Gregory, Patrice M. PhD, MPH*; Malka, Edmond S. AB; Kostis, John B. MD; Wilson, Alan C. PhD; Arora, Jasmine K. MD*; Rhoads, George G. MD, MPH
Background. In a highly competitive health care environment, even microgeographic differences in availability of tertiary services might affect access to care.
Objectives. To study the impact of (1) geographic distance from patient's residence to cardiac revascularization services and (2) the availability of cardiac revascularization services at the hospital nearest the patient's residence on utilization of these services in a geographically small, densely populated area.
Methods. Historical cohort study of 55,659 New Jersey residents hospitalized between 1992 and 1996 with primary diagnosis of acute myocardial infarction (AMI).
Main Study Outcomes. Use of percutaneous transluminal coronary angioplasty (PTCA) or coronary artery bypass graft surgery (CABG) within 90 days of initial hospitalization for AMI and in-hospital mortality. Distance from patient's residence to nearest hospital with cardiac revascularization services (PTCA and CABG) was a straight-line distance in miles, categorized as 0 to <2, 2 to <5, 5 to <10, 10 to <15, 15 to <20, 20 to <25, >=25 miles. Adjusted odds of PTCA or CABG use at each distance category were compared with odds at >=25 miles.
Results. A strong linear decline in adjusted odds ratios for PTCA use was found with increasing distance of this service from the patient's residence (p <0.05). Adjusted odds of PTCA use were 2.4, 2.1, 1.8, 1.5, 1.3, and 1.0 times higher for each increasing distance category in comparison with >=25 for patients aged <65 and 3.1, 2.7, 2.2, 1.9, 1.7, and 1.1 for patients aged >=65. Use of CABG was also higher for patients residing closer to cardiac revascularization services. The availability of these services at the hospital nearest to the patient's residence also increased utilization. In-hospital mortality was not associated with distance from services.
Conclusion. Even across a relatively small geographic area, shorter distance to services and availability of services at the nearest hospital were strongly related to increased utilization of cardiac revascularization services.
While the rationale for regionalization of high-technology services has been based on studies that have shown higher volume of procedures associated with lower mortality, morbidity, and unit cost,1,2 the argument against has centered on limited competition and access to services.3 Distance to health care services has long been recognized as an important factor in access to care. In 1961, the American Hospital Association and Public Health Service issued a joint report that suggested one consideration in planning of hospitals and other health facilities should be patient transportation and travel time.4 A review of published studies in the 1950s,5 as well as many more recent studies,6-12 have described inverse or negative relationships between distance to health care personnel or facilities and utilization of their services.
The effects of regionalization and geographic availability of tertiary services on access to care have become increasingly complex issues. In our highly competitive health care environment, there is a strong incentive within each hospital to maximize revenue-generating services and a disincentive to refer patients to competitors. Where a limited number of hospitals provide a tertiary service, this could lead to microgeographic differences in utilization related to hospital catchment areas.
We examined this issue in the context of cardiac revascularization services following acute myocardial infarction (AMI) in the geographically small and densely populated state of New Jersey. This area provided a unique setting in which distances to cardiac revascularization services were relatively short: no study patients were more than 13 miles from an acute-care hospital or more than 70 miles from a hospital with cardiac revascularization services. Our objectives were to study the effects of (1) geographic proximity of the patient's residence to cardiac revascularization services on utilization of percutaneous transluminal coronary angioplasty (PTCA) and coronary artery bypass graft surgery (CABG), and (2) the availability of cardiac revascularization services at the hospital nearest the patient's residence on utilization of these services.
A historical cohort study of patients hospitalized with a primary diagnosis of AMI was conducted. Patients were followed up to determine utilization of cardiac revascularization procedures for 90 days after initial hospitalization for AMI to allow for stabilization, transfers, referrals, and readmissions.
All hospital discharge records from January 1, 1992, to October 1, 1996, from acute-care, nonfederal hospitals in New Jersey were considered. The AUTOMATCH-Generalized Record Linkage System, Version 3.0, was used to link all hospitalizations for the same patient in the absence of unique identifying variables across files.13,14 The method of determining the sensitivity and specificity of this probabilistic record linkage procedure has been described in a previous report and was found to be 98% and 99%, respectively.15
Acute myocardial infarction, International Classification of Diseases Ninth Revision (ICD-9CM) code 410.0-410.9, was used as the main selection criterion because it is a diagnosis that is well documented in hospital discharge records whether or not cardiac revascularization procedures are used. The validity of the primary diagnosis of ICD-9CM 410 in these discharge data was documented previously through medical record review to be approximately 94%.16
The intent was to study service utilization surrounding an acute episode of illness as closely as possible, given the limited nature of administrative data. Only the first hospitalization with AMI as the primary diagnosis for each patient from 1992 to 1996 was considered and linked for subsequent hospitalizations within 90 days. Except for those who were discharged early in the year 1992 for whom we had no prior data, all the study patients had no evidence of cardiac procedure use in New Jersey hospitals in the 30 days preceding the index hospitalization.
Of the 66,340 patients with AMI as a primary diagnosis, only those with valid New Jersey zip codes of residence were included in this study (93.8%). Discharge records indicating transfers from other acute-care hospitals that represented the first occurrences of primary AMI in the study data set were excluded from eligibility as index hospitalizations (5.1%). Over 90% of the patients aged >=65 were Medicare beneficiaries. Because Medicare patients with coverage from health maintenance organizations or preferred provider organizations could not be identified in the hospital discharge data, all those aged >=65 with health insurance coverage other than Medicare fee-for-service were eliminated from the study. Medicare managed care had low penetration in New Jersey during this time. Other insurance types for patients aged >=65 were not numerous enough to be analyzed. In addition, patients aged <65 with Medicare coverage were excluded because of their marked increase in comorbid conditions and reduced likelihood of procedure use. The number of AMI patients remaining in the study group was 55,659.
The 20,468 AMI patients between the ages of 30 and 64 years were analyzed separately from the 35,191 AMI patients aged >=65 to control for differences in health insurance coverage, severity of illness, and likelihood of procedure use.
The utilizations of two cardiac revascularization procedures, PTCA (ICD-9CM 36.00-36.06) or CABG (ICD-9CM 36.10-36.16, 36.19), were considered separately as study outcomes. These outcomes were studied for the entire group of AMI patients and for the subgroup of patients who had received the diagnostic procedure of cardiac catheterization (CATH: ICD-9CM 37.22-37.23, 88.53, 88.55-88.56) within the 90 days after AMI. The rationale for this subgroup analysis was to study distance effects in patients who had already received this key evaluation and who might therefore have experienced fewer barriers to receiving these therapeutic cardiac procedures. The outcome for this subgroup analysis was utilization of either revascularization procedure (PTCA or CABG). In-hospital mortality was determined from vital status at discharge.
The geographic distance was based on the longitude and latitude of the centers of the zip code of the patient's billing address and nearest hospital in New Jersey with cardiac revascularization services. The straight-line distance in miles was calculated and categorized for ease of interpretation as follows: 0 to <2, 2 to <5, 5 to <10, 10 to <15, 15 to <20, 20 to <25, >=25 miles.
During the study period 1992 to 1996, 14 of the 84 acute-care hospitals in New Jersey were equipped to provide cardiac revascularization services. Because this study focused on evaluating revascularization procedures only, the additional 25 hospitals providing only diagnostic CATH services without revascularization were grouped with hospitals providing none of these services.
An additional variable was defined that indicated whether or not the hospital in New Jersey nearest the patient's residence was equipped to provide cardiac revascularization services.
Other variables used in the analyses were race/ethnicity (black or Hispanic compared with white) and type of health insurance for patients aged <65 (Medicaid, no insurance [self-pay, uninsured], managed care [health maintenance organization, preferred provider organization]) compared with private fee-for-service insurance. The anatomic location of the infarct had been found to be highly related to short-term mortality in our previous work and thus was used to adjust for differences in severity of illness.15 The locations were classified as follows and listed from greatest to lowest mortality: other, unspecified (ICD-9CM 410.8-410.9), anterior (ICD-9CM 410.0, 410.1, 410.5), inferior or inferolateral, (ICD-9CM 410.2-410.4, 410.6), or subendocardial (ICD-9CM 410.7) locations. Other diagnoses used to adjust for differences in the case mix of patients were cardiogenic shock, pulmonary edema, congestive heart failure, cardiac arrest, ventricular fibrillation, cerebrovascular disease, diabetes, cancer, and chronic liver, kidney, and lung diseases.
All hospital discharge records from New York and Pennsylvania with New Jersey zip codes of residence were obtained only for those with PTCA and/or CABG procedure codes. Similar information from other states was not available. Personal identifiers on the discharge records were not obtainable; therefore, linkage of out-of-state records to the New Jersey data was not possible. Thus, we could not account for out-of-state migration for revascularization procedures in our study group directly. We estimated the effect as described below.
Distance categories were based on the nearest hospital in New Jersey with revascularization services and the patient's zip code of residence in New Jersey. The number of out-of-state procedures that occurred among New Jersey residents in each distance category was totaled. However, because the records were selected on the basis of procedure use only, without consideration of AMI diagnosis or multiple discharges per patient, these totals were overestimates of the actual out-of-state utilization. To adjust the out-of-state PTCA totals, they were multiplied by the percentages of all hospital discharge records from New Jersey with PTCA codes that were represented in our AMI study cohort (20% for patients aged <65 and 15% for patients aged >=65). The same adjustment was done for CABG codes (17% for patients aged <65 and 15% for patients aged >=65). The estimated out-of-state total was then added to the actual in-state total and expressed as a percentage of AMI patients for each distance category.
Chi-square tests for categorical variables and Wilcoxon rank-sum tests for continuous variables were used to compare each distance category with the category >=25 miles from cardiac revascularization services at P < 0.05. Multiple logistic regression models with 95% maximum likelihood confidence intervals were used to adjust for differences in patient characteristics, severity of illness, and other medical conditions. The odds of cardiac revascularization at each distance category were compared with those for the category of patients living >=25 miles from a hospital with cardiac revascularization services. In other analyses, the odds of procedure use for patients whose nearest hospital was equipped to provide cardiac revascularization procedures were compared with those for patients whose nearest hospital was not so equipped. A test for a linear relationship between the likelihood of procedure use and distance to services was approximated with linear regression methods.17 The adjusted odds ratios for the distance categories (and their variances) were associated with the midpoints of the distance categories.
The characteristics of the study group are shown in Table 1 for the 20,468 AMI patients aged <65 (36.8%) and for the 35,191 AMI patients aged >=65 (63.2%) according to the distance from patient's residence to nearest available cardiac revascularization services in New Jersey. The demographic characteristics of patients in close geographic proximity to cardiac revascularization services reflected the urban locations of most hospitals with these services in New Jersey. The distinct features were the high percentages of black and Hispanic patients in both age strata, and Medicaid and self-pay patients aged <65. Although there were some differences among the distance categories in severity of illness and comorbid conditions, no consistent pattern in relation to distance was observed (see Table 1).
|Table 1. Characteristics of Patients With Acute Myocardial Infarction (AMI) According to Distance From Patient's Residence to Nearest Hospital With Cardiac Revascularization Services, New Jersey, 1992-1996|
The characteristics of the hospital services available in New Jersey in relation to distance to patient's residence are shown in Table 2. The study AMI patients were no more than 13.0 miles and on average 4.7 miles away from the nearest acute-care hospital in New Jersey. The mean distance from the hospital of first admission after AMI was ~3 miles for those in the <2-mile category and 7 or 8 miles for those in the >=25 mile category. Over one third of patients living within 2 miles of cardiac revascularization services did not have those services available at their nearest hospital in New Jersey, reflecting that many hospitals are in close geographic proximity. The proportion of AMI patients first presenting to New Jersey hospitals with cardiac revascularization services decreased sharply as distance to nearest hospital with these services increased.
|Table 2. Hospital Characteristics According to the Distance From Patient's Residence to the Nearest Hospital With Cardiac Revascularization Services (CRS)|
Through linkage of hospital discharge records, we were able to track cardiac revascularization procedures within 90 days of AMI admission regardless of where in New Jersey the procedure was done. The unadjusted percentages of cardiac revascularization procedure use according to distance from cardiac revascularization services in New Jersey are shown in Figure 1 for all AMI patients. Estimates of the out-of-state utilization are also shown in Figure 1. In both age groups, the percentages of patients undergoing PTCA procedures were higher for patients living closer to hospitals with cardiac revascularization services in New Jersey. Although use of CABG also showed a decline with increasing distance from a patient's residence to cardiac revascularization services, the decline was not as great as for PTCA use. Migration to border states of New York and Pennsylvania for revascularization procedures did not appear to change the overall trend of lower procedure use with increasing distance to services.
|Fig. 1. Distance to the nearest hospital with cardiac revascularization services and use of these services in New Jersey, 1992 to 1996. PTCA = percutaneous transluminal coronary angioplasty. CABG = coronary artery bypass graft surgery. Actual in-state utilization based on use of services in New Jersey (solid bars). Out-of-state utilization based on estimates of use of services by New Jersey residents in New York and Pennsylvania (open bars).|
After adjustment for differences in patient characteristics, severity of illness, and comorbidity, differences in in-state utilization of cardiac revascularization procedures according to distance to services still remained. Linear declines in adjusted odds ratios for PTCA use with increasing distance from patient's residence to cardiac revascularization services in New Jersey were observed for AMI patients in both age strata (Fig. 2). Compared with patients living >=25 miles from a hospital with cardiac revascularization services, those in the <2-mile group had almost 2.5 times increased odds of PTCA utilization for patients aged <65, and >3 times for those aged >=65. Every distance category <20 miles showed significantly increased likelihood of PTCA use compared with the category >=25 miles. There were statistically significant inverse linear relationships between the likelihood of PTCA use and distance category in both age groups (P < 0.05).
|Fig. 2. Adjusted effect of distance to the nearest hospital with cardiac revascularization services on use of these services in New Jersey, 1992-1996. PTCA = percutaneous transluminal coronary angioplasty. CABG = coronary artery bypass graft surgery. Odds ratio for use of PTCA or CABG within 90 days of hospitalization for AMI. Adjusted for age, gender, race, health insurance (if aged <65), anatomic location of primary infarct, presence of comorbid conditions using logistic regression. Odds ratios and maximum likelihood 95% confidence intervals are shown. Patients residing >=25 miles from hospitals in New Jersey with cardiac revascularization services composed the reference group.|
Similar but less sharp declines in the adjusted odds ratios for CABG use were seen with increasing distance to cardiac revascularization services. For AMI patients aged <65, adjusted odds of receiving CABG for all distances <10 miles were significantly greater by ~1.5 times compared with the distance category >=25 miles (Fig. 2). The older age group had significantly increased adjusted odds of CABG utilization between 1.2 and 2.0 times for all distance categories <20 miles compared with distance category >=25 miles (Fig. 2). The downward slope of adjusted odds ratios was statistically significant for patients aged >=65 (P < 0.05) and borderline significant for the younger patients.
Even in the subset of AMI patients who received diagnostic CATH, and who therefore might have experienced fewer barriers to undergoing PTCA or CABG, a downward trend in the adjusted odds ratios for revascularization with increasing distance was evident. The decline was more gradual than in the analysis of procedure utilization for the total study group. In both age strata, the adjusted odds of PTCA or CABG utilization among the subset of AMI patients who had received CATH were ~25% higher for distance groups <15 miles than for the distance group >=25 miles (data not shown).
The results for separate subgroup analyses among white, black, male, and female patients yielded similar inverse relationships between distance to services and service utilization. The number of Hispanic patients was insufficient for subgroup analysis.
The effect on in-state procedure utilization of availability of cardiac revascularization services at the hospital in New Jersey nearest to patient's residence is shown for the total study group and for male, female, white, and black patients in Figure 3. The utilization of PTCA was generally 35% greater for those aged <65 and 50% greater for those aged >=65 when the New Jersey hospital nearest the patient's residence had cardiac revascularization services than when it did not. For CABG use, the increases were ~15% for AMI patients aged <65 and 30% for those aged >=65.
|Fig. 3. Effect of availability of cardiac revascularization services at the nearest hospital on use of these services in New Jersey, 1992-1996. PTCA = percutaneous transluminal coronary angioplasty. CABG = coronary artery bypass graft surgery. Odds ratio for use of PTCA or CABG within 90 days of hospitalization for AMI. Adjusted for age, gender, race (if not stratified by gender or race), health insurance (if aged <65), anatomic location of primary infarct, presence of comorbid conditions using logistic regression. Odds ratios and maximum likelihood 95% confidence intervals are shown. Patients whose nearest hospital in New Jersey was not equipped to provide cardiac revascularization services composed the reference group.|
The percentages of AMI patients who died during initial hospitalization in New Jersey had no pattern in relation to distance from services (Table 3). After adjustment for differences in patient characteristics, in-hospital mortality was still not related in any consistent manner to distance of patient's residence from cardiac revascularization services in New Jersey (Table 3). Patients aged <65 residing <25 miles away seemed slightly less likely to die in the hospital than were those in the most distant category.
|Table 3. In-Hospital Mortality and Distance of Patient's Residence to the Nearest Hospital With Cardiac Revascularization Services|
Distance from patient's residence to hospitals with cardiac revascularization services was inversely related to utilization of these services after AMI. The closer the services, the higher was the use of PTCA (up to three times higher) and CABG (up to two times higher). Availability of cardiac revascularization services at the hospital nearest to the patient's residence was also associated with increased utilization.
Our findings not only substantiate those of some previous studies but showed an even greater effect of geographic proximity to services on use of revascularization procedures. In a national sample of veterans hospitalized for AMI, those living <20 miles from their admitting hospital were 20% more likely to receive ambulatory care within 30 days after discharge than those living >20 miles away.7 Another national study of Medicare patients with AMI in 1987 found that those living <=2.5 miles from a hospital with CATH or revascularization services were more likely to have received revascularization procedures than those living >2.5 miles away (8.6% vs. 6.9% for CABG and 6.4 vs. 4.3% for PTCA).8 After adjustments for gender and socioeconomic differences, wards in London that contained cardiothoracic units had CABG rates more than two times higher than wards without such units.9 A population-based study comparing CABG rates in New York, California, British Columbia, and Ontario found that rates of CABG in New York and California were actually lower in zip codes 0 to 5 miles from a CABG center than in those from 5 to 25 miles and 25 to 50 miles.12 Rates of CABG utilization were also reported to be much less related to distance from CABG services in Canada.12,18
The availability of revascularization services in close geographic proximity to a patient's residence increases the likelihood of availability of these services at first hospitalization, and thus increases the likelihood of procedure utilization. Over half of AMI patients in this study who lived <2 miles from a hospital with cardiac revascularization services in New Jersey were first admitted to hospitals with these services, compared with <10% for all those >10 miles away. The availability of cardiac revascularization services at first hospitalization increased utilization of PTCA by almost 2.5 times and CABG by >50% after 90-day follow-up in our study group (unpublished results). In our previous study of New Jersey residents, focusing only on 1993 data, a similar result was reported.15 Other researchers have also demonstrated the relationship between availability of cardiac revascularization services at first hospitalization for AMI and increased utilization.10,19-21
Hospital affiliations with other hospitals, payer groups, and community groups might also affect how distance to services impacts utilization. Physician affiliations with hospitals and the geographic distributions of physicians have been shown to influence hospital admission rates.22 New Jersey has a high proportion of specialist physicians. It is likely that the distribution of invasive cardiologists in New Jersey is similar to the distribution of hospitals with cardiac revascularization services. Patients at greater distance from cardiac revascularization services are likely to be cared for by noninvasive cardiologists and/or primary care physicians. Data from the Cooperative Cardiovascular Project and from Medicare claims records of patients hospitalized for AMI in 1992 showed that cardiologists had higher use of cardiac procedures and medications than primary care physicians.23 Even within a single tertiary care center, patients with similar clinical characteristics and utilization of CATH who were under the care of invasive cardiologists received PTCA more often than patients who were under the care of noninvasive cardiologists.24 Furthermore, the less marked inverse association of distance and use of CABG compared with PTCA might have reflected the referral process from primary care physicians and cardiologists to cardiac surgeons. Factors other than distance that could have affected access to additional referrals might have lessened the effect of distance on utilization of CABG.
The patients' previous experience with the health care system likely varied according to distance to revascularization services. Patients living closer to revascularization services not only had greater numbers of revascularization procedures after AMI but might also have had greater utilization of health care services before AMI, including encounters with specialists and primary care physicians. Such variations in previous experience could have affected subsequent patterns of service use. Research allowing more detailed study of this issue would be important to further understand the observed patterns of utilization.
Travel time and effort influence convenience of services and thus choice of hospital.25 In a study based on hospital discharge data from California, an independent predictor of hospital choice for both AMI patients and CABG patients was found to be proximity to patient's residence.26 However, the relationships between travel distance, time, and effort can be complex. Distances calculated for this study were straight-line distances based on the geographic centers of zip codes. Errors in estimating actual distance of patient's residence to nearest hospital with cardiac revascularization services would be greater in less populated, more distant regions where zip codes cover larger areas. Traffic congestion at certain times of the day and days of the week, road construction activities, and variability in access to public transportation are common problems that could affect ease of travel in this densely populated state. Ambulance services usually transport AMI patients for stabilization to the nearest acute care hospital. It is possible in this study, however, that different transport patterns existed based on convenience, affiliation, or preference.
It is conceivable that patients who lived farther away from cardiac revascularization services were sicker and less likely candidates for revascularization procedures than those who lived closer to such services. Two previous studies, one of veterans in Michigan 27 and the other of hospital discharges in Washington State,28 suggested higher levels of patient illness with increasing distance to health facilities. We observed no such evidence of greater severity of illness or comorbidity in relation to distance to cardiac revascularization services. The lack of clinical details in our hospital discharge data set, however, could have resulted in uncontrolled differences in severity of illness and comorbidity among the distance groups.
Differences in procedure use related to distance to revascularization services may reflect underutilization or overutilization. Appropriateness of procedure use in various geographic regions has been studied by other researchers,29-33 although not specifically in relation to geographic distance to revascularization services. In a study of 23 contiguous counties in one large populous state, rates of CATH use were directly related to the percentages of procedures judged inappropriate.34 Inappropriate and uncertain CATH procedures were combined to account for 28% of the variation in rates of CATH utilization across counties. In our study, use of CATH declined with increasing distance from hospitals with cardiac revascularization services in New Jersey, similar to the pattern reported for PTCA (data not shown). Further investigations of the appropriateness of procedure use would be necessary to ascertain the extent to which this geographic pattern of utilization in New Jersey was a function of overuse or underuse.
The extent to which these results for New Jersey may be generalized to other states is limited by its location in the highly urbanized Northeast. New Jersey is the most densely populated state in the United States and ranks 46th in land area. Studies have reported regional variations in rates of cardiac revascularization procedures across the United States 35-39 and other countries 38,39 as well as in rates of other treatments for AMI patients.40 According to the distances calculated in our study, few AMI patients were very far from the nearest acute-care hospital or from cardiac revascularization care. Revascularization services in border states were also relatively close for many New Jersey residents. Out-of-state migration did not seem to account for the effects in this study. Further research of migration patterns would require data containing personal identifiers to allow for record linkage and patient follow-up.
Despite the relative closeness of services for New Jersey residents, there were still distinct patterns of lower utilization associated with greater distances. The link between geographic location of services and utilization of services has been described in an access-to-care framework.41 The microgeographic differences in use of revascularization services reported here indicate that geographic proximity was a major factor in access to care and suggest that regionalization of revascularization services might not be effective in this area.
In-hospital mortality was studied to determine whether distance to revascularization services affected this outcome and thus the likelihood of subsequent procedure use. In-hospital mortality of these AMI patients did not seem to be related to the geographic proximity of cardiac revascularization services in New Jersey. No information was available to us to enable examination of patterns in out-of-hospital deaths, emergency room admissions, or transfers. In the national study of Medicare patients, there was lower mortality after 1 day for AMI patients residing <=2.5 miles than for patients residing >2.5 miles from a hospital with cardiac revascularization services, but no difference was reported in longer-term mortality between 1 and 4 years.8 Other studies of mortality in hospitals with and without on-site invasive cardiac services have found no relationship 19 or significantly lower mortality associated with on-site services.20,42 Longer-term mortality, morbidity, and functional status are important patient outcomes that deserve additional study.
In conclusion, geographic proximity of patient's residence to cardiac revascularization services and availability of these services at the hospital nearest the patient's residence were found to be strongly associated with utilization, even across a relatively small geographic area. The effect of distance from cardiac revascularization services on patient outcomes, however, requires further study.
4. US Public Health Service. Areawide planning for hospitals and related health facilities. PHS Publication No. 855. Washington, DC: US Government Printing Office, 1961. [Context Link]
5. Shannon GW, Bashshur RL, Metzner CA. The concept of distance as a factor in accessibility and utilization of health care. Med Care Rev 1969;26:143-161. [Context Link]
8. McClellan M, McNeil BJ, Newhouse JP. Does more intensive treatment of acute myocardial infarction in the elderly reduce mortality? Analysis using instrumental variables. JAMA 1994;272:859-866. Ovid Full Text Bibliographic Links [Context Link]
9. Ben-Shlomo Y, Chaturvedi N. Accessing equity in access to health care provision in the UK: Does where you live affect your chances of getting a coronary artery bypass graft? J Epidemiol Community Health 1995;49:200-204. [Context Link]
11. Weiss JE, Greenlick MG. Determinants of medical care utilization: The effect of social class and distance on contacts with the medical care system. Med Care 1970;8:456-462. Bibliographic Links [Context Link]
12. Grumbach K, Anderson GM, Luft HS, Roos LL, Brook R. Regionalization of cardiac surgery in the United States and Canada: Geographic access, choice and outcomes. JAMA 1995;274:1282-1288. Ovid Full Text Bibliographic Links [Context Link]
13. Jaro MA. AUTOMATCH Generalized Record Linkage System Version 3.0 Manual. 1995. [Context Link]
14. Jaro MA. Advances in record-linkage methodology as applied to matching the 1985 census of Tampa, Florida. J Am Stat Assoc 1989;84:414-420. [Context Link]
15. Gregory PM, Rhoads GG, Wilson AC, O'Dowd KJ, Kostis JB. Impact of availability of hospital-based invasive cardiac services on racial differences in the use of these services. Am Heart J 1999;138:507-517. Ovid Full Text Bibliographic Links [Context Link]
16. Kostis JB, Wilson AC, O'Dowd K, et al. Sex differences in the management and long-term outcome of acute myocardial infarction. A statewide study. Circulation 1994;90:1715-1730. Ovid Full Text Bibliographic Links [Context Link]
17. Rothman KJ. Modern epidemiology. Boston: Little, Brown and Company, 1986, p. 177-236. [Context Link]
19. Every NR, Larson, EB, Litwin PE. The association between on-site cardiac catheterization facilities and the use of coronary angiography after acute myocardial infarction. N Engl J Med 1993;329:546-551. Bibliographic Links [Context Link]
20. Wright SM, Daley J, Peterson ED, Thibault GE. Outcomes of acute myocardial infarcion in the department of veterans affairs: Does regionalization of health care work? Med Care 1997;35:128-141. [Context Link]
21. Mirvis DM, Graney MJ. Impact of race and age of the effect of regionalization of cardiac procedures in the Department of Veterans Affairs Health Care System. Am J Cardiol 1998;81:982-987. Bibliographic Links [Context Link]
22. Morrill RL, Earickson RJ, Rees P. Factors influencing distances traveled to hospitals. Economic Geography 1970;46:161-172. [Context Link]
23. Jollis JG, DeLong ER, Peterson ED, et al. Outcome of acute myocardial infarction according to the speciality of the admitting physician. N Engl J Med 1996;335:1880-1887. Bibliographic Links [Context Link]
24. Di Salvo TG, Paul SD, Lloyd-Jones D, Conrad Smith AJ, Villarreal-Levy G, Bamezai V. Care of acute myocardial infarction by noninvasive and invasive cardiologists: Procedure use, cost and outcome. J Am Coll Cardiol 1996;27:262-269. Bibliographic Links [Context Link]
27. Holloway JJ, Mendendorp SV, Bromberg J. Risk factors for early readmission among Veterans. Health Serv Res 1996;25:213-217. [Context Link]
28. Welch HG, Larson EB, Welch WP. Could distance be a proxy for severity-of-illness? A comparison of hospital costs in distant and local patients. Health Serv Res 1993;24:442-458. [Context Link]
29. Leape LL, Hilborne LH, Park RE, et al. The appropriateness of use of coronary artery bypass graft surgery in New York State. JAMA 1993;269:753-760. Ovid Full Text Bibliographic Links [Context Link]
30. Bengtson A, Herlitz J, Karlsson T, Brandrup-Wognsen G, Hjalmarson A. The appropriateness of performing coronary angiography and coronary artery revascularization in a Swedish population. JAMA 1994;271:1260-1265. [Context Link]
31. Hilborne LH, Leape LL, Bernstein SJ, et al. The appropriateness of use of percutaneous transluminal coronary angioplasty on New York State. JAMA 1993;269:761-765. Ovid Full Text Bibliographic Links [Context Link]
33. Chassin MR, Kosecoff J, Winslow CM, et al. Does inappropriate use explain geographic variations in the use of health care services? A study of three procedures. JAMA 1987;258:2533-2537. Bibliographic Links [Context Link]
34. Leape LL, Park RE, Solomon DH, Chassin MR, Kosecoff J, Brook RH. Does inappropriate use explain small-area variations in the use of health care services? JAMA 1990;263:669-672. Bibliographic Links [Context Link]
37. Wennberg JE, Cooper MM, eds. The Dartmouth Atlas of Health Care in the United States. Chicago: American Hospital Publishing Inc., 1998. [Context Link]
38. Pilote L, Granger C, Armstrong PW, Mark DB, Hlatky MA. Differences in the treatment of myocardial infarction between the United States and Canada: A survey of physicians in the GUSTO trial. Med Care 1995;33:598-610. Bibliographic Links [Context Link]
39. Van de Werf F, Topol EJ, Lee KL, et al. Variations in patient management and outcomes for acute myocardial infarction in the United States and other countries: Results from the GUSTO trial. JAMA 1995;273:1586-1591. Ovid Full Text Bibliographic Links [Context Link]
40. O'Connor GT, Quinton HB, Traven ND, et al. Geographic variation in the treatment of acute myocardial infarction. The Cooperative Cardiovascular Project. JAMA 1999;281:627-633. [Context Link]
42. Wright SM, Peterson LA, Daley J. Availability of cardiac technology: Trends in procedure use and outcomes for patients with acute myocardial infarction. Med Care Res Rev 1998;55:239-254. Bibliographic Links [Context Link]
Key words: health services accessibility; coronary artery bypass; angioplasty