Current Risk Adjustment and Comorbidity Index Underperformance in Predicting Post-Acute Utilization and Hospital Readmissions After Joint Replacements: Implications for Comprehensive Care for Joint Replacement Model

Validation of a Discharge Risk Calculator for Rural Patients Following Total Joint Arthroplasty

BACKGROUND

As the volume of total joint arthroplasty in the US continues to grow, new challenges surrounding appropriate discharge surface. Arthroplasty literature has demonstrated discharge disposition to postacute care facilities carries major risks regarding the need for revision surgery, patient comorbidities, and financial burden. To quantify, categorize, and mitigate risks, a decision tool that uses preoperative patient variables has previously been published and validated using an urban patient population. The aim of our investigation was to validate the same predictive model using patients in a rural setting undergoing total knee arthroplasty (TKA) and total hip arthroplasty.

METHODS

All TKA and THA procedures that were performed between January 2012 and September 2022 at our institution were collected. A total of 9,477 cases (39.6% TKA, 60.4% THA) were included for the validation analysis. There were 9 preoperative variables that were extracted in an automated fashion from the electronic medical record. Included patients were then run through the predictive model, generating a risk score representing that patient's differential risk of discharge to a skilled nursing facility versus home. Overall accuracy, sensitivity and specificity were calculated after obtaining risk scores.

RESULTS

Score cutoff equally maximizing sensitivity and specificity was 0.23, and the proportion of correct classifications by the predictive tool in this study population was found to be 0.723, with an area under the curve of 0.788 - both higher than previously published accuracy levels. With the threshold of 0.23, sensitivity and specificity were found to be 0.720 and 0.723, respectively.

CONCLUSIONS

The risk calculator showed very good accuracy, sensitivity, and specificity in predicting discharge location for rural patients undergoing TKA and THA, with accuracy even higher than in urban populations. The model provides an easy-to-use interface, with automation representing a viable tool in helping with shared decision-making regarding postoperative discharge plans.

Association Between Hospital Participation in Value-Based Programs and Timely Initiation of Post-Acute Home Health Care, Functional Recovery, and Hospital Readmission After Joint Replacement

OBJECTIVES

This study examined the association between hospital participation in Bundled Payments for Care Improvement (BPCI) or Comprehensive Care for Joint Replacement (CJR) and the timely initiation of home health rehabilitation services for lower extremity joint replacements. Furthermore, this study examined the association between the timely initiation of home health rehabilitation services with improvement in self-care, mobility, and 90-day hospital readmission.

METHOD

This retrospective cohort study used Medicare inpatient claims and home health assessment data from 2016 to 2017 for older adults discharged to home with home health following hospitalization after joint replacement. Multilevel multivariate logistic regression was used to examine the association between hospital participation in BPCI or CJR programs and timely initiation of home health rehabilitation service. A 2-staged generalized boosted model was used to examine the association between delay in home health initiation and improvement in self-care, mobility, and 90-day risk-adjusted hospital readmission.

RESULTS

Compared with patients discharged from hospitals that did not have BPCI or CJR, patients discharged from hospitals with these programs had a lower likelihood of delayed initiation of home health rehabilitation services for both knees and hip replacement. Using propensity scores as the inverse probability of treatment weights, delay in the initiation of home health rehabilitation services was associated with lower improvement in self-care (odds ratio [OR] = 1.23; 95% CI = 1.20-1.26), mobility (OR = 1.15; 95% CI = 1.13-1.18), and higher rate of 90-day hospital readmission (OR = 1.19; 95% CI = 1.15-1.24) for knee replacement. Likewise, delayed initiation of home health rehabilitation services was associated with lower improvement in self-care (OR = 1.16; 95% CI = 1.13-1.20) and mobility (OR = 1.26; 95% CI = 1.22-1.30) for hip replacement.

CONCLUSION

Hospital participation in BPCI or comprehensive CJR was associated with early home health rehabilitation care initiation, which was further associated with significant increases in functional recovery and lower risks of hospital readmission.

IMPACT

Policy makers may consider incentivizing health care providers to initiate early home health services and care coordination in value-based payment models.

Readmission rates and outcomes in adults with and without COVID-19 following inpatient chemotherapy admission: A nationwide analysis

BACKGROUND

The coronavirus disease 2019 (COVID-19) pandemic has received considerable attention in the scientific community due to its impact on healthcare systems and various diseases. However, little focus has been given to its effect on cancer treatment.

AIM

To determine the effect of COVID-19 pandemic on cancer patients' care.

METHODS

A retrospective review of a Nationwide Readmission Database (NRD) was conducted to analyze hospitalization patterns of patients receiving inpatient chemotherapy (IPCT) during the COVID-19 pandemic in 2020. Two cohorts were defined based on readmission within 30 d and 90 d. Demographic information, readmission rates, hospital-specific variables, length of hospital stay (LOS), and treatment costs were analyzed. Comorbidities were assessed using the Elixhauser comorbidity index. Multivariate Cox regression analysis was performed to identify independent predictors of readmission. Statistical analysis was conducted using Stata® Version 16 software. As the NRD data is anonymous and cannot be used to identify patients, institutional review board approval was not required for this study.

RESULTS

A total of 87755 hospitalizations for IPCT were identified during the pandemic. Among the 30-day index admission cohort, 55005 patients were included, with 32903 readmissions observed, resulting in a readmission rate of 59.8%. For the 90-day index admission cohort, 33142 patients were included, with 24503 readmissions observed, leading to a readmission rate of 73.93%. The most common causes of readmission included encounters with chemotherapy (66.7%), neutropenia (4.36%), and sepsis (3.3%). Comorbidities were significantly higher among readmitted hospitalizations compared to index hospitalizations in both readmission cohorts. The total cost of readmission for both cohorts amounted to 1193000000.00 dollars. Major predictors of 30-day readmission included peripheral vascular disorders [Hazard ratio (HR) = 1.09, P < 0.05], paralysis (HR = 1.26, P < 0.001), and human immunodeficiency virus/acquired immuno-deficiency syndrome (HR = 1.14, P = 0.03). Predictors of 90-day readmission included lymphoma (HR = 1.14, P < 0.01), paralysis (HR = 1.21, P = 0.02), and peripheral vascular disorders (HR = 1.15, P < 0.01).

CONCLUSION

The COVID-19 pandemic has significantly impacted the management of patients undergoing IPCT. These findings highlight the urgent need for a more strategic approach to the care of patients receiving IPCT during pandemics.

Can process mapping and a multisite Delphi of perioperative professionals inform our understanding of system-wide factors that may impact operative risk?

OBJECTIVES

To examine whether the use of process mapping and a multidisciplinary Delphi can identify potential contributors to perioperative risk. We hypothesised that this approach may identify factors not represented in common perioperative risk tools and give insights of use to future research in this area.

DESIGN

Multidisciplinary, modified Delphi study.

SETTING

Two centres (one tertiary, one secondary) in the UK during 2020 amidst coronavirus pressures.

PARTICIPANTS

91 stakeholders from 23 professional groups involved in the perioperative care of older patients. Key stakeholder groups were identified via process mapping of local perioperative care pathways.

RESULTS

Response rate ranged from 51% in round 1 to 19% in round 3. After round 1, free text suggestions from the panel were combined with variables identified from perioperative risk scores. This yielded a total of 410 variables that were voted on in subsequent rounds. Including new suggestions from round two, 468/519 (90%) of the statements presented to the panel reached a consensus decision by the end of round 3. Identified risk factors included patient-level factors (such as ethnicity and socioeconomic status), and organisational or process factors related to the individual hospital (such as policies, staffing and organisational culture). 66/160 (41%) of the new suggestions did not feature in systematic reviews of perioperative risk scores or key process indicators. No factor categorised as 'organisational' is currently present in any perioperative risk score.

CONCLUSIONS

Through process mapping and a modified Delphi we gained insights into additional factors that may contribute to perioperative risk. Many were absent from currently used risk stratification scores. These results enable an appreciation of the contextual limitations of currently used risk tools and could support future research into the generation of more holistic data sets for the development of perioperative risk assessment tools.

Rates, Reasons, and Independent Predictors of Readmissions in Portal Venous Thrombosis Hospitalizations in the USA

BACKGROUND

Portal vein thrombosis (PVT), generally considered rare, is becoming increasingly recognized with advanced imaging. Limited data exist regarding readmissions in PVT and its burden on the overall healthcare cost. This study aimed to outline the burden of PVT readmissions and identify the modifiable predictors of readmissions.

METHODS

The National Readmission Database (NRD) was used to identify PVT admissions from 2016 to 2019. Using the patient demographic and hospital-specific variables within the NRD, we grouped patient encounters into two cohorts, 30- and 90-day readmission cohorts. We assessed comorbidities using the validated Elixhauser comorbidity index. We obtained inpatient mortality rates, mean length of hospital stay (LOS), total hospital cost (THC), and causes of readmissions in both 30- and 90-day readmission cohorts. Using a multivariate Cox regression analysis, we identified the independent predictors of 30-day readmissions.

RESULTS

We identified 17,971 unique index hospitalizations, of which 2,971 (16.5%) were readmitted within 30 days. The top five causes of readmissions in both 30-day and 90-day readmission cohorts were PVT, sepsis, hepatocellular cancer, liver failure, and alcoholic liver cirrhosis. The following independent predictors of 30-day readmission were identified: discharge against medical advice (AMA) (adjusted hazard ratio (aHR) 1.86; P = 0.002); renal failure (aHR 1.44, P = 0.014), metastatic cancer (aHR 1.31, P = 0.016), fluid and electrolyte disorders (aHR 1.20, P = 0.004), diabetes mellitus (aHR 1.31, P = 0.001) and alcohol abuse (aHR 1.31, P ≤ 0.001).

CONCLUSION

The readmission rate identified in this study was higher than the national average and targeted interventions addressing these factors may help reduce the overall health care costs.

Validation of a Predictive Tool for Discharge to Rehabilitation or a Skilled Nursing Facility After TJA

BACKGROUND

Cost excess in bundled payment models for total joint arthroplasty (TJA) is driven by discharge to rehabilitation or a skilled nursing facility (SNF). A recently published preoperative risk prediction tool showed very good internal accuracy in stratifying patients on the basis of likelihood of discharge to an SNF or rehabilitation. The purpose of the present study was to test the accuracy of this predictive tool through external validation with use of a large cohort from an outside institution.

METHODS

A total of 20,294 primary unilateral total hip (48%) and knee (52%) arthroplasty cases at a tertiary health system were extracted from the institutional electronic medical record. Discharge location and the 9 preoperative variables required by the predictive model were collected. All cases were run through the model to generate risk scores for those patients, which were compared with the actual discharge locations to evaluate the cutoff originally proposed in the derivation paper. The proportion of correct classifications at this threshold was evaluated, as well as the sensitivity, specificity, positive and negative predictive values, number needed to screen, and area under the receiver operating characteristic curve (AUC), in order to determine the predictive accuracy of the model.

RESULTS

A total of 3,147 (15.5%) of the patients who underwent primary, unilateral total hip or knee arthroplasty were discharged to rehabilitation or an SNF. Despite considerable differences between the present and original model derivation cohorts, predicted scores demonstrated very good accuracy (AUC, 0.734; 95% confidence interval, 0.725 to 0.744). The threshold simultaneously maximizing sensitivity and specificity was 0.1745 (sensitivity, 0.672; specificity, 0.679), essentially identical to the proposed cutoff of the original paper (0.178). The proportion of correct classifications was 0.679. Positive and negative predictive values (0.277 and 0.919, respectively) were substantially better than those of random selection based only on event prevalence (0.155 and 0.845), and the number needed to screen was 3.6 (random selection, 6.4).

CONCLUSIONS

A previously published online predictive tool for discharge to rehabilitation or an SNF performed well under external validation, demonstrating a positive predictive value 79% higher and number needed to screen 56% lower than simple random selection. This tool consists of exclusively preoperative parameters that are easily collected. Based on a successful external validation, this tool merits consideration for clinical implementation because of its value for patient counseling, preoperative optimization, and discharge planning.

LEVEL OF EVIDENCE

Prognostic Level III . See Instructions for Authors for a complete description of levels of evidence.

Medicare Claim-Based National Institutes of Health Stroke Scale to Predict 30-Day Mortality and Hospital Readmission

BACKGROUND

The Centers for Medicare and Medicaid Services (CMS) penalizes hospitals for higher than expected 30-day mortality rates using methods without accounting for condition severity risk adjustment. For patients with stroke, CMS claims did not quantify stroke severity until recently, when the National Institutes of Health Stroke Scale (NIHSS) reporting began.

OBJECTIVE

Examine the predictive ability of claim-based NIHSS to predict 30-day mortality and 30-day hospital readmission in patients with ischemic stroke.

DESIGN

Retrospective cohort study of Medicare claims data.

PATIENTS

Medicare beneficiaries with ischemic stroke (N=43,241) acute hospitalization between October 2016 and November 2017.

MEASUREMENTS

All-cause 30-day mortality and 30-day hospital readmission. NIHSS score was derived from ICD-10 codes and stratified into the following: minor to moderate, moderate, moderate to severe, and severe categories.

RESULTS

Among 43,241 patients with ischemic stroke with NIHSS from 2,659 US hospitals, 64.6% had minor to moderate stroke, 14.3% had moderate, 12.7% had moderate to severe, and 8.5% had a severe stroke,10.1% died within 30 days, 12.1% were readmitted within 30 days. The NIHSS exhibited stronger discriminant property (C-statistic 0.83, 95% CI: 0.82-0.84) for 30-day mortality compared to Elixhauser (0.74, 95% CI: 0.73-0.75). A monotonic increase in the adjusted 30-day mortality risk occurred relative to minor to moderate stroke category: hazard ratio [HR]=2.92 (95% CI=2.59-3.29) for moderate stroke, HR=5.49 (95% CI=4.90-6.15) for moderate to severe stroke, and HR=7.82 (95% CI=6.95-8.80) for severe stroke. After accounting for competing risk of mortality, there was a significantly higher readmission risk in the moderate stroke (HR=1.11, 95% CI=1.03-1.20), but significantly lower readmission risk in the severe stroke (HR=0.84, 95% CI=0.74-0.95) categories.

LIMITATION

Timing of NIHSS reporting during hospitalization is unknown.

CONCLUSIONS

Medicare claim-based NIHSS is significantly associated with 30-day mortality in Medicare patients with ischemic stroke and significantly improves discriminant property relative to the Elixhauser comorbidity index.

Evaluation of risk adjustment performance of diagnosis-based and medication-based comorbidity indices in patients with chronic obstructive pulmonary disease

Objectives

This study assessed risk adjustment performance of six comorbidity indices in two categories of comorbidity measures: diagnosis-based comorbidity indices and medication-based ones in patients with chronic obstructive pulmonary disease (COPD).

Methods

This was a population–based retrospective cohort study. Data used in this study were sourced from the Taiwan National Health Insurance Research Database. The study population comprised all patients who were hospitalized due to COPD for the first time in the target year of 2012. Each qualified patient was individually followed for one year starting from the index date to assess two outcomes of interest, medical expenditures within one year after discharge and in-hospital mortality of patients. To assess how well the added comorbidity measures would improve the fitted model, we calculated the log-likelihood ratio statistic G2. Subsequently, we compared risk adjustment performance of the comorbidity indices by using the Harrell c-statistic measure derived from multiple logistic regression models.

Results

Analytical results demonstrated that that comorbidity measures were significant predictors of medical expenditures and mortality of COPD patients. Specifically, in the category of diagnosis-based comorbidity indices the Elixhauser index was superior to other indices, while the RxRisk-V index was a stronger predictor in the framework of medication-based codes, for gauging both medical expenditures and in-hospital mortality by utilizing information from the index hospitalization only as well as the index and prior hospitalizations.

Conclusions

In conclusion, this work has ascertained that comorbidity indices are significant predictors of medical expenditures and mortality of COPD patients. Based on the study findings, we propose that when designing the payment schemes for patients with chronic diseases, the health authority should make adjustments in accordance with the burden of health care caused by comorbid conditions.

Patients From Medically Underserved Areas Are at Increased Risk for Nonhome Discharge and Emergency Department Return After Total Joint Arthroplasty

BACKGROUND

Maryland Health Enterprise Zones (MHEZs) were introduced in 2012 and encompass underserved areas and those with reduced access to healthcare providers. Across the United States many underserved and minority populations experience poorer total joint arthroplasty (TJA) outcomes seemingly because they reside in underserved areas. The purpose of this study is to identify and quantify the relationship between living in an MHEZ and TJA outcomes.

METHODS

Retrospective review of 11,451 patients undergoing primary TJA at a single institution from July 1, 2014 to June 30, 2020 was conducted. Patients were classified based on whether they resided in an MHEZ. Statistical analyses were used to compare outcomes for TJA patients who live in MHEZ and those who do not.

RESULTS

Of the 11,451 patients, 1057 patients lived in MHEZ and 10,394 patients did not. After risk adjustment, patients who live in an MHEZ were more likely to return to the emergency department within 90 days postoperatively and were less likely to be discharged home than those patients who do not live in an MHEZ.

CONCLUSION

Total joint arthroplasty patients residing in MHEZ appear to present with poorer overall health as measured by increased American Society of Anesthesiologists and Hierarchical Condition Categories scores, and they are less likely to be discharged home and more likely to return to the emergency department within 90 days. Several factors associated with these findings such as socioeconomic factors, household composition, housing type, disability, and transportation may be modifiable and should be targets of future population health initiatives.

Impact of Hospital-Based Rehabilitation Services on Discharge to the Community by Value-Based Payment Programs After Joint Replacement Surgery

OBJECTIVE

The purpose of this study was to examine the impact of hospital-based rehabilitation services on community discharge rates after hip and knee replacement surgery according to hospital participation in value-based care models: bundled payments for care improvement (BPCI) and comprehensive care for joint replacement (CJR). The secondary objective was to determine whether community discharge rates after hip and knee replacement surgery differed by participation in these models.

METHODS

A secondary analysis of Medicare fee-for-service claims was conducted for beneficiaries 65 years of age or older who underwent hip and knee replacement surgery from 2016 to 2017. Independent variables were hospital participation in value-based programs categorized as: (1) BPCI, (2) CJR, and (3) non-BPCI/CJR; and total minutes per day of hospital-based rehabilitation services categorized into tertiles. The primary outcome variable was discharged to the community versus discharged to institutional post-acute care settings. The association between rehabilitation amount and community discharge among BPCI, CJR, and non-BPCI/CJR hospitals was adjusted for patient-level clinical and hospital characteristics.

RESULTS

Participation in BPCI or CJR was not associated with community discharge. This analysis found a dose-response relationship between the amount of rehabilitation services and odds of community discharge. Among those who received a hip replacement, this relationship was most pronounced in the BPCI group; compared with the low rehabilitation category, the medium category had odds ratio (OR) = 1.28 (95% CI = 1.17 to 1.41), and the high category had OR = 1.90 (95% CI = 1.71 to 2.11). For those who received a knee replacement, there was a dose-response relationship in the CJR group only; compared with the low rehabilitation category, the medium category had OR = 1.21 (95% CI = 1.15 to 1.28), and the high category had OR = 1.56 (95% CI = 1.46 to 1.66).

CONCLUSION

Regardless of hospital participation in BPCI or CJR models, higher amounts of rehabilitation services delivered during acute hospitalization is associated with a higher likelihood of discharge to community following hip and knee replacement surgery.

IMPACT

In the era of value-based care, frontloading of rehabilitation care is vital for improving patient-centered health outcomes in acute phases of lower extremity joint replacement.

The Update on Instruments Used for Evaluation of Comorbidities in Total Hip Arthroplasty

INTRODUCTION

It is a well-established fact that concomitant diseases can affect the outcome of total hip arthroplasty (THA). Therefore, careful preoperative assessment of a patient's comorbidity burden is a necessity, and it should be a part of routine screening as THA is associated with a significant number of complications. To measure the multimorbidity, dedicated clinical tools are used.

METHODS

The article is a systematic review of instruments used to evaluate comorbidities in THA studies. To create a list of available instruments for assessing patient's comorbidities, the search of medical databases (PubMed, Web of Science, Embase) for indices with proven impact on revision risk, adverse events, mortality, or patient's physical functioning was performed by two independent researchers.

RESULTS

The initial search led to identifying 564 articles from which 26 were included in this review. The measurement tools used were: The Charlson Comorbidity Index (18/26), Society of Anesthesiology classification (10/26), Elixhauser Comorbidity Method (6/26), and modified Frailty Index (5/26). The following outcomes were measured: quality of life and physical function (8/26), complications (10/26), mortality (8/26), length of stay (6/26), readmission (5/26), reoperation (2/26), satisfaction (2/26), blood transfusion (2/26), surgery delay or cancelation (1/26), cost of care (1/26), risk of falls (1/26), and use of painkillers (1/26). Further research resulted in a comprehensive list of eleven indices suitable for use in THA outcomes studies.

CONCLUSION

The comorbidity assessment tools used in THA studies present a high heterogeneity level, and there is no particular system that has been uniformly adopted. This review can serve as a help and an essential guide for researchers in the field.

A Comprehensive Monetary Analysis of Inpatient Total Hip and Knee Arthroplasties Billed to Medicare by Hospitals: 2011-2017

BACKGROUND

Total joint arthroplasty (TJA) has been a recent target of reimbursement reform. As such, the purpose of this study was to evaluate trends in Medicare reimbursement to hospitals for TJA patients from 2011 to 2017.

METHODS

The Inpatient Utilization and Payment Public Use File was queried for all primary total hip and knee arthroplasty episodes. This file includes all services billed to Medicare via the Inpatient Prospective Payment System. Extracted data included hospital charges and amount paid by Medicare. All data were adjusted for inflation to 2017 US dollars. Multiple linear mixed-model regression analyses were conducted to assess change over time, and geo-modelling was used to represent reimbursement by location.

RESULTS

A total of 3,368,924 primary TJA procedures were billed to Medicare by hospitals from 2011 to 2017 and included in the study. The mean inflation-adjusted Medicare payment to hospitals for DRG 469 decreased from $22,783.66 to $19,604.62 per procedure (-$3179.04; -14.0%; P < .001) and decreased from $13,290.79 to $11,771.54 for DRG 470 (-$1519.25; -11.4%, P = .011) from 2011 to 2017. Meanwhile, the mean charge submitted by hospitals increased by $6483.39 and $5115.60 for DRGs 469 and 470, respectively (+7.4% for 469, +9.3% for 470; P < .001). Medicare reimbursement to hospitals varied by state.

CONCLUSION

During the study period, the mean Medicare reimbursement to hospitals decreased for TJA from 2011 to 2017. Meanwhile, the average charge submitted by hospitals increased. As alternative payment models continue to undergo evaluation and development, these data are important for the advancement of more agreeable reimbursement models in arthroplasty care.

Body Mass Index, American Society of Anesthesiologists Score, and Elixhauser Comorbidity Index Predict Cost and Delay of Care During Total Knee Arthroplasty

BACKGROUND

Body mass index (BMI), American Society of Anesthesiologists (ASA) score, and Elixhauser Comorbidity Index are measures that are utilized to predict perioperative outcomes, though little is known about their comparative predictive effects. We analyzed the effects of these indices on costs, operating room (OR) time, and length of stay (LOS) with the hypothesis that they would have a differential influence on each outcome variable.

METHODS

A retrospective review of the institutional database was completed on primary TKA patients from 2015 to 2018. Univariable and multivariable models were constructed to evaluate the strength of BMI, ASA, and Elixhauser comorbidities for predicting changes to total hospital and surgical costs, OR time, and LOS.

RESULTS

In total, 1313 patients were included. ASA score was independently predictive of all outcome variables (OR time, LOS, total hospital and surgical costs). BMI, however, was associated with intraoperative resource utilization through time and cost, but only remained predictive of OR time in an adjusted model. Total Elixhauser comorbidities were independently predictive of LOS and total hospital cost incurred outside of the operative theater, though they were not predictive of intraoperative resource consumption.

CONCLUSION

Although ASA, BMI, and Elixhauser comorbidities have the potential to impact outcomes and cost, there are important differences in their predictive nature. Although BMI is independently predictive of intraoperative resource utilization, other measures like Elixhauser and ASA score were more indicative of cost outside of the OR and LOS. These data highlight the differing impact of BMI, ASA, and patient comorbidities in impacting cost and time consumption throughout perioperative care.

A Preoperative Risk Prediction Tool for Discharge to a Skilled Nursing or Rehabilitation Facility After Total Joint Arthroplasty

BACKGROUND

Discharge to rehabilitation or a skilled nursing facility (SNF) after total joint arthroplasty remains a primary driver of cost excess for bundled payments. An accurate preoperative risk prediction tool would help providers and health systems identify and modulate perioperative care for higher risk individuals and serve as a vital tool in preoperative clinic as part of shared decision-making regarding the risks/benefits of surgery.

METHODS

A total of 10,155 primary total knee (5,570, 55%) and hip (4,585, 45%) arthroplasties performed between June 2013 and January 2018 at a single institution were reviewed. The predictive ability of 45 variables for discharge location (SNF/rehab vs home) was tested, including preoperative sociodemographic factors, intraoperative metrics, postoperative labs, as well as 30 Elixhauser comorbidities. Parameters surviving selection were included in a multivariable logistic regression model, which was calibrated using 20,000 bootstrapped samples.

RESULTS

A total of 1786 (17.6%) cases were discharged to a SNF/rehab, and a multivariable logistic regression model demonstrated excellent predictive accuracy (area under the receiver operator characteristic curve: 0.824) despite requiring only 9 preoperative variables: age, partner status, the American Society of Anesthesiologists score, body mass index, gender, neurologic disease, electrolyte disorder, paralysis, and pulmonary circulation disorder. Notably, this model was independent of surgery (knee vs hip). Internal validation showed no loss of accuracy (area under the receiver operator characteristic curve: 0.8216, mean squared error: 0.0004) after bias correction for overfitting, and the model was incorporated into a readily available, online prediction tool for easy clinical use.

CONCLUSION

This convenient, interactive tool for estimating likelihood of discharge to a SNF/rehab achieves excellent accuracy using exclusively preoperative factors. These should form the basis for improved reimbursement legislation adjusting for patient risk, ensuring no disparities in access arise for vulnerable populations.

LEVEL OF EVIDENCE

III.

Preoperative Frailty Assessment, Operative Severity Score, and Early Postoperative Loss of Independence in Surgical Patients Age 65 Years or Older

BACKGROUND

Preoperative discussions around postoperative discharge planning have been amplified by the COVID pandemic. We wished to determine whether our preoperative frailty screen would predict postoperative loss of independence (LOI).

STUDY DESIGN

This single-institutional study included demographic, procedural, and outcomes data from patients 65 years or older who underwent frailty screening before a surgical procedure. Frailty was assessed using the Edmonton Frail Scale. The Operative Severity Score was used to categorize procedures. The Hierarchical Condition Category risk-adjustment score, as calculated by the Centers for Medicare and Medicaid Services, was included. LOI was defined as an increase in support outside of the home after discharge. Univariable, multivariable logistic regressions, and adjusted postestimation analyses for predictive probabilities of best fit were performed.

RESULTS

Five hundred and thirty-five patients met inclusion criteria and LOI was seen in 38 patients (7%). Patients with LOI were older, had a lower BMI, a higher Edmonton Frail Scale score (7 vs 3.0; p < 0.001), and a higher Hierarchical Condition Category score than patients without LOI. Being frail and undergoing a procedure with an Operative Severity Score of 3 or higher was independently associated with an increased risk of LOI. In addition, social dependency, depression, and limited mobility were associated with an increased risk for LOI. On multivariable modeling, frailty status, undergoing an operation with an Operative Severity Score of 3 or higher, and having a Hierarchical Condition Category score ≥1 were the most predictive of LOI (odds ratio 12.72; 95% CI, 12.04 to 13.44; p < 0.001). In addition, self-reported depression, weight loss, and limited mobility were associated with a nearly 11-fold increased risk of postoperative LOI.

CONCLUSIONS

This study was novel, as it identified clear, generalizable risk factors for LOI. In addition, our findings support the implementation of preoperative assessments to aid in care coordination and provide specific targets for intervention.

Improving evidence-based grouping of transitional care strategies in hospital implementation using statistical tools and expert review

BACKGROUND

As health systems transition to value-based care, improving transitional care (TC) remains a priority. Hospitals implementing evidence-based TC models often adapt them to local contexts. However, limited research has evaluated which groups of TC strategies, or transitional care activities, commonly implemented by hospitals correspond with improved patient outcomes. In order to identify TC strategy groups for evaluation, we applied a data-driven approach informed by literature review and expert opinion.

METHODS

Based on a review of evidence-based TC models and the literature, focus groups with patients and family caregivers identifying what matters most to them during care transitions, and expert review, the Project ACHIEVE team identified 22 TC strategies to evaluate. Patient exposure to TC strategies was measured through a hospital survey (N = 42) and prospective survey of patients discharged from those hospitals (N = 8080). To define groups of TC strategies for evaluation, we performed a multistep process including: using ACHIEVE'S prior retrospective analysis; performing exploratory factor analysis, latent class analysis, and finite mixture model analysis on hospital and patient survey data; and confirming results through expert review. Machine learning (e.g., random forest) was performed using patient claims data to explore the predictive influence of individual strategies, strategy groups, and key covariates on 30-day hospital readmissions.

RESULTS

The methodological approach identified five groups of TC strategies that were commonly delivered as a bundle by hospitals: 1) Patient Communication and Care Management, 2) Hospital-Based Trust, Plain Language, and Coordination, 3) Home-Based Trust, Plain language, and Coordination, 4) Patient/Family Caregiver Assessment and Information Exchange Among Providers, and 5) Assessment and Teach Back. Each TC strategy group comprises three to six, non-mutually exclusive TC strategies (i.e., some strategies are in multiple TC strategy groups). Results from random forest analyses revealed that TC strategies patients reported receiving were more important in predicting readmissions than TC strategies that hospitals reported delivering, and that other key co-variates, such as patient comorbidities, were the most important variables.

CONCLUSION

Sophisticated statistical tools can help identify underlying patterns of hospitals' TC efforts. Using such tools, this study identified five groups of TC strategies that have potential to improve patient outcomes.

Discharge Disposition Following Hematopoietic Cell Transplantation: Predicting the Need for Rehabilitation and Association with Survival

Many hematopoietic cell transplantation (HCT) recipients require rehabilitation due to deconditioning following intensive conditioning regimens and immune reconstitution. HCT recipients are preferentially discharged to home to avoid the risk of exposure to healthcare-associated infection in a rehabilitation facility (RF), with a caregiver who has been provided specific education about the complexity of post-HCT care. This study was conducted to determine the incidence of discharge to an RF following HCT, identify pre-HCT and peri-HCT risk factors for discharge to an RF, and estimate the effect of discharge disposition on overall survival (OS). This retrospective, matched 1:4 case-control study included 56 cases over a 10-year period from a single institution. Controls were matched by transplantation type (autologous versus allogeneic) and date of transplantation. The incidence of discharge to an RF was 2.2%. Controlling for disease, increasing age (odds ratio [OR], 1.09; 95% confidence interval [CI], 1.04 to 1.15; P < .001), female sex (OR, 3.11; 95% CI, 1.32 to 7.32; P = .01), high-risk HCT Comorbidity Index (HCT-CI) score (≥3) (OR, 3.44; 95% CI, 1.39 to 8.52; P = .008), decreasing pre-HCT serum albumin (OR, 2.60; 95% CI, 1.07 to 6.38; P = .037), and development of acute kidney injury during HCT (OR, 4.10; 95% CI, 1.36 to 12.40; P = .012) were associated with discharge to an RF. Discharge to an RF was associated with worse OS and higher nonrelapse mortality (NRM) compared with discharge to home (1-year OS, 70.5% [95% CI, 55.8% to 81.1%] versus 88.8% [95% CI, 83.6% to 92.4%], P < .001; 100-day NRM: 9.5% [95% CI, 3.5% to 19.2%] versus 1.8% [95% CI, 0.6% to 4.3%]; P = .03). Discharge to an RF following HCT is a rare event but associated with poor OS. Modifiable risk factors for discharge to an RF, including serum albumin as a measure of nutrition and reversible HCT-CI components, should be prospectively studied to determine the effect of mitigation on discharge disposition and survival.

Association of Pain on Hospital Discharge with the Risk of 30-Day Readmission in Patients with Total Hip and Knee Replacement

BACKGROUND

It is not clear if there is a risk of 30-day readmissions following total hip and knee arthroplasty in patients reporting high levels of pain at hospital discharge. We examined the relationship between post-surgical pain on the day of discharge and 30-day readmission in patients who received total knee and hip arthroplasty.

METHODS

Retrospective cohort study was conducted of patients who received total knee (n = 155,284) or hip arthroplasty (n = 89,283) from 2011 to 2018 using electronic health records from the Optum database. Four categories of pain at discharge were created, from none to severe. Multivariate logistic regression models to predict 30-day all-cause readmission were adjusted for patient and clinical characteristics and built separately for knee and hip arthroplasty patients.

RESULTS

Mean ages for hip and knee patients were 64.4 (standard deviation 11.3) and 65.7 (standard deviation 9.7) years, respectively. The majority of patients were female (hip: 54.4%; knee: 61.5%). The unadjusted rate of 30-day readmission was 3.54% for hip replacement and 3.66% for knee replacement. In models adjusted for patient and clinical characteristics, for patients with total hip replacement, the odds of 30-day readmission for those with severe pain score at discharge vs those with no pain at discharge were 1.60 (95% confidence interval 1.33-1.92). Similarly, readmission likelihood increased as pain at discharge increased (severe pain vs no pain) for patients with total knee arthroplasty (odds ratio 1.38, 95% confidence interval 1.19-1.59).

CONCLUSION

Our findings demonstrated that the pain scores on the day of discharge are associated with 30-day hospital readmission.

Observational study assessing changes in timing of readmissions around postdischarge day 30 associated with the introduction of the Hospital Readmissions Reduction Program

BACKGROUND

The Hospital Readmissions Reduction Program (HRRP) initially penalised hospitals for excess readmission within 30 days of discharge for acute myocardial infarction (AMI), congestive heart failure (CHF) or pneumonia (PNA) and was expanded in subsequent years to include readmissions for chronic obstructive pulmonary disease, elective total hip arthroplasty, total knee arthroplasty and coronary artery bypass graft surgery. We assessed whether HRRP was associated with delays in readmissions from immediately before the 30-day penalty threshold to just after it.

METHODS

We included Medicare fee-for-service beneficiaries discharged between 1 January 2007 and 31 October 2015. Readmissions were assessed until December 31, 2015. The study period was divided into three phases: January 2007 to March 2009 (pre-HRRP), April 2009 to September 2012 (implementation) and October 2012 to December 2015 (penalty). We estimated additional readmissions between postdischarge days 31-35 compared with days 26-30 using a negative binomial difference-in-differences model, comparing target HRRP versus non-HRRP conditions at the same hospital in the same month in the pre-HRRP and penalty phases.

RESULTS

HRRP was not associated with a significant difference in AMI readmissions between postdischarge days 31-35 versus postdischarge days 26-30 for each hospital in the penalty phase, as compared with non-HRRP conditions and the pre-HRRP phase (p=0.19). There were statistically significant increases in readmissions CHF (0.040%, 95% CI 0.024% to 0.056%, p<0.01), PNA (0.022%, 95% CI 0.002% to 0.042%, p=0.03) and stroke (0.035%, 95% CI 0.010% to 0.060%, p<0.01); however, these readmissions represent <0.01% of readmissions during this time period.

CONCLUSION

We did not identify consistently significant associations between HRRP and delayed readmissions, and importantly, any findings suggesting delayed readmissions were extremely small and unlikely to be clinically relevant.

Predicting Length of Stay and the Need for Postacute Care After Acute Myocardial Infarction to Improve Healthcare Efficiency

Background To improve value in the care of patients with acute myocardial infarction (MI), payment models increasingly hold providers accountable for costs. As such, providers need tools to predict length of stay (LOS) during hospitalization and the likelihood of needing postacute care facilities after discharge for acute MI patients. We developed models to estimate risk for prolonged LOS and postacute care for acute MI patients at time of hospital admission to facilitate coordinated care planning. Methods and Results We identified patients in the National Cardiovascular Data Registry ACTION registry (Acute Coronary Treatment and Intervention Outcomes Network) who were discharged alive after hospitalization for acute MI between July 1, 2008 and March 31, 2017. Within a 70% random sample (Training cohort) we developed hierarchical, proportional odds models to predict LOS and hierarchical logistic regression models to predict discharge to postacute care. Models were validated in the remaining 30%. Of 633 737 patients in the Training cohort, 16.8% had a prolonged LOS (≥7 days) and 7.8% were discharged to a postacute facility (extended care, a transitional care unit, or rehabilitation). Model discrimination was moderate in the validation dataset for predicting LOS (C statistic=0.640) and strong for predicting discharge to postacute care (C statistic=0.827). For both models, discrimination was similar in ST-segment-elevation MI and non-ST-segment-elevation MI subgroups and calibration was excellent. Conclusions These models developed in a national registry can be used at the time of initial hospitalization to predict LOS and discharge to postacute facilities. Prospective testing of these models is needed to establish how they can improve care coordination and lower costs.

A Weighted Index of Elixhauser Comorbidities for Predicting 90-day Readmission After Total Joint Arthroplasty

BACKGROUND

Evolving reimbursement models increasingly compel hospitals to assume costs for 90-day readmission after total joint arthroplasty. Although risk assessment tools exist, none currently reach the predictive performance required to accurately identify high-risk patients and modulate perioperative care accordingly. Although unlikely to perform adequately alone, the Elixhauser index is a set of 31 variables that may lend value in a broader model predicting 90-day readmission.

METHODS

Elixhauser comorbidities were examined in 10,022 primary unilateral total joint replacements, of which 4535 were hip replacements and 5487 were knee replacements, all performed between June 2013 and January 2018 at a single tertiary referral center. Data were extracted from electronic medical records using structured query language. After randomizing to derivation (80%) and validation (20%) subgroups, predictive models for 90-day readmission were generated and transformed into a system of weights based on each parameter's relative performance.

RESULTS

We observed 497 90-day readmissions (5.0%) during the study period, which demonstrated independent associations with 14 of the 31 Elixhauser comorbidity groups. A score created from the sum of each patient's weighted comorbidities did not lose substantial predictive discrimination (area under the curve: 0.653) compared to a comprehensive multivariable model containing all 31 unweighted Elixhauser parameters (area under the curve: 0.665). Readmission risk ranged from 3% for patients with a score of 0 to 27% for those with a score of 8 or higher.

CONCLUSIONS

The Elixhauser comorbidity score already meets or exceeds the predictive discrimination of available risk calculators. Although insufficient by itself, this score represents a valuable summary of patient comorbidities and merits inclusion in any broader model predicting 90-day readmission risk after total joint arthroplasty.

LEVEL OF EVIDENCE

III.

A Novel Risk Calculator Predicts 90-Day Readmission Following Total Joint Arthroplasty

BACKGROUND

A reliable prediction tool for 90-day adverse events not only would provide patients with valuable estimates of their individual risk perioperatively, but would also give health-care systems a method to enable them to anticipate and potentially mitigate postoperative complications. Predictive accuracy, however, has been challenging to achieve. We hypothesized that a broad range of patient and procedure characteristics could adequately predict 90-day readmission after total joint arthroplasty (TJA).

METHODS

The electronic medical records on 10,155 primary unilateral total hip (4,585, 45%) and knee (5,570, 55%) arthroplasties performed at a single institution from June 2013 to January 2018 were retrospectively reviewed. In addition to 90-day readmission status, >50 candidate predictor variables were extracted from these records with use of structured query language (SQL). These variables included a wide variety of preoperative demographic/social factors, intraoperative metrics, postoperative laboratory results, and the 30 standardized Elixhauser comorbidity variables. The patient cohort was randomly divided into derivation (80%) and validation (20%) cohorts, and backward stepwise elimination identified important factors for subsequent inclusion in a multivariable logistic regression model.

RESULTS

Overall, subsequent 90-day readmission was recorded for 503 cases (5.0%), and parameter selection identified 17 variables for inclusion in a multivariable logistic regression model on the basis of their predictive ability. These included 5 preoperative parameters (American Society of Anesthesiologists [ASA] score, age, operatively treated joint, insurance type, and smoking status), duration of surgery, 2 postoperative laboratory results (hemoglobin and blood-urea-nitrogen [BUN] level), and 9 Elixhauser comorbidities. The regression model demonstrated adequate predictive discrimination for 90-day readmission after TJA (area under the curve [AUC]: 0.7047) and was incorporated into static and dynamic nomograms for interactive visualization of patient risk in a clinical or administrative setting.

CONCLUSIONS

A novel risk calculator incorporating a broad range of patient factors adequately predicts the likelihood of 90-day readmission following TJA. Identifying at-risk patients will allow providers to anticipate adverse outcomes and modulate postoperative care accordingly prior to discharge.

LEVEL OF EVIDENCE

Prognostic Level IV. See Instructions for Authors for a complete description of levels of evidence.

Predictive Modeling for Blood Transfusion After Adult Spinal Deformity Surgery: A Tree-Based Machine Learning Approach

STUDY DESIGN

Retrospective cohort study.

OBJECTIVE

Blood transfusion is frequently necessary after adult spinal deformity (ASD) surgery. We sought to develop predictive models for blood transfusion after ASD surgery, utilizing both classification tree and random forest machine-learning approaches.

SUMMARY OF BACKGROUND DATA

Past models for transfusion risk among spine surgery patients are disadvantaged through use of single-institutional data, potentially limiting generalizability.

METHODS

This investigation was conducted utilizing the American College of Surgeons National Surgical Quality Improvement Program dataset years 2012 to 2015. Patients undergoing surgery for ASD were identified using primary-listed current procedural terminology codes. In total, 1029 patients were analyzed. The primary outcome measure was intra-/postoperative blood transfusion. Patients were divided into training (n = 824) and validation (n = 205) datasets. Single classification tree and random forest models were developed. Both models were tested on the validation dataset using area under the receiver operating characteristic curve (AUC), which was compared between models.

RESULTS

Overall, 46.5% (n = 479) of patients received a transfusion intraoperatively or within 72 hours postoperatively. The final classification tree model used operative duration, hematocrit, and weight, exhibiting AUC = 0.79 (95% confidence interval 0.73-0.85) on the validation set. The most influential variables in the random forest model were operative duration, surgical invasiveness, hematocrit, weight, and age. The random forest model exhibited AUC = 0.85 (95% confidence interval 0.80-0.90). The difference between the classification tree and random forest AUCs was nonsignificant at the validation cohort size of 205 patients (P = 0.1551).

CONCLUSION

This investigation produced tree-based machine-learning models of blood transfusion risk after ASD surgery. The random forest model offered very good predictive capability as measured by AUC. Our single classification tree model offered superior ease of implementation, but a lower AUC as compared to the random forest approach, although this difference was not statistically significant at the size of our validation cohort. Clinicians may choose to implement either of these models to predict blood transfusion among their patients. Furthermore, policy makers may use these models on a population-based level to assess predicted transfusion rates after ASD surgery.

LEVEL OF EVIDENCE

3.