Admission to Hospitals With On-Site Cardiac Catheterization Facilities
Impact on Long-Term Costs and Outcomes
Background—Admission to a hospital with a capability for cardiac procedures is associated with a higher likelihood of referral for a cardiac procedure but not with a better short-term clinical outcome. Whether there are differences in long-term mortality and resource consumption is not clear. We sought to determine whether elderly Medicare patients with acute myocardial infarction admitted to hospitals with on-site cardiac catheterization facilities have lower long-term hospital costs and better outcomes than patients admitted to hospitals without such facilities.
Methods and Results—As part of the Cooperative Cardiovascular Project pilot in Connecticut, we conducted a retrospective cohort study using data from medical charts and administrative files. The study sample included 2521 patients with acute myocardial infarction covered by Medicare from 1992 to 1993. The cardiac catheterization rate was higher in the hospitals with facilities (38.6% versus 26.9%; P<0.001), but the revascularization rate was similar (20.5% versus 19.5%) during the initial episode of care and at 3 years (29.7% versus 29.7%). Mortality rates were similar for patients admitted to the 2 types of hospitals at 30 days (OR, 1.08; 95% CI, 0.83 to 1.42) and at 3 years (OR, 1.02; 95% CI, 0.83 to 1.26). The adjusted readmission rates were significantly lower among patients admitted to hospitals with cardiac catheterization facilities (OR, 0.76; 95% CI, 0.61 to 0.94). However, the overall mean days in the hospital for the 3 years after admission was 25.9 for patients admitted to hospitals with facilities and 24.6 for the other patients (P=0.234). Adjusting for baseline patient characteristics, there was no significant difference in the 3-year costs between patients admitted to the 2 types of hospitals.
Conclusions—With higher rates of cardiac catheterization and lower readmission rates, patients admitted to hospitals with on-site cardiac catheterization facilities did not have significantly different hospital costs compared with patients admitted to hospitals without these facilities. There was also no significant difference in short- or long-term mortality rates.
Several studies have documented substantial geographic and hospital-based variation in the performance of cardiac catheterization after acute myocardial infarction (AMI).1 2 3 4 5 6 7 8 9 10 11 12 Studies have shown that the presence of an on-site cardiac catheterization facility in the admitting hospital is an important predictor of cardiac catheterization after an AMI.8 10 13 Investigators from the Myocardial Infarction Triage and Intervention Project (MITI) reported that patients admitted to a hospital with an on-site cardiac catheterization facility had a 3-fold increase in the likelihood of undergoing cardiac catheterization,7 exceeding the importance of any other clinical factor, including recurrent chest pain and infarct extension.
The MITI investigators reported that patients admitted to hospitals with on-site catheterization laboratories had approximately $2500 per patient higher 3-year cumulative costs despite having no better clinical outcomes.14 This study was conducted solely in the Seattle metropolitan area, an urban/suburban area in which hospitals are relatively close together and have a high rate of angiography. As a result, the authors provide a caution about the generalizability of their findings.
To address this issue in a different area of the country, we sought to examine the care and outcomes of Medicare beneficiaries throughout Connecticut to determine whether hospitals with on-site cardiac catheterization facilities have higher procedure rates and lower long-term readmission rates and hospital costs. We created longitudinal hospitalization records for Medicare beneficiaries discharged for AMI from June 1992 through May 1993. This project was conducted as a substudy of the Cooperative Cardiovascular Project (CCP) pilot, a Health Care Financing Administration (HCFA) initiative to improve the quality of care for Medicare beneficiaries with AMI.15
Connecticut is the only CCP pilot state in which a detailed economic substudy was conducted. The patient sample includes Medicare patients with confirmed AMI who were discharged between June 1, 1992, and May 20, 1993, from the 33 acute-care hospitals in Connecticut, as previously described.16 We excluded (1) admissions not related to the acute care of an AMI (the fifth digit of the ICD-9-CM code=2), (2) patients <65 years old, (3) patients admitted in transfer from another acute-care facility, and (4) non–Connecticut residents (because of difficulty in tracking subsequent admissions in the year after discharge).
Medical records were abstracted by trained nurses and medical record technicians who collected detailed demographic, clinical, and process-of-care information. A random subsample of cases was reabstracted, and reliability, as measured by the κ statistic, ranged from 72% to 97% for several key quality indicators.17 For each patient, the initial and final ECGs were interpreted by a physician.
Transfers were linked with the index admission to form 1 continuous episode of care. Patients were considered to have been transferred if the discharge date of the first hospitalization matched the admission date of a subsequent hospitalization. Thus, the “index” hospitalization included any hospital stays associated with transfer to another acute-care hospital over 3 years. If patients were transferred to another institution, procedures performed in the transfer institution were considered part of the index hospitalization.
The primary administrative data source for this analysis was the HCFA 1992 to 1996 MEDPAR file for Connecticut. This file contains discharge abstracts for all Medicare inpatients discharged from Connecticut hospitals between October 1991 and September 1996. Discharge abstracts include ICD-9-CM–coded diagnoses, procedures, and charges. To calculate hospital costs, ratios of costs to charges for each hospital at the time of the index admission were obtained from HCFA’s Hospital Cost Reports.
The independent variable of principal interest in this study was the type of hospital to which the patient was initially admitted during the index hospitalization. Hospitals were classified on the basis of having or not having cardiac catheterization facilities (“catheterization hospital” versus “noncatheterization hospital”).
Covariates in this study included age, sex, race, cardiac risk factors, cardiac history, admission characteristics, and hospital treatments and procedures (Table 1⇓). Systolic blood pressure, pulse, and respiratory rate were defined as the highest value recorded within the first 24 hours after admission. Terminal illness was defined as chart documentation of terminal illness, life expectancy of <6 months, palliative care only, or a do-not-resuscitate order written at the time of admission. Hospital treatment variables included the use of aspirin, β-blockers, and thrombolytic therapy in the emergency department or in the hospital on the first day. Left ventricular function was obtained from diagnostic tests in the following order: equilibrium radionuclide angiocardiography, cardiac catheterization, and echocardiography. Hospitalizations in the year before index admission were determined from the MEDPAR database.
The primary outcome for the study was hospital costs in the 3 years after the index admission (including the index admission). Because hospital charges are not a reliable measure of resource consumption,5 18 hospital costs were determined from each patient’s detailed hospital charge record and each hospital’s departmental Medicare cost-to-charge ratios derived from each hospital’s annual Medicare Cost Report5 as described in a previous report.19 All costs were converted to 1996 dollars at a discount rate of 5%.
Mortality, readmission, and cardiovascular procedures (including cardiac catheterization, PTCA, and CABG) were also outcomes. Readmissions and procedures in the 3 years after admission were ascertained from the MEDPAR database. Mortality information was obtained from Medicare’s Enrollment Database, which was derived from data from the Social Security Administration.
We compared baseline characteristics, use of cardiovascular procedures, mortality, and costs between patients admitted to catheterization and noncatheterization hospitals. We also compared costs between the 2 types of hospitals for patients who did and did not undergo cardiac catheterization as well as those who underwent PTCA, CABG, or both. Comparisons were performed by Student’s t test (for continuous variables and normally distributed data), the χ2 test (for categorical variables), or the Wilcoxon test (for continuous non–normally distributed variables). Hospital costs are presented both as a mean to represent the best estimate of the cost per patient and as a median to represent the best estimate of the cost of a typical patient.
To determine whether patients initially admitted to catheterization hospitals were more likely to undergo cardiac catheterization, we developed a series of multiple logistic regression models. First, we evaluated the unadjusted relationship between cardiac catheterization and the presence of an on-site facility. Second, we adjusted the model for variables associated with cardiac catheterization in bivariate comparisons (P<0.10), as well as those identified from previous studies7 8 and variables considered clinically relevant. We repeated the analysis for the revascularization procedures.
To evaluate whether patients initially admitted to catheterization hospitals had a higher likelihood of survival 3 years after the index admission than those admitted to hospitals without such facilities, we developed another series of multiple logistic regression models. We evaluated the unadjusted relationship between 3-year mortality and the type of hospital and then adjusted the model for variables associated with mortality in bivariate comparisons (P<0.10), as well as those from previous studies17 and variables considered clinically relevant. These variables included the patient characteristics listed in Table 1⇑. We repeated the models using readmission as the dependent variable.
To evaluate whether patients initially admitted to catheterization hospitals had a higher hospital cost in the 3 years after the index admission compared with those admitted to hospitals without such facilities, we developed a series of multiple linear regression models. We evaluated the unadjusted relationship between log-transformed 3-year hospital cost and the type of hospital. Then we adjusted the model for variables associated with cost in bivariate comparisons (P<0.10), as well as those considered clinically relevant. These covariates were the same as those listed above.
Of the 33 Connecticut hospitals, 17 have cardiac catheterization facilities, including 6 that also have facilities for cardiac surgery. All but 1 (94%) of the catheterization hospitals had medical house staff, compared with 14% of the noncatheterization hospitals.
Characteristics of the Patient Sample
A total of 2521 patients with AMI were in the study sample, with a mean age of 76.8 years (SD, 7.5 years); 50.5% were men, and 90.3% were white. The baseline characteristics of the 1730 patients admitted to catheterization hospitals were generally similar to those of the 791 patients admitted to hospitals without such facilities (Table 1⇑). However, there were significant differences in the proportion of patients by race, history of PTCA, history of CABG, chest pain, heart failure, ST-segment elevation, and LV functional status.
Use of Cardiovascular Procedures
The rate of cardiac catheterization during the initial admission was significantly higher among hospitals with cardiac catheterization facilities (Table 2⇓). After adjustment for potential confounders, the increased odds of the patient’s undergoing cardiac catheterization associated with admission to a catheterization hospital was 1.83 (95% CI, 1.48 to 2.26).
After adjustment for the same potential confounders, the odds of the patient’s undergoing PTCA associated with admission to a hospital with cardiac catheterization facilities was 1.48 (95% CI, 1.07 to 2.05), and the odds of undergoing CABG was 0.73 (95% CI, 0.55 to 0.97).
In the 3 years after admission, the early significant differences in cardiac catheterization and CABG rates persisted (Table 3⇓). There remained a trend toward more PTCA procedures in the hospitals with catheterization facilities.
The 3-year postadmission mortality rate for the sample was 45.1% for the catheterization hospitals and 44.5% for noncatheterization hospitals (Table 4⇓). After adjustment for potential confounders, initial admission to a catheterization hospital was not associated with an improved 3-year survival (OR, 1.02; 95% CI, 0.83 to 1.26).
The 3-year postdischarge readmission rates for survivors of the index hospitalization were 69.0% for patients initially admitted to catheterization hospitals and 73.6% for the other patients (Table 4⇑). After adjustment for potential confounders, admission to a hospital with an on-site cardiac catheterization facility was associated with a significantly lower odds of any-cause readmission (OR, 0.76; 95% CI, 0.61 to 0.94). Readmission rates for cardiovascular procedures were 14.5% for patients admitted to catheterization hospitals and 14.6% for the other patients (P=0.95).
Length of Stay and Facility Costs
The median length of stay for the index admission was 9 days for patients admitted to either catheterization hospitals or noncatheterization hospitals (Table 2⇑). The median cost per patient was $13 176 for the patients admitted to hospitals with facilities and $11 695 for the patients admitted to hospitals without facilities (P=0.03). The similarity of facility costs for the index admission between the groups was due, in part, to the costs associated with transfers to other institutions for cardiac catheterization (Table 5⇓). Patients admitted to noncatheterization hospitals who were transferred for cardiac catheterization had higher costs than patients admitted to catheterization hospitals who underwent the procedure.
The mean number of hospital days for the 3 years after index admission (including the index admission) was 25.9 for the patients admitted to catheterization hospitals and 24.6 for patients admitted to noncatheterization hospitals (P=0.234, Table 3⇑). The Figure⇓ shows the cumulative distribution of cost over the year by type of hospital. The median cost per patient was $25 030 for catheterization hospitals and $23 790 for noncatheterization hospitals (P=0.135). In multivariate regression analyses, the presence of an on-site cardiac catheterization facility was not independently associated with significantly higher 3-year hospital costs.
Despite the higher rate of cardiac catheterization at hospitals with on-site cardiac catheterization facilities, our findings do not support the hypothesis that elderly patients initially admitted to such hospitals incur significantly different costs, mortality rates, or long-term revascularization rates than those admitted to hospitals without such facilities. Our findings are consistent with those of other studies,9 12 and we extend their findings by showing no long-term economic advantage of a hospital with a more aggressive approach to managing older patients with an AMI.
Although we did not find an important difference in the short- or long-term costs of care at the 2 different types of hospitals, our findings do not imply that catheterization rates do not affect the cost of care. The higher cost of catheterization for patients who were transferred for the procedure contributed to equalizing the initial costs. Consequently, it is plausible that hospitals with on-site facilities could reduce costs by performing catheterization at a rate comparable to the hospitals without such facilities. These data also suggest that fewer catheterizations could be performed without jeopardizing mortality or economic outcomes.
We cannot determine with certainty why our study found more revascularization procedures per catheterization among the patients admitted to a hospital without a catheterization laboratory. However, this finding may be due to selection effects. The smaller number of patients referred for cardiac catheterization from these hospitals was different from the larger number of patients referred at the hospitals with facilities, and these patients may have been, on average, more suitable for revascularization. Another possibility is that a transfer into a hospital with facilities is associated with an expectation that a procedure will be done, if possible. Consequently, transfer patients may be more likely to be referred subsequently for a revascularization procedure because of the expectations of their physicians.
Our economic results contrast with the findings of Every and colleagues14 from the MITI Registry. They found that patients admitted to hospitals without on-site catheterization facilities were managed with fewer procedures at a lower cost. The MITI investigators evaluated patients admitted to Seattle hospitals with an AMI between 1988 and 1994. Our study sample was older and represented a statewide sample in Connecticut from 1992 to 1993. The cardiac procedure rates in the MITI study were higher than in our study, and the differential between the 2 types of hospitals was larger. The major difference in our results is that we show that although catheterization rates are different at the 2 types of hospitals, the long-term revascularization rates, costs, and outcomes were very similar. The discordant results between the studies suggest that the impact of an on-site catheterization laboratory on clinical practice may vary by geographic location and age of the patients. In regions or in patient groups with lower procedure rates, the presence of an on-site facility may not have a major influence on revascularization rates or cost.
Our study does have several limitations. First, there are a few special features that characterize Connecticut. The state limited its number of laboratories by requiring a certificate of need. Also, Connecticut is relatively small, the hospitals are not far apart, and there is a relatively high density of cardiologists. These factors may have been responsible for the similar patterns of care among the hospitals.
Second, we cannot determine whether there were important differences in functional outcomes between patients admitted to the 2 types of hospitals. Although this issue has been addressed previously, there is need for further study. One report compared use of procedures and outcomes in a high-volume state (Texas) and a low-volume state (New York).9 Despite substantial differences in the use of cardiac procedures after AMI, there were no differences in symptoms or functional status. In contrast, a study by Mark and associates5 comparing Canadian and US patients enrolled in a substudy of the GUSTO-1 trial found that Canadians underwent fewer procedures and had worse functional status 1 year after discharge.
In conclusion, the admission of elderly patients to a hospital with an on-site cardiac catheterization facility was associated with a higher likelihood of their undergoing coronary angiography during the index admission, but there was no clear economic or mortality benefit. Readmission rates were significantly lower for patients admitted to hospitals with catheterization facilities, but readmission for an AMI was higher. Consequently, in this cohort, we cannot identify a clear benefit, harm, or cost attributable to admission to a hospital with an on-site cardiac catheterization facility.
This study was supported in part by the Patrick and Catherine Weldon Donaghue Medical Research Foundation. Dr Krumholz is a Paul Beeson Faculty Scholar. Dr Cohen is the recipient of a Clinician-Scientist Award from the American Heart Association. Jersey Chen is the recipient of a Student Scholar Award from the American Heart Association.
Reprint requests to Harlan M. Krumholz, MD, Yale University School of Medicine, 333 Cedar St, PO Box 208025, New Haven, CT 06520-8025.
The analyses on which this publication is based were performed under Contract No. 500-96-P549, entitled “Utilization and Quality Control Peer Review Organization for the State of Connecticut,” sponsored by the Health Care Financing Administration, Department of Health and Human Services. The content of this publication does not necessarily reflect the views or policies of the Department of Health and Human Services, nor does mention of trade names, commercial products, or organizations imply endorsement by the US Government. The author assumes full responsibility for the accuracy and completeness of the ideas presented. This article is a direct result of the Health Care Quality Improvement Program initiated by the Health Care Financing Administration, which has encouraged identification of quality improvement projects derived from analysis of patterns of care, and therefore required no special funding on the part of this contractor.
- Received January 1, 1998.
- Revision received June 25, 1998.
- Accepted July 1, 1998.
- Copyright © 1998 by American Heart Association
Mark DB, Naylor CD, Hlatky MA, Califf RM, Topol EJ, Granger CB, Knight JD, Nelson CL, Lee KL, Clapp-Channing NE, Sutherland W, Pilote L, Armstrong PW. Use of medical resources and quality of life after acute myocardial infarction in Canada and the United States. N Engl J Med. 1994;331:1130–1135.
Every NR, Larson EB, Litwin PE, Maynard C, Fihn SD, Eisenberg MS, Hallstrom AP, Martin JS, Weaver WD, the Myocardial Infarction Triage and Intervention Project Investigators. 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.
Behar S, Hod H, Benari B, Narinsky R, Pauzner H, Rechavia E, Faibel HE, Katz A, Roth A, Goldhammer E, Freedberg NA, Rougin N, Kracoff O, Shapira C, Jafari J, Lotan C, Daka F, Gottlieb S, Weiss T, Kanetti M, Klutstein M, Rudnik L, Barasch E, Mahul N, Blondheim D, Gelvan A, Barbash G, Israeli Thrombolytic Survey Group. On-site catheterization laboratory and prognosis after acute myocardial infarction. Arch Intern Med. 1995;155:813–817.
Selby JV, Fireman BH, Lundstrom RJ, Swain BE, Truman AF, Wong CC, Froelicher ES, Barron HV, Hlatky MA. Variation among hospitals in coronary-angiography practices and outcomes after myocardial infarction in a large health maintenance organization. N Engl J Med. 1996;335:1888–1896.
Every NR, Parsons LS, Fihn SD, Larson EB, Maynard C, Hallstrom AP, Martin JS, Weaver D, The MITI Investigators. Long-term outcome in acute myocardial infarction patients admitted to hospitals with and without on-site cardiac catheterization facilities. Circulation. 1997;96:1770–1775.
Finkler SA. The distinction between cost and charges. Ann Intern Med. 1982;96:102–109.