Author + information
- Stephen W. Waldo, MDa,b,∗ (, )@CARTProgram@StephenWaldoMD,
- Thomas J. Glorioso, MSb,
- Anna E. Barón, PhDb,c,
- Mary E. Plomondon, PhDb,
- Javier A. Valle, MD, MSca,b,
- Richard Schofield, MDd,e and
- P. Michael Ho, MD, PhDa,b
- aUniversity of Colorado School of Medicine, Aurora, Colorado
- bDepartment of Medicine, Rocky Mountain Regional VA Medical Center, Aurora, Colorado
- cDepartment of Biostatistics and Informatics, Colorado School of Public Health, University of Colorado Anschutz Medical Campus, Aurora, Colorado
- dDepartment of Medicine, Veterans Affairs Medical Center, Gainesville, Florida
- eUniversity of Florida College of Medicine, Gainesville, Florida
- ↵∗Address for correspondence:
Dr. Stephen W. Waldo, Rocky Mountain Regional VA Medical Center, 1700 North Wheeling Street, Aurora, Colorado 80045.
The Department of Veterans Affairs (VA) has established and maintained the largest integrated health care system in the United States. Previous research has shown that this system provides high-quality health care but has been limited by concerns regarding patient access (1,2). Because of this, the organization has expanded access to clinical care outside of the integrated health care system with the development of a community care program that enables veterans to receive care from nonfederal facilities, with the expenses covered by the federal government. These services are primarily offered if the VA integrated health care system cannot provide access to specialist care in a timely manner near the home of a veteran (3). Recent changes in policy will expand the use of health care in the community for veterans (4), despite reports that community providers may struggle to meet the need for increased capacity (5).
Concurrent with these policy changes, the cases of veterans afflicted with cardiovascular disease have increased in complexity over the last decade (6). Despite this, the clinical outcomes of patients undergoing percutaneous coronary intervention (PCI) have remained constant or improved over the same time. Similar data regarding the quality of cardiovascular care for veterans treated outside the integrated health care system is more challenging to obtain. With this in mind, the present analysis compared the clinical outcomes of veterans undergoing elective percutaneous revascularization at VA and community hospitals after the significant expansion of the community care program.
The present project identified patients actively enrolled in the VA Healthcare System who underwent elective percutaneous revascularization for stable angina between October 1, 2015, and December 31, 2018. A patient was considered actively enrolled if he or she had 2 or more visits to a primary care physician or cardiologist in the 2 years before the procedure and filled at least 1 medication in the prior year. Patients who underwent revascularization for emergent indications, such as ST-segment elevation myocardial infarction (International Classification of Diseases-10th Revision [ICD-10]: I21.01, I21.02, I21.09, I21.11, I21.19, I21.21, I21.29, I21.3, I22.0, I22.1, I22.2, I22.8, I22.9), non-ST elevation myocardial infarction (ICD-10: I21.4), or unstable angina (ICD-10: I20.0) were excluded. Patients who underwent a diagnostic procedure in the VA Healthcare System only to undergo revascularization in the community within 1 month were also removed to reduce the chances of a higher-risk staged procedure to be attributed to 1 treatment locale over another. Among the remaining population, we limited analyses to patients living in U.S. congressional districts where no more than 4 out of 5 (80%) patients received PCI in 1 location to ensure that each veteran had a reasonable option for care in either setting. To further reduce the impact of residual confounding, we also limited the analytic cohort to those who received the dominant care (>50%) in their respective districts. In most congressional districts, the minority of patients were treated in the community, and this reduced the chances of biasing the results against this treatment venue. The analysis was performed in an operational capacity for the Department of Veterans Affairs and was confirmed as an operational evaluation through review by the Colorado Multiple Institutional Review Board.
Information for those treated within the VA Healthcare System was obtained from the electronic medical record, and the Fee-Basis Claim System and the Program Integrity Tool were used for community data.
The temporal trend in the use of community facilities was calculated. All-cause mortality was ascertained between the 2 groups. All outcomes were summarized as time-to-event variables with follow-up capped at 1 year and censored as of December 31, 2018.
A logistic regression model estimated the probability of receiving PCI in the community, adjusted for demographic information (age, sex, race, ethnicity, insurance status, urban vs. rural, distance to nearest VA primary care, and U.S. Census division) and medical comorbidities (atrial fibrillation, alcohol abuse, heart failure, chronic kidney disease, chronic obstructive lung disease, cerebrovascular disease, depression, diabetes mellitus, family history of coronary artery disease, hypertension, hyperlipidemia, peripheral artery disease, post-traumatic stress disorder, and sleep apnea). Additional covariates representing prior PCI, myocardial infarction, or surgical coronary revascularization were also included, as was prior hospitalization within 90 days of the index procedure. The inverse probability of treatment weights were calculated, and Kaplan-Meier curves were created using the weighted cohort for the death outcome. A Cox proportional hazards model using the inverse probability of treatment weights was then constructed to compare the hazards for mortality based on the location of treatment. All analyses were performed with R (R Core Team, Vienna, Austria). A p value of <0.05 was considered statistically significant.
We identified 30,310 patients who underwent percutaneous intervention for stable angina. Further restriction to only those who received the dominant form of care in a given location led to a final analytic cohort of 8,913 patients, of whom 5,942 (67%) received PCI in the VA Healthcare System and 2,971 (33%) in the community facilities. The largest number of exclusions were a result of patients being treated in geographic regions where there was not an equal opportunity to be seen at a community or VA facility, making this excluded population difficult to compare. The use of community care has been increasing over time among veterans undergoing PCI, with the probability of an elective intervention being performed in the community increasing from 38.9% to 51.8% over the study period (Figure 1).
Patients treated at community hospitals had a higher rate of death compared to similar patients treated within the VA Healthcare System, with early separation of the Kaplan-Meier curves (Figure 1). A Cox proportional hazards model for death demonstrated a 33% increase (hazard ratio: 1.33; 95% confidence interval [CI]: 1.09 to 1.62) in hazard for mortality among patients treated in the community and an absolute risk difference of 1.4% (95% CI: 0.2 to 2.6). A censored analysis examining the same relationship over the first month demonstrated similar findings, with a 143% increase (HR: 2.43; 95% CI: 1.50 to 3.94) in the hazard for death during that abbreviated time period and a 0.7% absolute risk difference (95% CI: 0.3% to 1.2%).
Using a hypothetical confounder with a difference in prevalence similar to prior myocardial infarction (VA: 15%; community: 22%) and association with death similar to heart failure (HR: 2.64), the hazard for death in the community would still be 1.22 times higher but only marginally significant (95% CI: 1.01 to 1.49). This would suggest that an unmeasured confounder would need to be prevalent (approximately 20%), have a higher prevalence among community patients (approximately 7%), and have a significant differential effect (approximately 164%) on mortality between those treated at community care or VA facilities to negate the observed discrepancy in clinical outcomes.
The use of community care facilities among veterans undergoing elective percutaneous revascularization has significantly increased over time. Previous studies suggested that extended wait times may have impeded access to care within the VA Healthcare System, particularly among patients located in rural areas without a tertiary care medical facility (1). As a response to these access concerns, legislation expanded the opportunities for veterans to use community care facilities outside the integrated health care system. The present study suggests that use of community facilities has also increased among patients undergoing elective coronary revascularization, with the probability of a patient receiving percutaneous intervention in the community nearly doubling over the last 3 years. Recent legislation has further expanded opportunities for veterans to receive care in the community, and, thus, community use will continue to increase (4).
Treatment in the community is associated with worse clinical outcomes, with a 33% increased hazard of death in a propensity weighted sample. The discrepancy in clinical outcomes is even more pronounced when the analysis is limited to the first month after the procedure, with a 143% increased hazard of death. The early separation of the mortality curves may reflect differences in the safety of the procedure and quality of the periprocedural care. The continued divergence over time also suggests potential deficiencies in the coordination of post-procedural care and possible differences in the medical management of coronary artery disease. Although residual confounding could be present, it would take a factor with a sizeable difference in prevalence between treatment groups (approximately 7%) associated with a significant increase in risk (approximately 164%) to completely nullify these findings.
The current community care program does not restrict access to facilities based on procedural volumes or quality, because patients have the liberty to choose a facility based on their own personal preferences. Because of this, the community facilities used under this program may constitute a heterogeneous collection of hospitals ranging from academic quaternary care medical centers to rural safety net facilities with differing levels of expertise. The Clinical Assessment, Reporting, and Tracking Program monitors the quality and safety of invasive cardiac procedures for those treated within the VA Healthcare System, where clinical outcomes have remained constant or improved despite treatment of a more comorbid population (6). A similar organization to monitor the same quality of care for veterans treated in the community does not currently exist on a national scale. Because the VA Healthcare System serves as an unrecognized national safety net (7), we must ensure that increased access to medical care is accompanied by mechanisms to ensure similar levels of quality regardless of treatment venue.
The present analysis should be interpreted in the context of several limitations. Data were derived from clinical documentation for care provided within the VA Healthcare System and administrative billing data for care provided in community hospitals. The administrative billing data provide diagnosis codes as well as procedural codes but have limited additional clinical information. Because limited data were available on the procedural characteristics for patients treated in the community, it is possible that more anatomically complex procedures were performed in that setting. Furthermore, the community care data are not sufficiently granular to make a distinction between the type of community facility that provided the care at this time. It is important to note that these findings are specific to the unique care delivery system established with the VA and community partners and are not likely generalizable to other health care systems.
The use of community facilities for percutaneous revascularization is increasing among patients with stable angina, with suggestions of increased mortality. Further investigations should focus on the most effective means to improve veteran access to medical care while also maintaining quality.
The views expressed in this article are those of the authors and do not necessarily reflect the position or policy of the Department of Veterans Affairs or the U.S. Government. Dr. Waldo has received unrelated investigator-initiated research support to the Denver Research Institute from Abiomed, Cardiovascular Systems Incorporated, Janssen Pharmaceuticals, and the National Institutes of Health; and serves as the National Director of the VA Clinical Assessment, Reporting, and Tracking (CART) Program, the national quality and safety oversight organization for invasive cardiac procedures within the Veterans Affairs (VA) Healthcare System. Dr. Valle has received unrelated consulting fees from Philips Medical, Medtronic, and Cardiovascular Systems Incorporated; and serves as the Associate Director of the VA CART Program. Dr. Schofield serves as the National Program Director for Cardiology within the VA Healthcare System. Dr. Ho is supported by grants from the National Heart, Lung, and Blood Institute, VA Health Services Research and Development, and University of Colorado School of Medicine; has received a research agreement with Bristol-Myers Squibb through the University of Colorado; and serves as the Deputy Editor for Circulation: Cardiovascular Quality and Outcomes. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
The authors attest they are in compliance with human studies committees and animal welfare regulations of the authors’ institutions and Food and Drug Administration guidelines, including patient consent where appropriate. For more information, visit the JACC author instructions page.
- Received May 6, 2020.
- Accepted May 10, 2020.
- Barnett P.G.,
- Hong J.S.,
- Carey E.,
- Grunwald G.K.,
- Joynt Maddox K.,
- Maddox T.M.
- Gellad W.F.
- Department of Veterans Affairs
- Tanielian T.,
- Farmer C.,
- Burns R.,
- Duffy E.,
- Messan Setodji C.
- Waldo S.W.,
- Gokhale M.,
- O’Donnell C.I.,
- et al.
- Wilson N.J.,
- Kizer K.W.