Hospital performance in a statewide commercial insurer episode-based incentive program.

OBJECTIVES: To evaluate hospital performance and behaviors in the first 2 years of a statewide commercial insurance episode-based incentive pay-for-performance (P4P) program. STUDY DESIGN: Retrospective cohort study of price- and risk-standardized episode-of-care spending from the Michigan Value Collaborative claims data registry. METHODS: Changes in hospital-level episode spending between baseline and performance years were estimated during the program years (PYs) 2018 and 2019. The distribution and hospital characteristics associated with P4P points earned were described for both PYs. A difference-in-differences (DID) analysis compared changes in patient-level episode spending associated with program implementation. RESULTS: Hospital-level episode spending for all conditions declined significantly from the baseline year to the performance year in PY 2018 (-$671; 95% CI, -$1113 to -$230) but was not significantly different for PY 2019 ($177; 95% CI, -$412 to $767). Hospitals earned a mean (SD) total of 6.3 (3.1) of 10 points in PY 2018 and 4.5 (2.9) of 10 points in PY 2019, with few significant differences in P4P points across hospital characteristics. The highest-scoring hospitals were more likely to have changes in case mix index and decreases in spending across the entire episode of care compared with the lowest-scoring hospitals. DID analysis revealed no significant changes in patient-level episode spending associated with program implementation. CONCLUSIONS: There was little evidence for overall reductions in spending associated with the program, but the performance of the hospitals that achieved greatest savings and incentives provides insights into the ongoing design of hospital P4P metrics.

Cardiac Rehabilitation Reduces 2-Year Mortality After Coronary Artery Bypass Grafting.

BACKGROUND: Cardiac rehabilitation (CR) is a supervised outpatient exercise and risk reduction program offered to patients who have undergone coronary revascularization procedures. Multiple professional societal guidelines support the use of CR after coronary artery bypass grafting (CABG) based on studies in combined percutaneous coronary intervention and CABG populations with surrogate outcomes. This statewide analysis of patients undergoing CABG evaluated the relationship between CR use and long-term mortality. METHODS: Medicare fee-for-service claims were linked to surgical data for patients discharged alive after isolated CABG from January 1, 2015, through September 30, 2019. Outpatient facility claims were used to identify any CR use within 1 year of discharge. Death within 2 years of discharge was the primary outcome. Mixed-effects logistic regression was used to predict CR use, adjusting for a variety of comorbidities. Unadjusted and inverse probability treatment weighting (IPTW) were used to compare 2-year mortality among CR users vs nonusers. RESULTS: A total of 3848 of 6412 patients (60.0%) were enrolled in CR for an average of 23.2 (SD, 12.0) sessions, with 770 of 6412 (12.0%) completing all recommended 36 sessions. Logistic regression identified increasing age, discharge to home (vs extended care facility), and shorter length of stay as predictors of postdischarge CR use (P < .05). Unadjusted and IPTW analyses showed significant reduction in 2-year mortality in CR users compared with CR nonusers (unadjusted: 9.4% reduction; 95% CI, 10.8%-7.9%; P < .001; IPTW: -4.8% reduction; 95% CI, 6.0%-3.5%; P < .001). CONCLUSIONS: These data suggest that CR use is associated with lower 2-year mortality. Future quality initiatives should consider identifying and addressing root causes of poor CR enrollment and completion.

Predictors and Variation in Cardiac Rehabilitation Participation After Transcatheter Aortic Valve Replacement.

Background: Cardiac rehabilitation (CR) is strongly recommended for a spectrum of cardiovascular conditions and procedures including aortic valve replacement. Objectives: The purpose of this study was to characterize patient and hospital factors associated with CR participation after transcatheter aortic valve replacement (TAVR) and determine which factors explain hospital-level variation in CR participation. Methods: We linked clinical and administrative claims data from patients who underwent TAVR at 24 Michigan hospitals between January 1, 2016 and June 30, 2020 and obtained rates of CR enrollment within 90 days of discharge. Sequential mixed models were fit to evaluate hospital variation in 90-day post-TAVR CR participation. Results: Among 3,372 patients, 30.6% participated in CR within 90-days after discharge. Several patient factors were negatively associated with CR participation after TAVR including older age, Medicaid insurance, atrial fibrillation/flutter, dialysis use, and slower baseline 5-m walk times. There was substantial hospital variation in CR participation after TAVR ranging from 5% to 60% across 24 hospitals. Patient case mix did not explain hospital variation in CR across hospitals with median OR numerically increasing from 2.11 (95% CI: 1.62-2.67) to 2.13 (95% CI: 1.61-2.68) after accounting for patient-level factors. Conclusions: Less than 1 in 3 patients who underwent TAVR in Michigan participated in CR within 90-days of discharge. Although several patient factors are associated with CR participation, hospital-level variation in CR participation after TAVR is not explained by patient case mix. Identifying hospital processes of care that promote CR participation after TAVR will be critical to improving CR participation after TAVR.

Evaluating the Feasibility of a Statewide Collaboration to Improve Cardiac Rehabilitation Participation: THE MICHIGAN CARDIAC REHAB NETWORK.

PURPOSE: Regional quality improvement collaboratives may provide one solution to improving cardiac rehabilitation (CR) participation through performance benchmarking and provider engagement. The objective of this study was to evaluate the feasibility of the Michigan Cardiac Rehab Network to improve CR participation. METHODS: Multipayer claims data from the Michigan Value Collaborative were used to identify hospitals and CR facilities and assemble a multidisciplinary advisory group. Univariate analyses described participating hospital characteristics and hospital-level rates of CR performance across eligible conditions including enrollment within 1 yr, mean days to first CR visit, and mean number of CR visits within 1 yr. Three diverse CR facilities were chosen for virtual site visits to identify areas of success and barriers to improvement. RESULTS: A total of 95 hospitals and 84 CR facilities were identified, with 48 hospitals (51%) providing interventional cardiology services and 33 (35%) providing cardiac surgical services. A 17-member multidisciplinary advisory group was assembled representing 13 institutions and diverse roles. Statewide CR enrollment across eligible admissions was 33.4%, with wide variation in CR performance measures across participating hospitals and eligible admissions. Virtual site visits revealed individual successes in improving CR participation but a variety of barriers to participation related to referrals, capacity and staffing constraints, and geographic and financial barriers. CONCLUSIONS: This study demonstrated the feasibility of creating a statewide collaboration of hospitals and CR facilities centered around the goal of equitably improving CR enrollment for all eligible patients in Michigan that is supported by a multidisciplinary advisory group and performance benchmarking.

Variation in Cardiac Rehabilitation Participation During Aortic Valve Replacement Episodes of Care.

BACKGROUND: Despite reported benefit in the setting of aortic valve replacement (AVR), cardiac rehabilitation (CR) utilization remains low, with few studies evaluating hospital and patient-level variation in CR participation. We explored determinants of CR variability during AVR episodes of care: transcatheter aortic valve replacement (TAVR) and surgical aortic valve replacement (SAVR). METHODS: A cohort of 10 124 AVR episodes of care (TAVR n=5121 from 24 hospitals; SAVR n=5003 from 32 hospitals) were identified from the Michigan Value Collaborative statewide multipayer registry (2015-2019). CR enrollment was defined as the presence of a single professional or facility claim within 90 days of discharge: 93 797, 93 798, G0422, G0423. Annual trends and hospital variation in CR were described for TAVR, SAVR, and all AVR. Multilevel logistic regression was used to estimate effects of predictors and hospital risk-adjusted rates of CR enrollment. RESULTS: Overall, 4027 (39.8%) patients enrolled in CR, with significant differences by treatment strategy: SAVR=50.9%, TAVR=28.9% (P<0.001). CR use after SAVR was significantly higher than after TAVR and increased over time for both modalities (P<0.001). There were significant differences in CR enrollment across age, gender, payer, and some comorbidities (P<0.05). At the hospital level, CR participation rates for all AVR varied 10-fold (4.8% to 68.7%) and were moderately correlated between SAVR and TAVR (Pearson r=0.56, P<0.01). CONCLUSIONS: Substantial variation exists in CR participation during AVR episodes of care across hospitals. However, within-hospital CR participation rates were significantly correlated across treatment strategies. These findings suggest that CR participation is the product of hospital-specific practice patterns. Identifying hospital practices associated with higher CR participation can help assist future quality improvement efforts to increase CR use after AVR.

Area Deprivation and Medicare Spending for Coronary Artery Bypass Grafting: Insights From Michigan.

BACKGROUND: Prior work has established that high socioeconomic deprivation is associated with worse short- and long-term outcomes for patients undergoing coronary artery bypass grafting (CABG). The relationship between socioeconomic status and 90-day episode spending is poorly understood. In this observational cohort analysis, we evaluated whether socioeconomically disadvantaged patients were associated with higher expenditures during 90-day episodes of care after isolated CABG. METHODS: We linked clinical registry data from 8728 isolated CABG procedures from January 1, 2012, to December 31, 2018, to Medicare fee-for-service claims data. Our primary exposure variable was patients in the top decile of the Area Deprivation Index. Linear regression was used to compare risk-adjusted, price-standardized 90-day episode spending for deprived against nondeprived patients as well as component spending categories: index hospitalization, professional services, post acute care, and readmissions. RESULTS: A total of 872 patients were categorized as being in the top decile. Mean 90-day episode spending for the 8728 patients in the sample was $55 258 (SD, $26 252). Socioeconomically deprived patients had higher overall 90-day spending compared with nondeprived patients ($61 579 vs $54 557; difference, $3003; P = .001). Spending was higher in socioeconomically deprived patients for index hospitalizations (difference, $1284; P = .005), professional services (difference, $379; P = .002), and readmissions (difference, $1188; P = .008). Inpatient rehabilitation was the only significant difference in post-acute care spending (difference, $469; P = .011). CONCLUSIONS: Medicare spending was higher for socioeconomically deprived CABG in Michigan, indicating systemic disparities over and above patient demographic factors.

Assessment of Patterns of Atherectomy Use.

Background Atherectomy has become the fastest growing catheter-based peripheral vascular intervention performed in the United States, and overuse has been linked to increased reimbursement, but the patterns of use have not been well characterized. Methods and Results We used Blue Cross Blue Shield of Michigan Preferred Provider Organization and Medicare fee-for-service professional claims data from the Michigan Value Collaborative for patients undergoing office-based laboratory atherectomy in 2019 to calculate provider-specific rates of atherectomy use, reimbursement, number of vessels treated, and number of atherectomies per patient. We also calculated the rate that each provider converted a new patient visit to an endovascular procedure within 90 days. Correlations between parameters were assessed with simple linear regression. Providers completing ≥20 office-based laboratory atherectomies and ≥20 new patient evaluations during the study period were included. A total of 59 providers performing 4060 office-based laboratory atherectomies were included. Median professional reimbursement per procedure was $4671.56 (interquartile range [IQR], $2403.09-$7723.19) from Blue Cross Blue Shield of Michigan and $14 854.49 (IQR, $9414.80-$18 816.33) from Medicare, whereas total professional reimbursement from both payers ranged from $2452 to $6 880 402 per year. Median 90-day conversion rate was 5.0% (IQR, 2.5%-10.0%), whereas the median provider-level average number of vessels treated per patient was 1.20 (IQR, 1.13-1.31) and the median provider-level average number of treatments per patient was 1.38 (IQR, 1.26-1.63). Total annual reimbursement for each provider was directly correlated with new patient-procedure conversion rate (R2=0.47; P<0.001), mean number of vessels treated per patient (R2=0.31; P<0.001), and mean number of treatments per patient (R2=0.33; P<0.001). Conclusions A minority of providers perform most procedures and are reimbursed substantially more per procedure compared with most providers. Procedural conversion rate, number of vessels, and number of treatments per patient represent potential policy levers to curb overuse.

High Socioeconomic Deprivation and Coronary Artery Bypass Grafting Outcomes: Insights From Michigan.

BACKGROUND: Cardiovascular outcomes are worse among individuals from areas with limited socioeconomic resources. This study evaluated the relationship between high socioeconomic deprivation and isolated coronary artery bypass grafting (CABG) outcomes. METHODS: We linked statewide Society of Thoracic Surgeons Adult Cardiac Surgery Database data to Medicare Fee-for-Service records for 10 423 Michigan residents undergoing isolated CABG between January 2012 and December 2018. High socioeconomic deprivation was defined as residing in the highest decile of the ZIP Code-level area deprivation index (ADI). Multivariable logistic regression estimated the relationship between top ADI decile and major morbidity, in-hospital mortality, and operative mortality. Survival analyses evaluated long-term survival comparing patients in the top vs not in the top ADI decile. RESULTS: A total of 1036 patients were in the top decile of ADI (ADI >82.4), and they were more likely to be female, Black, and have a higher predicted risk of mortality. Patients in the top ADI decile had significantly higher rates of major morbidity (17.4% vs 11.4%; adjusted odds ratio, 1.26; 95% CI, 1.04-1.54; P = .021) and in-hospital mortality (3.2% vs 1.3%, adjusted odds ratio, 1.84; 95% CI, 1.18-2.86, P = .007) but not operative mortality. The adjusted hazard of mortality was 16% higher for patients residing in the top ADI decile (95% CI, 1.01-1.33; P = .032). CONCLUSIONS: Isolated CABG patients residing in the highest areas of socioeconomic deprivation differed with respect to demographic and clinical characteristics and experienced worse short- and long-term outcomes compared with those not in the top ADI decile.

Hospital and Operator Variation in Cardiac Rehabilitation Referral and Participation After Percutaneous Coronary Intervention: Insights From Blue Cross Blue Shield of Michigan Cardiovascular Consortium.

BACKGROUND: Despite its established benefit and strong endorsement in international guidelines, cardiac rehabilitation (CR) use remains low. Identifying determinants of CR referral and use may help develop targeted policies and quality improvement efforts. We evaluated the variation in CR referral and use across percutaneous coronary intervention (PCI) hospitals and operators. METHODS: We performed a retrospective observational cohort study of all patients who underwent PCI at 48 nonfederal Michigan hospitals between January 1, 2012 and March 31, 2018 and who had their PCI clinical registry record linked to administrative claims data. The primary outcomes included in-hospital CR referral and CR participation, defined as at least one outpatient CR visit within 90 days of discharge. Bayesian hierarchical regression models were fit to evaluate the association between PCI hospital and operator with CR referral and use after adjusting for patient characteristics. RESULTS: Among 54 217 patients who underwent PCI, 76.3% received an in-hospital referral for CR, and 27.1% attended CR within 90 days after discharge. There was significant hospital and operator level variation in in-hospital CR referral with median odds ratios of 3.88 (95% credible interval [CI], 3.06-5.42) and 1.64 (95% CI, 1.55-1.75), respectively, and in CR participation with median odds ratios of 1.83 (95% CI, 1.63-2.15) and 1.40 (95% CI, 1.35-1.47), respectively. In-hospital CR referral was significantly associated with an increased likelihood of CR participation (adjusted odds ratio, 1.75 [95% CI, 1.52-2.01]), and this association varied by treating PCI hospital (odds ratio range, 0.92-3.75) and operator (odds ratio range, 1.26-2.82). CONCLUSIONS: In-hospital CR referral and 90-day CR use after PCI varied significantly by hospital and operator. The association of in-hospital CR referral with downstream CR use also varied across hospitals and less so across operators suggesting that specific hospitals and operators may more effectively translate CR referrals into downstream use. Understanding the factors that explain this variation will be critical to developing strategies to improve CR participation overall.

Percutaneous Coronary Intervention in Patients With a History of Gastrointestinal Bleeding (From the Blue Cross Blue Shield of Michigan Cardiovascular Consortium).

Potent antithrombotic agents are routinely prescribed after percutaneous coronary intervention (PCI) to reduce ischemic complications. However, in patients who are at an increased bleeding risk, this may pose significant risks. We sought to evaluate the association between a history of gastrointestinal bleeding (GIB) and outcomes after PCI. We linked clinical registry data from PCIs performed at 48 Michigan hospitals between 1/2013 and 3/2018 to Medicare claims. We used 1:5 propensity score matching to adjust for patient characteristics. In-hospital outcomes included bleeding, transfusion, stroke or death. Post-discharge outcomes included 90-day all-cause readmission and long-term mortality. Of 30,206 patients, 1.1% had a history of GIB. Patients with a history of GIB were more likely to be older, female, and have more cardiovascular comorbidities. After matching, those with a history of GIB (n = 312) had increased post-procedural transfusions (15.7% vs 8.4%; p < 0.001), bleeding (11.9% vs 5.2%; p < 0.001), and major bleeding (2.8% vs 0.6%; p = 0.004). Ninety-day readmission rates were similar among those with and without a history of GIB (34.3% vs 31.3%; p = 0.318). There was no significant difference in post-discharge survival (1 year: 78% vs 80%; p = 0.217; 5 years: 54% vs 51%; p = 0.189). In conclusion, after adjusting for baseline characteristics, patients with a history of GIB had increased risk of post-PCI in-hospital bleeding complications. However, a history of GIB was not significantly associated with 90-day readmission or long-term survival.

Determinants of Hospital Variation in Cardiac Rehabilitation Enrollment During Coronary Artery Disease Episodes of Care.

BACKGROUND: Cardiac rehabilitation (CR) is associated with improved outcomes for patients with coronary artery disease (CAD). However, CR enrollment remains low and there is a dearth of real-world data on hospital-level variation in CR enrollment. We sought to explore determinants of hospital variability in CR enrollment during CAD episodes of care: medical management of acute myocardial infarction (AMI-MM), percutaneous coronary intervention (PCI), and coronary artery bypass grafting (CABG). METHODS: A cohort of 71 703 CAD episodes of care were identified from 33 hospitals in the Michigan Value Collaborative statewide multipayer registry (2015 to 2018). CR enrollment was defined using professional and facility claims and compared across treatment strategies: AMI-MM (n=18 678), PCI (n=41 986), and CABG (n=11 039). Hierarchical logistic regression was used to estimate effects of predictors and hospital risk-adjusted rates of CR enrollment. RESULTS: Overall, 20 613 (28.8%) patients enrolled in CR, with significant differences by treatment strategy: AMI-MM=13.4%, PCI=29.0%, CABG=53.8% (P<0.001). There were significant differences in CR enrollment across age groups, comorbidity status, and payer status. At the hospital-level, there was over 5-fold variation in hospital risk-adjusted CR enrollment rates (9.8%-51.6%). Hospital-level CR enrollment rates were highly correlated across treatment strategy, with the strongest correlation between AMI-MM versus PCI (R2=0.72), followed by PCI versus CABG (R2=0.51) and AMI-MM versus CABG (R2=0.46, all P<0.001). CONCLUSIONS: Substantial variation exists in CR enrollment during CAD episodes of care across hospitals. However, within-hospital CR enrollment rates were significantly correlated across all treatment strategies. These findings suggest that CR enrollment during CAD episodes of care is the product of hospital-specific rather than treatment-specific practice patterns.