scholarly journals Predicting Prolonged Length of Hospital Stay for Peritoneal Dialysis–Treated Patients Using Stacked Generalization: Model Development and Validation Study

10.2196/17886 ◽  
2021 ◽  
Vol 9 (5) ◽  
pp. e17886
Author(s):  
Guilan Kong ◽  
Jingyi Wu ◽  
Hong Chu ◽  
Chao Yang ◽  
Yu Lin ◽  
...  
Keyword(s):  
Peritoneal Dialysis ◽  
High Risk ◽  
Prediction Model ◽  
Performance Measures ◽  
Prediction Models ◽  
Data Driven ◽  
Brier Score ◽  
Prolonged Length ◽  
Admission Data ◽  
Overall Performance

Background The increasing number of patients treated with peritoneal dialysis (PD) and their consistently high rate of hospital admissions have placed a large burden on the health care system. Early clinical interventions and optimal management of patients at a high risk of prolonged length of stay (pLOS) may help improve the medical efficiency and prognosis of PD-treated patients. If timely clinical interventions are not provided, patients at a high risk of pLOS may face a poor prognosis and high medical expenses, which will also be a burden on hospitals. Therefore, physicians need an effective pLOS prediction model for PD-treated patients. Objective This study aimed to develop an optimal data-driven model for predicting the pLOS risk of PD-treated patients using basic admission data. Methods Patient data collected using the Hospital Quality Monitoring System (HQMS) in China were used to develop pLOS prediction models. A stacking model was constructed with support vector machine, random forest (RF), and K-nearest neighbor algorithms as its base models and traditional logistic regression (LR) as its meta-model. The meta-model used the outputs of all 3 base models as input and generated the output of the stacking model. Another LR-based pLOS prediction model was built as the benchmark model. The prediction performance of the stacking model was compared with that of its base models and the benchmark model. Five-fold cross-validation was employed to develop and validate the models. Performance measures included the Brier score, area under the receiver operating characteristic curve (AUROC), estimated calibration index (ECI), accuracy, sensitivity, specificity, and geometric mean (Gm). In addition, a calibration plot was employed to visually demonstrate the calibration power of each model. Results The final cohort extracted from the HQMS database consisted of 23,992 eligible PD-treated patients, among whom 30.3% had a pLOS (ie, longer than the average LOS, which was 16 days in our study). Among the models, the stacking model achieved the best calibration (ECI 8.691), balanced accuracy (Gm 0.690), accuracy (0.695), and specificity (0.701). Meanwhile, the stacking and RF models had the best overall performance (Brier score 0.174 for both) and discrimination (AUROC 0.757 for the stacking model and 0.756 for the RF model). Compared with the benchmark LR model, the stacking model was superior in all performance measures except sensitivity, but there was no significant difference in sensitivity between the 2 models. The 2-sided t tests revealed significant performance differences between the stacking and LR models in overall performance, discrimination, calibration, balanced accuracy, and accuracy. Conclusions This study is the first to develop data-driven pLOS prediction models for PD-treated patients using basic admission data from a national database. The results indicate the feasibility of utilizing a stacking-based pLOS prediction model for PD-treated patients. The pLOS prediction tools developed in this study have the potential to assist clinicians in identifying patients at a high risk of pLOS and to allocate resources optimally for PD-treated patients.

scholarly journals Predicting Prolonged Length of Hospital Stay for Peritoneal Dialysis–Treated Patients Using Stacked Generalization: Model Development and Validation Study (Preprint)

2020 ◽  
Author(s):  
Guilan Kong ◽  
Jingyi Wu ◽  
Hong Chu ◽  
Chao Yang ◽  
Yu Lin ◽  
...  
Keyword(s):  
Peritoneal Dialysis ◽  
High Risk ◽  
Prediction Model ◽  
Performance Measures ◽  
Prediction Models ◽  
Data Driven ◽  
Brier Score ◽  
Prolonged Length ◽  
Admission Data ◽  
Overall Performance

BACKGROUND The increasing number of patients treated with peritoneal dialysis (PD) and their consistently high rate of hospital admissions have placed a large burden on the health care system. Early clinical interventions and optimal management of patients at a high risk of prolonged length of stay (pLOS) may help improve the medical efficiency and prognosis of PD-treated patients. If timely clinical interventions are not provided, patients at a high risk of pLOS may face a poor prognosis and high medical expenses, which will also be a burden on hospitals. Therefore, physicians need an effective pLOS prediction model for PD-treated patients. OBJECTIVE This study aimed to develop an optimal data-driven model for predicting the pLOS risk of PD-treated patients using basic admission data. METHODS Patient data collected using the Hospital Quality Monitoring System (HQMS) in China were used to develop pLOS prediction models. A stacking model was constructed with support vector machine, random forest (RF), and K-nearest neighbor algorithms as its base models and traditional logistic regression (LR) as its meta-model. The meta-model used the outputs of all 3 base models as input and generated the output of the stacking model. Another LR-based pLOS prediction model was built as the benchmark model. The prediction performance of the stacking model was compared with that of its base models and the benchmark model. Five-fold cross-validation was employed to develop and validate the models. Performance measures included the Brier score, area under the receiver operating characteristic curve (AUROC), estimated calibration index (ECI), accuracy, sensitivity, specificity, and geometric mean (Gm). In addition, a calibration plot was employed to visually demonstrate the calibration power of each model. RESULTS The final cohort extracted from the HQMS database consisted of 23,992 eligible PD-treated patients, among whom 30.3% had a pLOS (ie, longer than the average LOS, which was 16 days in our study). Among the models, the stacking model achieved the best calibration (ECI 8.691), balanced accuracy (Gm 0.690), accuracy (0.695), and specificity (0.701). Meanwhile, the stacking and RF models had the best overall performance (Brier score 0.174 for both) and discrimination (AUROC 0.757 for the stacking model and 0.756 for the RF model). Compared with the benchmark LR model, the stacking model was superior in all performance measures except sensitivity, but there was no significant difference in sensitivity between the 2 models. The 2-sided <i>t</i> tests revealed significant performance differences between the stacking and LR models in overall performance, discrimination, calibration, balanced accuracy, and accuracy. CONCLUSIONS This study is the first to develop data-driven pLOS prediction models for PD-treated patients using basic admission data from a national database. The results indicate the feasibility of utilizing a stacking-based pLOS prediction model for PD-treated patients. The pLOS prediction tools developed in this study have the potential to assist clinicians in identifying patients at a high risk of pLOS and to allocate resources optimally for PD-treated patients.

scholarly journals Development of a multivariable prediction model for plantar foot ulcer recurrence in high-risk people with diabetes

2020 ◽  
Vol 8 (1) ◽  
pp. e001207
Author(s):  
Wouter B aan de Stegge ◽  
Ameen Abu-Hanna ◽  
Sicco A Bus
Keyword(s):  
High Risk ◽  
Prediction Model ◽  
Prediction Models ◽  
Foot Ulcer ◽  
Receiver Operating Curve ◽  
Brier Score ◽  
Ulcer Recurrence ◽  
Increased Risk ◽  
A Minor ◽  
Receiver Operating

IntroductionForty per cent of people with diabetes who heal from a foot ulcer recur within 1 year. The aim was to develop a prediction model for plantar foot ulcer recurrence and to validate its predictive performance.Research design and methodsData were retrieved from a prospective analysis of 171 high-risk patients with 18 months follow-up. Demographic, disease-related, biomechanical and behavioral factors were included as potential predictors. Two logistic regression models were created. Model 1 for all recurrent plantar foot ulcers (71 cases) and model 2 for those ulcers indicated to be the result of unrecognized repetitive stress (41 cases). Ten-fold cross-validation, each including five multiple imputation sets, was used to internally validate the prediction strategy; model performance was assessed in terms of discrimination and calibration.ResultsThe presence of a minor lesion, living alone, increased barefoot peak plantar pressure, longer duration of having a previous foot ulcer and less variation in daily stride count were predictors of the first model. The area under the receiver operating curve was 0.68 (IQR 0.61–0.80) and the Brier score was 0.24 (IQR 0.20–0.28). The predictors of the second model were presence of a minor lesion, longer duration of having a previous foot ulcer and location of the previous foot ulcer. The area under the receiver operating curve was 0.76 (IQR 0.66–0.87) and the Brier score was 0.17 (IQR 0.15–0.18).ConclusionsThese validated prediction models help identify those patients that are at increased risk of plantar foot ulcer recurrence and for that reason should be monitored more carefully and treated more intensively.

scholarly journals A Prediction Model for Metachronous Peritoneal Carcinomatosis in Patients with Stage T4 Colon Cancer after Curative Resection

Cancers ◽  
2021 ◽  
Vol 13 (11) ◽  
pp. 2808
Author(s):  
Tzong-Yun Tsai ◽  
Jeng-Fu You ◽  
Yu-Jen Hsu ◽  
Jing-Rong Jhuang ◽  
Yih-Jong Chern ◽  
...  
Keyword(s):  
Colon Cancer ◽  
High Risk ◽  
Prediction Model ◽  
Prediction Error ◽  
Area Under The Curve ◽  
Low Risk ◽  
Brier Score ◽  
Individual Risk ◽  
T4 Colon Cancer ◽  
Testing Set

(1) Background: The aim of this study was to develop a prediction model for assessing individual mPC risk in patients with pT4 colon cancer. Methods: A total of 2003 patients with pT4 colon cancer undergoing R0 resection were categorized into the training or testing set. Based on the training set, 2044 Cox prediction models were developed. Next, models with the maximal C-index and minimal prediction error were selected. The final model was then validated based on the testing set using a time-dependent area under the curve and Brier score, and a scoring system was developed. Patients were stratified into the high- or low-risk group by their risk score, with the cut-off points determined by a classification and regression tree (CART). (2) Results: The five candidate predictors were tumor location, preoperative carcinoembryonic antigen value, histologic type, T stage and nodal stage. Based on the CART, patients were categorized into the low-risk or high-risk groups. The model has high predictive accuracy (prediction error ≤5%) and good discrimination ability (area under the curve >0.7). (3) Conclusions: The prediction model quantifies individual risk and is feasible for selecting patients with pT4 colon cancer who are at high risk of developing mPC.

scholarly journals Development and Validation of a Clostridium difficile Infection Risk Prediction Model

2011 ◽  
Vol 32 (4) ◽  
pp. 360-366 ◽  
Author(s):  
Erik R. Dubberke ◽  
Yan Yan ◽  
Kimberly A. Reske ◽  
Anne M. Butler ◽  
Joshua Doherty ◽  
...  
Keyword(s):  
High Risk ◽  
Clostridium Difficile ◽  
Prediction Model ◽  
Risk Prediction ◽  
Predictive Accuracy ◽  
Medical Center ◽  
Tertiary Care ◽  
Risk Prediction Model ◽  
Brier Score ◽  
Tertiary Care Medical Center

Objective.To develop and validate a risk prediction model that could identify patients at high risk for Clostridium difficile infection (CDI) before they develop disease.Design and Setting.Retrospective cohort study in a tertiary care medical center.Patients.Patients admitted to the hospital for at least 48 hours during the calendar year 2003.Methods.Data were collected electronically from the hospital's Medical Informatics database and analyzed with logistic regression to determine variables that best predicted patients' risk for development of CDI. Model discrimination and calibration were calculated. The model was bootstrapped 500 times to validate the predictive accuracy. A receiver operating characteristic curve was calculated to evaluate potential risk cutoffs.Results.A total of 35,350 admitted patients, including 329 with CDI, were studied. Variables in the risk prediction model were age, CDI pressure, times admitted to hospital in the previous 60 days, modified Acute Physiology Score, days of treatment with high-risk antibiotics, whether albumin level was low, admission to an intensive care unit, and receipt of laxatives, gastric acid suppressors, or antimotility drugs. The calibration and discrimination of the model were very good to excellent (C index, 0.88; Brier score, 0.009).Conclusions.The CDI risk prediction model performed well. Further study is needed to determine whether it could be used in a clinical setting to prevent CDI-associated outcomes and reduce costs.

Abstract 2701: Multicenter Study For Prediction And Validation Of Intracranial Hemorrhage Risk In The Natural History Of Brain Arteriovenous Malformation Patients

Stroke ◽  
2012 ◽  
Vol 43 (suppl_1) ◽  
Author(s):  
Helen Kim ◽  
Tony Pourmohamad ◽  
Charles E McCulloch ◽  
Michael T Lawton ◽  
Jay P Mohr ◽  
...  
Keyword(s):  
Risk Factors ◽  
High Risk ◽  
Natural History ◽  
Prediction Model ◽  
Intracranial Hemorrhage ◽  
Prediction Models ◽  
Statistical Significance ◽  
Venous Drainage ◽  
Cox Proportional Hazards ◽  
Natural Course

Background: BAVM is an important cause of intracranial hemorrhage (ICH) in younger persons. Accurate and reliable prediction models for determining ICH risk in the natural history course of BAVM patients are needed to help guide management. The purpose of this study was to develop a prediction model of ICH risk, and validate the performance independently using the Multicenter AVM Research Study (MARS). Methods: We used 3 BAVM cohorts from MARS: the UCSF Brain AVM Study Project (n=726), Columbia AVM Study (COL, n=640), and Scottish Intracranial Vascular Malformation Study (SIVMS, n=218). Cox proportional hazards analysis of time-to-ICH in the natural course after diagnosis was performed, censoring patients at first treatment, death, or last visit, up to 10 years. UCSF served as the model development cohort. We chose a simple model, including known risk factors that are reliably measured across cohorts (age at diagnosis, gender, initial hemorrhagic presentation, and deep venous drainage); variables were included without regard to statistical significance. Tertiles of predicted probabilities corresponding to low, medium, and high risk were obtained from UCSF and risk thresholds were validated in COL and SIVMS using Kaplan-Meier survival curves and log-rank tests (to assess whether the model discriminated between risk categories). Results: Overall, 82 ICH events occurred during the natural course: 28 in UCSF, 41 in COL, and 13 in SIVMS. Effects in the prediction model (estimated from UCSF data) were: age in decades (HR=1.1, 95% CI=0.9-1.4, P=0.41), initial hemorrhagic presentation (HR=3.6, 95% CI=1.5-8.6, P=0.01), male gender (HR=1.1, 95% CI=0.48-2.6; P=0.81), and deep venous drainage (HR=0.8, 95% CI=0.2-2.8 P=0.72). Tertiles of ICH risk are shown in the Figure , demonstrating good separation of curves into low, medium and high risk after 3 years in UCSF (left, log-rank P=0.05). The model validated well in the COL referral cohort with better discrimination of curves (middle, P<0.001). In SIMVS, a population-based study, the model separated curves in the earlier years but a consistent pattern was not observed (right, P=0.51), possibly due to the small number of ICH events. Conclusion: Our current prediction model for predicting ICH risk in the natural history course validates well in another referral population, but not as well in a population cohort. Inclusion of additional cohorts and risk factors after data harmonization may improve overall prediction and discrimination of ICH risk, and provide a generalizable model for clinical application.

Clinical Prediction Model and Tool for Assessing Risk of Persistent Pain After Breast Cancer Surgery

2017 ◽  
Vol 35 (15) ◽  
pp. 1660-1667 ◽  
Author(s):  
Tuomo J. Meretoja ◽  
Kenneth Geving Andersen ◽  
Julie Bruce ◽  
Lassi Haasio ◽  
Reetta Sipilä ◽  
...  
Keyword(s):  
Breast Cancer ◽  
High Risk ◽  
Prediction Model ◽  
Prediction Models ◽  
Persistent Pain ◽  
Cancer Surgery ◽  
Breast Cancer Surgery ◽  
Outcome Variable ◽  
Axillary Lymph ◽  
Operating Characteristics

Purpose Persistent pain after breast cancer surgery is a well-recognized problem, with moderate to severe pain affecting 15% to 20% of women at 1 year from surgery. Several risk factors for persistent pain have been recognized, but tools to identify high-risk patients and preventive interventions are missing. The aim was to develop a clinically applicable risk prediction tool. Methods The prediction models were developed and tested using three prospective data sets from Finland (n = 860), Denmark (n = 453), and Scotland (n = 231). Prediction models for persistent pain of moderate to severe intensity at 1 year postoperatively were developed by logistic regression analyses in the Finnish patient cohort. The models were tested in two independent cohorts from Denmark and Scotland by assessing the areas under the receiver operating characteristics curves (ROC-AUCs). The outcome variable was moderate to severe persistent pain at 1 year from surgery in the Finnish and Danish cohorts and at 9 months in the Scottish cohort. Results Moderate to severe persistent pain occurred in 13.5%, 13.9%, and 20.3% of the patients in the three studies, respectively. Preoperative pain in the operative area ( P < .001), high body mass index ( P = .039), axillary lymph node dissection ( P = .008), and more severe acute postoperative pain intensity at the seventh postoperative day ( P = .003) predicted persistent pain in the final prediction model, which performed well in the Danish (ROC-AUC, 0.739) and Scottish (ROC-AUC, 0.740) cohorts. At the 20% risk level, the model had 32.8% and 47.4% sensitivity and 94.4% and 82.4% specificity in the Danish and Scottish cohorts, respectively. Conclusion Our validated prediction models and an online risk calculator provide clinicians and researchers with a simple tool to screen for patients at high risk of developing persistent pain after breast cancer surgery.

scholarly journals The Impact Prediction Models of Neoplasia for Lung Nodules in High-Risk Patients

2020 ◽  
Vol 3 (3) ◽  
pp. 138-146
Author(s):  
Camilla Matos Pedreira ◽  
José Alves Barros Filho ◽  
Carolina Pereira ◽  
Thamine Lessa Andrade ◽  
Ricardo Mingarini Terra ◽  
...  
Keyword(s):  
High Risk ◽  
Prediction Model ◽  
Prediction Models ◽  
Lung Nodule ◽  
Smoking History ◽  
Predictive Values ◽  
Impact Prediction ◽  
High Risk Patients ◽  
Risk Patients ◽  
The Impact

Objectives: This study aims to evaluate the impact of using three predictive models of lung nodule malignancy in a population of patients at high-risk for neoplasia according to previous analysis by physicians, as well as evaluate the clinical and radiological malignancy-predictors of the images. Material and Methods: This is a retrospective cohort study, with 135 patients, undergone surgical in the period from 01/07/2013 to 10/05/2016. The study included nodules with dimensions between 5mm and 30mm, excluding multiple nodules, alveolar consolidation, pleural effusion, and lymph node enlargement. The main variables analyzed were age, sex, smoking history, extrathoracic cancer, diameter, location, and presence of spiculation. The calculation of the area under the ROC curve assessed the accuracy of each prediction model. Results: The study analyzed 135 individuals, of which 96 (71.1%) had malignant nodules. The areas under the ROC curves for each prediction model were: Swensen 0.657; Brock 0.662; and Herder 0.633. The models Swensen, Brock, and Herder presented positive predictive values in high-risk patients, corresponding to 83.3%, 81.8%, and 82.9%, respectively. Patients with the intermediate and low-risk presented a high malignant nodule rate, ranging from 69.3-72.5% and 42.8-52.6%, respectively. Conclusion: None of the three quantitative models analyzed in this study was considered satisfactory (AUC> 0.7) and should be used with caution after specialized evaluation to avoid underestimation of the risk of neoplasia. The pretest calculations might not contemplate other factors than those predicted in the regressions, that could present a role in the clinical decision of resection.

scholarly journals Prediction Models for Radiation-Induced Neurocognitive Decline in Adult Patients with Primary or Secondary Brain Tumors: A Systematic Review

2021 ◽  
Author(s):  
Fariba Tohidinezhad ◽  
Dario Di Perri ◽  
Catharina M.L. Zegers ◽  
Jeanette Dijkstra ◽  
Monique Anten ◽  
...  
Keyword(s):  
High Risk ◽  
Radiation Dose ◽  
Brain Tumors ◽  
Prediction Model ◽  
Prediction Models ◽  
Verbal Learning ◽  
Risk Of Bias ◽  
Neurocognitive Impairments ◽  
Radiation Induced ◽  
Neurocognitive Decline

Abstract Purpose: Although an increasing body of literature suggests a relationship between brain irradiation and deterioration of neurocognitive function, it remains as the standard therapeutic and prophylactic modality in patients with brain tumors. This review was aimed to abstract and evaluate the prediction models for radiation-induced neurocognitive decline in patients with primary or secondary brain tumors.Methods: MEDLINE was searched on October 31, 2021 for publications containing relevant truncation and MeSH terms related to “radiotherapy”, “brain”, “prediction model”, and “neurocognitive impairments”. Risk of bias was assessed using the Prediction model Risk Of Bias ASsessment Tool.Results: Of 3,580 studies reviewed, 23 prediction models were identified. Age, tumor location, education level, baseline neurocognitive score, and radiation dose to the hippocampus were the most common predictors in the models. The Hopkins verbal learning (n=7) and the trail making tests (n=4) were the most frequent outcome assessment tools. All studies used regression (n=14 linear, n=8 logistic, and n=4 Cox) as machine learning method. All models were judged to have a high risk of bias mainly due to issues in the analysis.Conclusion: Existing models have limited quality and are at high risk of bias. Following recommendations are outlined in this review to improve future models: develop a standardized instrument for neurocognitive assessment in patients with brain tumors; adherence to model development and validation guidelines; careful choice of candidate predictors according to the literature and domain expert consensus; and considering radiation dose to brain substructures as they can provide important information on specific neurocognitive impairments.

Predicting poor outcome in patients with suspected COVID-19 presenting to the Emergency Department (COVERED) – Development, internal and external validation of a prediction model

2021 ◽  
Vol 20 (1) ◽  
pp. 4-14
Author(s):  
K. Azijli ◽  
◽  
A.W.E. Lieveld ◽  
S.F.B. van der Horst ◽  
N. de Graaf ◽  
...  
Keyword(s):  
Emergency Department ◽  
Prediction Model ◽  
Clinical Decision Making ◽  
Prediction Models ◽  
External Validation ◽  
Model Performance ◽  
Area Under The Curve ◽  
Brier Score ◽  
Multi Centre Study ◽  
Poor Outcome

Background: A recent systematic review recommends against the use of any of the current COVID-19 prediction models in clinical practice. To enable clinicians to appropriately profile and treat suspected COVID-19 patients at the emergency department (ED), externally validated models that predict poor outcome are desperately needed. Objective: Our aims were to identify predictors of poor outcome, defined as mortality or ICU admission within 30 days, in patients presenting to the ED with a clinical suspicion of COVID-19, and to develop and externally validate a prediction model for poor outcome. Methods: In this prospective, multi-centre study, we enrolled suspected COVID-19 patients presenting at the EDs of two hospitals in the Netherlands. We used backward logistic regression to develop a prediction model. We used the area under the curve (AUC), Brier score and pseudo-R2 to assess model performance. The model was externally validated in an Italian cohort. Results: We included 1193 patients between March 12 and May 27 2020, of whom 196 (16.4%) had a poor outcome. We identified 10 predictors of poor outcome: current malignancy (OR 2.774; 95%CI 1.682-4.576), systolic blood pressure (OR 0.981; 95%CI 0.964-0.998), heart rate (OR 1.001; 95%CI 0.97-1.028), respiratory rate (OR 1.078; 95%CI 1.046-1.111), oxygen saturation (OR 0.899; 95%CI 0.850-0.952), body temperature (OR 0.505; 95%CI 0.359-0.710), serum urea (OR 1.404; 95%CI 1.198-1.645), C-reactive protein (OR 1.013; 95%CI 1.001-1.024), lactate dehydrogenase (OR 1.007; 95%CI 1.002-1.013) and SARS-CoV-2 PCR result (OR 2.456; 95%CI 1.526-3.953). The AUC was 0.86 (95%CI 0.83-0.89), with a Brier score of 0.32 and, and R2 of 0.41. The AUC in the external validation in 500 patients was 0.70 (95%CI 0.65-0.75). Conclusion: The COVERED risk score showed excellent discriminatory ability, also in external validation. It may aid clinical decision making, and improve triage at the ED in health care environments with high patient throughputs.

scholarly journals The HOSPITAL score and LACE index as predictors of 30 day readmission in a retrospective study at a university-affiliated community hospital

PeerJ ◽  
2017 ◽  
Vol 5 ◽  
pp. e3137 ◽  
Author(s):  
Robert Robinson ◽  
Tamer Hudali
Keyword(s):  
Retrospective Study ◽  
High Risk ◽  
Hospital Readmission ◽  
Patient Population ◽  
Goodness Of Fit ◽  
Operating Characteristic ◽  
Brier Score ◽  
Goodness Of Fit Test ◽  
C Statistic ◽  
Overall Performance

IntroductionHospital readmissions are common, expensive, and a key target of the Medicare Value Based Purchasing (VBP) program. Validated risk assessment tools such as the HOSPITAL score and LACE index have been developed to identify patients at high risk of hospital readmission so they can be targeted for interventions aimed at reducing the rate of readmission. This study aims to evaluate the utility of HOSPITAL score and LACE index for predicting hospital readmission within 30 days in a moderate-sized university affiliated hospital in the midwestern United States.Materials and MethodsAll adult medical patients who underwent one or more ICD-10 defined procedures discharged from the SIU-SOM Hospitalist service from Memorial Medical Center (MMC) from October 15, 2015 to March 16, 2016, were studied retrospectively to determine if the HOSPITAL score and LACE index were a significant predictors of hospital readmission within 30 days.ResultsDuring the study period, 463 discharges were recorded for the hospitalist service. The analysis includes data for the 432 discharges. Patients who died during the hospital stay, were transferred to another hospital, or left against medical advice were excluded. Of these patients, 35 (8%) were readmitted to the same hospital within 30 days. A receiver operating characteristic evaluation of the HOSPITAL score for this patient population shows a C statistic of 0.75 (95% CI [0.67–0.83]), indicating good discrimination for hospital readmission. The Brier score for the HOSPITAL score in this setting was 0.069, indicating good overall performance. The Hosmer–Lemeshow goodness of fit test shows a χ2value of 3.71 with apvalue of 0.59. A receiver operating characteristic evaluation of the LACE index for this patient population shows a C statistic of 0.58 (95% CI [0.48–0.68]), indicating poor discrimination for hospital readmission. The Brier score for the LACE index in this setting was 0.082, indicating good overall performance. The Hosmer–Lemeshow goodness of fit test shows a χ2value of 4.97 with apvalue of 0.66.DiscussionThis single center retrospective study indicates that the HOSPITAL score has superior discriminatory ability when compared to the LACE index as a predictor of hospital readmission within 30 days at a medium-sized university-affiliated teaching hospital.ConclusionsThe internationally validated HOSPITAL score may be superior to the LACE index in moderate-sized community hospitals to identify patients at high risk of hospital readmission within 30 days.

Sign in / Sign up

Export Citation Format

Share Document