Search

This study demonstrated the feasibility of a statewide collaboration of acute care hospitals and cardiac rehabilitation facilities centered around the goal of equitably improving cardiac rehabilitation enrollment for all eligible patients in Michigan, one that is supported by a diverse, multidisciplinary stakeholder group and is driven by performance-benchmarking data.

The enrollment (28.6%), engagement (mean number of visits = 25), and program completion rates (27.6%) remain low among Medicare beneficiaries in the United States, indicating that many patients did not fully benefit from a class I guideline-recommended therapy. Additional research and successful enrollment and engagement initiatives are needed, especially among identified populations.

Supplemental Digital Content is available in the text.

Supplemental Digital Content is available in the text.

Background:Cardiac rehabilitation (CR) participation following percutaneous coronary intervention (PCI) and coronary artery bypass grafting (CABG) decreases patient morbidity and mortality. Many studies are underpowered to evaluate gender differences. We aimed to look at gender differences in CR use and short term outcomes including 1-year mortality in a national Medicare sample.Methods:We identified Medicare beneficiaries over 65 years undergoing PCI and CABG from 07/2016-12/2018. Primary exposures included female vs. male sex and CR use within one year of discharge based on outpatient claims. Clinical outcomes included 1-year mortality and readmissions. We analyzed differences in characteristics, CR use, and associated outcomes between females and males using multivariable and instrumental variable analyses using distance to CR facility as the instrument.Results:414,730 beneficiaries underwent PCI and CABG, of which 145,512 (35.1%) were female. Females were more likely to be older, Black race, dual eligible, live in distressed communities, and have a CR facility in their zip code, and were less likely to be Hispanic, elective procedures, or undergo CABG. CR use was lower for females than males (32.2% vs. 40.5%, p<0.001), which remained significant after adjustment for patient covariates (OR=0.91, 95% CI: 0.89-0.92, p<0.001). One-year mortality for CR users vs. non-CR users was 2.1% / 2.4% and 15.4% / 14.5% for women and men, respectively. One-year readmissions for CR users and non users was 34.3% / 31.2% and 52.8% / 45.0% for women and men, respectively. The adjusted and IV-adjusted average absolute effect of CR use on mortality was similar for females and males but was stronger on 1-year readmissions for females vs. males (Figure).Conclusion:Female Medicare beneficiaries attend CR following PCI and CABG at lower rates than men, but have increased benefit with CR participation on 1-year readmissions with similar mortality benefit relative to men.

Introduction:Participation in cardiac rehabilitation (CR) is poor, with disparities across patient demographic and clinical subgroups. Lower CR participation and completion rates may be associated with patients residing in socioeconomically distressed communities.Hypothesis:Patients undergoing surgical and percutaneous coronary revascularization living in distressed communities have lower rates of CR participation and completion.Methods:Medicare beneficiaries undergoing coronary revascularization (July 2016-December 2018) were identified using 100% Medicare claims data. The primary exposure was the quintile of distressed community index (DCI), with 1st and 5th quintiles representing the least and most distressed communities, respectively. CR participation and completion were defined as attending ≥1 session and attending 36 sessions within 1 year of discharge based on CPT codes in outpatient files. Multivariable logistic regression was used to evaluate the relationship between beneficiary DCI quintile and CR participation and completion.Results:Among 414,730 beneficiaries, 23.4% lived in the 1st DCI quintile and 16.4% lived in the 5th quintile. Patients in the 5th DCI quintile tended to be younger, non-white, dual eligible for Medicare/Medicaid, and more comorbid. Overall, 155,872 (37.6%) patients participated in ≥1 CR session, and 45,591 (29.3%) of those patients completed all 36 sessions. CR participation decreased as DCI quintile increased, with participation 12.0 percentage-points lower in the 5th vs 1st DCI quintile (Table). There were no significant differences in completion of 36 sessions across DCI quintiles.Conclusions:CR participation was significantly lower in more socioeconomically distressed communities, but there were no differences in completion rates. Strategies to reduce disparities in CR participation in distressed communities should focus on barriers to initial enrollment after qualifying events.

IntroductionPredicting personalized risk for adverse events following percutaneous coronary intervention (PCI) remains critical in weighing treatment options, employing risk-mitigation strategies, and enhancing shared decision-making.HypothesisWe hypothesized that by employing machine learning models, we would be able to use pre-procedural factors to predict common post-PCI complications.MethodsA patient panel ranked the importance of common PCI complications. A separate group of 66 adults underwent a semiquantitative survey assessing for a preferred list of outcomes and model display. Our cohort included 71,963 PCI procedures performed at 48 hospitals in Michigan between 4/1/2018 and 9/30/2020 in the BMC2 registry divided into two groups for training (75%) and validation (25%). XGBoost and random forest models were created to predict the risk of in-hospital outcomes.ResultsThe overall rate of mortality was 1.76% (n=1,264), AKI 2.64% (n=1,712), new need for dialysis 0.41% (n=295), major bleeding 0.89% (n=631), and transfusion 2.43% (n=1,746). The XGBoost model was selected as the final model and demonstrated excellent discrimination and calibration for all outcomes including mortality (AUC: 0.950 [95%CI 0.939-0.962]), AKI (AUC: 0.889 [95%CI 0.874-0.904]), dialysis (AUC: 0.949 [95%CI 0.928-0.969]), transfusion (AUC: 0.918 [95%CI 0.905-0.931]), and major bleeding (AUC: 0.892 [95%CI 0.861-0.922]). Survey subjects preferred a comprehensive list of individually reported post-procedure outcomes.ConclusionUsing common pre-procedural risk factors, we designed an XGBoost machine learning model that accurately predicts individualized post-PCI outcomes. Utilizing patient feedback, we created a patient-centered tool to clearly display risks to patients and providers (Figure, shiny.bmc2.org/pci-prediction/). Enhanced risk prediction prior to PCI could help inform treatment selection and shared decision-making discussions.

Introduction:Patients who present to the hospital with acute coronary syndromes (ACS) often have poor control of cardiac risk factors. Multiple studies have demonstrated that lipid lowering therapy is not appropriately initiated or intensified in these patients.Hypothesis:We hypothesized that an interruptive alert delivered via the electronic medical record would increase the initiation and intensification of lipid lowering therapies.MethodsUsing Epic EHR (Epic Systems, Verona, WI), an interruptive alert was designed to trigger upon opening the patient’s chart when the patients were admitted with an ICD code that indicated an ACS (i.e., STEMI, NSTEMI, or Unstable Angina). The alert included an order set to obtain a cholesterol panel or if a cholesterol panel had already been collected it would be displayed and suggest modifications to the patient’s current medication regimen. These alterations included statin intensification or the addition of ezetimibe.ResultsBetween September 1, 2021 and May 30, 2022 172 unique patients were identified. The mean age was 64.8 ±13.8 years, 64.5% were male, 74% were white, and 12% were black. The identified patients had multiple comorbidities including cerebrovascular disease (23%), heart failure (47%), peripheral vascular disease (30%), and being a current or former smoker (53%). The triggered order set had a direct effect on 42 (24%) patients, with 10 initiating ezetimibe and 32 having their statins intensified. The most common indicated reason for overriding the alert (n=99) was not meeting criteria. Of these 99 patients, 4 were ultimately started on ezetimibe, 57 had their statin intensified, and 23 were appropriately excluded (including patients with type II myocardial infarctions who were not the target of this intervention). Of note, 15 patients should have received intensified therapy but had their alert over ridden.ConclusionsThis targeted alert led to direct intensification of lipid lowering therapy in 24% of ACS patients with an additional 35% of patients identified by the alert also having therapeutic intensification. This alert will remain in place to allow for further assessment of its effects and can be easily translated to other clinical systems.

Background:Despite the known benefits for patients undergoing coronary revascularization, cardiac rehabilitation (CR) use remains low, with few studies evaluating geographic barriers to care. In this analysis, we examined geographic access (i.e. distance) to a CR facility and its relationship to CR participation and completion.Methods:A cohort of 414,730 Medicare beneficiaries with coronary revascularization were identified from 2016-2018. Geographic access was defined as the presence of a CR facility within the beneficiary’s home zip code (yes vs. no) and the distance to the nearest CR facility: within the same zip code (0 miles), 0-10 miles, 10-20 miles, 20+ miles. We then used claims data to determine any CR participation within one year of discharge and completion of the full CR program (36 sessions). Multivariable logistic regression was used to identify the relationships, adjusting for patient factors.Results:A total of 155,872 (37.6%) beneficiaries participated in CR during the study period, with 45,591 (29.3%) completing 36 sessions. Nearly a quarter of beneficiaries had a CR facility located within their residential zip code (23.7%), with the majority of beneficiaries having a CR facility within 10 miles of the center of their home zip code (Table 1).Beneficiaries without a CR facility in their zip code were more likely to be younger in age, men, non-white race categories, dual eligible for Medicare/Medicaid, and increasingly comorbid. Adjusting for patient factors, increasing distance category was associated with an absolute reduction in CR participation and completion (Table 1).Conclusions:Distance to the nearest CR facility is an important predictor of CR participation and completion, with overall use declining as distance increased. Interventions to address geographic barriers, such as transportation programs or virtual/hybrid CR, may be necessary to improve CR participation in areas with limited facility access.

Background: Percutaneous ventricular assist devices (PVADs) offer an important but resource-intensive option for management of severe cardiogenic shock (CS). Optimal selection of patients for PVAD support remains undefined. We investigated outcomes, including characteristics associated with in-hospital survival, during PVAD support for CS.Methods: We established a prospective quality improvement program among patients undergoing TandemHeart PVAD implantation for CS at Brigham and Women’s Hospital (Boston, MA). We evaluated 65 consecutive patients between 2006 and 2014, analyzing demographic, clinical, laboratory, hemodynamic, and survival data.Results: Thirty-two patients (49.2%) survived to hospital discharge, of which 12 received destination surgical therapy. Baseline characteristics associated with survival included younger age (47 ± 15 years vs 61 ± 11 years; p<0.001), non-ischemic cardiomyopathy (NICMP) vs ischemic CMP (survival 70.4% vs 34.2%, p=0.004), and, paradoxically, lower presenting left ventricular ejection fraction (LVEF) (survival 66.7% for LVEF ?15%, 41.2% for LVEF 16–25%, 25.0% for LVEF >25%; p=0.010). Younger age (p=0.026) and NICMP (p=0.034) remained independent predictors of survival. Twenty-four hours after PVAD placement, a more modest increase in cardiac index (?0.75 L/min/m2) was associated with higher in-hospital mortality (OR 6.3, 95% CI 1.8–22.1), as was lack of improvement in serum anion gap (?2 mEq/L; OR 5.1, 95% CI 1.6–16.6).Conclusions: Despite intensive care and provision of circulatory support, survival is poor in severe CS. Patients in CS with younger age and NICMP were more likely to survive to hospital discharge. Less robust hemodynamic improvement and persistent acidosis after 24 hours of PVAD support also identified patients less likely to survive.

Introduction:Cardiac rehabilitation (CR) reduces morbidity and mortality in patients with heart failure with reduced ejection fraction (HFrEF). In an effort to improve enrollment, Medicare expanded coverage for CR to include HFrEF patients in 2014, but the impact of this policy change is unknown.Methods:A 20% sample of Medicare inpatient, outpatient, and carrier claim files (2008 to 2017) were used to identify 849,054 patients with HFrEF, defined as having ≥ 2 outpatient claims or ≥ 1 inpatient hospitalization for chronic systolic heart failure during a calendar year. Patients who died within the calendar year of diagnosis were excluded. CR participation was identified using CPT and revenue center codes. We performed an interrupted time-series analysis to estimate annual rates of CR use over the ten year period, annual changes in CR use in the pre- (2008-2013) and post-policy (2014-2017), and the differences in annual trends pre- vs. post-policy. Results were risk adjusted by age, sex, race, dual eligibility, and Charlson comorbidities.Results:A total of 34,179 (4.0%) beneficiaries with HFrEF attended CR during the study period. Annual trends in CR enrollment are shown in Figure 1.From 2008 to 2014, CR utilization increased from 3.30% to 4.30% at an annual absolute change of 0.15% per year (p<0.001). After Medicare expanded coverage, rates of CR enrollment went from 4.30% in 2014 to 5.54% in 2017, at an annual absolute change of 0.31% (p<0.001). Comparing trends in participation before and after the change in policy, the annual rate of change in CR enrollment was 0.16% higher on an absolute scale (p<0.001).Conclusions:Although the absolute enrollment rate of patients with HFrEF in CR remains low, expanding coverage in 2014 was associated with a doubling in the annual rate of increase of CR use. Interventions beyond coverage changes, at the individual, provider, and system level, are needed to improve CR participation in HFrEF patients.

Introduction:Precise risk-prediction of post-procedural adverse events and outcomes after percutaneous coronary intervention (PCI) is critical to weighing treatment options and shared decision-making. Post-PCI complications are becoming less coming with quality improvement initiatives, however, there remains a significant risk for acute kidney injury (AKI), bleeding, and death. We hypothesized that with the use of random forest algorithms, we would be able to create accurate models to predict post-PCI complications.Methods:Our study cohort included 71,963 PCI procedures from the BMC2 registry performed at 48 hospitals in Michigan between 4/1/2018 and 9/30/2020. Random forest models were created using 74 pre-procedural clinical and laboratory variables to estimate the risk of each in-hospital outcome including mortality, AKI, major bleeding, and the need for a blood transfusion. A reduced model was created with the 20 most influential variables. The entire cohort was randomly separated into two groups for training (n=35,981) and validation (n=35,982). Model performance was evaluated in an independent validation data set using area under the receiver-operating characteristic curve (AUC).Results:The overall mortality was 1.76% (n = 1,264), there were 1,743 (2,42%) AKI events, 665 (0.92%) major bleeds, and 1,745 (2.42%) transfusions. The reduced models demonstrated excellent discrimination for all measured outcomes (Fig. 1) including mortality (AUC: 0.936 [95%CI 0.931-0.941]), AKI (AUC: 0.857 [95%CI 0.848-0.865]), major bleeding (AUC: 0.859 [95%CI 0.846-0.872]), and need for transfusion (AUC: 0.881 [95%CI 0.874-0.888]).Conclusions:The risk of post-PCI complications including mortality, AKI, major bleeding, and need for transfusion can be accurately calculated using a random forest algorithm of common pre-procedural factors to help better inform clinicians of a patient’s risk for better treatment selection and shared decision-making.

Introduction:The COVID-19 pandemic led to temporary restrictions on in-person cardiac rehabilitation (CR) services to mitigate high-risk exposure.Hypothesis:In order to better understand the impact of COVID-19 restrictions on CR services, we evaluated changes in the frequency and characteristics of CR visits in 2020 compared with 2019 in a commercially-insured population.Methods:We queried the Michigan Value Collaborative statewide claims data registry for outpatient CR visits between 01/01/2019 and 12/31/2020 using CPT codes (93797, 93798) among patients with diagnoses of ischemic heart disease, heart failure, percutaneous coronary intervention (PCI), coronary artery bypass grafting (CABG), or heart valve procedure. We described monthly trends in CR visits, and used bivariate analyses to compare changes in the demographics and medical diagnoses for CR visits in 2020 compared with 2019.Results:We identified a total of 45,553 CR visits in 2019 and 30,236 visits in 2020, representing a 33.6% relative decline. Monthly CR visits reached a nadir of 138 visits in April 2020 (96.5% reduction compared with April 2019), recovered to 3072 visits in September (-17.9%), and fell again to 2434 visits in December (-34.7%) (Figure). Relative changes in CR visits from 2019 to 2020 varied by diagnosis (p<0.001): -50.3% for heart failure (2931 to 1456, -37.5% for CABG (10121 to 6329), -35.3% for PCI (19669 to 12729), -28.9% for valve surgery (5621 to 3994), and -20.6% for ischemic heart disease (7211 to 5728). There were no significant differences in patient mean age (68.0 vs. 68.1 years, p=0.62) or gender (39.8% vs. 39.9% female, p=0.839) in 2019 compared with 2020.Conclusions:Total cardiac rehabilitation visits in 2020 were 33.6% lower compared with 2019, with heterogeneity in declines by underlying eligible diagnosis. Continued monitoring is needed to understand the public health impact of reduced CR use due to the COVID-19 pandemic.

Background:Black Americans comprise 14% of Michigan’s population, but 30% of the COVID-19 cases and 40% of deaths. The accumulating national quantitative data on decreased presentation and increased pre-hospital delay during the pandemic confirmed our Flint, MI community partners’ impression of a decline in stroke presentations. Thus, we set out to understand the community’s perception of seeking acute stroke care during the pandemic which will inform the development and dissemination of public health messaging in a predominantly Black American community.Method:To honor social distancing orders, we conducted semi-structured interviews based on the Theory of Planned Behavior via HIPAA-approved teleconferencing with community members. Due to the clinical and public health implications of the pandemic, we employed a rapid assessment approach to streamline qualitative data analysis. Results were used to inform the creation of a music video. Lyrics were written by the academic team, set to a soundtrack and sung by a community partner. The music video theme was conceptualized by the academic team and performed by both academic and community partners.Results:We reached thematic saturation after completing 15 semi-structured interviews with Flint, MI community members. Mean duration of interviews was 40 minutes. Eighty percent of participants were Black; median age was 50; 74% were women and 47% reported some college or above. There was an unfavorable attitude towards seeking emergent stroke care via ambulance and at the hospital, due to concerns for viral transmission at the hospital, hospital capacity and ability to triage, and quality of care. Community and academic partners co-created a music video with verses addressing the community-identified barriers: “stroke is an emergency all the time, even with COVID-19 / get to the hospital as soon as symptoms start, it’s so important to do your part, be Stroke Ready”(https://youtu.be/lKefAiUM2W0) The video reached over 1,200 users on our community-academic partner Stroke Ready Facebook page.Conclusion:We found that community members’ attitudes and perceived behavioral control to seek emergent stroke care were impacted during the COVID-19 pandemic. We addressed these barriers in an academic-community partner created music video. Academic and community partnerships facilitated a timely, innovative response to seeking acute stroke care in the setting of the COVID-19 pandemic.

Supplemental Digital Content is available in the text.

Background:Rates of transradial percutaneous coronary intervention (PCI) are increasing. There is concern that with declining rates of femoral access, operators and patient care teams may become less proficient with femoral access, thereby leading to increased transfemoral complications. We used the Blue Cross Blue Shield of Michigan Cardiovascular Consortium (BMC2) PCI registry to define the optimal volume threshold to maintain proficiency in transfemoral PCI.Methods:All PCIs from 2013-2017 from the BMC2 registry were included. To determine the impact of recent operator transfemoral experience on vascular complications, as of the date of each PCI procedure, we determined the total number of transfemoral procedures the operator had performed within the previous 365 days. Logistic regression models adjusting for patient and procedural characteristics were then fitted to the data to evaluate the impact on vascular complications of prior operator experience dichotomized at each integer threshold between 1st and the 75thpercentile of operator experience.Results:The study cohort comprised of 152,073 treated by 372 operators (97,540 (64.1%) femoral and 54,533 (35.9%) radial PCIs). During the study period transradial PCI increased from 25.9% to 45.2%. A nonlinear inverse relationship was noted between the total annual volume of transfemoral PCI volume and risk-adjusted vascular complication rate. A threshold volume of 30 annual transfemoral PCIs was identified as being associated with a statistically significant lowering of the risk for vascular complications (maximal effect at 38, Figure).Conclusions:Although transradial PCI has reduced vascular access site complications compared to transfemoral, operators need to maintain proficiency with both approaches. The number of transfemoral cases per year to maintain proficiency is 30. The optimal balance of transradial and transfemoral PCI to ensure the best patient outcome remains unknown.

Introduction:Cardiac rehabilitation (CR) confers significant clinical benefit after coronary artery disease (CAD) events, yet data on its use across and within hospitals is sparse.Hypothesis:Post-discharge CR use after CAD events varies across hospitals, but is correlated within hospitals by CAD event type: percutaneous coronary intervention (PCI), coronary artery bypass grafting (CABG) and acute myocardial infarction with medical management alone (AMI-MM).Methods:We identified 71,064 90-day episodes of CAD from 33 hospitals in the Michigan Value Collaborative multipayer registry (2014-2017), and compared CR use across CAD event type with a chi-square test: AMI-MM (n=17,779), PCI (n=42,832) and CABG (n=10,453). We defined CR use as having at least one facility or professional claim within 90 days of discharge using CPT (93797, 93798) and HCPCS (G0422, G0423) codes. We used hierarchical logistic regression to estimate hospital risk-adjusted CR rates, and the percent of variation attributed to the hospital-level (intraclass correlation coefficient, ICC). We also estimated the Pearson correlation coefficient (r) for within-hospital CR use across CAD event types.Results:Overall, 18,792 patients (26%) used any CR, with significant differences by CAD event type: CABG=52%, PCI=27%, AMI-MM=11% (p<0.001). We found over 10-fold variation in risk-adjusted CR use across hospitals (4% to 51%), with 12% of the variation attributed to the hospital (Figure). We observed similar variation in AMI-MM and PCI, but less variation in CABG. Within hospitals, CR use by event type was highly correlated, with the strongest correlation between AMI-MM vs. PCI (r=0.85), followed by PCI vs. CABG (r=0.79) and AMI-MM vs. CABG (r=0.70, all p<0.001).Conclusions:We found dramatic variation in CR use across hospitals and CAD event types. However, use within hospitals by CAD event type was similar. Work is ongoing to identify the determinants of CR use.

Supplemental Digital Content is available in the text.

ObjectiveTo evaluate the association between hospital penalization in the US Hospital Acquired Condition Reduction Program (HACRP) and subsequent changes in clinical outcomes.DesignRegression discontinuity design applied to a retrospective cohort from inpatient Medicare claims.Setting3238 acute care hospitals in the United States.ParticipantsMedicare fee-for-service beneficiaries discharged from acute care hospitals between 23 July 2014 and 30 November 2016 and eligible for at least one targeted hospital acquired condition (n=15 470 334).InterventionHospital receipt of a penalty in the first year of the HACRP.Main outcome measuresEpisode level count of targeted hospital acquired conditions per 1000 episodes, 30 day readmissions, and 30 day mortality.ResultsOf 724 hospitals penalized under the HACRP in fiscal year 2015, 708 were represented in the study. Mean counts of hospital acquired conditions were 2.72 per 1000 episodes for penalized hospitals and 2.06 per 1000 episodes for non-penalized hospitals; 30 day readmissions were 14.4% and 14.0%, respectively, and 30 day mortality was 9.0% for both hospital groups. Penalized hospitals were more likely to be large, teaching institutions, and have a greater share of patients with low socioeconomic status than non-penalized hospitals. HACRP penalties were associated with a non-significant change of −0.16 hospital acquired conditions per 1000 episodes (95% confidence interval −0.53 to 0.20), −0.36 percentage points in 30 day readmission (−1.06 to 0.33), and −0.04 percentage points in 30 day mortality (−0.59 to 0.52). No clear patterns of clinical improvement were observed across hospital characteristics.ConclusionsPenalization was not associated with significant changes in rates of hospital acquired conditions, 30 day readmission, or 30 day mortality, and does not appear to drive meaningful clinical improvements. By disproportionately penalizing hospitals caring for more disadvantaged patients, the HACRP could exacerbate inequities in care.

Supplemental Digital Content is available in the text.

Supplemental Digital Content is available in the text.

Supplemental Digital Content is available in the text.

Supplemental Digital Content is available in the text.

Supplemental Digital Content is available in the text.