Predictors Of Readmission Research Articles

Analysis of clinical features and prognosis in cardiac amyloidosis patients from Spanish hospitals (2016-2021).

Cardiac amyloidosis (CA) is a prevalent yet underdiagnosed heart condition characterized by the abnormal accumulation of amyloid fibres, frequently resulting in heart failure (HF), particularly in older people. Despite advancements in non-invasive diagnostic techniques and treatments, the epidemiology of CA patients remains inadequately understood. This nationwide retrospective observational study sought to comprehensively investigate CA patients' characteristics, mortality, and readmission patterns. A retrospective observational study encompassed all patients hospitalized with CA between 2016 and 2021 across Spanish hospitals. Standardized incidence rates were calculated using age and sex-adjusted methods, utilizing the Spanish population as the reference. The investigation delved into demographic variables, comorbidities, mortality during the index episode, and 30 and 365-day readmissions for circulatory system diseases. Predictors of readmission were also examined. A total of 5,739 index episodes were identified, with CA being the primary cause of admission in 14.1% of cases. The mean age was 81.4 ± 9.9 years, predominantly males (70.3%). The age and sex-standardized hospital attendance rate was 3.90 admissions per 100,000 population (95% CI: 3.82-3.98), higher in males. Common comorbidities included HF (96.4%), atrial fibrillation (46.3%), and renal failure (44.4%). The mortality rate during the index episode was 11.7%, with cardiogenic shock (OR: 9.03; 95% CI: 4.22-19.32) and major psychiatric disorders (OR: 3.36; 95% CI: 1.42-7.94) identified as predictors. Over the follow-up period, 13.1% were readmitted at 30 days and 36.6% at 365. Nephritis (IRR: 2.05; 95% CI: 1.42-2.96) and asthma (IRR: 1.52; 95% CI 1.11-2.07) were associated with increased 30-day readmissions, while renal failure (IRR: 1.43; 95% CI: 1.28-1.59) and chronic pulmonary disease (IRR: 1.40; 95% CI: 1.18-1.67) were linked to higher 365-day readmissions. Predictors of mortality risk in 365-day readmissions included advanced cancer (HR: 1.31; 95% CI: 1.00-1.71), cardiogenic shock in the index episode (HR: 2.72; 95% CI: 1.33-5.57), and a higher number of readmissions during that period (HR: 1.64; 95% CI: 1.56-1.73). This study contributes valuable insights into the significant prevalence of CA, which is often overlooked and underestimated. Afflicting predominantly elderly males with concomitant HF and multiple comorbidities, CA poses a significant clinical challenge. The findings underscore the need for enhanced awareness, early detection, and a multidisciplinary approach to manage this complex cardiac condition.

Predictors of COVID-19 Readmission Among Patients Previously Hospitalized for SARS-CoV-2.

Predictors of coronavirus disease 2019 (COVID-19)-related rehospitalization remain underexplored. This study aims to identify the main risk factors associated with rehospitalizations due to severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) reinfections among residents of Lombardy, northern Italy. A retrospective observational study was conducted using two linked administrative databases covering demographic data, comorbidities, hospital records, and COVID-19 data of Lombardy residents. The study population included patients hospitalized for COVID-19 between February 2020 and August 2021. Rehospitalization was defined as a second COVID-19-related hospitalization occurring at least 90days after the first admission. The Fine-Gray subdistribution hazard model was used to identify risk factors, accounting for death as a competing risk. Out of 98,369 patients hospitalized for COVID-19 between February 1, 2020 and August 31, 2021, 72,593 were alive 90days after admission and 610 of these (0.8%) were rehospitalized. A higher rehospitalization risk was observed in older male patients with multiple comorbidities. Renal failure, liver disease, and use of diuretics were significantly associated with rehospitalization risk, while female biological sex and the use of lipid-lowering drugs were associated with a lower risk. This is the first study conducted on regional administrative databases to investigate COVID-19 rehospitalizations. Through the availability of a huge cohort, it provides a groundwork for optimizing care for individuals at higher risk for COVID-19-related rehospitalizations. It underlines the need for patient-management approaches that extend beyond the initial recovery. This stresses the importance of ongoing monitoring and personalized interventions for those at heightened risk not only of SARS-CoV-2 reinfection but also related rehospitalizations.

Comparing prediction accuracy for 30-day readmission following primary total knee arthroplasty: the ACS-NSQIP risk calculator versus a novel artificial neural network model

BackgroundUnplanned readmission, a measure of surgical quality, occurs after 4.8% of primary total knee arthroplasties (TKA). Although the prediction of individualized readmission risk may inform appropriate preoperative interventions, current predictive models, such as the American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP) surgical risk calculator (SRC), have limited utility. This study aims to compare the predictive accuracy of the SRC with a novel artificial neural network (ANN) algorithm for 30-day readmission after primary TKA, using the same set of clinical variables from a large national database.MethodsPatients undergoing primary TKA between 2013 and 2020 were identified from the ACS-NSQIP database and randomly stratified into training and validation cohorts. The ANN was developed using data from the training cohort with fivefold cross-validation performed five times. ANN and SRC performance were subsequently evaluated in the distinct validation cohort, and predictive performance was compared on the basis of discrimination, calibration, accuracy, and clinical utility.ResultsThe overall cohort consisted of 365,394 patients (trainingN = 362,559; validationN = 2835), with 11,392 (3.1%) readmitted within 30 days. While the ANN demonstrated good discrimination and calibration (area under the curve (AUC)ANN = 0.72, slope = 1.32, intercept = −0.09) in the validation cohort, the SRC demonstrated poor discrimination (AUCSRC = 0.55) and underestimated readmission risk (slope = −0.21, intercept = 0.04). Although both models possessed similar accuracy (Brier score: ANN = 0.03; SRC = 0.02), only the ANN demonstrated a higher net benefit than intervening in all or no patients on the decision curve analysis. The strongest predictors of readmission were body mass index (> 33.5 kg/m2), age (> 69 years), and male sex.ConclusionsThis study demonstrates the superior predictive ability and potential clinical utility of the ANN over the conventional SRC when constrained to the same variables. By identifying the most important predictors of readmission following TKA, our findings may assist in the development of novel clinical decision support tools, potentially improving preoperative counseling and postoperative monitoring practices in at-risk patients.

Association between preoperative exercise tolerance and unplanned readmission in patients who underwent pancreatectomy for pancreatic ductal adenocarcinoma: a retrospective analysis

BackgroundDespite advances in treatment, the incidence of postoperative complications following pancreatectomy remains high, leading to frequent hospital readmissions. Therefore, this study aimed to investigate the relationship between preoperative exercise tolerance and the likelihood of unplanned readmission in patients with pancreatic ductal adenocarcinoma.MethodsThis retrospective analysis included 88 patients who underwent pancreatectomy at a single institution between July 2019 and September 2022 and focused on patients with pancreatic ductal adenocarcinoma. Patients’ preoperative exercise tolerance was assessed using the 6-minute walk distance (6 MWD).ResultsThe study found a 22.7% readmission rate within 1 year, with the median 6 MWD being significantly lower in readmitted patients than in those who were not readmitted (390.0 m versus 436.5 m; p < 0.01). A 6 MWD cut-off of 425 m was a strong predictor of readmission, with lower preoperative exercise tolerance associated with a higher risk of readmission.ConclusionsEnhancing preoperative physical reserves through exercise therapy may reduce readmission rates and improve patient outcomes. Further research with larger sample sizes is required to confirm these findings.

External validation of the "simplified hospital" scale as a predictor of 30-day readmission after hospitalisation in osi araba medical services.

A readmission is defined as a new admission to hospital for at least one night in the 30 days following a discharge. Some are caused by events unrelated to the reason for the initial admission (unavoidable) and others caused by the same pathology (related). Different predictive models allow the identification of patients at higher risk of readmission. The "HOSPITAL simplified" score stands out for its performance and simplicity. To validate the "HOSPITAL simplified" scale in a cohort of patients at the University Hospital of Alava. Review of patients discharged from OSI Araba Medical Services in June and November 2023. The scale was applied to each of them and re-admission was studied in the 30 days following discharge. The Brier Score, discrimination and calibration capacity were analysed, using the guidelines of the TRIPOD declaration. In 1849 discharges there were 240 unscheduled readmissions: 73 for causes unrelated to the previous admission and 151 related. For predicting related readmissions the scale obtained a Brier score = 0.0703. The mean score for cases with related readmission was 4.7 versus 2.66 for cases without readmission (P < 0.001). The area under the ROC curve = 0.747. Related readmission occurred in 29.3% of patients at high risk, compared to 17.3% predicted by the scale. The "HOSPITAL simplified" scale showed in OSI Araba a higher discrimination capacity than the original series although it underestimated the real risk. The scale offers good portability.

Diabetes is associated with a higher incidence of short-term mortality risk and readmission in patients who undergo surgical but not transcatheter aortic valve replacement

BackgroundTranscatheter aortic valve replacement (TAVR) is increasingly used for aortic valve replacement instead of surgical aortic valve replacement (sAVR). We aimed to examine the impact of diabetes on 30-day mortality,...

TECHNOLOGY TRANSFER IN HEALTHCARE: LEVERAGING PREDICTIVE MODELS TO OPTIMIZE MEDICAL OUTCOMES

Objective: This study aims to apply a predictive model using a decision tree algorithm on real patient data to identify patterns related to hospital readmission rates, contributing to improved clinical decision-making and advancing technology transfer in healthcare. Theoretical Framework: Grounded in predictive analytics and healthcare management, the study leverages the decision tree algorithm's ability to provide interpretable outputs, enabling healthcare professionals to make informed decisions based on data-driven insights. Method: Real patient data was used to construct a decision tree model. The algorithm was trained to identify factors associated with clinical resubmission rates. Performance measures, such as accuracy, precision, and recall, were applied to validate the model's robustness and predictive capabilities. Results and Discussion: The decision tree successfully identified critical predictors of hospital readmissions, including patient characteristics and clinical factors. The model demonstrated high predictive accuracy, providing actionable insights for stratifying at-risk patients. These findings underline the model’s practical utility in supporting evidence-based interventions. Research Implications: For clinical practice, the model facilitates targeted care and resource allocation, reducing readmission rates. For healthcare management, it demonstrates the potential of technology transfer to integrate predictive models into routine workflows, enhancing efficiency and outcomes. Originality/Value: This study contributes to the growing body of research on predictive analytics in healthcare by showcasing the application of a decision tree model in real-world scenarios. Its emphasis on technology transfer highlights its practical significance, bridging the gap between research and implementation in clinical settings.

Incidence and predictors of readmission following hospitalization for thoracic disc herniation.

Symptomatic thoracic disc herniations (TDH) are relatively rare and can be discovered incidentally on neuroimaging. Surgical interventions for TDH represent only 4 % of all surgeries performed for intervertebral disc pathologies, which are most commonly indicated for myelopathy and radiculopathy. Given the absence of publications on rates of readmissions following hospitalization for TDH, we aim to establish baseline metrics for the 90-day all-cause readmission rates and pertinent risk factors. The 2020 Nationwide Readmissions Database was used to screen for patients with a primary diagnosis of TDH and disc degeneration within the first 9 months of the year. Demographic information, admission details, clinical data, comorbidities, and surgical treatment were extracted. Patients were divided into two groups based on readmission status. A sub-analysis was performed by treatment. Overall, 970 patients met our inclusion criteria. Of these, 183 patients (18.9 %) were readmitted within a mean of 34.58 days. The readmission group was significantly older and more likely to have been admitted non-electively. Surgical treatment was associated with a lower readmission rate. Eight comorbidities differed significantly between the groups. Independent risk factors for readmission included non-surgical treatment, Medicare insurance, hypertension, and depression. We established a 90-day all-cause readmission rate of 18.9 % for TDH. There was no difference in readmission based on patients' initial neurological presentation. Non-surgical treatment was identified as an independent risk factor for readmission, suggesting that timely surgical interventions may reduce the risk of readmission. Medicare insurance, hypertension and depression were also identified as independent risk factors.

Symptomatic Progression, Recurrence, and Long-Term Follow-Up of Patients With Intracranial Epidermoid Cysts.

Intracranial epidermoid cysts are rare, slow-growing but highly recurrent tumors with incompletely understood symptoms, progression, complications, and outcomes. The aim of the study was to characterize the symptomatology, surgical management, and long-term outcomes of these tumors. This single-center retrospective analysis identified patients with pathologically confirmed intracranial epidermoid cysts from 1989 to 2023. Demographic and clinical variables were collected at diagnosis, before and after each resection, and at the latest follow-up. Differences across time points were evaluated using analysis of variance with post hoc Tukey's honestly significant difference tests and Pearson χ2 test with posthoc Bonferroni-corrected z-tests for independent proportions (significance: P ≤ .05). Kaplan-Meier, Cox regression, and multivariate logistic regression analyses were used to investigate predictors for recurrence-free survival and 30-day readmission following first surgery, respectively. Of 146 patients in the final cohort, the average age at diagnosis was 40.3 years, and 52.1% of patients were female. The mean maximum tumor dimension at presentation was 4.3 cm. Most presenting symptoms were significantly reduced at the latest follow-up, including headaches (presentation: 40.4%/latest follow-up: 8.8%, P < .001), vertigo/nausea/vomiting (33.1%/3.7%, P < .001), ataxia/incoordination (22.8%/4.4%, P < .001), and seizures (11.8%/2.9%, P < .005). Similarly, significant improvements were observed in cranial nerve (CN) II deficits (11.0%/3.7%, P = .020) and CN V (11.8%/3.7%, P = .012) and CN VIII excitatory symptoms (11.8%/2.9%, P = .005). Following the first resection, 12.5% of patients were readmitted within 30 days; lumbar drain placement was associated with higher odds of readmission (adjusted odds ratio [95% CI] = 8.26 [1.29-88.19]). Evidence of tumor remnant on the immediate postoperative MRI was associated with a shorter time to recurrence following the first surgery (P = .023). The clinical progression of intracranial epidermoid cysts is marked by a spectrum of troublesome presenting symptoms and a notable tendency for recurrence, particularly following subtotal resection. Nevertheless, symptomatic improvement is the rule even if reoperations are needed, and mortality is exceedingly rare.

Comprehensive analysis of clinical characteristics, management, and prognosis in patients with dilated cardiomyopathy discharged from Spanish hospitals.

Dilated cardiomyopathy (DCM) is a leading cause of heart failure (HF) characterized by left ventricular dilatation and systolic dysfunction not explained by abnormal loading conditions. Despite its prevalence, DCM's epidemiology and prognosis remain poorly studied in our country. A retrospective observational study encompassed patients discharged from all Spanish public hospitals between 2016 and 2021 diagnosed with DCM. Data were extracted from the Minimum Basic Data Set. The study focused on hospital admissions, comorbidities, in-hospital mortality, and readmission rates for circulatory system diseases at 30 and 365days. Among 27,402 index episodes, DCM was the primary diagnosis in 12.4%, predominantly affecting men (72.5%). In-hospital mortality was 8.7%, with significant predictors including cardiogenic shock (OR: 12.4, 95%CI: 9.6-15.9), advanced or metastatic cancer (OR: 4.3, 95%CI: 3.8-5.0), renal failure (OR: 2.4, 95%CI: 2.2-2.7), and chronic liver disease (OR: 2.4, 95%CI: 2.1-2.8). Readmission rates were 7.9% at 30days and 25.5% at 365days, predominantly due to HF. Multivariate analysis identified age (IRR: 1.02, 95%CI: 1.01-1.02), female sex (IRR: 0.87, 95%CI: 0.79-0.96), severe hematological diseases (IRR: 2.12, 95%CI: 1.45-3.10), and metastatic cancer (IRR: 1.65, 95%CI: 1.31-2.07) as predictors of 30-day readmissions. At 365days, predictors included age (IRR: 1.02, 95%CI: 1.01-1.02), female sex (IRR: 0.80, 95%CI: 0.74-0.86), severe hematological diseases (IRR: 2.43, 95%CI: 1.66-3.56), and renal failure (IRR: 1.42, 95%CI: 1.31-1.55). This study highlights the substantial hospitalization burden and mortality risk among DCM patients, emphasizing the necessity for advanced management strategies and specialized cardiac care.

Prediction model of ICU readmission in Chinese patients with acute type A aortic dissection: a retrospective study.

Readmission to the intensive care unit (ICU) remains a severe challenge, leading to higher rates of death and a greater financial burden. This study aimed to develop a nomogram-based prediction model for individuals with acute type A aortic dissection (ATAAD). A total of 846 ATAAD patients were retrospectively enrolled between May 2014 and October 2021. Logistic regression was employed to identify the independent risk factors. The prediction model was evaluated using the Hosmer-Lemeshow (H-L) test, the calibration curve, and the area under the receiver operating characteristic curve (AUC). Decision curve analysis (DCA) was used to assess the clinical utility. 57 (6.7%) ATAAD patients were readmitted to ICU following their release from the ICU. ICU readmission was predicted with age ≥ 65years old, body mass index (BMI) ≥ 28kg/m2, tracheotomy, continuous renal replacement therapy (CRRT), and the length of initial ICU stay were predictors of ICU readmission. The AUC was 0.837 (95%CI: 0.789-0.884) and the model fit the data well (H-L test, P = 0.519). DCA also demonstrated good clinical practicability. This prediction model may be helpful for clinicians to assess the risk of ICU readmission, and facilitate the early identification of ATAAD patients at high risk.

Predictive Analytics in Heart Failure Risk, Readmission, and Mortality Prediction: A Review.

Heart failure is a leading cause of death among people worldwide. The cost of treatment can be prohibitive, and early prediction of heart failure would reduce treatment costs to patients and hospitals. Improved readmission prediction would also greatly help hospitals, allowing them to manage their treatment programs and budgets better. This literature review aims to summarize recent studies of predictive analytics models that have been constructed to predict heart failure risk, readmission, and mortality. Random forest, logistic regression, neural networks, and XGBoost were among the most common modeling techniques applied. Most selected studies leveraged structured electronic health record data, including demographics, clinical values, lifestyle, and comorbidities, with some incorporating unstructured clinical notes. Preprocessing through imputation and feature selection were frequently employed in building the predictive analytics models. The reviewed studies exhibit demonstrated promise for predictive analytics in improving early heart failure diagnosis, readmission risk stratification, and mortality prediction. This review study highlights rising research activities and the potential of predictive analytics, especially the implementation of machine learning, in advancing heart failure outcomes. Further rigorous, comprehensive syntheses and head-to-head benchmarking of predictive models are needed to derive robust evidence for clinical adoption.

Abstract A026: Assessing the impact of obesity on postoperative outcomes in RCC using machine learning

Abstract Introduction Obesity is associated with an increased risk of developing renal cell carcinoma (RCC) but paradoxically correlates with improved outcomes in metastatic cases. Moreover, the data on the relationship between obesity and postoperative outcomes following nephrectomy are inconsistent. Accurate prediction of postoperative outcomes, including complications, 30-day readmissions, and mortality, is essential for improving patient outcomes in surgical procedures. In this study, we utilized ML models to predict such outcomes in RCC patients undergoing nephroureterectomy, radical nephrectomy, partial nephrectomy, and other excision procedures on the kidney using data from the National Surgical Quality Improvement Program (NSQIP; 2016- 2021). Methods A gradient-boosted tree (GBT) ML model was developed and trained to predict the primary outcomes of interest: major complications, minor complications, 30-day readmission, and mortality. Patients with a BMI ≥30 kg/m2 were considered obese, and patients with a BMI &amp;lt;30 kg/m2 were used as controls. The model's performance was rigorously evaluated using various measures, such as AUROC, generalized R-square, and misclassification rate. The model's predictions were validated using separate training (70%) and validation (30%) cohorts to ensure generalizability and applicability in diverse patient populations. Results In the analysis, 36,284 cases were included. Bivariate analysis revealed that obesity was associated with an increased rate of minor complications (5.48% vs. 4.41%, p&amp;lt;0.0001) compared to non-obese individuals with similar major complications (3.33 vs. 3.21, p=0.49), mortality (0.54 vs. 0.64, p=0.22) and 30-day readmission rates (5.74 vs. 5.83, p=0.70). The GBT model demonstrated substantial predictive power across all four outcomes. In predicting mortality, the model achieved an AUROC of 0.80 in the validation set, with a misclassification rate of 0.68%. The prediction of major complications was also robust, with an AUROC of 0.91 in the validation set and a misclassification rate of 2.2%. Similarly, the model's performance in predicting minor complications and 30-day readmissions was strong, with AUROC values of 0.70 and 0.66, respectively, in the validation sets. Conclusions Obesity was linked to a higher rate of minor complications but did not affect other postoperative outcomes in RCC patients undergoing nephrectomy. By integrating a wide range of patient-specific variables, ML models are powerful tools for clinicians to identify high-risk patients and tailor perioperative care accordingly. The interaction between RCC and the surrounding adipose tissue microenvironment may have clinical relevance, warranting further research to elucidate the role of BMI in postoperative complications. Our findings advocate for the expanded use of ML techniques in surgical decision-making, which could enhance patient safety, optimize resource use, and improve overall surgical outcomes. Citation Format: Atulya A. Khosla, Manas Pustake, Sufal Chhabra, Yanjia Zhang, Mukesh Roy, Muni Rubens, Venkataraghavan Ramamoorthy, Anshul Saxena, Ishmael A. Jaiyesimi. Assessing the impact of obesity on postoperative outcomes in RCC using machine learning [abstract]. In: Proceedings of the AACR Special Conference in Cancer Research: Tumor-body Interactions: The Roles of Micro- and Macroenvironment in Cancer; 2024 Nov 17-20; Boston, MA. Philadelphia (PA): AACR; Cancer Res 2024;84(22_Suppl):Abstract nr A026.

Risk factors for readmission following esophagectomy and gastrectomy for cancer.

Hospital readmission after surgery is a key quality indicator. This nationwide cohort study aimed to assess readmission rates following esophagectomy and gastrectomy for cancer and identify associated risk factors. Data were extracted from the Dutch Upper GI Cancer Audit (DUCA) for patients with esophagogastric cancer who underwent esophagectomy or gastrectomy with curative intent between January 2011 and June 2016. Logistic regression analysis identified risk factors for 30-day readmission. In total, 5566 patients were included. Readmission within 30days occurred in 483 of 3488 (13.8%) patients after esophagectomy and 243 of 2078 patients (11.7%) after gastrectomy. Both minor (Clavien Dindo 1-2) and major (Clavien Dindo ≥3) postoperative complications were independent predictors of readmission after esophagectomy (OR 2.99; 95%CI 2.23-4.02; p< 0.001 and OR 5.20; 95%CI 3.82-7.09; p< 0.001). Specific complications included pulmonary (OR 1.49; 95%CI 1.20-1.85; p< 0.001), gastrointestinal (OR 2.43; 95%CI 1.94-3.05; p< 0.001), and infectious (OR 2.27; 95%CI 1.60-3.22; p< 0.001). Prolonged length of stay (pLOS) was associated with higher readmission rates in patients without complications following esophagectomy (OR 1.91 95% CI 1.19-3.07; p= 0.008), but lower rates in those with complications (OR 0.65 95% CI 0.51-0.83; p< 0.001). For gastrectomy, postoperative complications were also linked to readmission (OR 3.18; 95%CI 2.30-4.40; p< 0.001), particularly gastrointestinal (OR 2.16; 95%CI 1.40-3.32; p< 0.001), and infectious (OR 3.80; 95%CI 2.53-5.71; p< 0.001). Readmission after esophagogastric resection is common, particularly among patients with both minor and major postoperative complications. Prolonged stay after esophagectomy impacts readmission risk differently based on the presence of complications.

An interpretable machine learning scoring tool for estimating time to recurrence readmissions in stroke patients

BackgroundStroke recurrence readmission poses an additional burden on both patients and healthcare systems. Risk stratification aims to accurately divide patients into groups to provide targeted interventions at reducing readmission. To accurately predict short and intermediate-term risks of readmission and provide information for further temporal risk stratification, we developed and validated an interpretable machine learning risk scoring tool. MethodsIn this retrospective study, all stroke admission episodes from January 1st 2015 to December 31st 2019 were obtained from the Shanghai Health and Health Development Research Centre database, which covers medical records of all patients hospitalized in 436 medical institutes in Shanghai. The outcome was time to stroke recurrence readmission within 90 days post discharge. The Score for Stroke Recurrence Readmission Prediction (SSRRP) tool was derived via an interpretable machine learning-based system for time-to-event outcomes. SSRRP as six-variable survival score includes sequelae, length of stay, type of stroke, random plasma glucose, medical expense payment, and number of hospitalizations. ResultsA total of 339,212 S admission episodes were finally included in the whole cohort. Among them, 217,393 episodes were included in the training dataset, 54,347 episodes in the internal validation dataset, and 67,472 in the temporal validation dataset. Readmission within 90 days was documented in 33922(9.97 %) episodes, with a median time to emergency readmission of 19 days (Interquartile range: 8–43). In the temporal validation dataset, the SSRRP achieved an integrated area under the curve of 0.730(95 % CI, 0.724–0.737). In addition, SSRRP demonstrated good calibration and clinical benefit rate. ConclusionsIn this retrospective cohort study, the SSRRP, a parsimonious and point-based scoring tool, was developed to predict the risk of recurrent readmission for stroke. It also provided accurate information on the time to stroke readmission, enabling further temporal risk stratification and informed clinical decision-making.

Development and validation of a machine learning model to predict the risk of readmission within one year in HFpEF patients: Short title: Prediction of HFpEF readmission

BackgroundHeart failure with preserved ejection fraction (HFpEF) is associated with elevated rates of readmission and mortality. Accurate prediction of readmission risk is essential for optimizing healthcare resources and enhancing patient outcomes. MethodsWe conducted a retrospective cohort study utilizing HFpEF patient data from two institutions: the First Affiliated Hospital Zhejiang University School of Medicine for model development and internal validation, and the Affiliated Hospital of Xuzhou Medical University for external validation. A machine learning (ML) model was developed and validated using 53 variables to predict the risk of readmission within one year. The model’s performance was assessed using several metrics, including the area under the receiver operating characteristic curve (AUC), accuracy, sensitivity, specificity, F1 score, model training time, model prediction time and brier score. SHAP (SHapley Additive exPlanations) analysis was employed to enhance model interpretability, and a dynamic nomogram was constructed to visualize the predictive model. ResultsAmong the 766 HFpEF patients included in the study, 203 (26.5%) were readmitted within one year. The LightGBM model exhibited the highest predictive performance, with an AUC of 0.88 (95% confidence interval (CI):0.84–0.91), an accuracy of 0.79, a sensitivity of 0.81, and a specificity of 0.78. Key predictors included the E/e’ ratio, NYHA classification, LVEF, age, BNP levels, MLR, history of atrial fibrillation (AF), use of ACEI/ARB/ARNI, and history of myocardial infarction (MI). External validation also demonstrated strong predictive performance, with an AUC of 0.87 (95%CI:0.83–0.91). ConclusionsThe LightGBM model exhibited robust performance in predicting one-year readmission risk among HFpEF patients, providing a valuable tool for clinicians to identify high-risk individuals and implement timely interventions.

Survival models and longitudinal medical events for hospital readmission forecasting

BackgroundThe rate of 30-day all-cause hospital readmissions can affect the funding a hospital receives. An accurate and reliable readmission prediction model could save money and increase quality-of-care. Few projects have explored formulating this task as a survival prediction problem, where models can exploit a real-valued time-to-readmission target. This paper demonstrates the effectiveness of a survival-inspired readmission model, especially when paired with a longitudinal patient representation that is agnostic to disease-cohort and predictive task.MethodsWe forecast readmissions for a population-level cohort of 421,088 patients discharged in 2015 and 2016 from hospitals in Alberta, Canada. Clinical features and sequences of historical medical codes (calculated from at least four full years prior to discharge) from linked administrative sources serve as model inputs. We trained binary 30-day readmission models (XGBoost and a Deep Neural Network) and time-to-event readmission models (CoxPH and N-MTLR) with and without machine-learned medical knowledge at initialization, then compared against the popular LACE-based model using the AUROC score at 30 days (AUROC@30). Survival models are additionally evaluated using concordance, Integrated Brier, and L1-loss scores.ResultsAll models that utilize sequence features markedly out-perform even the best models trained on only clinical features. Further, a time-to-event target improves predictive performance at 30 days, given the same model inputs and architecture. N-MTLR, using solely sequence inputs and initialized with pre-learned medical knowledge, achieves an average AUROC@30 of 0.8460 over five folds with a standard deviation of 0.003. All trained models match or out-perform the LACE baseline of 0.6587±0.003.ConclusionSequences of administrative medical codes contain rich predictive information for forecasting readmissions, and embedding medical knowledge a priori using machine learning provides readmission models an advantageous foundation for training. When combined with a model that can leverage a time-to-event target, excellent performance is possible on the 30-day all-cause readmission task using only administrative data.

Abstract 4134692: In-Hospital Mortality Rate and Predictors of 30-Day Readmission in Cancer Patients with MI Undergoing PCI -A Cross Sectional Study From Nationwide Readmission Database

Background and objectives: Data regarding readmission rates and predictors of readmission in cancer patients undergoing PCI are sparce. With the increasing survival rates and prevalence of cardiovascular complications in cancer patients, understanding the patterns and predictors of readmission in this population is paramount for optimizing their outcomes. Cancer patients pose unique clinical challenges due to their combined prothrombotic state and propensity for bleeding. We attempted to identify factors associated with readmission in cancer patients. Methods: We utilized the Nationwide Readmission Database from 2016 to 2020 and included patients more than 18 years of age with primary diagnosis of myocardial infarction(MI) who underwent percutaneous coronary intervention(PCI) and have a preexisting diagnosis of cancer. We used International Classification of Disease, Tenth Revision, Clinical Modification (ICD10 CM) codes to define MI, PCI, and cancer. The primary outcome was the 30-day readmission rate, and secondary outcomes were mortality rates, predictors of readmission, and common causes of readmission. The independent predictors of readmission were analyzed using cox regression analysis. Results: Of the 52,307 cancer patients who underwent PCI, 7,767 were readmitted within a 30-day period. The readmission rate for these patients was 15.70%. The mortality rate was 6.05% for index admission and 6.80% for readmitted cases. Among the readmitted patients in the strongest independent predictor for readmission were leaving against medical advice(AMA), anemia, congestive heart failure, and discharge to a skilled nursing facility or home health. Common causes of readmission within this time included hypertensive heart disease with concomitant CKD stage I-IV and heart failure (6.21%), sepsis (6.12%), NSTEMI (5.60%), hypertensive heart disease with concomitant heart failure (4.62%) and acute kidney injury (1.98%). Conclusions: Thirty-day readmission rate was 15.70%. Independent predictors of readmission were anemia, diabetes mellitus, congestive heart failure, malnutrition, peripheral artery disease, leaving against medical advice, and discharge to facility. Most common cause of readmission was hypertensive heart and kidney disease with heart failure, which comprised 6.21%.

Abstract 4145842: The Incidence, Predictors and Impact of 30-day Readmission of Takotsubo Cardiomyopathy Hospitalization: A Retrospective Nationwide Analysis

Objective: This study aimed to describe the recent readmission and mortality rates of unplanned 30-day all cause readmission after takotsubo cardiomyopathy hospitalization. Methods: We conducted an analysis of the 2020 Nationwide Readmission Database. Using the ICD-10 code, we identified hospitalizations for takotsubo cardiomyopathy. We excluded hospitalizations for age&lt;18years and planned or elective readmissions. To compare baseline characteristics between readmissions and index hospitalizations, Pearson χ2 tests were employed. We utilized multivariate Cox regression analyses to identify independent predictors of readmissions. Results: A total of 7554 patients were hospitalized for index admission for takotsubo cardiomyopathy, and 633 of them were readmitted. The 30-day all-cause readmission rate was 8.5% while specific takotsubo readmission rate was 0.4%. The mortality rate in the index admission was 1.9% and 5.5% in readmitted patients. The mean age of index admission was 67.6years vs 70.7 years in readmission, about 89.0% were female in index admission vs 85.0% in readmission cohorts. The most common reasons for readmission were hypertensive heart disease with heart failure, hypertensive heart disease and chronic kidney disease with heart failure, takotsubo syndrome, sepsis, NSTEMI, COPD exacerbation, acute kidney failure, pneumonia, atrial fibrillation. Readmissions had more acute kidney injury (20.6% vs 11.3%, p-value &lt;0.001) and atrial fibrillation (23.0% vs 13.0%, p &lt;0.001) compared to the index admission, however the index admission had more ventricular tachycardia (4.6% versus 1.8%, p=0.02) compared to readmission cohorts. Readmission had longer length of hospital stay (4.9 days vs 3.7 days, p&lt; 0.001), but less mean total hospital charge of (US$51039 vs US$60617, P =0.008). Older age, acute renal failure, pulmonary disease, coronary dissection, longer length of hospital stays and transfer to rehabilitation facility after discharge were identified as statistically significant predictors of readmissions. Conclusion: About one in twelve patients hospitalized with takotsubo cardiomyopathy were readmitted within 30days. Readmissions had a higher mortality rate; acute renal failure, pulmonary disease, coronary dissection were notable reasons for readmission. Understanding the predictors for readmission highlighted in this study would enable clinicians implement preventive strategies to improve the rate of takotsubo cardiomyopathy readmission

Abstract 4119129: PREDICTING 30-DAY ALL-CAUSE READMISSION AFTER STROKE: THE APPLICATION OF MACHINE LEARNING TO LINKED REGISTRY DATA

Background: Thirty-day readmission following hospital discharge for stroke is an important quality measure for US hospitals. Current US prediction models for post stroke readmission based on electronic medical records from single healthcare systems or hospitals have modest discrimination (AUC range 0.64 - 0.74). Aim: To develop 30-day all-cause readmission prediction model using a machine learning (ML) based method trained on linked stroke registry and administrative claims data. Methods: Using probabilistic linking, we matched acute stroke (ICD-10 I61-I63) discharges from 31 hospitals participating in the Michigan Acute Stroke registry between 2016-2020 to multipayer administrative claims data provided by the Michigan Value Collaborative for Medicare and Blue Cross Blue Shield of Michigan commercial beneficiaries. Stroke registry data included patient demographics, clinical characteristics, past medical history, and treatments. Claims data was used to identify readmissions within 30 days of discharge. We used multivariable LASSO logistic regression- a simple ML technique to predict 30-day all-cause-readmission and evaluated the prediction accuracy using a hospital-split internal validation scheme to generate hospital-specific and pooled AUC estimates with 95% confidence intervals (Figure 1). Results: Of 19,382 linked stroke discharges, 2,724 (14.1%) were readmitted within 30-days. Readmitted patients were older, more likely to be male, black, and have higher stroke severity (NIHSS &gt;5). Registry hospitals were either primary (64%) or comprehensive (26%) stroke centers. Hospital-specific 30-day readmission ranged between 9.9%-23.1% ( P &lt;.001) with an average of 14.1% (95% CI:13.6%-14.5%) (Figure 2). Hospital specific AUC estimates ranged between 0.60-0.80 with a pooled AUC of 0.68 (95% CI:0.65-0.70) (Figure 2). Conclusions: ML prediction model fitted to linked registry data can be used to predict hospital-specific and statewide readmissions post stroke.