Abstract 14590: Cardiac Rehabilitation Utilization After Percutaneous Coronary Intervention Remains Exceedingly Low and is Strongly Associated With the Treating Hospital and Operator: Insights From the Blue Cross Blue Shield of Michigan Cardiovascular Consortium

Introduction: Despite its proven benefits, cardiac rehabilitation (CR) use remains low. Identifying the sources of variation in CR referral and downstream use may help direct policies and quality improvement efforts. Objectives: We sought to quantify the magnitude of hospital and operator-level variation in CR referral and utilization after percutaneous coronary intervention (PCI). Methods: We used clinical registry data from patients (pts) discharged after PCI between 1/2012 and 10/2016 at 32 Michigan hospitals performing at least 100 PCIs linked to Blue Cross Blue Shield or Medicare claims. Registry and claims data were used to identify (i) CR referral prior to discharge and (ii) any CR use within 90 days after discharge. We calculated hospital-level risk and reliability adjusted rates of CR referral and utilization with Bayesian hierarchical regression models adjusted for patient characteristics and clustering by hospital and operator. Median odds ratios (MORs) with 95% credible intervals (CI) measured variation attributable to hospitals and operators. Results: Among 33,593 pts with a valid zip code and discharged home after PCI, 76.4% received an in-hospital referral for CR. Overall CR use was 26.7% and 31.3% among those who received an in-hospital referral. Adjusted rates of CR referral and use across hospitals ranged from 1.0%-100% and 14.9%-73.9%, respectively (Fig). After adjusting for pt factors, there was significant hospital and operator-level variation in CR referral with MORs of 5.91 (95% CI 4.02-9.34) and 1.71 (95% CI 1.60-1.85), respectively. There was also significant hospital and operator-level variation in CR use with MORs of 1.99 (95% CI 1.67 - 2.46) and 1.36 (95% CI 1.29 - 1.44), respectively. Conclusions: We found significant hospital and operator-level variation not only in post-PCI CR referral, but also CR utilization. Multi-faceted interventions targeting hospitals and PCI operators are needed to improve CR use after PCI.

Abstract 14449: Not So Harmless: Marijuana Use and In-hospital Outcomes After Percutaneous Coronary Intervention: Insights From the Blue Cross Blue Shield of Michigan Cardiovascular Consortium

Introduction: Marijuana use is increasing as more states are legalizing it for both recreational and medicinal uses. Although studies have suggested an association between marijuana use and cardiovascular risk, there are limited data on outcomes after percutaneous coronary intervention (PCI) among marijuana users. Hypothesis: We hypothesize that marijuana users are at an increased risk of adverse outcomes after PCI compared with non-users. Methods: Between 1/1/2013 and 10/1/2016, self-reported marijuana use was collected by the Blue Cross Blue Shield Michigan Cardiovascular Consortium (BMC2), a state-wide registry of all patients who underwent PCI at 48 non-federal hospitals in Michigan. In-hospital outcomes included post-procedural stroke, transfusion, bleeding, acute kidney injury (AKI), and death. We used 1:1 propensity matching and multivariable logistic regression techniques to adjust for differences in baseline characteristics between marijuana users and non-users. Results: Among 113,477 patients, 3,970 (3.5%) self-identified as marijuana users. Compared with non-users, marijuana users were more likely to be younger (53.9 vs. 65.8), male (79.2% vs 66.8%), cigarette smokers (73.0% vs. 26.8%), present with ST-elevation myocardial infarction (27.3% vs. 15.9%), and have fewer cardiovascular comorbidities. After matching (n=3,803 per group), compared with non-users, marijuana users had significantly increased risks of bleeding (5.2% vs. 3.4%; aOR 1.54 [1.20-1.97], p<0.001) and stroke (0.3% vs 0.1%; aOR 11.01 [1.32-91.67]; p=0.026), and a decreased risk of AKI (2.2% vs 2.9%; aOR 0.61 [0.42-0.87]; p = 0.007). There were no differences in the risks of transfusion or death. Conclusions: Marijuana use was associated with significantly increased risks of stroke and bleeding after PCI. As marijuana use continues to grow, clinicians and patients should be aware of the increased risks of post-PCI complications in these patients.