Actual Length Of Stay Research Articles

Adopting combination machine learning models could reduce hospital length of stay for oncology patients.

11084 Background: Integra Connect previously created in-patient (IP) admission prediction model based on OCM data. One of the challenges for practices in a value-based care (VBC) program is to provide continuous care-coordination during and after an IP admission. Our objective is to show how adopting combination machine learning (ML) models can predict IP length of stay (LOS). Potential benefits include an overall reduction in LOS which could minimize the risk of hospital acquired conditions. Methods: ML models were trained on 5 major cancer types (Lung Cancer, Multiple Myeloma, Lymphoma, Small Intestine / Colorectal Cancer, High-risk Breast Cancer') from 7 OCM practices of PP4-PP9 data, excluding surgery IP admission. The top 5 cancer types accounted for ~50% of total in-patient admissions and ~50% of total LOS in days. The IP admissions were divided into 4 major cohorts in terms of LOS in days (1-3: class 1, 4-8: class 2, 9-15: class 3, and more than 15: class 4). To reduce the overall LOS, we adopted two ML models: 1) To classify LOS, a multi-classification model (Model1, an eXtreme Gradient Boosted Trees Classifier) and 2) To predict the LOS for classes 1-3, a regression-based model (Model2, an eXtreme Gradient Boosted Trees Regressor with Early Stopping). Class 4 cohort had a wide range (16- 90+ days) for LOS and therefore was excluded from Model2. The other three cohorts belonged to Q1, IQR, and Q4, respectively. Model1 was trained on 4,280 randomly selected sample IP admissions to balance each class and Model2 was trained on 19,636 IP admissions, both with 102 features. Results: The models were tested on PP10-PP11 claims excluding 0 days LOS. Model1 predicted 296 (3,945) IP admissions in class 4 of which 88 were true positives. Assuming at least a 10% reduction in associated total LOS translated to lower LOS of 1-3% for 5 (7) practices. The remaining 3,649 IP admissions were used to predict LOS by Model2. The difference between actual LOS and Predicted LOS was found as 18%-24%. To address model prediction errors, we defined the target LOS to be predicted LOS with an upper bound of at least 10%. This led to reductions of 6% -16% across practices. Finally, combining the results of both ML models we determined that the potential to lower LOS was at least 6% and at most 19%. Conclusions: Our ML models identified opportunities to reduce LOS across multiple OCM cancer practices for 5 major cancer types. We also identified a cohort of patients with critical condition (class 4) that is vital for practice transformation initiatives. These identified reduction in LOS for oncology patients provides cost reduction and quality improvement opportunities in VBC programs. In the future, more opportunities to lower LOS will be explored with a re-admission prediction model.

Length of Stay Prediction Models for Oral Cancer Surgery: Machine Learning, Statistical and ACS-NSQIP.

Accurate prediction of hospital length of stay (LOS) following surgical management of oral cavity cancer (OCC) may be associated with improved patient counseling, hospital resource utilization and cost. The objective of this study was to compare the performance of statistical models, a machine learning (ML) model, and The American College of Surgeons National Surgical Quality Improvement Program's (ACS-NSQIP) calculator in predicting LOS following surgery for OCC. A retrospective multicenter database study was performed at two major academic head and neck cancer centers. Patients with OCC who underwent major free flap reconstructive surgery between January 2008 and June 2019 surgery were selected. Data were pooled and split into training and validation datasets. Statistical and ML models were developed, and performance was evaluated by comparing predicted and actual LOS using correlation coefficient values and percent accuracy. Totally 837 patients were selected with mean patient age being 62.5 ± 11.7 [SD] years and 67% being male. The ML model demonstrated the best accuracy (validation correlation 0.48, 4-day accuracy 70%), compared with the statistical models: multivariate analysis (0.45, 67%) and least absolute shrinkage and selection operator (0.42, 70%). All were superior to the ACS-NSQIP calculator's performance (0.23, 59%). We developed statistical and ML models that predicted LOS following major free flap reconstructive surgery for OCC. Our models demonstrated superior predictive performance to the ACS-NSQIP calculator. The ML model identified several novel predictors of LOS. These models must be validated in other institutions before being used in clinical practice. 3 Laryngoscope, 134:3664-3672, 2024.

A Digital Respiratory Ward in Leicester, Leicestershire, and Rutland, England, for Patients With COVID-19: Economic Evaluation of the Impact on Acute Capacity and Wider National Health Service Resource Use.

The COVID-19 pandemic stressed global health care systems' acute capacity and caused a diversion of resources from elective care to the treatment of acute respiratory disease. In preparing for a second wave of COVID-19 infections, England's National Health Service (NHS) in Leicester, Leicestershire, and Rutland sought to protect acute capacity in the winter of 2020-2021. Their plans included the introduction of a digital ward where patients were discharged home early and supported remotely by community-based respiratory specialists, who were informed about patient health status by a digital patient monitoring system. The objective of the digital ward was to maintain acute capacity through safe, early discharge of patients with COVID-19 respiratory disease. The study objective was to establish what impact this digital ward had on overall NHS resource use. There were no expected differences in patient outcomes. A cost minimization was performed to demonstrate the impact on the NHS resource use from discharging patients into a digital COVID-19 respiratory ward, compared to acute care length of stay (LOS). This evaluation included all 310 patients enrolled in the service from November 2020 (service commencement) to November 2021. Two primary methods, along with sensitivity analyses, were used to help overcome the uncertainty associated with the estimated comparators for the observational data on COVID-19 respiratory acute LOS, compared with the actual LOS of the 279 (90%) patients who were not discharged on oxygen nor were in critical care. Historic comparative LOS and an ordinary least squares model based on local monthly COVID-19 respiratory median LOS were used as comparators. Actual comparator data were sourced for the 31 (10%) patients who were discharged home and into the digital ward for oxygen weaning. Resource use associated with delivering care in the digital ward was sourced from the digital system and respiratory specialists. In the base case, the digital ward delivered estimated health care system savings of 846.5 bed-days and US $504,197 in net financial savings across the 2 key groups of patients-those on oxygen and those not on oxygen at acute discharge (both P<.001). The mean gross and net savings per patient were US $1850 and US $1626 in the base case, respectively, without including any savings associated with a potential reduction in readmissions. The 30-day readmission rate was 2.9%, which was below comparative data. The mean cost of the intervention was US $223.53 per patient, 12.1% of the estimated gross savings. It was not until the costs were increased and the effect reduced simultaneously by 78.4% in the sensitivity analysis that the intervention was no longer cost saving. The digital ward delivered increased capacity and substantial financial savings and did so with a high degree of confidence, at a very low absolute and relative cost.

Prediction of Intensive Care Unit Length of Stay in the MIMIC-IV Dataset

Accurately estimating the length of stay (LOS) of patients admitted to the intensive care unit (ICU) in relation to their health status helps healthcare management allocate appropriate resources and better plan for the future. This paper presents predictive models for the LOS of ICU patients from the MIMIC-IV database based on typical demographic and administrative data, as well as early vital signs and laboratory measurements collected on the first day of ICU stay. The goal of this study was to demonstrate a practical, stepwise approach to predicting patient’s LOS in the ICU using machine learning and early available typical clinical data. The results show that this approach significantly improves the performance of models for predicting actual LOS in a pragmatic framework that includes only data with short stays predetermined by a prior classification.

Validation for Abbreviated Hospital Stay in Pediatric Patients with Solid Organ Injury.

In 2000, the American Pediatric Surgical Association (APSA) published guidelines for the management of pediatric solid organ injury, recommending a hospital length of stay (LOS) of grade of injury plus 1day. Since the publication of these guidelines, several studies have suggested that it is safe to discharge patients sooner based upon hemodynamic and clinical factors. The results of several of these studies have been confounded by the existence of other injuries. The aim of this study was to examine LOS and outcomes in children with strictly isolated solid organ injuries. This is a 12-year retrospective review of pediatric patients with isolated trauma to the kidney, liver, or spleen to determine LOS. Patients were excluded for associated intracranial, neurologic, orthopedic, or pulmonary injuries which would impact length of stay. Documented hemodynamic parameters were reviewed as determinants of patient stability. A total of 156 patients were included in the study. The projected average LOS for all patients based on the 2000 APSA guidelines would have been 3.71 ± 0.98days. The actual average LOS for all patients 2.85 ± 3.32days. Need for operation, ICU stay, and transfusion all contributed to increased LOS. The number of episodes of abnormal vitals positively correlated with increased LOS. This study validates that management of isolated solid organ injuries based upon hemodynamic parameters and clinical status is safe and decreases hospital length of stay. Consistently normal vital signs indicate these children can be safely discharged sooner.

Prospective Cohort Study on the Impact of Early Versus Late Inpatient Palliative Care on Length of Stay and Cost of Care.

Aim: To evaluate the impact of early vs late palliative care on (1) length of stay (LOS) in the context of expected LOS measures and (2) total cost of care to the hospital for each patient. Methods: A prospective cohort study was performed at a single large academic medical center on patients who received an inpatient palliative care consultation. The two cohorts were early palliative care (within 3days of admission) and late palliative care (after 3days of admission). Comparisons were made between patients' actual LOS, expected LOS, and total hospital costs between both cohorts. Results: Compared to the late palliative care cohort (N = 126), patients who received early palliative care (N = 68) had a significantly shorter LOS (P < .001) and also performed better compared to CMS-Expected LOS standards (Observed/Expected 3.1 vs 1.5 respectively; P < .001). Early palliative care patients also saw an average decline of $1431 in total costs 1-day pre/post consult as opposed to a more modest $403 decline in the later palliative care cohort (P < .001). Similarly, patients who received early palliative care had a $5839 decline in aggregated total 3-day costs, as opposed to a $1478 decline in those who received late palliative care (P < .001). Conclusions: In the competitive and rapidly evolving healthcare system, the opportunity to suppress costs and lower patient LOS has increasing importance. Our study strongly supports the implementation of earlier palliative care intervention to assist hospitals in approaching LOS targets and reducing patient costs.

Changes in inpatient healthcare utilization at public hospitals during the Covid-19 pandemic

Objective: Public hospitals play an important role in reducing health inequality among socially marginalized groups and local communities. The objective of this study was to investigate changes in healthcare utilization at public hospitals during the COVID-19 pandemic in Korea.Methods: The unit of analysis was public hospitals. This study adopted a retrospective longitudinal study design and used inpatient health insurance claim data from January 1, 2019 to December 31, 2020. In total, 206 public hospitals were analyzed. For statistical analysis, a generalized linear model using the number of health insurance claims was constructed.Results: The total number of patients who received COVID-19-related care was 44,949, with a length of stay (LOS) of 657,717 days. After controlling for hospital covariates, healthcare utilization at public hospitals decreased by 22.5% (relative risk [RR], 0.775; p=0.029) and 16.8% (RR, 0.832; p=0.045) for the actual number and total LOS of national health insurance patients, respectively. For beneficiaries of the Medical Aid program, the actual number and total LOS were reduced by 29.6% (RR, 0.704; p=0.001) and 35.8% (RR, 0.642; p=0.001), respectively. Interestingly, most of this reduction occurred at public hospitals providing general healthcare services.Conclusion: Despite significant changes in healthcare utilization, public hospitals played a critical role in providing healthcare regarding COVID-19. Healthcare policymakers need to pay more attention to the role of public hospitals in the next infectious disease pandemic.

Usage of a Value-based Triaging Methodology for Assessing Improvements in Value for Hip Fracture Inpatient Episodes of Care From 2014 to 2019: A Pilot Study.

Introduction:The purpose of this study was to demonstrate a novel technology used to measure improvements in quality and value of care for treatment of hip fracture patients.Methods:A novel value-based triaging methodology uses a risk prediction (risk M) and inpatient cost prediction (risk C) algorithm and has been demonstrated to accurately predict high-risk:high-cost episodes of care. Two hundred twenty-nine hip fracture patients from 2014 to 2016 were used to establish baseline length of stay (LOS) and total inpatient cost for each (16) risk:cost quadrants. Two hundred sixty-five patients between 2017 and 2019 with hip fractures were input into the algorithm, and historical LOS and cost for each patient were calculated. Historical values were compared with actual values to determine whether the value of the inpatient episode of care differed from the 2014 to 16 cohort.Results:When evaluated without risk or cost stratification, the mean actual LOS and cost of the baseline cohort compared with the 2017 to 2019 cohort were 8.0 vs 7.5 days (P = 0.43) and $25,446 vs $29,849 (P = 0.15), respectively. This analysis demonstrates that there was only a small change in value of care provided to patients based on LOS/cost over the studied period; however, risk:cost analysis using the novel methodology demonstrated that for select risk:cost quadrants, value of care measured by LOS/cost improved, whereas for others it decreased and for others there was no change.Conclusion:Risk-cost–adjusted analysis of inpatient episodes of care rendered by a value-based triaging methodology provides a robust method of assessing improvements and/or decreases in value-based care when compared with a historical cohort. This methodology provides the tools to both track hospital interventions designed to improve quality and decrease cost as well as determine whether these interventions are effective in improving value.

Predicting Patient Length of Stay Using Artificial Intelligence to Assist Health Care Professionals in Resource Planning and Scheduling Decisions

Artificial Intelligence (AI) significantly revolutionizes and transforms the global healthcare industry by improving outcomes, increasing efficiency, and enhancing resource utilization. The applications of AI impact every aspect of healthcare operation, particularly resource allocation and capacity planning. This study proposes a multi-step AI-based framework and applies it to a real dataset to predict the length of stay (LOS) for hospitalized patients. The results show that the proposed framework can predict the LOS categories with an AUC of 0.85 and their actual LOS with a mean absolute error of 0.85 days. This framework can support decision-makers in healthcare facilities providing inpatient care to make better front-end operational decisions, such as resource capacity planning and scheduling decisions. Predicting LOS is pivotal in today’s healthcare supply chain (HSC) systems where resources are scarce, and demand is abundant due to various global crises and pandemics. Thus, this research’s findings have practical and theoretical implications in AI and HSC management.

Validation of a risk prediction calculator in Australian patients undergoing head and neck microsurgery reconstruction

The American College of Surgeons (ACS) National Surgical Quality Improvement Program (NSQIP) surgical risk calculator (SRC) is an open access calculator predicting patients' risk of postoperative complications. This study aims to assess the validity of the SRC in patients undergoing microsurgical free flap reconstruction at an Australian tertiary referral centre. This is a retrospective cohort study of 200 consecutive patients treated up to November 2020. SRC-predicted rates of postoperative complications and hospital length of stay (LOS) were compared to those observed for the ablative and reconstructive components of the procedure. The performance of the SRC was assessed using Brier scores, area under the receiver operating characteristic (ROC) curve (AUC), and the Hosmer-Lemeshow test. For both ablative and reconstructive components, the SRC discriminates well for pneumonia and urinary tract infection, and it is calibrated well for readmission and sepsis, but it does not discriminate and calibrate well for any single outcome. SRC-predicted hospital LOS and actual LOS did not correlate well for the reconstructive component, but they correlated strongly for the ablative component. The SRC is a poor predictor of postoperative complication rates and hospital LOS in patients undergoing head and neck microsurgical reconstruction.

Relationship between pediatric intensive care unit length of stay and 24-h unplanned readmission rate.

To evaluate the relationship between pediatric intensive care unit (PICU) severity-adjusted length of stay (LOS) and 24-h unplanned readmission rate. Data were obtained from a 10-year cohort from 2009 to 2018 from the Virtual Pediatric Systems (VPS, LLC) database. In this retrospective study, standardized LOS ratio was computed for each PICU as the ratio of the sum of actual LOS divided by the predicted LOS for each PICU using VPS predictive LOS model. Correlation between standardized LOS ratios and 24-h unplanned readmission rates were computed using Pearson's correlation coefficient. There was practically no relationship between standardized LOS ratio and 24-h readmission rate (R2 =0.05). Not Applicable. Severity-adjusted LOS has no relationship with 24-h unplanned readmission rate. These findings suggest that the relationship between PICU severity-adjusted LOS and 24-h unplanned readmission rate should not be used as a balancing quality measure.

The relationship between Intensive Care Unit Length of Stay information and its operational performance

The increasing demand for Intensive Care Unit (ICU) beds requires efficient admission, discharge, and care processes. These efforts require predictions of Length of Stay (LOS) values; however, it is unclear how accurate these predictions need to be. This study investigates the relationship between the accuracy level of LOS predictions and operational performance indicators. A discrete event simulation model is developed to model the ICU patient flows. A linear function of actual and simulated LOS values is used to measure the accuracy level of the predictions. Multiple configurations of patient mix and patient waiting threshold were included in the simulation scenarios. Performance indicators are the average waiting time of patients for an ICU bed and overall admission ratios. Further statistical tests were carried out to evaluate the significance of the results. Results suggest that inaccurate LOS predictions overestimated both the average waiting time of patients for an ICU bed and overall admission rates which can have several quality and performance implications for hospitals. The gaps increased when more elective patients were included in the patient mix. Moreover, higher waiting thresholds (i.e., the maximum amount of time a patient will wait for an ICU bed) yielded higher values in both performance indicators.

Risk Factors for Longer Pediatric Intensive Care Unit Length of Stay Among Children Who Required Escalation of Care Within 24 Hours of Admission.

Children who require early escalation of care (EOC) to the pediatric intensive care unit (PICU) after floor admission have higher mortality and increased hospital length of stay (LOS) as compared with direct emergency department (ED) admissions. This study was designed to identify subgroups of patients within this cohort (EOC to PICU within 24 hours of hospital admission) who have worse outcomes (actual PICU LOS [aLOS] > predicted PICU LOS [pLOS]). This was a retrospective single-center cohort study. Patients who required EOC to PICU from January 2015 to December 2019 within 24 hours of admission were included. Postoperative patients, missing cause of EOC, and mortality were excluded. Predicted LOS was calculated based on Pediatric Risk of Mortality scores. Patients with aLOS > pLOS (group A) were compared with patients with aLOS ≤ pLOS (group B). Multivariable logistic regression was performed to adjust for confounders. Of 587 patients transferred to PICU after hospital admission during the study period, 286 patients met the study criteria (group A, n = 69; group B, n = 217). The 2 groups were similar in age, race, the severity of illness, and ED vitals and therapies. A higher proportion of patients in group B had EOC ≤ 6 hours of admission (51.1% vs 36.2%, P = 0.03), and a higher proportion in group A required Mechanical ventilation (56% vs 34%, P = 0.01). On multivariable regression, patients who required EOC to PICU after 6 hours after admission (adjusted odds ratio, 2.27; 95% confidence interval [CI] 1.2, 4.0), p,<0.01) and patients admitted to the floor from referral hospitals (adjusted odds ratio, 1.8; 95% confidence interval, 1.0-3.2), P = 0.04) had higher risk of greater than PLOS. Among patients who required EOC to PICU, risk factors associated with aLOS > pLOS were patients who required EOC to PICU longer than 6 hours after admission to the hospital and patients admitted to the floor as a transfer from referral hospitals.

Perioperative Risk Stratification: A Need for an Improved Assessment in Surgery and Anesthesia-A Pilot Study.

Background and Objectives: Numerous scoring systems have been introduced into modern medicine. None of the scoring systems assessed both anesthetic and surgical risk of the patient, predict the morbidity, mortality, or the need for postoperative intensive care unit admission. The aim of this study was to compare the anesthetic and surgical scores currently used, for a better evaluation of perioperative risks, morbidity, and mortality. Material and Methods: This is a pilot, prospective, observational study. We enrolled 50 patients scheduled for elective surgery. Anesthetic and surgery risk was assessed using American Society of Anesthesiologists (ASA) scale, Physiological and Operative Severity Score for the enumeration of Mortality and morbidity (P-POSSUM), Acute Physiology and Chronic Health Evaluation (APACHE II), and Surgical APGAR Score (SAS) scores. The real and the estimated length of stay (LOS) were registered. Results: We obtained several statistically significant positive correlations: ASA score–P-POSSUM (p < 0.01, r = 0.465); ASA score–SAS, (p < 0.01, r = −0.446); ASA score–APACHE II, (p < 0.01 r = 0.519); predicted LOS and ASA score (p < 0.01, r = 0.676); predicted LOS and p-POSSUM (p < 0.01, r = 0.433); and predicted LOS and APACHE II (p < 0.01, r = 0.454). A significant negative correlation between predicted LOS, real LOS, ASA class, and SAS (p < 0.05) was observed. We found a statistically significant difference between the predicted and actual LOS (p < 001). Conclusions: Anesthetic, surgical, and severity scores, used together, provide clearer information about mortality, morbidity, and LOS. ASA scale, associated with surgical scores and severity scores, presents a better image of the patient’s progress in the perioperative period. In our study, APACHE II is the best predictor of mortality, followed by P-POSSUM and SAS. P-POSSUM score and ASA scale may be complementary in terms of preoperative physiological factors, providing valuable information for postoperative outcomes.

Centers for Medicare & Medicaid Services' 2018 Removal of Total Knee Arthroplasty From the Inpatient-only List Led to Broad Changes in Hospital Length of Stays.

Centers for Medicare & Medicaid Services (CMS) removed total knee arthroplasty (TKA) from the "inpatient-only" list from January 1, 2018. The impact of this change on actual hospital length of stay (LOS) and patient coding is of interest. Patients undergoing TKA were abstracted from the 2015 to 2018 National Surgical Quality Improvement Program database. Patient characterization as "inpatient" or "outpatient" and actual LOS were assessed. Ordinal and categorical data comparisons were done with Pearson chi-squared tests. Continuous variables were tested for normality, and nonparametric analyses were conducted using the Mann-Whitney test. Significance was set at P < 0.05. In total, 125,613 TKA patients from 2017 to 2018 were identified (232,269 TKA patients from 2015 to 2018). Most patients undergoing TKA were of Medicare eligibility (≥65 years old; 60.78% in 2017 and 62.42% in 2018). Overall, LOS decreased significantly from 2017 to 2018 (2.31 ± 1.56 days versus 2.05 ± 1.57 days; P < 0.001), and more patients were discharged the same day (5.09% versus 2.28%; P < 0.001). In 2017, patients were coded as "outpatient" 1.66% of the time (those with LOS = 0 days were 22.85%, LOS = 1 day were 1.80%, LOS = 2 days were 0.79%, and LOS ≥3 days were 0.85%). In 2018, patients were coded as "outpatient" 17.14% of the time (those with LOS = 0 days were 78.2%, LOS = 1 day were 29.75%, LOS = 2 days were 6.96%, and LOS ≥3 days were 3.05%). This represented a significant change for each LOS day (P < 0.001). These results remained true when stratifying by Medicare eligibility (P < 0.001 for those <65 years old and those ≥65 years old). After the 2018 removal of TKA from the CMS "inpatient-only" list, patients were more likely to be discharged the same day and be considered "outpatients." Patients with more prolonged LOS and those younger than 65 years were more likely to have been coded as "outpatient" in 2018 compared with 2017. These data demonstrate that national changes in CMS policies can have broad impact on overall practice patterns. Retrospective cohort study.

Length of Stay After Joint Arthroplasty is Less Than Predicted Using Two Risk Calculators

BackgroundPredicting the length of stay (LOS) after total joint arthroplasty (TJA) has become more important with their recent removal from inpatient-only designation. The American College of Surgeons (ACS) National Surgical Quality Improvement Program surgical risk calculator and the CMS’ diagnosis-related group (DRG) calculator are two common LOS predictors. The aim of our study was to determine how our actual LOS compared with those predicted by both the ACS and DRG. Methods99 consecutive TJA (49 hips and 50 knee procedures) were reviewed in Medicare-eligible patients from four fellowship-trained arthroplasty surgeons. Predicted LOS was calculated using the DRG and ACS risk calculators for each patient using demographics, medical histories, and comorbidities. LOS was compared between the predicted and the actual LOS for both total hip arthroplasty (THA) and total knee arthroplasty (TKA) using paired t-tests. ResultsActual LOS was shorter in the THA group vs the TKA group (1.29 days vs 1.46 days, P < .05). The actual LOS of patients at our institution was significantly shorter than both DRG and ACS predictions for both THA and TKA (P < .05). In both the THA and TKA patients, the actual LOS (1.29 and 1.46 day) was significantly shorter than the DRG-predicted LOS (2.15 and 2.15 days) which was significantly shorter than the ACS-predicted LOS (2.9 and 3.14 days). ConclusionWe found the actual LOS was significantly shorter than that predicted by both the DRG and ACS risk calculators. Current risk calculators may not be accurate for contemporary fast-track protocols and newer tools should be developed.

Optimizing discharge decision-making in colorectal surgery: a prospective cohort study of discharge practices in a recently implemented enhanced recovery pathway.

The objectives of this project were (1) to compare time to readiness for discharge by set criteria and actual length of stay (LOS) in a newly implemented colorectal enhanced recovery pathway and (2) to identify reasons for delayed hospital discharge. We conducted a prospective cohort study of 73 adult patients (age 67±14years, 56% men, 51% laparoscopic, 13% stoma creation) undergoing elective colorectal surgery in a university hospital with a recently implemented recovery pathway (<2years). Time to readiness for discharge (oral intake, flatus, pain control, ability to walk, and no complications) was compared to actual LOS using a correlation-adjusted log-rank test. The treating team was interviewed, and thematic analysis was used to identify reasons for patients remaining in hospital after discharge criteria (DC) were achieved. Median LOS was 6 (4-8) days and median time to readiness for discharge was 5 (3-8) days (P<0.001). Twenty-eight patients (37%) remained in hospital after DC were achieved. Although some delayed discharges were medically justified (e.g., workup [13%] or treatment of complications not captured by DC [2.6%]), unnecessary hospital stays were common (e.g., perceived need for observation [16%], or patients not willing to be discharged [11%]). Unnecessary hospital stays were common within a recently implemented enhanced recovery pathway and represent a target for quality improvement. Efforts should be directed at optimizing patient education regarding discharge expectations, early consultation of the discharge planning team and improving discharge decision-making using standardized DC.

726. Investigation of Epidemiological Characteristics and Outcomes in Enteropathogenic Escherichia coli positive patients

BackgroundNew generation multiplex polymerase chain reaction (PCR) panels have led to the ability of rapid detection of Enteropathogenic Escherichia coli (EPEC). Although many studies have looked at the pathogenesis of this organism in the pediatric populations, fewer studies include adults. We aimed to determine if EPEC is pathogenic on its own, or has a predilection for certain populations based upon risk factors. This was achieved by comparing risk factors in patients positive for EPEC on the BioFire Gastrointestinal (GI) Pathogen Panel versus patients negative for all targets on the panel.MethodsThis is a single center case control study that was performed using a retrospective chart review from January 1, 2016 thru August 31, 2019. All patients were symptomatic with diarrheal illness and had a GI pathogen panel performed. The study group were patients that were EPEC positive. The control group were patients with a negative PCR panel that were matched three to one based upon age and gender. Chi-squared statistical analysis was used.Results792 patients were evaluated. In the adult group the EPEC positive actual length of stay (LOS) was 13.5 rather than 9.3 in the control. Although this was not statically significant this trended towards a longer LOS as seen in Table 1. In the pediatric population, both expected and actual LOS were statically longer viewed in Table 2. This may be due to these patients being more chronically ill at baseline, with a greater number of pediatric coronary artery disease and ulcerative colitis in the control. Males were more likely to test positive for EPEC at 69.86% in the pediatric population than in the adult population with statistical significance at p=0.0035. The pediatric population had more co-infections with EPEC at 57.53% with p< 0.0001. This can be further seen in Table 3.Table 1 Table 2 Table 3 ConclusionIn the adult EPEC positive patients there was no statistically significant difference in length of stay in comparison to negative control. This raises the question to if this is truly a pathogen in adults. In the pediatric population, our control group was likely more ill, which made it difficult interpret the significance of differences in LOS.Disclosures Kenneth Rand, MD, BioFire Diagnostics, Inc (Consultant, Grant/Research Support, Advisor or Review Panel member, Research Grant or Support) Stacy Beal, MD, FilmArray BioFire (Grant/Research Support)

Abstract P104: Impact Of Orthostatic Hypotension In Inpatient Rehabilitation

Inpatient rehabilitation is crucial in the management of patients after injuries, surgery, or those with debilitating or neurological conditions. Orthostatic hypotension (OH) is commonly associated with these conditions, and therefore, the inability to stand and participate in rehabilitation therapy can negatively impact their recovery. To test this hypothesis, we study a cohort of 8352 patients who underwent rehabilitation at Vanderbilt Stallworth Rehabilitation Hospital between 2014-2019 to assess the prevalence of OH and its impact on the length of stay (LOS), functional independence measure (FIM) change (discharge FIM-admission FIM), and FIM efficiency (FIM change/LOS), compared to patients without OH admitted during the same period. We found that the prevalence of OH was 3.8% (n = 314; 60±18 yrs; 227 men). OH was commonly associated with patients admitted for neurological conditions (29%, n = 91), spinal cord dysfunction (22%, n = 71), orthopedic disorders (13%, n = 13), and stroke (10%, n = 33). On admission, patients with OH had a longer predicted LOS compared to those without OH (18±6 vs. 16±5 days, respectively, p = 0.001), reflecting greater medical complexity. Moreover, the actual LOS at discharge was significantly longer in OH patients compared to those without OH (19±12 vs. 13±9 days; p = 0.001) as was the difference between actual-predicted LOC (1±11 vs. 2±8; p = 0.001). The change in FIM was similar in both patients with and without OH (30±26 vs. 33±20, respectively; p = 0.52) but FIM efficiency was lower in OH patients (2±2.2 vs. 3±2.2 without OH; p = 0.001). OH prolonged LOS regardless of the admission diagnosis (stroke, other neurological conditions, spinal cord dysfunction and orthopedic disorders). In conclusion, OH is present in a significant proportion of patients undergoing acute rehabilitation and adversely impacts the recovery of these patients as measured by lower FIM efficacy and a prolonged rehabilitation period to achieve the same functional gain as patients without OH.

Case-mix adjusted postanaesthesia care unit length of stay and business intelligence dashboards for feedback to anaesthetists

BackgroundDespite advances in business intelligence software and evidence that feedback to doctors can improve outcomes, objective feedback regarding patient outcomes for individual anaesthetists is hampered by lack of useful benchmarks. We aimed to address this issue by producing case-mix and risk-adjusted postanaesthesia care unit (PACU) length of stay (LOS) benchmarks for integration into modern reporting tools. MethodsWe extended existing hospital information systems to calculate predicted PACU LOS using a neural network trained on patient age, surgery duration, sex, operating specialty, urgency, weekday, and insurance status (n=100 511). We then calculated the difference between observed mean and predicted PACU LOS for individual doctors, and compared the results with and without case-mix adjustment. We report practical implications of using visual analytics dashboards displaying the difference between observed and predicted PACU LOS to provide feedback to anaesthetic doctors. ResultsThe neural network accounted for over half of observed variation in individual doctors' mean PACU LOS (mean predicted and mean actual LOS Spearman's r2=0.57). Account for case-mix reduced apparent spread, with 80% of individual doctors falling in a band of 4.3 min after case-mix adjusting, compared with a range of 24 min without adjustment. Case-mix adjusting also identified different individual doctors as outliers (Weighted Cohen's kappa [κ]=0.27). Finally, we demonstrated that we were able to integrate the adjusted metrics into routine reporting tools. ConclusionWith caution, case-mix adjustment of anaesthetic outcome measures such as PACU LOS potentially provides a useful continuous quality improvement tool. Unadjusted outcome measures are imprecise at best and misleading at worst.