30-day Unplanned Readmission Research Articles

Abstract 4134687: Takotsubo Cardiomyopathy Complicating Stroke

Background: There have been studies suggesting Takotsubo cardiomyopathy (TCM) as a potential complication following cerebrovascular accident (CVA), but the association, contributing factors, and prognosis significance have not been well-characterized. Aims: 1) To assess the incidence, factors and prognosis associated with TCM-complicated CVA, and 2) to determine the 30-day unplanned readmission rates for TCM after discharge for CVA and the associated factors. Methods: The U.S. Nationwide Readmission Database was queried using ICD-codes to gather information on discharges with primary diagnosis of CVA between January 2010 and November 2020. Incidence and factors associated with TCM-complicated CVA and 30-day readmissions with primary diagnosis of TCM were assessed. Results: Of 3,696,202 patients hospitalized with primary diagnosis of CVA, 5,526 (0.15%) individuals developed TCM during the admission. Patients with TCM were more likely to be women (80% vs. 52%, P<0.001), present with hemorrhagic (as opposed to ischemic stroke, P<0.001), and have a higher comorbidity burden, adverse events, and longer hospital stay. Mortality rate was significantly higher among those with TCM vs no-TCM (20% vs. 6.4%, P<0.001). Of 3,690,676 patients who were discharged without experiencing TCM during index admission for CVA, 255,947 (6.9%) patients had unplanned readmissions within 30 days of discharge, and 250 (0.1%) of these were with a primary diagnosis of TCM. These individuals with TCM were more likely to be women (78% vs. 51%, P<0.001), had a higher comorbidity burden, and experienced higher mortality rate during readmission compared to those without TCM (12% vs. 5.7%, P=0.009). Conclusions: Although rare, a TCM-complicated CVA can be highly fatal. Female sex, presentation with hemorrhagic stroke and pre-existing health conditions were strong factors associated with TCM-complicated CVA. Readmission for TCM after discharge for CVA is also uncommon but carries a significantly elevated risk of mortality. Our study presents an opportunity to improve processes and clinical outcomes by monitoring and recognizing TCM post-CVA, both during and after hospitalization.

Rates, Diagnoses, and Predictors of Unplanned 30-Day Readmissions of Critical Care Survivors Hospitalized for Lung Involvement in Systemic Lupus Erythematosus: An Analysis of National Representative US Readmissions Data

Rates, Diagnoses, and Predictors of Unplanned 30-Day Readmissions of Critical Care Survivors Hospitalized for Lung Involvement in Systemic Lupus Erythematosus: An Analysis of National Representative US Readmissions Data

The Impact of Socioeconomic Factors on 90-day Postoperative Readmissions and Cost in Shoulder Arthroplasty Patients

Understanding the impact of Social Drivers of Health (SDOH) on shoulder arthroplasty is pivotal for the development of equitable value-based payment models that enhance the quality of patient care. This investigation aims to understand the influence of Area Deprivation Index (ADI) on hospital admissions, readmissions, and associated costs post-shoulder arthroplasty. We conducted an analysis using US Medicare claims data from 2019 to 2021, identifying patients who received shoulder arthroplasty in either an inpatient or outpatient setting using Current Procedural Terminology (CPT) codes. Our primary outcomes were 90-day unplanned readmissions and post-procedure 90-day care costs. The ADI was our primary exposure variable, calculated at the census block level. Our analysis adjusted for multiple factors using a stepwise modeling approach including patient demographics, 29 Elixhauser comorbid conditions, Medicaid-dual eligibility, and end-stage renal disease status using logistic regression models. A total of 145,435 patients were included in our study. The average age of patients in our cohort was 74.5 years old, with 59.5% being female. Patients in the high ADI group had a higher readmission rate than the lowest ADI group (8.5% vs. 6.0%, p<0.001).When controlling for confounding factors, there was an indepdentent association between high ADI and readmission (OR: 1.28 [95% CI 1.12, 1.46]). Overall, we saw a dose-dependent relationship between ADI and readmission, with the association growing stronger as ADI increased. Conversely, we found a negative association with ADI and high-costs. (High ADI Group OR: 0.80 [95% CI 0.70, 0.91]). Patients in the high SES group had higher healthcare contact days during the 90-day follow-up period, with a median of 16 visits (IQR 8, 23), compared to 13 visits (IQR 6, 22) in the middle SES group and 10 visits (IQR 5, 20) in the low SES group (p<0.001). Socioeconomic disparities significantly influence the outcomes of primary shoulder arthroplasty, as indicated by higher readmission rates for low SES patients. Notably, our analysis shows a strong, independent association between ADI and readmission. Moreover, patients with higher SES incurred greater costs within a 90-day period potentially due to increased resource utilization.

Comparison of a Risk Calculator With Frailty Indices in Patients Undergoing Lung Cancer Resection.

While frailty has gained attention for its utility in risk stratification, no studies have directly compared them to existing risk calculators. The objective of this study was to compare the risk stratification of the American College of Surgeons Surgical Risk Calculator (ACS-SRC), the Revised Risk Analysis Index (RAI-rev), and the Modified Frailty Index (5-mFI). The primary outcomes were 30-day postoperative morbidity, 30-day postoperative mortality, unplanned readmission, unplanned reoperation, and discharge disposition other than home. Patients undergoing anatomic lung resection for primary, nonsmall cell lung cancer were identified within the ACS National Quality Improvement Program (ACS NSQIP) database. Tools were compared for discrimination in the primary outcomes. 9663 patients undergoing anatomic lung resection for cancer between 2012 and 2014 were included. The cohort was 53.1% female. Median age at diagnosis was 67 (IQR 59-74) years. Perioperative morbidity and mortality rates were 10.9% (n = 1048) and 1.6% (n = 158). Rates of 30-day postoperative unplanned readmission and reoperation were 7.5% (n = 725) and 4.8% (n = 468). The ACS-SRC had the highest discrimination for all measured outcomes, as measured by the area under the receiver operating curve (AUC) and corresponding confidence interval (95% CI). This included perioperative mortality (AUC 0.74, 95% CI 0.71-0.78), compared to RAI-rev (AUC 0.66, 95% CI 0.62-0.69) and 5-mFI (AUC 0.61, 95% CI 0.57-0.65; p < 0.001). The RAI-rev and 5-mFI had similar discrimination for all measured outcomes. ACS-SRC was the perioperative risk stratification tool with the highest predictive discrimination for adverse, 30-day, postoperative events for patients with cancer treated with anatomic lung resection.

Incidence and predictors of unplanned 30-day hospital readmissions among heart failure patients in Ethiopia: a 5-year retrospective cohort study

The burden of heart failure increases over time and is a leading cause of unplanned readmissions worldwide. In addition, its impact has doubled in countries with limited health resources, including Ethiopia. Identifying and preventing the possible contributing factors is crucial to reducing unplanned hospital readmissions and improving clinical outcomes. The study aimed to assess the incidence and predictors of 30-day unplanned readmission among heart failure patients at selected South Wollo general hospitals in 2022. A hospital-based retrospective cohort study design was employed from January 1, 2016, to December 30, 2020. The data was collected from 572 randomly selected medical records using data extraction checklists. Data were entered in Epi-Data version 4.6 and analyzed with Stata version 17. The Kaplan–Meier and log-rank tests were used to estimate and compare the survival failure time. A Cox proportional hazard analysis was used to identify the predictors of readmission. The statistical significance level was declared at a p-value < 0.05 with an adjusted odds ratio and a 95% confidence interval. A total of 151 (26.40%) heart failure patients were readmitted within 30 days of discharge. Among the study participants, 302 (52.8%) were male, and 370 (64.7%) were rural residents. The mean age was 45.8 ± 14.1 SD years. In the multivariate Cox proportional hazards analysis being an age (> 65 years) (AHR: 3.172, 95% CI:.21, 4.55, P = 0.001), rural in residency (AHR: 2.47, 95%CI: 1.44, 4.24, P = 0.001), Asthma or Chronic Obstructive Pulmonary Disease (AHR: 1.62, 95% CI: 1.11, 2.35, P = 0.012), HIV/AIDS (AHR: 1.84, 95%CI: 1.24, 2.75, P = 0.003), Haemoglobin level 8–10.9 g/dL (AHR: 6.20, 95% CI: 3.74, 10.28, P = 0.001), and Mean platelet volume > 9.1 fl (AHR: 2.08, 95% CI: 1.27, 3.40, P = 0.004) were identified as independent predictors of unplanned hospital readmission. The incidence of unplanned hospital readmission was relatively high among heart failure patients. Elderly patients, rural residency, comorbidity, a higher mean platelet volume, and a low hemoglobin level were independent predictors of readmission. Working on these factors will help reduce the hazards of unplanned hospital readmission.

The 30-day readmission rate of patients with an overlap of probable sarcopenia and malnutrition undergoing major oncological surgery.

Overlapping sarcopenia and malnutrition may increase the risk of readmission in surgical oncology. Overlapping probable sarcopenia/malnutrition was found in 4.6% of 238 patients and the 30-day unplanned readmission rate was 9.0%. In multivariate analysis, the overlap of probable sarcopenia and malnutrition was a significant predictor for the 30-day unplanned readmission (OR= 8.10, 95%CI= 1.20-0.55; p=0.032). ■ Probable sarcopenia plus malnutrition was significantly associated with unplanned readmission. ■ Overlap of probable sarcopenia and malnutrition was an independent risk factor for readmission. ■ Certification of whether the patient is malnourished and/or sarcopenic preoperatively is necessary. ■ SARC-F and subjective global assessment can effectively and easily assess sarcopenia and malnutrition at admission. To assess the 30-day unplanned readmission rate and its association with overlapping probable sarcopenia and malnutrition after major oncological surgery. A prospective bicentric observational cohort study performed with adult oncological patients undergoing major surgery. The primary outcome was unplanned readmission within 30 days after discharge and the association with probable sarcopenia and malnutrition. Nutritional status and probable sarcopenia were assessed just prior to surgery. Patients classified using subjective global assessment, as B and C were malnourished. Probable sarcopenia was defined using SARC-F (strength, assistance with walking, rise from a chair, climb stairs, falls) questionnaire ≥4 points and low HGS (handgrip strength) <27kg for males and <16kg for females. Two hundred and thirty-eight patients (51.7% female) with a median age of 60 years were included. The 30-day readmission rate was 9.0% (n=20). Univariate analysis showed an association of malnutrition (odds ratio (OR) = 4.84; p=0.024) and probable sarcopenia (OR = 4.94; p=0.049) with 30-day readmission. Furthermore, when both conditions were present, the patient was almost nine times more likely to be readmitted (OR = 8.9; p=0.017). Multivariable logistic regression analysis showed that overlapping probable sarcopenia and malnutrition was an independent predictor of 30-day unplanned readmission (OR = 8.10, 95% confidence interval (95%CI) 1.20-0.55; p=0.032). The 30-day unplanned readmission rate was 9.0%, and the overlap of probable sarcopenia and malnutrition is an independent predictor for the 30-day unplanned readmission after major oncologic surgery.

Comparing Deprivation vs Vulnerability Index Performance Using Medicare Beneficiary Surgical Outcomes

Health care researchers, professionals, payers, and policymakers are increasingly relying on publicly available composite indices of area-level socioeconomic deprivation to address health equity. Implications of index selection, however, are not well understood. To compare the performance of 2 frequently used deprivation indices using policy-relevant outcomes among Medicare beneficiaries undergoing 3 common surgical procedures. This cross-sectional study examined outcomes among Medicare beneficiaries (65 to 99 years old) undergoing 1 of 3 common surgical procedures (hip replacement, knee replacement, or coronary artery bypass grafting) between 2016 and 2019. Index discriminative performance was compared for beneficiaries residing in tracts with high- and low-deprivation levels (deciles) according to each index. Analyses were conducted between December 2022 and August 2023. Tract-level deprivation was operationalized using 2020 releases of the area deprivation index (ADI) and the social vulnerability index (SVI). Binary outcomes were unplanned surgery, 30-day readmissions, and 30-day mortality. Multivariable logistic regression models, stratified by each index, accounted for beneficiary and hospital characteristics. A total of 2 433 603 Medicare beneficiaries (mean [SD] age, 73.8 [6.1] years; 1 412 968 female beneficiaries [58.1%]; 24 165 Asian [1.0%], 158 582 Black [6.5%], and 2 182 052 White [89.7%]) were included in analyses. According to both indices, beneficiaries residing in high-deprivation tracts had significantly greater adjusted odds of all outcomes for all procedures when compared with beneficiaries living in low-deprivation tracts. However, compared to ADI, SVI resulted in higher adjusted odds ratios (adjusted odds ratios, 1.17-1.31 for SVI vs 1.09-1.23 for ADI), significantly larger outcome rate differences (outcome rate difference, 0.07%-5.17% for SVI vs outcome rate difference, 0.05%-2.44% for ADI; 95% CIs excluded 0), and greater effect sizes (Cohen d, 0.076-0.546 for SVI vs 0.044-0.304 for ADI) for beneficiaries residing in high- vs low-deprivation tracts. In this cross-sectional study of Medicare beneficiaries, SVI had significantly better discriminative performance-stratifying surgical outcomes over a wider range-than ADI for identifying and distinguishing between high- and low-deprivation tracts, as indexed by outcomes for common surgical procedures. Index selection requires careful consideration of index differences, index performance, and contextual factors surrounding use, especially when informing resource allocation and health care payment adjustment models to address health equity.

Improving transitional care after acute myocardial infarction: A scoping review.

Cardiovascular disease remains the leading cause of morbidity and mortality, posing a significant challenge to healthcare systems worldwide. Transitional care interventions, which ensure coordination and continuity of care as patients move between different levels of healthcare, have been shown to reduce unnecessary healthcare utilization and improve patient outcomes. While much attention has been given to transitional care in heart failure, this review aims to map the interventions implemented for patients following an acute myocardial infarction (AMI). A scoping review was conducted following the Joanna Briggs Institute (JBI) methodology, with literature searches performed in the Cochrane, CINAHL, MEDLINE, JBI, and SciELO databases, focusing on publications from 2013 onwards in both Portuguese and English. Seventy-five studies were included, with most combining multiple interventions that contributed to improved cardiovascular health outcomes, including increased adherence to healthy lifestyle behaviors, enhanced medication compliance, and better healthcare self-management. These interventions were effective in reducing cardiovascular-related Emergency Department visits, unplanned 30-day readmissions, and mortality following a first-time myocardial infarction. Key strategies identified included discharge planning, digital health solutions, outpatient care, and healthcare coordination. The findings of this review underscore the need for developing methodologies that enhance the transition of care from hospital to primary care following an AMI. There is an urgent need to design and implement new healthcare programs that integrate discharge interventions, digital health, outpatient care, and healthcare coordination to ensure continuity of care and optimize patient outcomes post-discharge.

S1534 National Landscape of Unplanned 30-Day Readmissions in Patients Undergoing ERCP: Insights from National Readmission Database

S1534 National Landscape of Unplanned 30-Day Readmissions in Patients Undergoing ERCP: Insights from National Readmission Database

Risk factors for early readmission to hospital in patients with malignancy-related ascites: a retrospective cohort study.

Malignancy-related ascites (MRA) is a common serious complication of many advanced malignant tumors with high morbidity and mortality. The high hospital expenditures induced by unplanned readmission in patients with MRA have become an urgent issue to the public. We aimed to overall assess the unplanned early readmission rate of patients with MRA and explore the potential risk factors for such readmission. A retrospective cohort study based on 2018 Nationwide Readmissions Database was performed and patients with MRA were recruited into the analysis. The primary outcome was unplanned 30-day readmission rate and inpatient outcomes. The multivariate logistic regression analysis was performed to evaluate the potential risk factors for such early readmission. Data obtained from 32,457 patients with MRA were analyzed, and of these 7,799 individuals (24.03%) were unplanned readmitted within 30-day follow-up. The mortality rate in the readmitted population was 15.15%. Patients at younger age were at a higher risk of readmission. The morbidities including hypertension (OR=1.117, 95%CI: 1.054-1.184), hyperlipemia (OR=1.075, 95%CI: 1.009-1.146) and diabetes (OR=1.118, 95%CI: 1.053-1.188), gastrointestinal malignancies and peritoneal procedure significantly increased the risk of 30-day readmission in patients with MRA. More than one in five patients with MRA was unplanned readmitted within 30-day follow-up. The above risk factors should be timely intervened and the corresponding medical care should be strengthened in patients with MRA to lessen the unplanned readmission and improve the readmission outcomes.

S1533 National Landscape of Unplanned 30-Day Readmissions in Liver Transplant Patients Undergoing ERCP: Insights from National Readmission Database

S1533 National Landscape of Unplanned 30-Day Readmissions in Liver Transplant Patients Undergoing ERCP: Insights from National Readmission Database

Quality of care after a horizontal merger between two large academic hospitals

BackgroundHospital mergers remain common, but their influence on healthcare quality varies. Data on effects of European hospital mergers are ill defined, and academic hospitals in particular. This case study assesses early quality of care changes in two formerly competing Dutch academic hospitals that merged on June 6, 2018. MethodsStatistical process control and interrupted time series analysis were performed. All adult, non-psychiatric patients, admitted between 01-03–2016 and 01-10-2022 were eligible for analysis. Primary outcome measure was all cause in-hospital mortality (or hospice), secondary outcomes were unplanned 30-day readmissions to same hospital, length of stay, and patients’ hospital rating. Data were obtained from electronic health records, and patient experience surveys. FindingsThe mean (SD) age of the 573 813 included patients was 54·3 (18·9) years. The minority was female (277 817, 48·4 %), and most admissions were acute (308 597, 53·8 %). No merger related change in mortality was found in the first 20 months post-merger (limited to the pre-Covid-19 era). For this same period, the 30-day readmission incidence changed to a downward slope post-merger, and the length of stay shortened (immediate level-change −3·796 % (95 % CI, −5·776 % to −1·816 %) and trend-change −0·150 % per month (95 % CI, −0·307 % to 0·007 %)). Patients’ hospital ratings seemed to improve post-merger. InterpretationIn this quality improvement study, a full- and gradual post-merger integration strategy for a Dutch academic hospital merger was not associated with changes in in-hospital mortality, and yielded slight improved results for secondary quality of care outcomes.

A pilot study testing a new transition of care model from hospital to the community for Hispanic/Latino adults with diabetes to reduce emergency department visits and hospital re-admissions

BackgroundHispanic/Latino populations have the second highest prevalence of diabetes (12.5%) among ethnic minority groups in the USA. They also have higher rates of uncontrolled diabetes and diabetes-related complications. Approximately 29% of diabetes care costs are attributed to inpatient hospital care. To reduce hospital length of stay and re-admission rates for diabetes, the American Diabetes Association (ADA) recommends a “structured discharge plan tailored to the individual patient with diabetes.” However, limited research exists on the feasibility and applicability of a transition of care model specifically tailored for the Hispanic/Latino population.MethodsWe conducted a 2-year pilot study to develop a practical, patient-centered, and culturally competent transition of care (TOC) model for Hispanic/Latino adults with diabetes discharged from the hospital to the community. Feasibility outcomes included recruitment rates, questionnaire completion rates, adherence to a 30-day post-discharge phone call, and resource needs and utilization for study implementation. Participant-centered outcomes included 30-day post-discharge emergency department (ED) visits, 30-day post-discharge unplanned re-admissions, follow-up visits within 2 weeks of discharge, and patient satisfaction with the TOC model.ResultsTwelve participants were enrolled over the study period, with weekly enrollment ranging from 0 to 4 participants. Participants’ average age in years was 47 (± 11.6); the majority were male (85%), and 75% had type 2 diabetes. Recruitment involved the support of 4 bilingual staff. The estimated time to review the chart, approach participants, obtain informed consent, complete questionnaires, and provide discharge instructions was approximately 2.5 h. Of the 10 participants who completed the 30-day post-discharge phone call, none had ED visits or unplanned hospital re-admissions within 30 days post-discharge, and all had a follow-up with a medical provider within 2 weeks.ConclusionsImplementing a patient-centered and culturally competent TOC model for Hispanic/Latino adults with diabetes discharged from the hospital to the community is feasible when considering key resources for success. These include a bilingual team with dedicated and funded time, alignment with existing discharge process and integration into the Electronic Medical Record (EMR) systems.

The safety of same-day discharge following percutaneous coronary intervention in regional Australia.

Same-day discharge (SDD) following percutaneous coronary intervention (PCI) has proven safe, and global adoption of this strategy has been increasing rapidly. These data are predominantly derived from high-volume, metropolitan centres with a relative paucity of data from regional and remote settings. The primary objective of this study was to evaluate the outcomes of a same-day, criteria-led discharge strategy following elective transradial PCI in a regional setting. This is a retrospective, single-centre cohort study. Consecutive outpatients aged ≥18 years presenting for elective transradial invasive coronary angiography between March 2019 and February 2024 were included in the analysis. We report the primary composite outcome of 30-day all-cause mortality and unplanned hospital readmission and compare proportions between those who were discharged on the day of their procedure with those admitted overnight in hospital and discharged the next day. A total of 555 eligible patients were identified, of which 330 (60%) were discharged on the day of their procedure. The composite primary end-point occurred in seven (2%) of the SDD patients and in six (3%) of the overnight admission patients (relative risk = 0.80; 95% confidence interval = 0.27-2.34; P = 0.68). No significant differences were seen in rates of readmission, or in other clinical outcomes assessed, including death, myocardial infarction, stroke and vascular complications or bleeding. In a regional setting, SDD following elective transradial PCI, in select patients, is a safe approach which was not associated with higher rates of unplanned readmission or adverse clinical outcomes.

Impact of wearable device data and multi-scale entropy analysis on improving hospital readmission prediction.

Unplanned readmissions following a hospitalization remain common despite significant efforts to curtail these. Wearable devices may offer help identify patients at high risk for an unplanned readmission. We conducted a multi-center retrospective cohort study using data from the All of Us data repository. We included subjects with wearable data and developed a baseline Feedforward Neural Network (FNN) model and a Long Short-Term Memory (LSTM) time-series deep learning model to predict daily, unplanned rehospitalizations up to 90 days from discharge. In addition to demographic and laboratory data from subjects, post-discharge data input features include wearable data and multiscale entropy features based on intraday wearable time series. The most significant features in the LSTM model were determined by permutation feature importance testing. In sum, 612 patients met inclusion criteria. The complete LSTM model had a higher area under the receiver operating characteristic curve than the FNN model (0.83 vs 0.795). The 5 most important input features included variables from multiscale entropy (steps) and number of active steps per day. Data available from wearable devices can improve ability to predict readmissions. Prior work has focused on predictors available up to discharge or on additional data abstracted from wearable devices. Our results from 35 institutions highlight how multiscale entropy can improve readmission prediction and may impact future work in this domain. Wearable data and multiscale entropy can improve prediction of a deep-learning model to predict unplanned 90-day readmissions. Prospective studies are needed to validate these findings.

WP7.6 - Unplanned Re-admisisons - How are we doing; Phase 1 study

Abstract Introduction 30-day unplanned re-admissions provide an indicator of performance in the NHS Framework. Average readmission rate is around 15 to 17%, however surgical re-admission is less published. Therefore, a study was set up to assess the factors influencing surgical re-admissions in a Tertiary Centre. Methods Retrospective study of prospectively collected data from the tertiary centre database including all ‘Unplanned Re-admissions’(URA) aged &amp;gt;15 years in general surgery between January and December 2022. URA &amp;lt;15 years old and to other surgical specialties were excluded. ANOVA test and descriptive statistics were used to analyse the data. Results 289 (58%) of 498 readmissions were URA with 150 (52%) female and 139 (48%) male. 57 (19%) were aged 15-30 years, 121 (42%) 31-60 years, 109 (38%) 61-90 years and 2 &amp;gt; 90 years (Median age 54 years; p&amp;lt;0.05). 191 (66%) were from ED, 84 (29%) were GP referrals, 10 (3%) from clinics and 4 (2%) transferred from network. 178 (62%) URA were within 10 days of discharge, 72 (25%) between 11-20 days and 39 (13%) &amp;gt; 20 days (p &amp;lt; 0.05). The commonest cause of URA was biliary diseases (n = 40, 14%) followed by lower (n=32, 11%) and upper abdominal pain (n=17, 5%) and nonspecific abdominal pain (n=22, 7%). Other causes included appendicitis (n=8, 3%), pancreatitis (n=8, 3%), post-op complications (n=14, 5%) and obstruction (n=24, 8%). Conclusions URA significantly impacts health care provision and patients’ confidence. An initial study has identified reasons and patterns of re-admissions, the next phase involves analysing specific reasons to mitigate factors leading to URA.

1492P The role of LACE index in predicting 30-day unplanned readmission in patients receiving immunotherapy

1492P The role of LACE index in predicting 30-day unplanned readmission in patients receiving immunotherapy

Comparison of a risk calculator with frailty indices in patients undergoing lung cancer resection.

Perioperative risk stratification is an essential component of preoperative planning for cancer surgery. While frailty has gained attention for its utility in risk stratification, no studies have directly compared it to existing risk calculators. Therefore, the objective of this study was to compare the risk stratification of the American College of Surgeons Surgical Risk Calculator (ACS-SRC), the Revised Risk Analysis Index (RAI-rev), and the Modified Frailty Index (5-mFI). The primary outcomes were 30-day postoperative morbidity, 30-day postoperative mortality, unplanned readmission, unplanned reoperation, and discharge disposition other-than-home. Patients undergoing anatomic lung resection for primary, non-small cell lung cancer were identified within the American College of Surgeons National Quality Improvement Program (ACS NSQIP) database. The ACS-SRC, RAI-rev, and 5-mFI tools were used to predict adverse postoperative events. Tools were compared for discrimination in the primary outcomes. 9663 patients undergoing anatomic lung resection for cancer between 2012 and 2014 were included. The cohort was 53.1% female. Median age at diagnosis was 67 (interquartile range = 59-74) years. Cardiothoracic surgeons performed 89% and general surgeons performed 11.0% of the operations. Perioperative morbidity and mortality rates were 10.9% (n = 1048) and 1.6% (n = 158). Rates of 30-day postoperative unplanned readmission and reoperation were 7.5% (n = 725) and 4.8% (n = 468). The ACS-SRC had the highest discrimination for all measured outcomes, as measured by the area under the receiver operating curve (AUC) and corresponding confidence interval (95% confidence interval [CI]). This included perioperative mortality (AUC = 0.74, 95% CI = 0.71-0.78), compared to RAI-rev (AUC = 0.66, 95% CI = 0.62-0.69) and 5-mFI (AUC = 0.61, 95% CI = 0.57-0.65; p < 0.001). The RAI-rev and 5-mFI had similar discrimination for all measured outcomes. ACS-SRC was the perioperative risk stratification tool with the highest predictive discrimination for adverse, 30-day, postoperative events for patients with cancer treated with anatomic lung resection.

Predictors and outcomes of delays between diagnosis and definitive surgery for rectal cancer

BackgroundThe National Accreditation Program for Rectal Cancer (NAPRC) recommends definitive treatment of rectal cancer commence within 60 days from diagnosis. This study aimed to assess predictors of >60 days delay between diagnosis and definitive surgery of rectal cancer and the impact on survival and short-term outcomes. MethodsRetrospective cohort analysis of patients with stage I-III rectal adenocarcinoma who underwent proctectomy without preoperative neoadjuvant treatment from the National Cancer Database (2015–2019). Based on the time interval between diagnosis and definitive surgery, patients were divided into timely non-adherent (>60 days) and timely-adherent (≤60 days) groups. Multivariate analysis determined predictors of delayed definitive surgery. Results9479 patients (57.5 % males; mean age: 63.7 years) had a 41-day median time between diagnosis and definitive surgery. Non-adherence was noted in 27.9 % of patients. Independent predictors of non-adherence were male sex (Odds ratio [OR]: 1.25; p < 0.001), Black (OR: 1.65; p < 0.001) or Asian (OR: 1.33; p = 0.014) race, Charlson score 2 (OR: 1.33; p = 0.005) or 3 (OR: 1.55; p < 0.001), urban residence (OR: 1.21; p = 0.003), abdominoperineal resection (OR: 1.69; p < 0.001), pelvic exenteration (OR: 1.7; p = 0.002), and robotic-assisted surgery (OR: 1.22; p = 0.001). Medicare (OR: 0.725; p = 0.003) and private insurance (OR: 0.711; p < 0.001) were associated with better adherence. 30-day and 90-day mortality, unplanned readmission, and overall survival were similar. ConclusionsMale Black or Asian patients with high Charlson scores, and undergoing abdominoperineal resection, pelvic exenteration, and robotic-assisted surgery were more likely non-adherent with NAPRC standards with >60 days delay before definitive surgery for rectal cancer. Hopefully, recognition for these reasons for delay of definitive surgery will lead to an improvement in adherence to the standards.

Comparison of machine-learning and logistic regression models for prediction of 30-day unplanned readmission in electronic health records: A development and validation study.

It is expected but unknown whether machine-learning models can outperform regression models, such as a logistic regression (LR) model, especially when the number and types of predictor variables increase in electronic health records (EHRs). We aimed to compare the predictive performance of gradient-boosted decision tree (GBDT), random forest (RF), deep neural network (DNN), and LR with the least absolute shrinkage and selection operator (LR-LASSO) for unplanned readmission. We used EHRs of patients discharged alive from 38 hospitals in 2015-2017 for derivation and in 2018 for validation, including basic characteristics, diagnosis, surgery, procedure, and drug codes, and blood-test results. The outcome was 30-day unplanned readmission. We created six patterns of data tables having different numbers of binary variables (that ≥5% or ≥1% of patients or ≥10 patients had) with and without blood-test results. For each pattern of data tables, we used the derivation data to establish the machine-learning and LR models, and used the validation data to evaluate the performance of each model. The incidence of outcome was 6.8% (23,108/339,513 discharges) and 6.4% (7,507/118,074 discharges) in the derivation and validation datasets, respectively. For the first data table with the smallest number of variables (102 variables that ≥5% of patients had, without blood-test results), the c-statistic was highest for GBDT (0.740), followed by RF (0.734), LR-LASSO (0.720), and DNN (0.664). For the last data table with the largest number of variables (1543 variables that ≥10 patients had, including blood-test results), the c-statistic was highest for GBDT (0.764), followed by LR-LASSO (0.755), RF (0.751), and DNN (0.720), suggesting that the difference between GBDT and LR-LASSO was small and their 95% confidence intervals overlapped. In conclusion, GBDT generally outperformed LR-LASSO to predict unplanned readmission, but the difference of c-statistic became smaller as the number of variables was increased and blood-test results were used.