WB Weeks, Director, VISN 1 Patient Safety Center of Inquiry; Hub Site Senior Scholar, VA Quality Scholars Program; LB Ryder, White River Junction VAMC; DJ O'Rourke, White River Junction VAMC

Objectives: Veterans frequently have a choice of receiving health care through the VA or through a third party. We sought to understand veterans’ site-of-care preferences for coronary artery bypass grafting surgery (CABG), a service that is accessible locally through the private sector or regionally through the VA system. Additionally, we wanted to examine quality and cost issues with respect to site-of-care.

Methods: Between 7/1/97 – 4/1/2000, the charts of all veterans referred for CABG surgery were reviewed to determine whether their surgery occurred within or outside the VA system and whether the veteran had insurance coverage. Out-of-pocket cost estimates for local care were obtained and provided to a subset of veterans prior to their making a final decision of where they sought their care.

Comparisons between in system and out-of-system care were performed using t-test analysis. Seven core variables that were previously shown to be associated with inpatient mortality from CABG were used in a logistic regression model to determine whether site-of-care was related to outcomes.

Results: 143 patients were referred for CABG (94 within the VA system; 47 through the private sector). 116 patients had some type of insurance coverage, of which, 60% obtained their care within the VA system. Of those without insurance, 93% obtained care within the VA system (p=0.000). Out-of-pocket cost estimates for local care were much higher ($5,900, 95% CI $2450-$9300) for veterans who stayed within the VA system compared to those who obtained surgery through the private sector ($1,750, 95% CI $860-$2625, p=0.006). Patients who obtained surgery within the VA system had a nonsignificant longer length of stay (12 vs. 9 days, p=0.06). Patients who stayed within the VA system were slightly younger (67 vs. 71 years old, p=0.01), had a lower left ventricular ejection fraction (46% vs. 58%, p=0.000), had a greater number of coronary arteries with >70% stenosis (2.8 vs. 2.5, p=0.015), and were more likely to have significant (>50% stenosis) left main disease (38% vs. 20%, p=0.02). Gender, height, and weight were comparable across systems (p>0.3 for all). There were no differences in outcomes (frequency of returning to the operating room, stroke or infection rate, and in-hospital mortality) between the two systems (p=0.18 for return to the operating room, and p>0.59 for all others). In multivariate regression analysis that incorporated the seven core variables, site-of-care was not significantly associated with mortality or other complications.

Conclusions: A substantial proportion of veterans in a rural setting chose to obtain CABG outside of the VA system. Having insurance and low estimated out-of-pocket costs were associated with obtaining care outside of the VA system. Veterans who undergo CABG within the VA system have more severe coronary disease and longer lengths of stay, however, outcomes across systems are similar.

Impact: Site-of-care preferences may impact resource allocation. As the VA begins to understand veterans’ use of multiple systems of care, it will be important to evaluate access, quality, and out-of-pocket costs of collateral systems.