Readmission Measures Research Articles

The Externalisation Gamble: Italy and Spain at the Forefront of Maritime Irregular Migration Governance

Irregular migration is rapidly becoming a permanent feature of the twenty-first century. Amid the European “refugee crisis”, Italy and Spain represent two major destination countries that are affected by substantial irregular flows. Despite this comparable condition, and notwithstanding the significance of their relevant policies, they have rarely been compared, a gap in the literature that this research addresses through a novel comparative analysis of their irregular migration governance. Following a broad contextualisation of destination countries’ strategies against irregular migration, this article delves into the two case studies’ external dimension of control policies from the 1990s to 2024 inclusive. In doing so, it assesses the nature of their foreign policy on irregular migration, their specific measures, and their geographical scope. It finds that there is a noticeable convergence towards readmission measures and externalisation, along with the growing use of informal deals. In terms of differences, the article highlights Italy’s much higher number of irregular arrivals compared to Spain, and the impact that this has had on domestic debates, priorities, and relevant policy formulation processes. As a novel investigation of two influential case studies, this article therefore contributes to the literature on both externalisation and Italy’s and Spain’s foreign policy.

Associations between readmission and patient-reported measures in acute psychiatric inpatients: a multicenter prospective longitudinal study.

This study examined whether patient-reported measures (PRMs) addressing quality of life, personal agency, functional impairment, and treatment satisfaction at hospital discharge were associated with future readmission during a 12-month follow-up period. The study also examined whether readmission influenced changes in the same measures. A multicenter prospective cohort study was conducted at 21 psychiatric hospitals in Japan. Participants completed the EuroQol-five-dimensions-five-level (EQ-5D), the Five-item Subjective Personal Agency Scale, and the Sheehan Disability Scale (SDS) at the time of index admission (T1), discharge from index admission (T2), and 6months (T3) and 12months (T4) after discharge. Inpatient treatment satisfaction was assessed at T2. Readmission and variables potentially associated with hospitalization and PRMs were evaluated using mixed-effects logistic regression models and mixed models for repeated measures. A total of 491 participants were followed for 12months (attrition rate: 19.4%), and 480 were included in the EQ-5D analysis. The most common diagnoses were schizophrenia (59%), depression (14%), and bipolar disorder (13%). No patient-reported measures were significantly associated with readmission over the follow-up period. Interaction of readmission and time did not significantly affect changes in EQ-5D. Readmission did significantly influence SDS score changes between T2 and T3 (B = 1.78, 95% CI = 0.30-3.25, p = 0.018) and between T3 and T4 (B = 1.43, 95% CI = 0.14-2.72, p = 0.029).The same influence of readmission on SDS score changes was not observed in the model which adjusted for all potential covariates. Readmission was potentially associated with changes in self-reported functional impairment. Findings highlight the potential role of intensive post-discharge services in preventing readmission, rather than relying on time-of-discharge PRMs in order to predict readmission risk. This study was registered in UMIN Clinical Trials Registry (UMIN000034220).

Assessing racial bias in healthcare predictive models: Practical lessons from an empirical evaluation of 30-day hospital readmission models

ObjectiveDespite increased availability of methodologies to identify algorithmic bias, the operationalization of bias evaluation for healthcare predictive models is still limited. Therefore, this study proposes a process for bias evaluation through an empirical assessment of common hospital readmission models. The process includes selecting bias measures, interpretation, determining disparity impact and potential mitigations. MethodsThis retrospective analysis evaluated racial bias of four common models predicting 30-day unplanned readmission (i.e., LACE Index, HOSPITAL Score, and the CMS readmission measure applied as is and retrained). The models were assessed using 2.4 million adult inpatient discharges in Maryland from 2016 to 2019. Fairness metrics that are model-agnostic, easy to compute, and interpretable were implemented and apprised to select the most appropriate bias measures. The impact of changing model’s risk thresholds on these measures was further assessed to guide the selection of optimal thresholds to control and mitigate bias. ResultsFour bias measures were selected for the predictive task: zero-one-loss difference, false negative rate (FNR) parity, false positive rate (FPR) parity, and generalized entropy index. Based on these measures, the HOSPITAL score and the retrained CMS measure demonstrated the lowest racial bias. White patients showed a higher FNR while Black patients resulted in a higher FPR and zero-one-loss. As the models’ risk threshold changed, trade-offs between models’ fairness and overall performance were observed, and the assessment showed all models’ default thresholds were reasonable for balancing accuracy and bias. ConclusionsThis study proposes an Applied Framework to Assess Fairness of Predictive Models (AFAFPM) and demonstrates the process using 30-day hospital readmission model as the example. It suggests the feasibility of applying algorithmic bias assessment to determine optimized risk thresholds so that predictive models can be used more equitably and accurately. It is evident that a combination of qualitative and quantitative methods and a multidisciplinary team are necessary to identify, understand and respond to algorithm bias in real-world healthcare settings. Users should also apply multiple bias measures to ensure a more comprehensive, tailored, and balanced view. The results of bias measures, however, must be interpreted with caution and consider the larger operational, clinical, and policy context.

Hospital readmissions among adults living with and without HIV in the US: findings from the Nationwide Readmissions Database

Thirty-day hospital readmission measures quality of care, but there are limited data among people with HIV (PWH) and people without HIV (PWoH) in the era of universal recommendation for antiretroviral therapy. We descriptively compared 30-day all-cause, unplanned readmission risk between PWH and PWoH. A retrospective cohort study was conducted using the 2019 Nationwide Readmissions Database (2019/01/01-2019/12/31), an all-payer database that represents all US hospitalizations. Index (initial) admissions and readmissions were determined using US Centers for Medicare & Medicaid Services definitions. Crude and age-adjusted risk ratios (aRR) comparing the 30-day all-cause, unplanned readmission risk between PWH to PWoH were estimated using random effect logistic regressions and predicted marginal estimates. Survey weights were applied to all analyses. We included 24,338,782 index admissions from 18,240,176 individuals. The median age was 52(IQR=40-60) years for PWH and 61(IQR=38-74) years for PWoH. The readmission risk was 20.9% for PWH and 12.2% for PWoH (age-adjusted-RR:1.88 [95%CI=1.84-1.92]). Stratified by age and sex, young female (age 18-29 and30-39 years) PWH had a higher readmission risk than young female PWoH (aRR=3.50 [95%CI=3.11-3.88] and aRR=4.00 [95%CI=3.67-4.32], respectively). While the readmission risk increased with age among PWoH, the readmission risk was persistently high across all age groups among PWH. The readmission risk exceeded 30% for PWH admitted for hypertensive heart disease, heart failure, and chronic kidney disease. PWH have a disproportionately higher risk of readmission than PWoH, which is concerning given the aging profile of PWH. More efforts are needed to address readmissions among PWH. US National Institutes of Health.

Incorporating Medicare Advantage Admissions Into the CMS Hospital-Wide Readmission Measure

Medicare Advantage (MA) enrollment is rapidly expanding, yet Centers for Medicare & Medicaid Services (CMS) claims-based hospital outcome measures, including readmission rates, have historically included only fee-for-service (FFS) beneficiaries. To assess the outcomes of incorporating MA data into the CMS claims-based FFS Hospital-Wide All-Cause Unplanned Readmission (HWR) measure. This cohort study assessed differences in 30-day unadjusted readmission rates and demographic and risk adjustment variables for MA vs FFS admissions. Inpatient FFS and MA administrative claims data were extracted from the Integrated Data Repository for all admissions for Medicare beneficiaries from July 1, 2018, to June 30, 2019. Measure reliability and risk-standardized readmission rates were calculated for the FFS and MA cohort vs the FFS-only cohort, overall and within specialty subgroups (cardiorespiratory, cardiovascular, medicine, surgery, neurology), then changes in hospital performance quintiles were assessed after adding MA admissions. Risk-standardized readmission rates. The cohort included 11 029 470 admissions (4 077 633 [37.0%] MA; 6 044 060 [54.8%] female; mean [SD] age, 77.7 [8.2] years). Unadjusted readmission rates were slightly higher for MA vs FFS admissions (15.7% vs 15.4%), yet comorbidities were generally lower among MA beneficiaries. Test-retest reliability for the FFS and MA cohort was higher than for the FFS-only cohort (0.78 vs 0.73) and signal-to-noise reliability increased in each specialty subgroup. Mean hospital risk-standardized readmission rates were similar for the FFS and MA cohort and FFS-only cohorts (15.5% vs 15.3%); this trend was consistent across the 5 specialty subgroups. After adding MA admissions to the FFS-only HWR measure, 1489 hospitals (33.1%) had their performance quintile ranking changed. As their proportion of MA admissions increased, more hospitals experienced a change in their performance quintile ranking (147 hospitals [16.3%] in the lowest quintile of percentage MA admissions; 408 [45.3%] in the highest). The combined cohort added 63 hospitals eligible for public reporting and more than 4 million admissions to the measure. In this cohort study, adding MA admissions to the HWR measure was associated with improved measure reliability and precision and enabled the inclusion of more hospitals and beneficiaries. After MA admissions were included, 1 in 3 hospitals had their performance quintile changed, with the greatest shifts among hospitals with a high percentage of MA admissions.

Measuring Equity in Readmission as a Distinct Assessment of Hospital Performance

Equity is an essential domain of health care quality. The Centers for Medicare & Medicaid Services (CMS) developed 2 Disparity Methods that together assess equity in clinical outcomes. To define a measure of equitable readmissions; identify hospitals with equitable readmissions by insurance (dual eligible vs non-dual eligible) or patient race (Black vs White); and compare hospitals with and without equitable readmissions by hospital characteristics and performance on accountability measures (quality, cost, and value). Cross-sectional study of US hospitals eligible for the CMS Hospital-Wide Readmission measure using Medicare data from July 2018 through June 2019. We created a definition of equitable readmissions using CMS Disparity Methods, which evaluate hospitals on 2 methods: outcomes for populations at risk for disparities (across-hospital method); and disparities in care within hospitals' patient populations (within-a-single-hospital method). Hospital patient demographics; hospital characteristics; and 3 measures of hospital performance-quality, cost, and value (quality relative to cost). Of 4638 hospitals, 74% served a sufficient number of dual-eligible patients, and 42% served a sufficient number of Black patients to apply CMS Disparity Methods by insurance and race. Of eligible hospitals, 17% had equitable readmission rates by insurance and 30% by race. Hospitals with equitable readmissions by insurance or race cared for a lower percentage of Black patients (insurance, 1.9% [IQR, 0.2%-8.8%] vs 3.3% [IQR, 0.7%-10.8%], P < .01; race, 7.6% [IQR, 3.2%-16.6%] vs 9.3% [IQR, 4.0%-19.0%], P = .01), and differed from nonequitable hospitals in multiple domains (teaching status, geography, size; P < .01). In examining equity by insurance, hospitals with low costs were more likely to have equitable readmissions (odds ratio, 1.57 [95% CI, 1.38-1.77), and there was no relationship between quality and value, and equity. In examining equity by race, hospitals with high overall quality were more likely to have equitable readmissions (odds ratio, 1.14 [95% CI, 1.03-1.26]), and there was no relationship between cost and value, and equity. A minority of hospitals achieved equitable readmissions. Notably, hospitals with equitable readmissions were characteristically different from those without. For example, hospitals with equitable readmissions served fewer Black patients, reinforcing the role of structural racism in hospital-level inequities. Implementation of an equitable readmission measure must consider unequal distribution of at-risk patients among hospitals.

Abstract 12884: Impact of an Addiction Consult Team on the Management and Outcomes of Patients With Surgically Managed Drug Use Associated Infective Endocarditis

Background: Recurrent injection drug use is a major driver of poor long term outcomes in patients who undergo surgery for drug use associated infective endocarditis (DUA-IE). The effect of an inpatient addiction consult team (ACT) on this population is not well described. Methods: To assess the impact of an ACT on DUA-IE patients, we performed a single center retrospective review of all adults who underwent surgery for active DUA-IE and survived to discharge (DC) between 2012-2022 at a large academic teaching institution. An ACT was established in 9/2017. We compared measures of substance use disorder (SUD) care, readmission, reinfection, and death pre- and post- ACT implementation using logistic regression models. Results: The pre- and post-ACT groups were similar at baseline (Table 1). In a univariate analysis, post-ACT patients were more likely to receive medications for opiate use disorder (MOUD) pre-operatively, to receive MOUD at DC, and to have MOUD follow up appointments scheduled at DC. They were less likely to receive non-MOUD opiates at DC or to be readmitted at 30 or 90 days. Overall rates of SUD follow up care scheduled prior to DC were low. There were no differences in harm reduction measures, recurrent IE, or death. In a multivariate analysis, the association of the ACT with decreased 90 day readmissions remained significant (Table 2). Conclusion: Among surgically managed DUA-IE patients, ACT implementation was associated with improved metrics of SUD care and readmissions at 90 days, but not recurrent IE or death. Future efforts should focus on strengthening harm reduction interventions and ensuring SUD follow up to sustain longitudinal engagement in care.

Hospital Performance Under Alternative Readmission Measures Incorporating Observation Stays.

To determine the extent to which counting observation stays changes hospital performance on 30-day readmission measures. This was a retrospective study of inpatient admissions and observation stays among fee-for-service Medicare enrollees in 2017. We generated 3 specifications of 30-day risk-standardized readmissions measures: the hospital-wide readmission (HWR) measure utilized by the Centers for Medicare and Medicaid Services, which captures inpatient readmissions within 30 days of inpatient discharge; an expanded HWR measure, which captures any unplanned hospitalization (inpatient admission or observation stay) within 30 days of inpatient discharge; an all-hospitalization readmission (AHR) measure, which captures any unplanned hospitalization following any hospital discharge (observation stays are included in both the numerator and denominator of the measure). Estimated excess readmissions for hospitals were compared across the 3 measures. High performers were defined as those with a lower-than-expected number of readmissions whereas low performers had higher-than-expected or excess readmissions. Multivariable logistic regression identified hospital characteristics associated with worse performance under the measures that included observation stays. Our sample had 2586 hospitals with 5,749,779 hospitalizations. Observation stays ranged from 0% to 41.7% of total hospitalizations. Mean (SD) readmission rates were 16.6% (5.4) for the HWR, 18.5% (5.7) for the expanded HWR, and 17.9% (5.7) in the all-hospitalization readmission measure. Approximately 1 in 7 hospitals (14.9%) would switch from being classified as a high performer to a low performer or vice-versa if observation stays were fully included in the calculation of readmission rates. Safety-net hospitals and those with a higher propensity to use observation would perform significantly worse. Fully incorporating observation stays in readmission measures would substantially change performance in value-based programs for safety-net hospitals and hospitals with high rates of observation stays.

Association of short-term hospital-level outcome metrics with 1-year mortality and recurrence for US Medicare beneficiaries with ischemic stroke.

Whether stroke patients treated at hospitals with better short-term outcome metrics have better long-term outcomes is unknown. We investigated whether treatment at US hospitals with better 30-day hospital-level stroke outcome metrics was associated with better 1-year outcomes, including reduced mortality and recurrent stroke, for patients after ischemic stroke. This cohort study included Medicare fee-for-service beneficiaries aged ≥65 years discharged alive from US hospitals with a principal diagnosis of ischemic stroke from 07/01/2015 to 12/31/2018. We categorized patients by the treating hospital's performance on the CMS hospital-specific 30-day risk-standardized all-cause mortality and readmission measures for ischemic stroke from 07/01/2012 to 06/30/2015: Low-Low (both CMS mortality and readmission rates for the hospital were <25th percentile of national rates), High-High (both >75th percentile), and Intermediate (all other hospitals). We balanced characteristics between hospital performance categories using stabilized inverse probability weights (IPW) based on patient demographic and clinical factors. We fit Cox models assessing patient risks of 1-year all-cause mortality and ischemic stroke recurrence across hospital performance categories, weighted by the IPW and accounting for competing risks. There were 595,929 stroke patients (mean age 78.9±8.8 years, 54.4% women) discharged from 2,563 hospitals (134 Low-Low, 2288 Intermediate, 141 High-High). For Low-Low, Intermediate, and High-High hospitals, respectively, 1-year mortality rates were 23.8% (95% confidence interval [CI] 23.3%-24.3%), 25.2% (25.1%-25.3%), and 26.5% (26.1%-26.9%), and recurrence rates were 8.0% (7.6%-8.3%), 7.9% (7.8%-8.0%), and 8.0% (7.7%-8.3%). Compared with patients treated at High-High hospitals, those treated at Low-Low and Intermediate hospitals, respectively, had 15% (hazard ratio 0.85; 95% CI 0.82-0.87) and 9% (0.91; 0.89-0.93) lower risks of 1-year mortality but no difference in recurrence. Ischemic stroke patients treated at hospitals with better CMS short-term outcome metrics had lower risks of post-discharge 1-year mortality, but similar recurrent stroke rates, compared with patients treated at other hospitals.

Annual Variation in 30-Day Risk-Adjusted Readmission Rates in U.S. Children's Hospitals.

Reducing pediatric readmissions has become a national priority; however, the use of readmission rates as a quality metric remains controversial. The goal of this study was to examine short-term stability and long-term changes in hospital readmission rates. Data from the Pediatric Health Information System were used to compare annual 30-day risk-adjusted readmission rates (RARRs) in 47 US children's hospitals from 2016 to 2017 (short-term) and 2016 to 2019 (long-term). Pearson correlation coefficients and weighted Cohen's Kappa statistics were used to measure correlation and agreement across years for hospital-level RARRs and performance quartiles. Median (IQR) 30-day RARRs remained stable from 7.7% (7.0-8.3) in 2016 to 7.6% (7.0-8.1) in 2019. Individual hospital RARRs in 2016 were strongly correlated with the same hospital's 2017 rate (R2=0.89 [95% confidence interval (CI) 0.80-0.94]) and moderately correlated with those in 2019 (R2=0.49 [95%CI 0.23-0.68]). Short-term RARRs (2016 vs 2017) were more highly correlated for medical conditions than surgical conditions, but correlations between long-term medical and surgical RARRs (2016 vs 2019) were similar. Agreement between RARRs was higher when comparing short-term changes (0.73 [95%CI 0.59-0.86]) than long-term changes (0.45 [95%CI 0.27-0.63]). From 2016 to 2019, RARRs increased by ≥1% in 7 (15%) hospitals and decreased by ≥1% in 6 (13%) hospitals. Only 7 (15%) hospitals experienced reductions in RARRs over the short and long-term. Hospital-level performance on RARRs remained stable with high agreement over the short-term suggesting stability of readmission measures. There was little evidence of sustained improvement in hospital-level performance over multiple years.

Reducing Postoperative Length of Stay for Idiopathic Scoliosis Patients using Quality Improvement Methodology.

The aim was to increase the percentage of idiopathic scoliosis patients with a procedure to discharge LOS of less than or equal to 4 days after posterior spinal fusion from 39.13% to 90%. Interventions included implementing a new pain management protocol, a daily checklist, education on expectations of postoperative pain, and updated order sets. Interventions improved patients discharged in less than 4 days from 39.13% to 93.48% (P ≤ 0.001), reducing the average postprocedure LOS from 4.93 to 2.59 (P ≤ 0.001) days. A key process measure tracked was the percentage of patients off the patient-control analgesia pump by postoperative day 2, which increased from 13% to 97.75% (P ≤ 0.001). These improvements did not affect the balancing measure of readmissions or Emergency Department visits for pain. By implementing a more standardized pathway, including a patient-focused daily checklist for providers and families, we established expectations for LOS and pain. This checklist and updates to the pain management protocol successfully reduced the length of stay in idiopathic scoliosis patients after posterior spinal fusion.

Achieving Equity in Hospital Performance Assessments Using Composite Race-Specific Measures of Risk-Standardized Readmission and Mortality Rates for Heart Failure.

The contemporary measures of hospital performance for heart failure hospitalization and 30-day risk-standardized readmission rate (RSRR) and risk-standardized mortality rate (RSMR) are estimated using the same risk adjustment model and overall event rate for all patients. Thus, these measures are mainly driven by the care quality and outcomes for the majority racial and ethnic group, and may not adequately represent the hospital performance for patients of Black and other races. Fee-for-service Medicare beneficiaries from January 2014 to December 2019 hospitalized with heart failure were identified. Hospital-level 30-day RSRR and RSMR were estimated using the traditional race-agnostic models and the race-specific approach. The composite race-specific performance metric was calculated as the average of the RSRR/RMSR measures derived separately for each race and ethnicity group. Correlation and concordance in hospital performance for all patients and patients of Black and other races were assessed using the composite race-specific and race-agnostic metrics. The study included 1 903 232 patients (75.7% White [n=1 439 958]; 14.5% Black [n=276 684]; and 9.8% other races [n=186 590]) with heart failure from 1860 hospitals. There was a modest correlation between hospital-level 30-day performance metrics for patients of White versus Black race (Pearson correlation coefficient: RSRR=0.42; RSMR=0.26). Compared with the race-agnostic RSRR and RSMR, composite race-specific metrics for all patients demonstrated stronger correlation with RSRR (correlation coefficient: 0.60 versus 0.74) and RSMR (correlation coefficient: 0.44 versus 0.51) for Black patients. Concordance in hospital performance for all patients and patients of Black race was also higher with race-specific (versus race-agnostic) metrics (RSRR=64% versus 53% concordantly high-performing; 61% versus 51% concordantly low-performing). Race-specific RSRR and RSMR metrics (versus race-agnostic) led to reclassification in performance ranking of 35.8% and 39.2% of hospitals, respectively, with better 30-day and 1-year outcomes for patients of all race groups at hospitals reclassified as high-performing. Among patients hospitalized with heart failure, race-specific 30-day RSMR and RSRR are more equitable in representing hospital performance for patients of Black and other races.

Effect of Medicare Advantage on Hospital Readmission and Mortality Rankings.

Medicare links hospital performance on readmissions and mortality to payment solely on the basis of outcomes among fee-for-service (FFS) beneficiaries. Whether including Medicare Advantage (MA) beneficiaries, who account for nearly half of all Medicare beneficiaries, in the evaluation of hospital performance affects rankings is unknown. To determine if the inclusion of MA beneficiaries in readmission and mortality measures reclassifies hospital performance rankings compared with current measures. Cross-sectional. Population-based. Hospitals participating in the Hospital Readmissions Reduction Program or Hospital Value-Based Purchasing Program. Using the 100% Medicare files for FFS and MA claims, the authors calculated 30-day risk-adjusted readmissions and mortality for acute myocardial infarction, heart failure, chronic obstructive pulmonary disease, and pneumonia on the basis of only FFS beneficiaries and then both FFS and MA beneficiaries. Hospitals were divided into quintiles of performance based on FFS beneficiaries only, and the proportion of hospitals that were reclassified to a different performance group with the inclusion of MA beneficiaries was calculated. Of the hospitals in the top-performing quintile for readmissions and mortality based on FFS beneficiaries, between 21.6% and 30.2% were reclassified to a lower-performing quintile with the inclusion of MA beneficiaries. Similar proportions of hospitals were reclassified from the bottom performance quintile to a higher one across all measures and conditions. Hospitals with a higher proportion of MA beneficiaries were more likely to improve in performance rankings. Hospital performance measurement and risk adjustment differed slightly from those used by Medicare. Approximately 1 in 4 top-performing hospitals is reclassified to a lower performance group when MA beneficiaries are included in the evaluation of hospital readmissions and mortality. These findings suggest that Medicare's current value-based programs provide an incomplete picture of hospital performance. Laura and John Arnold Foundation.

Pay-for-performance programs effectiveness in healthcare: the case of the end-stage renal disease quality incentive program.

This paper focuses on Medicare's End-Stage Renal Disease Quality Incentive Program (QIP). QIP aims to promote high-quality services in outpatient dialysis facilities by tying their payments to their performance on pre-specified quality measures. In this paper, employing principal-agent theory, we examine the effectiveness of QIP by exploring the changes in various clinical/operational measures when they become a part of the program as a performance measure. We study five QIP quality measures; two are operational: hospitalization and readmission. And three others are clinical: blood transfusion, hypercalcemia, and dialysis adequacy. Overall, we observe a significant improvement in all QIP quality measures after being included in the program, except for readmission. We recommend adjusting the weight and redesigning the readmission measure for Medicare to incentivize providers to reduce readmission. We also discuss establishing care coordination and employing data-driven clinical decision support systems as opportunities for dialysis facilities to improve the care delivery process.

Evaluating Algorithmic Bias in 30-Day Hospital Readmission Models: Retrospective Analysis.

The adoption of predictive algorithms in health care comes with the potential for algorithmic bias, which could exacerbate existing disparities. Fairness metrics have been proposed to measure algorithmic bias, but their application to real-world tasks is limited. This study aims to evaluate the algorithmic bias associated with the application of common 30-day hospital readmission models and assess the usefulness and interpretability of selected fairness metrics. We used 10.6 million adult inpatient discharges from Maryland and Florida from 2016 to 2019 in this retrospective study. Models predicting 30-day hospital readmissions were evaluated: LACE Index, modified HOSPITAL score, and modified Centers for Medicare & Medicaid Services (CMS) readmission measure, which were applied as-is (using existing coefficients) and retrained (recalibrated with 50% of the data). Predictive performances and bias measures were evaluated for all, between Black and White populations, and between low- and other-income groups. Bias measures included the parity of false negative rate (FNR), false positive rate (FPR), 0-1 loss, and generalized entropy index. Racial bias represented by FNR and FPR differences was stratified to explore shifts in algorithmic bias in different populations. The retrained CMS model demonstrated the best predictive performance (area under the curve: 0.74 in Maryland and 0.68-0.70 in Florida), and the modified HOSPITAL score demonstrated the best calibration (Brier score: 0.16-0.19 in Maryland and 0.19-0.21 in Florida). Calibration was better in White (compared to Black) populations and other-income (compared to low-income) groups, and the area under the curve was higher or similar in the Black (compared to White) populations. The retrained CMS and modified HOSPITAL score had the lowest racial and income bias in Maryland. In Florida, both of these models overall had the lowest income bias and the modified HOSPITAL score showed the lowest racial bias. In both states, the White and higher-income populations showed a higher FNR, while the Black and low-income populations resulted in a higher FPR and a higher 0-1 loss. When stratified by hospital and population composition, these models demonstrated heterogeneous algorithmic bias in different contexts and populations. Caution must be taken when interpreting fairness measures' face value. A higher FNR or FPR could potentially reflect missed opportunities or wasted resources, but these measures could also reflect health care use patterns and gaps in care. Simply relying on the statistical notions of bias could obscure or underplay the causes of health disparity. The imperfect health data, analytic frameworks, and the underlying health systems must be carefully considered. Fairness measures can serve as a useful routine assessment to detect disparate model performances but are insufficient to inform mechanisms or policy changes. However, such an assessment is an important first step toward data-driven improvement to address existing health disparities.

Associations Between Long-Term Air Pollutant Exposure and 30-Day All-Cause Hospital Readmissions in US Patients With Stroke.

Long-term exposure to air pollutants is associated with increased stroke incidence, morbidity, and mortality; however, research on the association of pollutant exposure with poststroke hospital readmissions is lacking. We assessed associations between average annual carbon monoxide (CO), nitrogen dioxide (NO2), ozone (O3), particulate matter 2.5, and sulfur dioxide (SO2) exposure and 30-day all-cause hospital readmission in US fee-for-service Medicare beneficiaries age ≥65 years hospitalized for ischemic stroke in 2014 to 2015. We fit Cox models to assess 30-day readmissions as a function of these pollutants, adjusted for patient and hospital characteristics and ambient temperature. Analyses were then stratified by treating hospital performance on the Centers for Medicare and Medicaid Services risk-standardized 30-day poststroke all-cause readmission measure to determine if the results were independent of performance: low (Centers for Medicare and Medicaid Services rate for hospital <25th percentile of national rate), high (>75th percentile), and intermediate (all others). Of 448 148 patients with stroke, 12.5% were readmitted within 30 days. Except for tropospheric NO2 (no national standard), average 2-year CO, O3, particulate matter 2.5, and SO2 values were below national limits. Each one SD increase in average annual CO, NO2, particulate matter 2.5, and SO2 exposure was associated with an adjusted 1.1% (95% CI, 0.4-1.9%), 3.6% (95% CI, 2.9%-4.4%), 1.2% (95% CI, 0.2%-2.3%), and 2.0% (95% CI, 1.1%-3.0%) increased risk of 30-day readmission, respectively, and O3 with a 0.7% (95% CI, 0.0%-1.5%) decrease. Associations between long-term air pollutant exposure and increased readmissions persisted across hospital performance categories. Long-term air pollutant exposure below national limits was associated with increased 30-day readmissions after stroke, regardless of hospital performance category. Whether air quality improvements lead to reductions in poststroke readmissions requires further research.

Assessing Hospital Quality Scores By Proportion Of Patients Dually Eligible For Medicare And Medicaid.

The Centers for Medicare and Medicaid Services has been reporting hospital star ratings since 2016. Some stakeholders have criticized the star ratings methodology for not adjusting for social risk factors. We examined the relationship between 2021 star rating scores and hospitals' proportion of Medicare patients dually eligible for Medicaid. We found that, on average, hospitals caring for a greater proportion of dually eligible patients had lower star ratings, but there was significant overlap in performance among hospitals when we stratified them by quintile of dually eligible patients. Hospitals in the highest quintile (those with the greatest proportion of dually eligible patients) had the best mean mortality scores (0.28) but the worst readmission (-0.44) and patient experience (-0.78) scores. We assigned star ratings after stratifying the readmission measure group by proportion of dually eligible patients and found that a total of 142 hospitals gained a star and 161 hospitals lost a star, of which 126 (89percent) and 1 (<1percent) were in the highest quintile, respectively. Adjusting public reporting tools such as star ratings for social risk factors is ultimately a policy decision, and views on the appropriateness of accounting for factors such as proportion of dually eligible patients are mixed, depending on the organization and stakeholder.

Abstract 15739: Utility of 30-Day Risk Adjusted Home Time as a Hospital Performance Metric for Transcatheter Edge-to-Edge Mitral Valve Repair

Introduction: While the utilization of mitral valve transcatheter edge-to-edge repair (TEER) procedures continues to increase over time, a validated, patient-centered, hospital-level performance metric for TEER is lacking Hypothesis: We examined the utility of 30-risk adjusted home time (HT) as a hospital-level performance metric for TEER procedure and compared it with the established volume, 30-day readmission, and 30-day mortality-based measures. Methods: Medicare patients who received TEER from 2013 to 2019 were identified. HT is defined as time spent alive and out of a hospital, skilled nursing facility, or long-term facility 30 days after TEER. Hospital-level risk-standardized measures of 30-day HT, readmission (RSRR), and mortality (RSMR) were developed using standard risk-adjustment models and correlation between these performance metrics was estimated by Pearson correlation. Longer-term patient-level outcomes of 1-year mortality and readmission were also compared across quartiles of hospitals on 30-day HT. Results: Overall, 24,420 patients who underwent TEER in 344 hospitals were included (mean age 80 y, 54% men, 7% Black race). At the hospital level, the median risk-adjusted 30-day HT was 24 days. There was significant inverse correlation between risk adjusted 30-day HT and 30-day RSRR (r=-0.43), 30-day RSMR (r=-0.5), and modest correlation with hospital’s annual TEER volume (r=0.11, P=0.05). The 1-year readmission and mortality among patients undergoing TEER decreased across increasing 30-day HT categories (Figure). 30-day HT (vs. annual TEER volume) meaningfully reclassified hospital performance in 34% of the hospitals with lower 1-year mortality and readmission among hospitals that were up classified (vs. no change) in performance ranking. Conclusions: Risk-adjusted 30-day HT is a patient-centered, comprehensive quality metric to assess hospital performance in TEER that correlates with important short- and long-term outcomes.

Ninety-Day Risk-Standardized Home Time as a Performance Metric for Cardiac Surgery Hospitals in the United States.

Assessing hospital performance for cardiac surgery necessitates consistent and valid care quality metrics. The association of hospital-level risk-standardized home time for cardiac surgeries with other performance metrics such as mortality rate, readmission rate, and annual surgical volume has not been evaluated previously. The study included Medicare beneficiaries who underwent isolated or concomitant coronary artery bypass graft, aortic valve, or mitral valve surgery from January 1, 2013, to October 1, 2019. Hospital-level performance metrics of annual surgical volume, 90-day risk-standardized mortality rate, 90-day risk-standardized readmission rate, and 90-day risk-standardized home time were estimated starting from the day of surgery using generalized linear mixed models with a random intercept for the hospital. Correlations between the performance metrics were assessed using the Pearson correlation coefficient. Patient-level clinical outcomes were also compared across hospital quartiles by 90-day risk-standardized home time. Last, the temporal stability of performance metrics for each hospital during the study years was also assessed. Overall, 919 698 patients (age 74.2±5.8 years, 32% women) were included from 1179 hospitals. Median 90-day risk-standardized home time was 71.2 days (25th-75th percentile, 66.5-75.6), 90-day risk-standardized readmission rate was 26.0% (19.5%-35.7%), and 90-day risk-standardized mortality rate was 6.0% (4.0%-8.8%). Across 90-day home time quartiles, a graded decline was observed in the rates of in-hospital, 90-day, and 1-year mortality, and 90-day and 1-year readmission. Ninety-day home time had a significant positive correlation with annual surgical volume (r=0.31; P<0.001) and inverse correlation with 90-day risk-standardized readmission rate (r=-0.40; P <0.001) and 90-day risk-standardized mortality rate (r=-0.60; P <0.001). Use of 90-day home time as a performance metric resulted in a meaningful reclassification in performance ranking of 22.8% hospitals compared with annual surgical volume, 11.6% compared with 90-day risk-standardized mortality rate, and 19.9% compared with 90-day risk-standardized readmission rate. Across the 7 years of the study period, 90-day home time demonstrated the most temporal stability of the hospital performance metrics. Ninety-day risk-standardized home time is a feasible, comprehensive, patient-centered metric to assess hospital-level performance in cardiac surgery with greater temporal stability than mortality and readmission measures.

Measuring health disparities using a continuous social risk factor.

To propose and evaluate a novel approach for measuring hospital-level disparities according to the effect of a continuous, polysocial risk factor on those outcomes. Our cohort consisted of Medicare Fee-for-Service (FFS) patients 65 years and older admitted to acute care hospitals for one of six common conditions or procedures. Medicare administrative claims data for six hospital readmission measures including hospitalizations from July 2015 to June 2018 were used. We adapted existing methodologies that were developed to report hospital-level disparities using dichotomous social risk factors (SRFs). The existing methods report disparities within and across hospitals; we developed and tested modified approaches for both methods using the Agency for Healthcare Research and Quality Socioeconomic Status Index. We applied the adapted methodologies to six 30-day hospital readmission measures included in the Centers for Medicare & Medicaid Services Hospital Readmissions Reduction Program measures. We compared the within- and across-hospital results for each to those obtained from using the original methods and dichotomizing the AHRQ SES Index into "low" and "high" scores. We used Medicare FFS administrative claims data linked to U.S. Census data. For all six readmission measures we find that, when compared with the existing methods, the methods for continuous SRFs provide disparity results for more facilities though across a narrower range of values. Measures of disparity based on this approach are moderately to highly correlated with those based on a dichotomous version of the same risk factor, while reflecting a fuller spectrum of risk. This approach represents an opportunity for detection of provider-level results that more closely align with underlying social risk. We have demonstrated the feasibility and utility of estimating hospital disparities of care using a continuous, polysocial risk factor. This approach expands the potential for reporting hospital-level disparities while better accounting for the multifactorial nature of social risk on hospital outcomes.