Risk Of Unplanned Readmission Research Articles

The burden of readmission after discharge from necrotizing soft tissue infection.

The need for extensive surgical debridement with necrotizing soft tissue infections (NSTIs) may put patients at high risk for unplanned readmission. However, there is a paucity of data on the burden of readmission in patients afflicted with NSTI. We hypothesized that unplanned readmission would significantly contribute to the burden of disease after discharge from initial hospitalization. The Nationwide Readmission Database was used to identify adults undergoing debridement for NSTI hospitalizations from 2010 to 2017. Risk factors for 90-day readmission were assessed by Cox proportional hazards regression. There were a total of 82,738 NSTI admissions during the study period, of which 25,076 (30.3%) underwent 90-day readmissions. Median time to readmission was 25 days (interquartile range, 9-49 days). Fragmentation of care, longer length of index stay (>2 weeks), and Medicaid status were independent risk factors for readmission. Median cost of a readmission was US $10,543. Readmission added 174,640 hospital days to episodes of care over the study period, resulting in an estimated financial burden of US $1.4 billion. Unplanned readmission caused by NSTIs is common and costly. Interventions that target patients at risk for readmission may help decrease the burden of disease. Economic/Epidemiological, level IV.

The Prognostic Value of Serum Zinc Levels in Acutely Hospitalized Patients: a Systematic Review.

Several small studies have identified possible associations between low serum zinc levels and worse outcomes in patients acutely hospitalized for a variety of diseases. This study systematically evaluated all published literature to determine whether serum zinc might independently predict important outcomes in hospitalized patients. PubMed, Embase, and Cochrane Libraries databases were searched from 1970 to 2019 to identify all citations having the keyword "zinc" with hospital outcomes. Studies were included if they measured the association between serum zinc levels in non-electively hospitalized patients with survival, length of stay, or unplanned readmission. Data were double-abstracted to evaluate the association between zinc levels and these outcomes. Our search returned 1091 citations of which 12 studies met the inclusion criteria. Studies were small (median 112.5 patients) using unspecified sampling methods. Seven studies were restricted to critically ill patients. Mean zinc levels ranged from 27.5 to 85.3 μg/dL. Baseline mean zinc levels were significantly lower (p < 0.05) in patients who eventually died in 1 of 7 studies. Five of seven studies found significantly increased risk of death in hospital with lower zinc levels. Associations between zinc levels and critical care or hospital length of stay were unclear. In 1 study, lower zinc levels were associated with a significantly increased risk of unplanned readmission. Current studies measuring the association between serum zinc levels and outcomes in acutely hospitalized patients are limited by their sample sizes, select patient populations, and limited statistical analyses. The association of zinc levels with hospital patient outcomes is unclear.

618. Post-discharge Care Coordination Telephone Call for Patients Enrolled in the UNC OPAT Program

BackgroundAt hospital discharge, patients enter a transition period that carries risks for their health and access to care, particularly those receiving home infusion. In April 2019, to enhance care coordination at discharge, our OPAT program implemented a structured telephone outreach program.MethodsStarting in April 2019, the UNC OPAT program rolled out a patient outreach intervention. Within approximately 3 days after discharge, a pharmacist or nurse coordinator called the patient or caregiver to discuss UNC’s OPAT program, review all medications and potential adverse effects, discuss home infusion and line care, schedule relevant follow-up appointments and transportation, and address patient/caregiver concerns. To evaluate dissemination and impact, we analyzed data from UNC OPAT patients from April 2019-May 2020. We abstracted EHR data for unplanned readmissions and adverse events (hepatotoxicity, nephrotoxicity, neutropenia, eosinophilia, thrombocytopenia, creatine kinase elevation, rash, Clostridioides difficile infection, line complications). We estimated risk differences (RD) to compare outcomes between contacted versus uncontacted patients. To provide context, we also assessed unplanned readmissions and adverse events in a historical control period from April 2018-March 2019, before implementation of the outreach program.ResultsAfter rollout of the outreach program, 374 patients completed their OPAT course and all were targeted for outreach. The success rate of contacting patients was 61%. Median age was similar between contacted and uncontacted patients (52 versus 55). Unplanned readmissions occurred less frequently for contacted patients (14% versus 21%; RD -7%; 95%CI -15%, 1%). Risk of adverse events was similar between contact groups (58% versus 54%; RD 4%; 95%CI -6%, 15%). In the historical control period (n=287, median age 56), unplanned readmission risk was 22% and adverse event risk was 63%.ConclusionPatients who completed the outreach phone call had lower risk of readmission compared to those who could not be reached. This intervention may be a simple, low-cost way to reduce readmissions for OPAT patients, but further study and a sustainable reimbursement structure are needed.Disclosures All Authors: No reported disclosures

607. Identifying Intervention Opportunities to Prevent Readmissions during OPAT

BackgroundPatients receiving outpatient parenteral antimicrobial therapy (OPAT) experience high rates of unplanned readmissions. To inform interventions that may reduce risk of unplanned readmissions during OPAT, we examined the frequency and reasons for readmission in a large cohort of OPAT patients.MethodsWe analyzed data on all patients enrolled in UNC’s OPAT program from February 2015-February 2020. Patients were evaluated by an infectious diseases (ID) physician prior to OPAT enrollment, discharged with >14 remaining days of prescribed therapy, and received care coordination and systematic monitoring by an ID pharmacist. We abstracted EHR data into a REDCap database to ascertain information on each patient’s OPAT course and readmission details: length of stay, primary ICD-9-CM/ICD-10-CM diagnosis code associated with readmission, and reason for readmission from clinical notes. Diagnosis codes and notes were adjudicated and summarized by a multidisciplinary team.ResultsAmong 1,165 OPAT courses, 19% resulted in at least one readmission during therapy, lasting for a median length of stay of 5 days. Among those patients who were readmitted during OPAT, the median time from OPAT start to readmission was 17 days (interquartile range, IQR: 8-29 days). 66% of readmissions preceded the scheduled follow-up appointment during OPAT (median time to scheduled follow-up was 27 days, IQR: 15-35 days). 55% of readmissions were unrelated to OPAT diagnosis. Based on ICD-9-CM/ICD-10-CM code classifications, the most common infectious diseases-related reasons for readmission were worsening OPAT infection (18%), OPAT-related adverse drug reaction (12%), and new infection (11%).ConclusionOne-fifth of OPAT courses resulted in readmission during therapy. Half of readmissions were associated with OPAT or other infection, and half were for other reasons. Earlier post-discharge follow-up by a multidisciplinary team (including primary care providers, case management, and OPAT) might prevent infection-related readmissions for OPAT patients. Future work should also address the need for enhanced care coordination with non-infectious disease providers to manage OPAT patients.Disclosures All Authors: No reported disclosures

One-Year Readmission and Emergency Department Presentation After an Epilepsy-Related Admission in People with Intellectual Disability: A Registry-Linkage Study

IntroductionIntellectual disability (ID) is a neurodevelopmental condition that affects approximately 1-2% of the population, and epilepsy is a common comorbidity in people with ID. Although hospital admission for epilepsy is also common, little is known about the impact of ID on healthcare use following an epilepsy admission.&#x0D; Objectives and ApproachUsing linked administrative health datasets that included hospital admissions and disability service data, we aimed to examine whether the presence of ID led to greater or lesser use of healthcare services following an epilepsy admission, as represented by all-cause unplanned readmissions and emergency department (ED) presentations. Comparing the rate of readmissions and ED presentations within 30, 90 and 365 days of the first epilepsy admission during the study period, the effect of ID was assessed using Poisson regression.&#x0D; ResultsA total of 18,987 patients had an epilepsy admission between 2005 and 2014, and of these, 3,256 (17.1%) had ID. Compared to patients without ID, patients with ID had a higher risk of unplanned readmissions within each follow-up period (adjusted incidence rate ratio (IRR) with 95% CI: 30 days: 1.48 (1.34, 1.65); 90 days: 1.42 (1.31, 1.54); 365 days: 1.49 (1.40, 1.59)). Differences were also found in the reasons for readmission, including more readmissions for neurological disorders. Similarly, the ED presentation risk was elevated in patients with ID (adjusted IRR: 30 days: 1.34 (1.23, 1.46); 90 days: 1.33 (1.24, 1.42); 365 days: 1.38 (1.30, 1.46)).&#x0D; Conclusion / ImplicationsIn patients with epilepsy, the presence of ID appears to increase the chance of a readmission or ED presentation following a hospital admission for epilepsy, with the reasons for readmission also potentially different. This suggests the potential for improvements in post-admission screening and management.

Unplanned readmission and survival after video-assisted thoracic surgery and open thoracotomy in patients with non-small-cell lung cancer: a 12-month nationwide cohort study.

To compare outcomes at 12 months between video-assisted thoracic surgery (VATS) and open thoracotomy (OT) in patients with non-small-cell lung cancer (NSCLC) using real-world evidence. We did a nationwide propensity-matched cohort study. We included all patients who had a diagnosis of NSCLC and who benefitted from lobectomy between 1 January 2015 and 31 December 2017. We divided this population into 2 groups (VATS and OT) and matched them using propensity scores based on patients' and hospitals' characteristics. Unplanned readmission, mortality, complications, length of stay and hospitalization costs within 12 months of follow-up were compared between the 2 groups. A total of 13027 patients from 180 hospitals were included, split into 6231 VATS (47.8%) and 6796 OT (52.2%). After propensity score matching (5617 patients in each group), VATS was not associated with a lower risk of unplanned readmission compared with OT [20.7% vs 21.9%, hazard ratio 1.03 (0.95-1.12)] during the 12-months follow-up. Unplanned readmissions at 90 days were mainly due to pulmonary complications (particularly pleural effusion and pneumonia) and were associated with higher mortality at 12 months (13.4% vs 2.7%, P < 0.0001). VATS and OT were both associated with high incidence of unplanned readmissions within 12 months, requiring a better identification of prognosticators of unplanned readmissions. Our study highlights the need to improve prevention, early diagnosis and treatment of pulmonary complications in patients with VATS and OT after discharge. These findings call for improving the dissemination of systematic perioperative care pathway including efficient pulmonary physiotherapy and rehabilitation.

Dynamic individual vital sign trajectory early warning score (DyniEWS) versus snapshot national early warning score (NEWS) for predicting postoperative deterioration

AimsInternational early warning scores (EWS) including the additive National Early Warning Score (NEWS) and logistic EWS currently utilise physiological snapshots to predict clinical deterioration. We hypothesised that a dynamic score including vital sign trajectory would improve discriminatory power. MethodsMulticentre retrospective analysis of electronic health record data from postoperative patients admitted to cardiac surgical wards in four UK hospitals. Least absolute shrinkage and selection operator-type regression (LASSO) was used to develop a dynamic model (DyniEWS) to predict a composite adverse event of cardiac arrest, unplanned intensive care re-admission or in-hospital death within 24 h. ResultsA total of 13,319 postoperative adult cardiac patients contributed 442,461 observations of which 4234 (0.96%) adverse events in 24 h were recorded. The new dynamic model (AUC = 0.80 [95% CI 0.78−0.83], AUPRC = 0.12 [0.10−0.14]) outperforms both an updated snapshot logistic model (AUC = 0.76 [0.73−0.79], AUPRC = 0.08 [0.60−0.10]) and the additive National Early Warning Score (AUC = 0.73 [0.70−0.76], AUPRC = 0.05 [0.02−0.08]). Controlling for the false alarm rates to be at current levels using NEWS cut-offs of 5 and 7, DyniEWS delivers a 7% improvement in balanced accuracy and increased sensitivities from 41% to 54% at NEWS 5 and 18% to –30% at NEWS 7. ConclusionsUsing an advanced statistical approach, we created a model that can detect dynamic changes in risk of unplanned readmission to intensive care, cardiac arrest or in-hospital mortality and can be used in real time to risk-prioritise clinical workload.

Preoperative simplified geriatric assessment in planned hip and knee arthroplasty.

Hip and knee arthroplasties are the most common planned orthopedic surgical procedures in older persons. It would be useful to identify frailties before surgery to improve the outcome of older patients. The objective of this work was to identify the criteria of a simplified comprehensive geriatric assessment (mini-CGA) that were associated with unplanned hospital readmission and postoperative complications within 3months after the planned hip and/or knee arthroplasty in patients ≥ 65years. This prospective study was carried out in the orthopedic department of Marseille University Hospital from January to May 2019. A mini-CGA was performed preoperatively. One hundred four patients were included in the study. The rate of early readmission within 3months after surgery was 12.5% and the rate of postoperative complications was 40.4%. In multivariate analysis, dependence in the activities of daily living (ADL ≤ 5) was the only factor associated with unplanned readmission (aOR = 9.9, 95% CI 1.9-50.8), and living alone was the only factor associated with postoperative complications (aOR = 3.2, 95% CI 1.2-8.8). We found that the ADL score was associated with the risk of unplanned readmission in older patients undergoing planned arthroplasty, and that living alone was associated with postoperative complications. A preoperative mini-CGA appears essential to limit postoperative morbidity.

Feasibility of Early Discharge after Open Hypopharyngeal Surgery for Dysphagia.

To determine the rate of surgical complications, unplanned readmission, and functional status after open hypopharyngeal surgery for dysphagia with early return to oral diet and hospital discharge. Retrospective chart review of patients who underwent open hypopharyngeal surgery for management of dysphagia from March 2013 to June 2018 at a single academic institution. A clear liquid diet is restarted the day of surgery and is advanced to a soft diet on postoperative day one. Univariate and multivariate logistic regression was performed to identify risk factors for postoperative complications and unplanned readmission. Ninety patients met eligibility criteria; 62 patients underwent open cricopharyngeal myotomy, 13 underwent Zenker's diverticulectomy, 15 had a Zenker's diverticulopexy. Mean inpatient length of stay was 2.0 ± 2.0 days (range 0-11 days); 57 patients (63.3%) were discharged on the same day as surgery or on postoperative day one. Seven patients (7.8%) had postoperative complications; the most common complication was esophageal leak (n = 6). Six patients (6.7%) had unplanned readmissions within 30 days. Mean time to unplanned readmission was 12 days (range 2-19 days). Open diverticulectomy was associated with an increased risk of unplanned readmission with an OR = 7.63 (95% CI 1.29, 45.45, P = .025). At last follow-up, 70% of patients had an increased functional oral intake by at least one scale score (FOIS) after surgery (mean follow-up 0.8 ± 1.1 years, range 0.02-5.1 years). An early diet and discharge may be safe for patients undergoing open surgery without a mechanical or suture closure of their esophageal mucosa, whereas in those undergoing diverticulectomy, NPO status and a radiological check before resuming drinking and eating are advised. 4.

Unplanned Readmission within 28 Days of Hospital Discharge in a Longitudinal Population-Based Cohort of Older Australian Women.

This study aimed to estimate the incidence of 28-day unplanned readmission among older women, and associated factors. Data were used from the 1921–1926 birth cohort of the Australian Longitudinal Study on Women’s Health. Linkage of self-reported survey data with the Admitted Patient Data Collection allowed the identification of hospital admissions for each woman and the corresponding baseline characteristics. The Cox proportional-hazards model was used to identify factors associated with time to unplanned readmission, using SAS software V 9.4. (SAS Institute, Cary, NC, USA). Of 2056 women with index unplanned admission, 363 (17.5%) were readmitted within 28 days of discharge, and of these 229 (11.14%) had unplanned readmission. Among women with unplanned readmission, 24% were for the same condition as for the index hospitalisation. Cardiovascular diseases were the main diagnoses for the index admission and readmission. Unplanned readmission risk was higher if not partnered (hazard ratio (HR) = 1.43, 95% confidence interval (CI): 1.05–1.95), of non-English speaking background (HR = 1.62%, 95% CI: 1.07–2.47), more than three days length of stay on index admission (HR = 1.41%, 95% CI: 1.04–1.90) and one or two of the assessed chronic diseases (HR = 1.68, 95% CI: 1.19–2.36). At least one in ten women had unplanned readmission at some time between ages 75–95 years. Women who are not partnered, not of English-speaking background, with longer hospital stay and those with multi-morbidity, may need further efforts during their stay and on discharge to mitigate unplanned readmission.

ST2 Predicts Risk of Unplanned Readmission Within 1 Year After Pediatric Congenital Heart Surgery

Approximately 10% to 20% of children are readmitted after congenital heart surgery. Very little is known about biomarkers as predictors of risk of unplanned readmission after pediatric congenital heart surgery. Novel cardiac biomarker ST2 may be associated with risk of unplanned readmission. ST2 concentrations are believed to reflect cardiovascular stress and fibrosis. Our objective was to explore the relationship between pre- and postoperative ST2 biomarker levels and risk of readmission within 1 year after congenital heart surgery. We prospectively enrolled pediatric patients aged < 18 years who underwent at least 1 congenital heart operation at Johns Hopkins Hospital from 2010 to 2014. Plasma samples were collected immediately before surgery and at the end of bypass. We used Kaplan-Meier survival analysis and multivariable Cox regression models to adjust for variables used in The Society of Thoracic Surgeons Congenital Heart Surgery Database mortality risk model. Of our cohort of 145 patients, we found 39 children with readmissions within 365 days. The median time to unplanned readmission was 54 days (interquartile range, 10-153). Kaplan-Meier analysis demonstrated a significant difference across terciles of pre- and postoperative ST2 biomarker levels. After adjustment, elevated serum levels of ST2 measured preoperatively and postoperatively were associated with increased risk of readmission (hazard ratio, 2.5-3.7; all P < .05). Elevated levels of ST2 are significantly associated with increased risk of unplanned readmission within 1 year after pediatric congenital heart surgery. Novel serum biomarker ST2 can be used for risk stratification or estimating postsurgical prognosis.

Comparison of Appendectomy for Perforated Appendicitis With and Without Abscess: A National Surgical Quality Improvement Program Analysis

BackgroundOutcomes of appendectomy stratified by type of complicated appendicitis (CA) features are poorly researched, and the evidence to guide operative versus nonoperative management for CA is lacking. This study aimed to determine laparoscopic-to-open conversion risk, postoperative abscess risk, unplanned readmission risk, and length of hospital stay (LOS) associated with appendectomy in patients with perforated appendicitis without abscess (PA) and perforated appendicitis with abscess (PAWA) compared with a control cohort of nonperforated appendicitis (NPA). MethodsThe 2016-2017 National Surgical Quality Improvement Program Appendectomy-targeted database identified 12,537 (76.1%) patients with NPA, 2142 (13.0%) patients with PA, and 1799 (10.9%) patients with PAWA. Chi-squared analysis and analysis of variance were used to compare categorical and continuous variables. Binary logistic and linear regression models were used to compare risk-adjusted outcomes. ResultsCompared with NPA, PA and PAWA had higher rates of conversion (0.8% versus 4.9% and 6.5%, respectively; P < 0.001), postoperative abscess requiring intervention (0.6% versus 4.8% and 7.0%, respectively; P < 0.001), readmission (2.8% versus 7.7% and 7.6%, respectively; P < 0.001), and longer median LOS (1 day versus 2 days and 2 days, respectively; P < 0.001). PA and PAWA were associated with increased odds of postoperative abscess (odds ratio [OR]: 7.18, 95% confidence interval [CI]: 5.2-9.8 and OR: 9.94, 95% CI: 7.3-13.5, respectively), readmission (OR: 2.70, 95% CI: 2.1-3.3 and OR: 2.66, 95% CI: 2.2-3.3, respectively), and conversion (OR: 5.51, 95% CI: 4.0-7.5 and OR: 7.43, 95% CI: 5.5-10.1, respectively). PA was associated with an increased LOS of 1.7 days and PAWA with 1.9 days of LOS (95% CI: 1.5-1.8 and 1.7-2.1, respectively). ConclusionsIndividual features of CA were independently associated with outcomes. Further research is needed to determine if surgical management is superior to nonoperative management for CA.

Development and Validation of a Web-Based Pediatric Readmission Risk Assessment Tool.

Accurately predicting and reducing risk of unplanned readmissions (URs) in pediatric care remains difficult. We sought to develop a set of accurate algorithms to predict URs within 3, 7, and 30 days of discharge from inpatient admission that can be used before the patient is discharged from a current hospital stay. We used the Children's Hospital Association Pediatric Health Information System to identify a large retrospective cohort of 1 111 323 children with 1 321 376 admissions admitted to inpatient care at least once between January 1, 2016, and December 31, 2017. We used gradient boosting trees (XGBoost) to accommodate complex interactions between these predictors. In the full cohort, 1.6% of patients had at least 1 UR in 3 days, 2.4% had at least 1 UR in 7 days, and 4.4% had at least 1 UR within 30 days. Prediction model discrimination was strongest for URs within 30 days (area under the curve [AUC] = 0.811; 95% confidence interval [CI]: 0.808-0.814) and was nearly identical for UR risk prediction within 3 days (AUC = 0.771; 95% CI: 0.765-0.777) and 7 days (AUC = 0.778; 95% CI: 0.773-0.782), respectively. Using these prediction models, we developed a publicly available pediatric readmission risk scores prediction tool that can be used before or during discharge planning. Risk of pediatric UR can be predicted with information known before the patient's discharge and that is easily extracted in many electronic medical record systems. This information can be used to predict risk of readmission to support hospital-discharge-planning resources.

Complications and Readmissions After Total Hip Replacement in Octogenarians and Nonagenarians.

Introduction:As the US population ages, the need for total hip arthroplasty (THA) is predicted to increase by 174% by 2030. The purpose of our study was to examine the rate and risks of 30-day complications and unplanned readmission in patients over the age of 80 years old.Materials and Methods:The National Surgical Quality Improvement Project database for the years 2008 to 2014 was queried for patients over the age of 80 undergoing THA. The risks of major complications, minor complications, bleeding occurrences, unplanned readmissions, and deaths over a 30-day period were examined using univariate and multivariate analyses.Results:A total of 7730 patients were included for analysis; 324 (4.2%) patients experienced a major complication, 1944 (25.1%) patients experienced a minor complication, 1776 (22.9%) patients had a bleeding occurrence requiring transfusion, and 376 (4.9%) patients experienced an unplanned readmission to the hospital within 30 days. A total of 33 (0.4%) patient deaths were recorded within 30 days postoperatively. Patients with congestive heart failure were at increased risk for developing a major complication, odds ratio (OR) 3.618 (1.052-12.437), and postoperative death, OR 11.920 (1.362-104.322). Patients with an American Society of Anesthesiologists >2, OR: 1.351 (1.131-1.613), and an operative time greater than 120 minutes, OR: 1.346 (1.084-1.670), had increased risks of minor complications. Patients with a body mass index greater than 30 had reduced rates of minor complications, OR: 0.609 (0.486-0.763). Increased risk of unplanned readmission was seen in patients with chronic obstructive pulmonary disease, OR: 2.403 (1.324-4.359).Discussion/Conclusion:Elderly patients undergoing THA have high complication and readmission rates. Surgeons should convey these increased risk factors and rates of complications in elective THAs to their elderly patients and work with primary care physicians to mitigate these risks.

792 Assessment of use of Beta-blockers Post PCI for Difference in Adherence by Sex and Effect on Outcome. Results from a National Multicenter Outcome Registry in Australia (GCOR)

Adherence to secondary prevention medication has a potential significant benefit on outcome and mortality. We examined whether adherence to betablockers differ by sex and the effect of adherence to beta-blockers on outcome using data from the Genesis Cardiovascular Outcome Registry (GCOR). Data was prospectively collected on 8255 patients (1866 women and 6386 men). Patients were assessed for 1-year adherence and effect on outcome measured by death, MACE, MI, TVR, and unplanned readmission. Drug adherence at 1-year was significantly lower in men compared to women (47.9% vs 51.3%, p=0.01). Patients adherent to beta-blockers at 1-year had a significant reduction in the risk of death (HR 0.30; 95% CI 0.14-0.61, p=0.001), with the benefit similar for both women (HR 0.17; 95% CI 0.04-0.79, p=0.02) and men (HR 0.37; 95% CI 0.16-0.86, p=0.02). A significant increase in risk for myocardial infarction at 1-year was noted for the whole cohort (HR 2.72; 95% CI 1.44-5.15, p=0.002), with the highest risk of myocardial infarction noted in men (HR 3.30; 95% CI 1.55-7.02, p=002). An increased risk for unplanned readmission was noted (HR 1.20; 95% CI 1.03-1.39, p=0.02), with similar rates seen for women and men. Adherence to betablockers at 1-year was associated with a significant reduction in death. The increase risk of MI in men and higher rate of unplanned readmission needs further investigation.

Factors Associated With 90-Day Readmission After Stroke or Transient Ischemic Attack: Linked Data From the Australian Stroke Clinical Registry.

Background and Purpose- Readmissions after stroke are common and appear to be associated with comorbidities or disability-related characteristics. In this study, we aimed to determine the patient and health-system level factors associated with all-cause and unplanned hospital readmission within 90 days after acute stroke or transient ischemic attack (TIA) in Australia. Methods- We used person-level linkages between data from the Australian Stroke Clinical Registry (2009-2013), hospital admissions data and national death registrations from 4 Australian states. Time to first readmission (all-cause or unplanned) for discharged patients was examined within 30, 90, and 365 days, using competing risks regression to account for deaths postdischarge. Covariates included age, stroke severity (ability to walk on admission), stroke type, admissions before stroke/TIA and the Charlson Comorbidity Index (derived from International Statistical Classification of Diseases and Related Health Problems, Tenth Revision, [Australian modified] coded hospital data in the preceding 5 years). Results- Among the 13 594 patients discharged following stroke/TIA (45% female; 65% ischemic stroke; 11% intracerebral hemorrhage; 4% undetermined stroke; and 20% TIA), 25% had an all-cause readmission and 15% had an unplanned readmission within 90 days. In multivariable analyses, the factors independently associated with a greater risk of unplanned readmission within 90 days were being female (subhazard ratio, 1.13 [95% CI, 1.03-1.24]), greater Charlson Comorbidity Index scores (subhazard ratio, 1.11 [95% CI, 1.09-1.12]) and having an admission ≤90 days before the index event (subhazard ratio, 1.85 [95% CI, 1.59-2.15]). Compared with being discharged to rehabilitation or aged care, those who were discharged directly home were more likely to have an unplanned readmission within 90 days (subhazard ratio, 1.44 [95% CI, 1.33-1.55]). These factors were similar for readmissions within 30 and 365 days. Conclusions- Apart from comorbidities and patient-level characteristics, readmissions after stroke/TIA were associated with discharge destination. Greater support for transition to home after stroke/TIA may be needed to reduce unplanned readmissions.

Association between same day discharge total knee and total hip arthroplasty and risks of cardiac/pulmonary complications and readmission: a population-based observational study

ObjectiveTo determine if same-day discharge total knee arthroplasty (TKA) or total hip arthroplasty (THA) is not associated with increased risk of unplanned readmission and adverse outcomes within 30 days of...

Machine Learning Algorithms in Predicting Hospital Readmissions in Sickle Cell Disease

Background: Sickle cell disease (SCD) is the most common inherited hemoglobinopathy worldwide. The pathophysiology of the disease results in end organ damage which leads to morbidity and mortality. In a subset of patients, SCD-related complications have resulted in prolonged hospitalizations and increased frequency of 30-day hospital readmissions. In the era of value-based health care, hospital quality metrics and reimbursements are generated based on strategic health care utilization. Therefore, being able to identify early unplanned hospital readmissions is critical in managing health care expenditure. Objective: To develop machine learning algorithms for predicting the 30-day unplanned readmission risk of SCD patients and to compare the predictive power of machine learning models against standard hospital readmission scoring systems. Methods: We analyzed retrospective real-world electronic health records (EHR) data for patients with SCD at our institution from January 1, 2013- November 1, 2018. The raw data set contained 2824 unique SCD patients from across 5 hospitals within the University of Pittsburgh Medical Center. After preprocessing using our inclusion criteria, our cohort included 3299 admissions comprising of 446 adult SCD patients. Features extracted from the EHR data were reduced and regrouped using both data-driven methods and clinical knowledge, resulting in 486 unique features. Logistic Regression (LR), Support Vector Machine (SVM), and Random Forest (RF) were applied to predict for 30-day unplanned hospital readmissions in SCD. Prediction performance was evaluated using the area under the receiver operating characteristic curve (AUC), sensitivity and specificity. We compared our results against standard hospital readmission prediction tools such as LACE and HOSPITAL indices. Results: We randomly selected the inpatient admissions incurred by 30 percent of the 195 return patients and 251 nonreturn patients to be included in the testing set (n = 134); the training set contained the inpatient admissions incurred by the remaining 211 patients. Thus, our training and testing sets contained similar demographic information, predictors, and outcomes. The average number of admissions was 7.40 (12.90) for the 446 patients, and 14.47 (16.97) for the 195 patients who had 30-day readmissions. Since the number of samples in our study is comparatively small, our results might be sensitive to the training and testing splits. To address this problem, we performed 100 different training and testing splits and averaged the resulting 100 AUCs. Figure 1 summarizes the two performance metrics of each model. The two benchmark prediction tools, LACE and HOSPITAL, have AUCs of 0.56 (95%CI 0.52-0.60) and 0.63 (95%CI 0.59-0.67), respectively. Notably, all three machine learning algorithms outperformed both benchmarks. The RF was the best machine learning model in prediction of hospital readmissions, as reported in similar machine learning studies (Deschepper et al. 2019), with an AUC of 0.73 (95%CI 0.69-0.76). Table 1 summarizes the sensitivity and specificity of our RF model. Conclusion: Machine learning algorithms outperformed the standard hospital readmission risk scoring systems, LACE and HOSPITAL, by a large margin in a real world data set of SCD patients at a single institution. In particular, machine learning algorithms were able to identify important variables that are underrepresented in the traditional risk scoring systems (Figure 2). The use of machine learning algorithms can be a powerful tool in providing valuable insight towards health care expenditure and resource allocation in high risk patient groups. Disclosures No relevant conflicts of interest to declare.

Association between potentially inappropriate medications at discharge and unplanned readmissions among hospitalised elderly patients at a single centre in Japan: a prospective observational study

ObjectiveTo determine the prevalence of potentially inappropriate medication (PIM) use at admission and discharge among hospitalised elderly patients and evaluate the association between PIMs at discharge and unplanned readmission in...

American College of Surgeons National Surgical Quality Improvement Program assessment of risk factors for 30-day unplanned readmission in patients undergoing head and neck surgery requiring free tissue reconstruction.

Unplanned readmissions have become a metric for measuring quality of care. We analyzed the factors associated with 30-day unplanned readmission (30dUR) following head and neck cancer resections that included free tissue reconstruction (FTR). The 2012-2014 ACS-National Surgical Quality Improvement Program (NSQIP) data set was queried. Univariate and multivariate logistic regression analyses were performed. Out of 1114 cases, 121 had a 30dUR. The most common reasons were wound complications, including incisional infections, hematoma, and hemorrhage. A significant independent risk factor for 30dUR included a clean/contaminated wound class (odds ratio [OR], 2.27; 95% confidence interval [CI], 1.20-4.76). Patients receiving an osseous FTR had lower readmission rates (OR, 0.51; CI, 0.27-0.91). Discharge destination had no statistical significance. Based on the NSQIP data set, 10.9% of patients receiving an FTR for head and neck malignancy had a 30dUR. Although large, population-based data sets have limitations, these results elucidate that these patients are at an increased risk for unplanned readmissions, which can guide patient expectations and discharge planning.