Quality of Care and In-Hospital Outcomes in Patients With Coronary Heart Disease in Rural and Urban Hospitals (from Get With the Guidelines–Coronary Artery Disease Program)

Article Outline

• Abstract
• Methods
• Results
• Discussion
• References
• Copyright

Previous studies have suggested that patients with coronary artery disease (CAD) in rural areas may have worse outcomes due to limited availability of specialists, fewer resources, and less institutional funding. Data were collected from hospitals participating in the Get With the Guidelines–Coronary Artery Disease Program (GWTG-CAD) from January 2000 to December 2008. In-hospital outcomes and quality of care were stratified by care at rural versus urban hospitals. Multivariate logistic regression analysis was used to determine the association of rural locale with in-hospital mortality, length of stay, and compliance with the GWTG-CAD performance measurements including (1) early aspirin use, (2) smoking cessation counseling and discharge prescriptions of (3) aspirin, (4) angiotensin-converting enzyme inhibitors, or angiotensin receptor blockers for left ventricular systolic dysfunction, (5) β-blockers, and (6) lipid-lowering therapy and a composite of all 6 measurements. Data were collected from 22,096 patients at 71 rural centers and 329,938 patients at 477 urban centers. Unadjusted rates of compliance with performance measurements were lower in rural (range 82.4% to 90.5%) compared to urban (range 81.3% to 95.0%) hospitals including the composite (74.7% vs 80.6%, p <0.0001). In multivariate analysis, rural status was not independently associated with lower compliance with any of the performance measurements. Unadjusted mortality rates were higher in rural versus urban hospitals (5.7% vs 4.4%, p <0.0001), but this was not significant in multivariate analysis (odds ratio 1.05, 95% confidence interval 0.87 to 1.26). In conclusion, within the GWTG-CAD quality improvement initiative, patients with CAD treated at rural hospitals receive similar quality of care and have similar outcomes as those at urban centers.

Based on 2000 US census data, 6.4% of the US population resides in rural areas within towns of <10,000 residents and another 10.1% live in areas of 10,000 to 50,000 residents.¹ Previous studies have shown that this rural population is more vulnerable than its urban counterparts in that they are older, less educated, more likely to be unemployed, and more susceptible to economic downturn.², ³ Moreover, delivery of medical care to rural populations can be challenging because there are fewer hospitals, providers, and specialists and greater distances between individual residences and hospitals.², ³ Given this background, the present study sought to determine the characteristics, treatments, quality of care, and in-hospital outcomes of patients with coronary artery disease (CAD) treated in rural versus urban hospitals participating in the Get With the Guidelines–Coronary Artery Disease Program (GWTG-CAD). We hypothesized that known differences in care by geography may be minimized or eliminated within this national quality improvement system.

Back to Article Outline

Methods 

The GWTG-CAD is a national, ongoing, prospective observational data collection and quality improvement initiative overseen by the American Heart Association. Details of this program have been described elsewhere.⁴, ⁵, ⁶ Hospitals participating in this registry include institutions from all regions of the United States and represent community hospitals and tertiary referral centers. Trained individuals at each site submitted clinical information regarding medical history, hospital care, and outcomes for consecutive patients hospitalized for CAD-related diagnoses using an online, interactive patient management tool (Outcome Sciences, Inc., Cambridge, Massachusetts). Variables entered included demographic and clinical characteristics, medical history, contraindications to evidence-based therapies, quality and performance measurements, and in-hospital outcomes.

All participating institutions were required to comply with local regulatory and privacy guidelines and to submit the GWTG-CAD protocol for review and approval by their institutional review boards. Because data were used primarily at each local site for quality improvement, sites were granted a waiver of informed consent under the common rule. Outcome Sciences, Inc., served as the clinical coordinating center for GWTG.

This study includes patients enrolled from 548 hospitals participating in the GWTG-CAD program who were hospitalized with confirmed clinical diagnoses of CAD—including patients with acute coronary syndromes, those with stable CAD hospitalized for revascularization, and those with documented CAD hospitalized for reasons other than heart failure. Case finding was based on clinical identification of patients with CAD diagnoses or Joint Commission International Classification of Disease, Ninth Revision, identification of CAD diagnoses with clinical verification for data abstraction.

The analysis cohort included patients reported in the GWTG-CAD database from January 2, 2000, to December 30, 2008, who were >18 years old. The data elements collected had written definitions and were gathered using common specifications for all participants. Using the Internet-based system, data quality was monitored and reports were generated to ensure the completeness and accuracy of the submitted data.

Patients were categorized as receiving treatment in hospitals located in rural or urban areas. A rural hospital was defined by location outside a core-based statistical area (CBSA), and an urban hospital was located within a CBSA. CBSAs are identified by having ≥1 urbanized area with a population of ≥10,000 residents.¹ CBSAs were derived from the 2000 US census data and cross-referenced with the participating hospitals' postal codes.

Demographics, clinical characteristics, and hospital characteristics were stratified by whether the patient was treated at a rural versus urban center. Quality of care, rates of revascularization in eligible patients, in-hospital mortality, and length of stay were then assessed. Quality-of-care assessments examined hospital compliance with several evidence-based performance measurements and were analyzed only for eligible patients in whom no therapeutic contraindication was identified. The GWTG-CAD performance measurements have been previously described⁵ and included (1) aspirin therapy within 24 hours of acute myocardial infarction (AMI), (2) use of an angiotensin-converting enzyme inhibitor or angiotensin receptor blocker in patients with left ventricular systolic dysfunction, (3) aspirin therapy at discharge, (4) β blocker therapy at discharge, (5) smoking cessation counseling for eligible patients, and (6) lipid-lowering therapy for patients with low-density lipoprotein cholesterol levels >100 mg/dl. A composite performance measurement was also assessed and was defined as a defect-free measurement of all 6 GWTG-CAD performance measurements.

Assessment of 6 additional performances included (1) the proportion of patients with ST-segment elevation MI who received fibrinolytic therapy within a door-to-thrombolytics time of ≤30 minutes, (2) the proportion of patients with ST-segment elevation MI receiving primary percutaneous coronary intervention (PCI) within a door-to-balloon time of ≤90 minutes, (3) documentation of appropriate blood pressure control (systolic blood pressure <140 mm Hg and diastolic blood pressure <90 mm Hg) at discharge, (4) patients with AMI or PCI discharged on clopidogrel, (5) low-density lipoprotein cholesterol level recorded in the medical record, and (6), physical activity recommendations given to a patient or referral made for cardiac rehabilitation.

Because the GWTG-CAD database included several CAD-related diagnoses, prespecified subgroup analysis was also performed limiting the population to only those patients a diagnosis of AMI. Differences in characteristics, performance measurements, rates of revascularization, and in-hospital outcomes between patients treated at rural and urban centers within this population were analyzed in a similar fashion.

All statistical analyses were performed centrally at the Duke Clinical Research Institute (Durham, North Carolina). Data were reported as means for continuous variables and as percentages of nonmissing values for categorical variables. Univariate comparisons between patients from rural versus urban centers were performed using Pearson chi-square tests for categorical variables and Wilcoxon rank-sum tests for continuous variables.

Potential confounders that may have affected outcomes were identified in univariate analysis and included demographics (age, gender, race), insurance status, medical co-morbidities, and clinical characteristics (weight, systolic blood pressure, cardiac diagnosis). The generalized estimating equation method was applied to provide valid inference accounting for within-site correlation. Other potential confounders included specific hospital characteristics—bed size, academic versus community status, and availability of on-site PCI, cardiac surgery, and cardiac transplantation. Logistic multivariable regression models were then used to adjust for these hospital confounders in the analysis of in-hospital outcomes and quality-of-care measurements. Estimated adjusted odds ratio for death, primary and secondary performance measurements, and estimated mean ratios for length of stay were then computed. SAS 9.1 (SAS Institute, Cary, North Carolina) was used for all statistical analyses.

Back to Article Outline

Results 

A total of 352,034 patients were identified among 548 hospitals from January 2, 2000, to December 30, 2008. Most patients (270,847 patients, or 76.9%) were admitted with a diagnosis of acute coronary syndrome defined as AMI or unstable angina. There were 71 rural centers (13.0%) accounting for 22,096 patients (6.3%) and 477 urban centers (87.0%) accounting for 329,938 patients (93.7%). As presented in Table 1, compared to urban counterparts, rural hospitals had fewer hospital beds, were less likely to be academic/teaching centers, were less likely to have on-site interventional cardiology, cardiac surgery, and cardiac transplantation capabilities, and were less likely to perform percutaneous transluminal coronary angioplasty for AMI.

Table 1. Baseline hospital characteristics between rural and urban centers

Patients from rural centers were older, more often women, more Caucasian, and were more likely to have Medicare insurance, whereas urban patients were more likely to be uninsured (Table 2). Atrial fibrillation and heart failure were more prevalent at rural hospitals. Clinical laboratory data and admission diagnoses were similar between patients from rural and urban hospitals.

Table 2. Baseline patient characteristics among rural and urban centers

The comparison of GWTG-CAD performance measurements by rural and urban locations is presented in Table 3. Patients at rural centers had lower unadjusted rates of compliance with the GWTG-CAD performance measurements of aspirin within 24 hours of AMI, aspirin therapy at discharge, β blocker at discharge, smoking cessation counseling, and lipid-lowering therapy. The overall composite measurement of compliance with all GWTG-CAD performance measurements was lower for rural centers. After adjusting for patient characteristics, rural status was associated only with lower rates of smoking cessation counseling. After adjusting for patient and hospital characteristics, rural status was independently associated only with higher rates of compliance with angiotensin-converting enzyme inhibitor or angiotensin receptor blocker therapy for left ventricular systolic dysfunction.

Table 3. Hospital outcomes and quality of care with univariate and multivariate analysis

Variable	Rural Centers	Urban Centers	Unadjusted Odds Ratio (95% CI)	Adjusted Odds Ratio for Patient Factors (95% CI)⁎	Adjusted Odds Ratio for Patient and Hospital Factors (95% CI)†	Multivariate p Value
Get With the Guidelines–Coronary Artery Disease Program performance measurements (percent compliance of eligible patients)
Aspirin within 24 hours for acute myocardial infarction	89.6%	92.1%	0.62(0.42–0.91)	1.20(0.67–2.14)	1.61(0.59–4.40)	0.349
Angiotensin-converting enzyme inhibitor/angiotensin receptor blocker at discharge for left ventricular systolic dysfunction	82.4%	81.3%	1.02(0.82–1.26)	1.19(0.97–1.45)	1.25(1.03–1.53)	0.025
Aspirin at discharge	90.5%	95.0%	0.58(0.45–0.75)	0.79(0.55–1.13)	0.80(0.56–1.16)	0.242
β blocker at discharge	86.2%	91.3%	0.62(0.47–0.83)	0.88(0.63–1.24)	0.96(0.69–1.33)	0.785
Smoking cessation counseling	85.6%	89.6%	0.46(0.33–0.65)	0.61(0.41–0.89)	0.74(0.53–1.03)	0.076
Lipid-lowering therapy for low-density lipoprotein >100 mg/dl	83.4%	86.5%	0.60(0.43–0.84)	1.02(0.73–1.42)	1.12(0.83–1.52)	0.455
Composite (defect free)	74.7%	80.6%	0.75(0.60–0.92)	0.85(0.63–1.16)	0.84(0.61–1.15)	0.274
Additional performance measurements (percent compliance of eligible patients)
Door to thrombolytics for ST-segment elevation myocardial infarction ≤30 minutes	33.2%	33.1%	1.23(0.92–1.64)	1.11(0.81–1.51)	1.11(0.81–1.51)	0.520
Door to balloon for ST-segment elevation myocardial infarction ≤90 minutes	49.6%	49.5%	0.94(0.66–1.33)	0.94(0.66–1.32)	0.94(0.65–1.38)	0.763
Blood pressure <140/90 mm Hg at discharge	74.8%	78.5%	0.88(0.77–1.01)	0.98(0.86–1.13)	1.07(0.93–1.24)	0.338
Acute myocardial infarction or percutaneous coronary intervention discharged on clopidogrel	76.2%	73.3%	0.52(0.34–0.78)	1.12(0.57–2.21)	1.23(0.61–2.50)	0.563
Low-density lipoprotein recorded	59.4%	62.9%	0.69(0.53–0.89)	0.92(0.64–1.33)	0.96(0.69–1.34)	0.798
Cardiac rehabilitation or physical activity recommendations	72.5%	64.4%	0.93(0.67–1.28)	1.19(0.61–2.32)	1.47(0.81–2.66)	0.211
Revascularization (percentage of eligible patients)
Percutaneous coronary intervention	38.3%	57.4%	0.23(0.14–0.38)	0.54(0.33–0.87)	0.58(0.36–0.94)	0.026
Coronary artery bypass grafting	7.6%	13.3%	0.29(0.14–0.58)	0.58(0.28–1.20)	0.69(0.35–1.36)	0.287
Outcomes
In-hospital mortality	5.7%	4.5%	1.15(0.89–1.47)	1.11(0.91–1.36)	1.05(0.87–1.26)	0.624
Hospital length of stay(days)	4.8	5.3	0.86(0.77–0.96)	0.88(0.79–0.98)	0.90(0.81–1.00)	0.059

CI = confidence interval.

With regard to additional performance measurements, rural centers were more compliant with appropriate clopidogrel use after PCI or AMI but were less compliant with recording of low-density lipoprotein cholesterol levels (Table 3). In multivariate analysis, rural status was not independently associated with differences in the any of the additional performance measurements.

Rates of revascularization and in-hospital outcomes are also listed in Table 3. There were significantly fewer revascularization procedures performed at rural centers—including PCI and coronary artery bypass grafting (CABG) surgeries. In multivariate analysis, rural status was independently associated with lower rates of PCI. Patients at rural centers did appear to have greater unadjusted in-hospital mortality than patients at urban centers (5.7% vs 4.4%, p <0.0001). However, in multivariate analysis, there were no significant differences between in-hospital mortality and length of stay.

Within the population of patients with AMI, similar differences were noted between patients treated at rural versus urban centers. At rural centers, unadjusted rates of revascularization for PCI and CABG were lower, in-hospital mortality was higher (8.1% vs 5.7%, p <0.0001), and length of stay was shorter (5.4 vs 5.9 days, p <0.0001). In multivariate analysis, rural status was independently associated only with shorter length of stay (odds ratio 0.90, 95% confidence interval 0.85 to 0.96). There was no difference in adjusted mortality rates. In addition, rural status was not independently associated with differences for any of the performance measurements.

Back to Article Outline

Discussion 

We investigated a broad cohort of US patients admitted with CAD to evaluate the influence of rural versus urban status on quality of care and outcomes. This study has 3 main findings. First, despite inherently limited resources and geographic isolation in rural US hospitals, there were no independent differences for most GWTG-CAD performance measurements (including the composite of all GWTG-CAD performance measurements) and all additional performance measurements. Second, there are more revascularization procedures (PCI and CABG) performed at urban centers, and the differences in PCI procedure rates remained significant in multivariate analysis. Third, overall risk adjusted in-hospital mortality was similar in those treated at rural versus urban centers.

Previous studies have suggested that rural centers have worse outcomes and lower quality of care by performance measurements. In a cohort from Kansas, patients with AMI at rural hospitals received less aspirin, less reperfusion therapy, and less β blockers. These differences remained after adjustment for age.⁷ In the Cooperative Cardiovascular Project database, patients with AMI from rural centers were less likely to receive aspirin and to undergo coronary reperfusion, but more likely to receive angiotensin-converting enzyme inhibitor and β-blocker therapy at discharge.⁸ Most notably, rural patients had a worse 30-day mortality compared to their urban counterparts. A more contemporary study of US hospitals showed that rural hospitals were less likely to initiate aspirin therapy, β blockers, and smoking cessation counseling.⁹

Our study differs from previous studies for several reasons. First, this study included a large group of patients collected over an 8-year period who were managed with contemporary therapies. In contrast, similar publications to date have had smaller cohorts assembled in a relatively short time frame and are not nearly as current as this study. Second, although participation by the GWTG hospitals was voluntary, many centers of various types were included. Most other studies assessed patients from a narrower geographic region. Moreover, previous studies have been largely limited to patients with AMI; by comparison, we included several CAD-related diagnoses—although subanalysis of patients with AMI yielded consistent results. Third—and most importantly—previous studies have not consistently examined confounders. Thus, it has been unclear whether it is the rural nature of a hospital itself that underlies the higher mortality rates and lesser compliance with performance measurements or whether geographic variation in patient disease severity or hospital infrastructure accounts for the differences.

In the present study, univariate analysis demonstrated worse in-hospital mortality and lower compliance to most performance measurements for rural centers; however, these observations were no longer significant after multivariate adjustment. This suggests that CAD care in rural hospitals is on par with care at better-resourced urban medical centers and that differences in mortality and care process instead are primarily related to patient characteristics (age, cardiac history, and other co-morbidities) and hospital features (bed size, academic status, cardiac revascularization capabilities).

Furthermore, it is possible that a program like GWTG-CAD can neutralize the impact of rural status by helping all hospitals improve quality of care. The GTWG-CAD program features a Web-based patient management tool, clinical decision support, and real-time hospital data feedback. Indeed, participation in the GWTG-CAD has been previously shown to be independently associated with higher rates of guideline adherence.¹⁰

Overall rates of compliance with the GWTG-CAD performance measurements in this study were 81.3% to 95.0%, with a composite rate of 74.7% to 80.6%. Compared to a more limited number of GWTG-CAD participating hospitals (223 hospitals) over a more discrete period (6 months in 2004), 1 study showed performance adherence rates of 80.9% to 95.0% with a composite rate of 91.8%.¹⁰ This suggests that sustaining high-quality care for longer periods among larger numbers of hospitals remains a challenge. Further work is needed to develop ways to ensure more uniform application of guideline-based care that is sustainable in a variety of institutions over time.

We acknowledge some limitations to this study. The GWTG-CAD is a voluntary program and could over-represent high-performing hospitals. Data were collected by chart review and thus depend on the accuracy and completeness of documentation. Rural centers were defined as being located outside a CBSA, so only areas of <10,000 residents were considered rural in this analysis. CBSAs are based on population density, and data regarding physician practice patterns, patient transfer patterns, and per-capita specialist availability are lacking. We did not ascertain critical-access hospital status. Critical-access hospitals have very limited resources and are only beginning the accreditation process, so greater disparities could exist in this setting.¹¹ In addition, the GWTG-CAD database does not track inpatient provider specialty, and this may influence mortality, length of stay, and quality of care.¹², ¹³, ¹⁴

Multivariate analysis was performed to attempt to adjust for confounding between measured differences between rural and urban centers, but it is possible that there were unmeasured confounders. Only in-hospital measurements were tracked, and thus long-term follow-up outcomes are unknown.

Despite these limitations, the present study has several advantages over other studies. This study was conducted from a large, contemporary, prospectively collected database from multiple diverse hospitals across the United States. It also reflects real-world experience compared to studies involving highly selected patients enrolled in a clinical trial. This suggests that, despite previous conflicting reports, CAD care among the nation's rural hospitals is equivalent to that in urban centers.

Back to Article Outline

References 

1. U.S. Census Bureau. Metropolitan and micropolitan statistical area estimates (Population and population change by CBSA status). http://www.census.gov/popest/metro/CBSA-est2007-CBSAstatus.htmlAccessed February 25, 2009
   • View In Article
2. Hart LG, Salsberg E, Phillips DM, Lishner DM. Rural health care providers in the United States. J Rural Health. 2002;18(suppl):211–232
   • View In Article
   • CrossRef
3. Moscovice I, Stensland J. Rural hospitals: trends, challenges, and a future research and policy analysis agenda. J Rural Health. 2002;18(suppl):197–210
   • View In Article
   • CrossRef
4. LaBresh KA, Gliklich R, Liljestrand J, Peto R, Ellrodt AG. Using “Get With the Guidelines” to improve cardiovascular secondary prevention. Jt Comm J Qual Saf. 2003;29:539–550
   • View In Article
   • MEDLINE
5. LaBresh KA, Ellrodt AG, Gliklich R, Liljestrand J, Peto R. Get with the guidelines for cardiovascular secondary prevention: pilot results. Arch Intern Med. 2004;164:203–209
   • View In Article
   • MEDLINE
   • CrossRef
6. Hong Y, LaBresh KA. Overview of the American Heart Association “Get With The Guidelines” programs: coronary heart disease, stroke, and heart failure. Crit Pathways Cardiol. 2006;5:179–186
   • View In Article
7. Sheikh K, Bullock C. Urban–rural differences in the quality of care for Medicare patients with acute myocardial infarction. Arch Intern Med. 2001;161:737–743
   • View In Article
   • MEDLINE
   • CrossRef
8. Baldwin LM, MacLehose RF, Hart LG, Beaver SK, Every N, Chan L. Quality of card for acute myocardial infarction in rural and urban US hospitals. J Rural Health. 2004;20:99–108
   • View In Article
   • MEDLINE
   • CrossRef
9. Lutfiyya MN, Bhat DK, Gandhi SR, Nguyen C, Weidenbacher-Hoper VL, Lipsky MS. A comparison of quality of care indicators in urban acute care hospitals and rural critical access hospitals in the United States. Int J Qual Health Care. 2007;19:141–149
   • View In Article
   • MEDLINE
   • CrossRef
10. Lewis WR, Peterson ED, Cannon CP, Super DM, LaBresh KA, Quealy K, et al. An organized approach to improvement in guideline adherence for acute myocardial infraction: results with the Get With Guidelines quality improvement program. Arch Intern Med. 2008;168:1813–1819
    • View In Article
    • CrossRef
11. Lutfiyya MN, Sikka A, Mehta S, Lipsky MS. Comparison of US accredited and non-accredited rural critical access hospitals. Int J Qual Health Care. 2009;21:112–118
    • View In Article
    • CrossRef
12. Ahmed A, Allman RM, Kiefe CI, Person SD, Shaneyfelt TM, Sims RV, et al. Association of consultation between generalists and cardiologists with quality and outcomes of heart failure care. Am Heart J. 2003;145:1086–1093
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (144 KB)
    • CrossRef
13. Masoudi FA, Rathore SS, Wang Y, Havranek EP, Curtis JP, Foody JM, et al. National patterns of use and effectiveness of angiotensin-converting enzyme inhibitors in older patients with heart failure and left ventricular systolic dysfunction. Circulation. 2004;110:724–731
    • View In Article
    • CrossRef
14. Reis SE, Holubkov R, Edmundowicz D, McNamara DM, Zell KA, Detre KM, et al. Treatment of patients admitted to the hospital with congestive heart failure: specialty-related disparities in practice patterns and outcomes. J Am Coll Cardiol. 1997;30:733–738
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (384 KB)
    • CrossRef