Search

Background: Although rapid response teams have been widely promoted as a strategy to reduce unexpected hospital deaths, most studies of rapid response teams have not adjusted for secular trends in mortality before their implementation. We examined whether implementation of a rapid response team was associated with a reduction in hospital mortality after accounting for preimplementation mortality trends. Methods: Among 56 hospitals in Get With The Guidelines-Resuscitation linked to Medicare, we calculated the annual rates of case mix–adjusted mortality for each hospital during 2000 to 2014. We constructed a hierarchical log-binomial regression model of mortality over time (calendar-year), incorporating terms to capture the effect of rapid response teams, to determine whether implementation of rapid response teams was associated with reduction in hospital mortality that was larger than expected based on preimplementation trends, while adjusting for hospital case mix index. Results: The median annual number of Medicare admissions was 5214 (range, 408–18 398). The median duration of preimplementation and postimplementation period was 7.6 years (≈2.5 million admissions) and 7.2 years (≈2.6 million admissions), respectively. Hospital mortality was decreasing by 2.7% annually during the preimplementation period. Implementation of rapid response teams was not associated with a change in mortality during the initial year (relative risk for model intercept, 0.98 [95% CI, 0.94–1.02]; P =0.30) or in the mortality trend (relative risk for model slope, 1.01 per year [95% CI, 0.99–1.02]; P =0.30). Among individual hospitals, implementation of a rapid response team was associated with a lower-than-expected mortality at only 4 (7.1%) and higher-than-expected mortality at 2 (3.7%) hospitals. Conclusions: Among a large and diverse sample of US hospitals, we did not find implementation of rapid response teams to be associated with reduction in hospital mortality. Studies are needed to understand best practices for rapid response team implementation, to ensure that hospital investment in these teams improves patient outcomes.

Background: Direct oral anticoagulants (DOACs) are superior to warfarin in preventing ischemic stroke from nonvalvular atrial fibrillation (AF). Transitioning from a DOAC to warfarin is rarely medically indicated. The frequency and determinants of switching from DOAC to warfarin, including the potential influence of cost, are not known. Objective: Examine the impact of financial assistance available through Medicaid prescription coverage or low-income subsidy programs on likelihood of switching to warfarin after initiating a DOAC. Methods: We identified patients in the Medicare 5% sample who initiated a DOAC from 2015 to 2017. Patients were eligible if they were age 65 or older, had AF diagnosis, had no oral anticoagulant use during the prior 12 months, and were enrolled in a Medicare Prescription Plan for at least 12 months prior to first oral anticoagulant. Patients who switched to warfarin after an initial DOAC were identified, and their characteristics were evaluated in bivariable and multivariable analyses using Cox regression to adjust for potential confounders and censor for death or end of the observation period. Results: Of 27,151 eligible patients,16,599 (61%), 8,930 (33%), and 1,622 (6%) initiated apixaban, rivaroxaban, and dabigatran, respectively. 30% (8,110) initiated a low dose DOAC. A total of 1,481 (5.5%) switched to warfarin sometime during the follow-up period. The median time to first warfarin dose among patients who switched was 153 days. Overall, 14% of patients were eligible for Medicaid prescription coverage and an additional 8.8% were eligible for other drug subsidies; these patients were significantly less likely to switch to warfarin, compared to other patients (3.4%, 4.1%, 5.8% of patients with Medicaid coverage, other drug subsidies, or neither). After risk adjustment, the relative hazard of switching to warfarin was 0.57 (95% confidence interval (CI): 0.47-0.68) and 0.69 (95% CI: 0.56-0.85), for patients with Medicaid coverage or drug subsidy, respectively, compared to other patients. Conclusions: The financial burden of DOACs may contribute to patients switching to warfarin, despite guideline recommendations. Financial assistance may facilitate evidenced-based best practices in stroke prevention.

Background: Rapid response teams (RRT) have been promoted as a strategy to reduce unexpected hospital deaths, as they are designed to evaluate and treat patients experiencing sudden decline. However, evidence to support their effectiveness in reducing in-hospital mortality remains uncertain. Methods: Using data from 56 hospitals participating in Get With The Guidelines Resuscitation linked to Medicare, we calculated annual rates of case-mix adjusted mortality for each hospital during 2000-2014. We constructed a hierarchical interrupted time series model to determine whether implementation of a RRT was associated with a reduction in mortality that was larger than expected based on pre-implementation trends alone. Results: Over the study period, the median annual number of Medicare admissions across study hospitals was 5214 (range: 408-18,398). The median duration of the pre-implementation period was 7.6 years comprising ~2.5 million admissions, and the median duration of the post-implementation period was 7.2 years comprising ~2.6 million admissions. Before implementation of RRTs, hospital mortality was already decreasing by 2.7% annually (Figure). Implementation of RRTs was not associated with change in mortality in the initial year of implementation (RR for model intercept: 0.98; 95% CI 0.94-1.02; P= 0.30) or in the mortality trend over time (RR for model slope: 1.01 per-year; 95% CI 0.99-1.02; P =0.30). Within individual hospitals, a RRT was associated with a significantly lower than expected mortality at 4 (7.1%) of hospitals, and significantly higher than expected mortality at 2 (3.6%), when compared to pre-implementation trends. Conclusion: Among a diverse sample of U.S. hospitals, we found that the implementation of a RRT was not associated with a significant reduction in hospital mortality. Given their prevalence in most U.S. hospitals, further studies are needed to understand best practices in composition, design, and implementation of RRTs.

Background: Socioeconomic (SE) disadvantage is a strong determinant of adverse outcomes in patients with HF. The contribution of SE disparities to adverse outcomes in HF may differ among Black vs. White patients and has not been well studied. Methods: Using the 100% CMS MedPAR data, Black and White patients hospitalized with HF between 2014 and 2017 were identified and stratified based on the distressed community index (DCI)—a measure of the SE disadvantage of residential ZIP codes on a continuous scale (range 0-100, see Fig. legend)—into two groups: SE distressed (Q5) vs. non-distressed (Q1-4). The rates of 30-day and 1-year mortality and readmission were compared across the distressed vs. non-distressed race groups. The adjusted association between DCI and risk of adverse outcomes was assessed separately across the race groups using adjusted hierarchical logistic regression models with restricted cubic splines. Results: The study included 1,238,537 White (14.8% distressed) and 190,721 Black (44.4% distressed) patients. White patients living in SE distressed communities had a significantly higher risk of adverse outcomes at 30-days and 1-year f/u (Fig. A). In contrast, among Black patients, the risk of adverse outcomes among those living in distressed vs. non-distressed communities were not meaningfully different at 30-days and became more prominent by 1-year f/u. Similar results were noted in the restricted cubic spline analysis with stronger and more graded association between DCI score and risk of adverse outcomes in White (vs. Black) patients (Fig. B). Conclusion: SE distress is strongly associated with risk of adverse outcomes in White patients with HF. Among Black patients, SE distress is more common, but its adverse effects are less evident during short-term f/u and are better highlighted in the long-term. Other societal factors such as structural racism and poor access to care may be important prognostic determinants in Black patients with HF.

We aimed to estimate antibiotic use during the last 6 months of life for hospitalized patients under hospice or palliative care and identify potential targets (i.e. time points) for antibiotic stewardship during the end-of-life period.We conducted a retrospective cohort study of nationwide Veterans Affairs (VA) patients who died between January 1, 2014 and December 31, 2019 and who had been hospitalized within 6 months prior to death. Data from the VA's integrated electronic medical record were collected, including demographics, comorbid conditions, and duration of inpatient antibiotics administered, along with outpatient antibiotics dispensed. A propensity score-matched cohort analysis was conducted to compare antibiotic use between hospitalized patients placed into palliative care or hospice matched to hospitalized patients not receiving palliative care or hospice.There were 9808 and 40 796 propensity score-matched patient pairs in the hospice and palliative care groups, respectively. Within 14 days of placement or consultation, 41% (4040/9808) of hospice patients and 48% (19 735/40 796) of palliative care patients received at least one antibiotic, while 25% (2420/9808) matched nonhospice and 27% (10 991/40 796) matched nonpalliative care patients received antibiotics. Entry into hospice was independently associated with a 12% absolute increase in antibiotic prescribing, and entry into palliative care was associated with a 17% absolute increase during the 14 days post-entry vs. pre-entry period.We observed that patients receiving end-of-life care had high levels of antibiotic exposure across this VA population, particularly during admissions when they received hospice or palliative care consultation.

Background Research has identified a group of hospitals with consistently poor performance for acute myocardial infarction (AMI) care as compared to other US hospitals. Factors influencing patients' selection of these hospitals are unknown. Methods We used 2004-06 Hospital Compare data to identify hospitals reporting AMI process performance measures. We calculated composite AMI performance scores and classified hospitals as low performing (lowest score decile for all 3 years, n = 100), top performing (top score decile for all 3 years, n = 64) and intermediate (all others, n = 2595). We used 2004-05 Medicare data to identify AMI patients living within 5 miles of low performing hospitals and alternative hospitals within 30 miles of patient residence. We used conditional multinomial logit models to evaluate the likelihood of choosing a low performing hospital over alternative hospitals. Models accounted for distance from patient residence to available hospitals, provision of revascularization (based on Medicare CABG volumes), and patient characteristics including age, race and residential social disadvantage (derived from US Census data). Results AMI patients admitted to low performing hospitals were older (mean age 80 vs. 79, p less likely to be admitted to low performing hospitals over alternative hospitals (HR 0.28, 95%CI 0.25 - 0.31, p < .01). However, patients living in areas with high disadvantage were relatively more likely to be admitted to low performing hospitals as compared to patients from intermediate and low disadvantage areas (HR 2.72 95% CI 2.24 - 3.30, p < .01), even after adjusting for age, race, and sex. Conclusion AMI patients are less likely to seek care at low performing hospitals except patients who live in zip code areas with high social disadvantage. Disincentives for poor performance against low performing hospitals could adversely impact care for this vulnerable population.