinuation (adjusted IRR: 1.59; 95% CI: 1.16-2.18); comorbidity (adjusted IRR: 1.90; 95% CI: 1.49-2.43); acute brain-related damage (adjusted IRR: 13.32; 95% CI: 9.22-19.24), acute cardiac-related damage (adjusted IRR: 7.40; 95% CI: 4.90-11.16); and acute kidney injury (adjusted IRR: 7.64; 95% CI: 5.46-10.69) were predictors of LOS. Thus, it is necessary to develop strategies that allow early screening and follow-up of patients at risk.

A Machine Learning Framework for Length of Stay Minimization in Healthcare Emergency Department

The emergency departments (EDs) in most hospitals, especially in middle-and-low-income countries, need techniques for minimizing the waiting time of patients. The application and utilization of appropriate methods can enhance the number of patients treated, improve patients’ satisfaction, reduce healthcare costs, and lower morbidity and mortality rates which are often associated with poor healthcare facilities, overcrowding, and low availability of healthcare professionals. Modeling the length of stay (LOS) of patients in healthcare systems is a challenge that must be addressed for sound decision-making regarding capacity planning and resource allocation. This paper presents a machine learning (ML) framework for predicting a patient’s LOS within the ED. A study of the services in the ED of a tertiary healthcare facility in Uyo, Nigeria was conducted to gain insights into its operational procedures and evaluate the impact of certain parameters on LOS. Then, a computer simulation of the system was performed in R programming language using data obtained from records in the hospital. Finally, the performance of four ML classifiers involved in patients’ LOS prediction: Classification and Regression Tree (CART), Random Forest (RF), K-Nearest Neighbour (K-NN), and Support Vector Machine (SVM), were evaluated and results indicate that SVM outperforms others with the highest coefficient of determination (R2) score of 0.986984 and least mean square error (MSE) value of 0.358594. The result demonstrates the capability of ML techniques to effectively assess the performance of healthcare systems and accurately predict patients’ LOS to mitigate the low physician-patient ratio and improve throughput.

Can machine learning models predict prolonged length of hospital stay following primary total knee arthroplasty based on a national patient cohort data?

The total length of stay (LOS) is one of the biggest determinators of overall care costs associated with total knee arthroplasty (TKA). An accurate prediction of LOS could aid in optimizing discharge strategy for patients in need and diminishing healthcare expenditure. The aim of this study was to predict LOS following TKA using machine learning models developed on a national-scale patient cohort. The ACS-NSQIP database was queried to acquire 267,966 TKA cases from 2013 to 2020. Four machine learning models-artificial neural network (ANN), random forest, histogram-based gradient boosting, and k-nearest neighbor were trained and tested on the dataset for the prediction of prolonged LOS (LOS exceeded the 75th of all values in the cohort). The model performance was assessed by discrimination (area under the receiver operating characteristic curve [AUC]), calibration, and clinical utility. ANN delivered the best performance among the four models. ANN distinguished prolonged LOS in the study cohort with an AUC of 0.71 and accurately predicted the probability of prolonged LOS for individual patients (calibration slope: 0.82; calibration intercept: 0.03; Brier score: 0.089). All models demonstrated clinical utility by generating positive net benefits in decision curve analyses. Operation time, pre-operative transfusion, pre-operative laboratory tests (hematocrit, platelet count, and white blood cell count), and BMI were the strongest predictors of prolonged LOS. ANN demonstrated modest discrimination capacity and excellent performance in calibration and clinical utility for the prediction of prolonged LOS following TKA. Clinical application of the machine learning models has the potential to improve care coordination and discharge planning for patients at high risk of extended hospitalization after surgery. Incorporating more relevant patient factors may further increase the models' prediction strength.

Activity Measure for Postacute Care “6-Clicks” Mobility Scores Predict Length of Stay Following Hip Resurfacing

AbstractNo studies have identified length of stay (LOS) predictors following hip resurfacing arthroplasty (HRA). The activity measure for postacute care (AM-PAC) score is a readily available tool that may allow physicians to predict LOS in patients undergoing HRA. We retrospectively reviewed patients who underwent HRA between February 2017 and April 2021. All patients completed AM-PAC mobility score on postoperative day 0. Age, American Society of Anesthesiology class, body mass index, surgical time, and Charlson comorbidity index were included as possible predictors of LOS. Predictive models were generated predictive for LOS times of more than 24 and more than 48 hours. Analysis was used to determine c. We identified 262 male patients who completed AM-PAC scores following HRA. Patients with a LOS more than 24 hours had significantly lower AM-PAC scores than patients with a LOS less than 24 hours (23.43 ± 1.37 vs. 19.76 ± 2.78; p &lt; 0.001). Similarly, patients with a LOS more than 48 hours had significantly lower AM-PAC scores than patients with a LOS less than 48 hours (22.08 ± 2.56 vs. 18.17 ± 2.23; p &lt; 0.001). AM-PAC scores were 84.5% accurate for predicting LOS more than 24 hours and 86.3% for LOS more than 48 hours. Patients with scores less than 23 were at higher risk for LOS more than 24 hours and less than 19 were likely to have a LOS more than 48 hours. AM-PAC scores following HRA were predictive of LOS. Providers may use the cutoffs in this analysis to determine how mobility influences LOS following HRA. This is a Level III evidence, retrospective cohort study.

Preoperative Prediction and Risk Factor Identification of Hospital Length of Stay for Total Joint Arthroplasty Patients Using Machine Learning.

The aim of this study was to improve understanding of hospital length of stay (LOS) in patients undergoing total joint arthroplasty (TJA) in a high-efficiency, hospital-based pathway. We retrospectively reviewed 1401 consecutive primary and revision TJA patients across 67 patient and preoperative care characteristics from 2016 to 2019 from the institutional electronic health records. A machine learning approach, testing multiple models, was used to assess predictors of LOS. The median LOS was 1day; outpatients accounted for 16.5%, 1-day inpatient stays for 38.0%, 2-day stays for 26.4%, and 3-days or more for 19.1%. Patients characteristically fell into 1 of 3 broad categories that contained relatively similar characteristics: outpatient (0-day LOS), short stay (1- to 2-day LOS), and prolonged stay (3days or greater). The random forest models suggested that a lower Risk Assessment and Prediction Tool score, unplanned admission or hospital transfer, and a medical history of cardiovascular disease were associated with an increased LOS. Documented narcotic use for surgery preparation prior to hospitalization and preoperative corticosteroid use were factors independently associated with a decreased LOS. After TJA, most patients have either an outpatient or short-stay hospital episode. Patients who stay 2days do not differ substantially from patients who stay 1day, while there is a distinct group that requires prolonged admission. Our machine learning models support a better understanding of the patient factors associated with different hospital LOS categories for TJA, demonstrating the potential for improved health policy decisions and risk stratification for centers caring for complex patients.

Machine learning-based prediction of hospital prolonged length of stay admission at emergency department: a Gradient Boosting algorithm analysis.

This study aims to develop and compare different models to predict the Length of Stay (LoS) and the Prolonged Length of Stay (PLoS) of inpatients admitted through the emergency department (ED) in general patient settings. This aim is not only to promote any specific model but rather to suggest a decision-supporting tool (i.e., a prediction framework). We analyzed a dataset of patients admitted through the ED to the "Sant"Orsola Malpighi University Hospital of Bologna, Italy, between January 1 and October 26, 2022. PLoS was defined as any hospitalization with LoS longer than 6 days. We deployed six classification algorithms for predicting PLoS: Random Forest (RF), Support Vector Machines (SVM), Gradient Boosting (GB), AdaBoost, K-Nearest Neighbors (KNN), and logistic regression (LoR). We evaluated the performance of these models with the Brier score, the area under the ROC curve (AUC), accuracy, sensitivity (recall), specificity, precision, and F1-score. We further developed eight regression models for LoS prediction: Linear Regression (LR), including the penalized linear models Least Absolute Shrinkage and Selection Operator (LASSO), Ridge and Elastic-net regression, Support vector regression, RF regression, KNN, and eXtreme Gradient Boosting (XGBoost) regression. The model performances were measured by their mean square error, mean absolute error, and mean relative error. The dataset was randomly split into a training set (70%) and a validation set (30%). A total of 12,858 eligible patients were included in our study, of whom 60.88% had a PloS. The GB classifier best predicted PloS (accuracy 75%, AUC 75.4%, Brier score 0.181), followed by LoR classifier (accuracy 75%, AUC 75.2%, Brier score 0.182). These models also showed to be adequately calibrated. Ridge and XGBoost regressions best predicted LoS, with the smallest total prediction error. The overall prediction error is between 6 and 7 days, meaning there is a 6-7 day mean difference between actual and predicted LoS. Our results demonstrate the potential of machine learning-based methods to predict LoS and provide valuable insights into the risks behind prolonged hospitalizations. In addition to physicians' clinical expertise, the results of these models can be utilized as input to make informed decisions, such as predicting hospitalizations and enhancing the overall performance of a public healthcare system.