Most comparative studies to date have used administrative data, which sharply limits the ability of researchers to characterize patients having CABG surgery in either the United States or Canada [3]. One exception was a RAND-Institute for Clinical Evaluative Sciences audit in two Canadian provinces (Ontario and British Columbia) and New York State [8], wherein records of patients who had CABG surgery between 1989 and 1990 were assessed individually using "appropriateness" criteria developed by Delphi expert panels. This study showed similarly high rates of appropriate utilization in Canada and New York despite a much lower rate of CABG surgery in Canada. As such, it provided prima facie evidence for underuse in Canada. However, the total sample size was less than 2000 patients, fewer than half of the cardiac surgery centers in New York were included, and rates of CABG surgery use have increased significantly in both New York State and Ontario since that time [8].

We analyzed the use of CABG surgery in New York State and Ontario, drawing on 1993 data from population-based cardiac surgery registries and other sources in both areas [9-12]. We sought to determine whether the clinical characteristics of patients having CABG surgery in New York and Ontario differed significantly. We also determined the rates of isolated CABG surgery (without concomitant valve surgery) in the two areas in light of the relative rates of coronary angiography and percutaneous transluminal angioplasty (PTCA). In particular, we compared the relative rates of CABG surgery according to coronary anatomy because type of coronary disease is the strongest determinant of life-expectancy gains from CABG surgery [13].

Methods

Top Methods Results Discussion Author & Article Info References

Cardiac Surgery Registries

Most of the data in our study come from cardiac surgery registries in New York State and Ontario; these registries have been described in detail elsewhere [9-12]. New York's registry, the Cardiac Surgery Reporting System, was created in 1989 and is used to disseminate data on risk-adjusted outcomes of cardiac surgery at both the hospital and the surgeon level to the general public as part of a quality improvement program [9, 10]. Ontario's registry, the Cardiac Care Network of Ontario, was initially set up to monitor cardiac surgery waiting lists; the registry has also been used to analyze outcomes of cardiac surgery [11, 12]. The data for New York came from all 16 690 patients who had isolated CABG surgery in New York in calendar year 1993; the data for Ontario came from all 5517 patients who had isolated CABG surgery in Ontario in fiscal year 1993 (1 April 1993 to 31 March 1994). In New York, data are collected during each patient's admission by the cardiac surgery programs at each hospital in New York; in Ontario, data are collected from the referring cardiologist at the time patients are put on the waiting list for CABG surgery. Data in both surgical registries are audited by chart reviews to ensure completeness and accuracy. These auditing processes are described in detail elsewhere [9, 10, 12].

Definitions of Risk Factors

Definitions of clinical risk factors in the two CABG surgery registries were compared; only variables that were considered to be objective and to have similar definitions were selected for the study. We excluded such variables as urgency of surgery, which was defined differently in the two registries. Noncardiac comorbid conditions could not be compared because they were not collected for the Ontario registry. For these reasons, we could not fairly compare the risk-adjusted short-term outcomes of patients having CABG surgery in New York and Ontario.

Grade of left ventricular function was defined on the basis of ejection fraction, as assessed by either angiography or echocardiography. Reoperation was defined as previous open heart surgery in the New York registry and as previous CABG surgery in the Ontario registry. In both registries, coronary anatomy was defined as follows: Left main disease was considered present if there was 50% or greater stenosis of the left main coronary artery; disease of other coronary vessels (left anterior descending, left circumflex, or right coronary vessels) was considered present if stenosis was 70% or greater. Proximal left anterior descending (PLAD) disease was considered present when stenosis of 70% or greater existed proximal to the first septal perforator. Coronary anatomy was divided into three categories (left main disease, three-vessel or two-vessel disease with PLAD disease, and one-vessel or two-vessel disease without PLAD disease). The first two categories were defined as severe coronary artery disease, and the latter category was defined as limited coronary artery disease.

Coronary Angiography and Percutaneous Transluminal Coronary Angioplasty

In Ontario, data used to determine coronary angiography rates were obtained from physician billing data from the Ontario Health Insurance Program and data on PTCA were obtained from the Canadian Institute for Health Information administrative database. In New York, data on coronary angiography were obtained from the New York State Department of Health [14] and data on PTCA were obtained from New York's angioplasty reporting system [15].

Rates of Cardiac Procedures

Age-specific rates of CABG use in different coronary anatomy strata per 100 000 adults were determined using only patients who were residents of New York State or Ontario; residency was established by use of ZIP codes in New York and health card numbers in Ontario. For rate calculations, patient ages were divided into three age strata (20 to 64 years, 65 to 74 years, and 75 years). The numerators in these rates were based on the crude number of CABG procedures done in each stratum. The denominators in the rate calculations were based on the number of adults in each age stratum, estimated from government census data for 1993 [16, 17]. Rates of coronary angiography and PTCA use in New York and Ontario were calculated in a similar manner. Overall procedure rates in Ontario were adjusted for age by using the direct method; the age distribution in New York was the standard population [18].

Increasing the Rate of Coronary Artery Bypass Graft Surgery in Ontario

The number of additional CABG procedures required in Ontario to increase that province's CABG surgery rate to the rate in New York was estimated on the basis of the assumptions that the prevalence and severity of coronary artery disease are similar and that New York's current distribution of coronary anatomy would apply. The absolute difference in CABG surgery rates between New York and Ontario within each age and coronary anatomy group was calculated; these differences were then multiplied by the corresponding total population in Ontario in 1993 within each age group to determine the total number of additional CABG procedures required.

Statistical Analysis

We compared all continuous variables using unpaired t-tests and compared all categorical variables using the appropriate chi-square statistic. All P values were two sided. Relative rates of cardiac procedure use in New York compared with Ontario were determined; 95% CIs of these rates were calculated using a Taylor series expansion [18]. Stata statistical software (Stata Corp., College Station, Texas) was used for statistical analysis.

Results

Top Methods Results Discussion Author & Article Info References

Number of Cardiac Surgeons and Hospitals and Volume of Coronary Artery Bypass Graft Surgery

Table 1 shows demographic characteristics of the cardiac surgery programs in New York State and Ontario. After adjustment for differences in population size, New York had approximately twice as many cardiac surgery hospitals and cardiac surgeons as Ontario. On average, cardiac surgeons in Ontario did 14% more CABG procedures annually than did their colleagues in New York (mean ±SD, 131 ± 45 per year in Ontario compared with 115 ± 81 per year in New York).

Rates of Cardiac Procedures

Age-adjusted rates of cardiac procedures in New York and Ontario in 1993 are shown in Table 2. Centers in New York did 2.20 times (95% CI, 2.17 to 2.24) as many coronary angiographies, 2.23 times (CI, 2.16 to 2.31) as many PTCAs, and 1.79 times (CI, 1.74 to 1.85) as many CABG operations in 1993. These findings show that lower rates of CABG surgery use in Ontario were not a function of higher relative rates of angioplasty. In New York, the rate of PTCA use was slightly greater than the rate of CABG surgery use; in Ontario, the rate of CABG surgery use was slightly higher than the rate of PTCA use.

Clinical Characteristics

Selected clinical characteristics of patients having CABG surgery in New York State and Ontario are shown in Table 3. The mean age of patients having CABG surgery was higher in New York, as was the proportion of patients 80 years of age and older. New York had a higher proportion of female patients than did Ontario; this difference was found in all age strata (data not shown). Patients in Ontario were more likely than patients in New York to have had left ventricular dysfunction with an ejection fraction of 50% or less (57% compared with 49%; P < 0.001). The percentage of patients who had had myocardial infarction in the 3 weeks before CABG surgery was much higher in New York than in Ontario (23% compared with 8%; P < 0.001). This finding is consistent with those of recent U.S.-Canadian studies showing that procedure rates after acute myocardial infarction are higher in the United States [19, 20].

Rates of Coronary Artery Bypass Graft Surgery

Age-specific rates of CABG surgery in New York State and Ontario, stratified by coronary anatomy, are shown in Table 4. The overall age-adjusted rate of CABG surgery in 1993 was 120.6 per 100 000 adults in New York and 67.4 per 100 000 adults in Ontario. The difference in CABG surgery rates was greatest among persons 75 years of age and older; 3.34 times (CI, 3.01 to 3.71) more CABG surgeries were done in New York than in Ontario. Overall, 6% of patients in Ontario and 30% of patients in New York had limited coronary artery disease (one-vessel or two-vessel disease without PLAD disease) (P < 0.001), although more patients in New York had left main disease (23% in New York compared with 16% in Ontario; P < 0.001). Ninety percent of the patients with limited coronary artery disease in New York had Canadian Cardiovascular Society class 3 or 4 angina before surgery [21]. In New York and Ontario, the proportion of patients with more severe coronary artery disease increased with patient age.

Figure 1 shows the relative rate of CABG surgery in New York compared with that in Ontario in different coronary anatomy groups stratified by age. The difference in rates was most striking for persons who had limited coronary artery disease: In New York, 16.75 times as many CABG surgeries were done among patients 75 years of age and older as were done in Ontario. Compared with Ontario, New York had 8.97 times as many CABG surgeries in patients with one-vessel or two-vessel disease without PLAD disease, 1.08 times as many surgeries in patients with two-vessel and PLAD disease or three-vessel disease, and 2.53 times as many surgeries in patients with left main disease.

View larger version (25K): [in this window] [in a new window]   Figure 1. Relative rate of isolated coronary artery bypass graft (CABG) surgery in 1993 by age and coronary anatomy in New York State (NY) compared with Ontario (ON). PLAD = proximal left anterior descending artery.

Increasing the Rate of Coronary Artery Bypass Graft Surgery in Ontario

An additional 4312 procedures would need to be done in Ontario each year if it were to achieve New York's higher rate of CABG surgery (Table 5). Fifty-seven percent of these procedures would be done in patients with limited coronary artery disease, and 29% would be done in patients with left main disease. The difference in the relative rate of CABG surgery is greatest among the elderly; however, Table 5 shows that 40% of the additional procedures would be required among persons younger than 65 years of age because these persons represent the largest population group.

Discussion

Top Methods Results Discussion Author & Article Info References

Population-based rates of CABG surgery have previously been shown to be much higher in the United States than in Canada [3]. Our study confirms these findings. In particular, although New York regulates the provision of CABG surgery (through limits on the total number of hospitals certified to do CABG surgery) more strictly than do most U.S. states, in 1993 CABG surgery was done 1.79 times more often (CI, 1.74 to 1.85) in New York than in Ontario. Higher angioplasty rates were also found in New York, highlighting that the observed findings did not reflect a substitution of angioplasty for CABG surgery in Ontario. Both jurisdictions have rationalized the provision of cardiac surgery through policies that-directly or indirectly-limit provision of the procedure primarily to high-volume centers and cardiac surgeons [22, 23]. Thus, New York has approximately twice as many hospitals and cardiac surgeons doing CABG surgery per capita as does Ontario.

Obvious socioeconomic and cultural similarities can be seen between Ontario and New York; thus, it is unlikely that the increased use of CABG surgery in New York reflects major differences in the prevalence and severity of coronary artery disease between the two areas. Moreover, Anderson and colleagues [24] found similar rates of hospital discharge for diagnoses of ischemic heart disease in Canada and the United States, providing indirect evidence that the prevalence of cardiovascular disease is similar in the two countries. Rates of death from ischemic heart disease are also similar in Canada and the United States [25].

These findings lead logically to the question, Why is the CABG surgery rate so much higher in New York State? As in earlier studies, we found that the actual difference in rates of CABG surgery use continues to be most pronounced among the elderly and that rates of CABG surgery use were higher in women and in patients who had recently had myocardial infarction [3, 8]. However, the most striking finding in our study was the use of CABG surgery among patients in New York and Ontario according to anatomical indications for surgery. The higher rate of CABG surgery in patients with left main disease in New York indirectly suggests underservicing in Ontario. Conversely, however, we saw a much higher rate of CABG surgery in patients from New York who had limited coronary disease; in these patients, the long-term benefits of surgery are less well established [13].

At first glance, these findings appear to contrast with those of other studies that have suggested a minimal relation between rates of procedure use and clinical appropriateness [8]. However, it is entirely possible that most patients in New York who had surgery for limited coronary disease would be considered "appropriate" candidates for surgery on the basis of symptoms, impaired left ventricular function, or positive results on noninvasive tests for myocardial ischemia. Quality of life may also be more important than increased longevity when revascularization is being considered in elderly patients. We suggest, therefore, than one "right" rate of CABG surgery is unlikely; rather, there appear to be trade-offs. Canadian funding restrictions and concomitantly lower rates of CABG surgery lead to more efficient use of available resources, and surgery is provided primarily to those patients who are the most likely to have the greatest life-expectancy gains [13]. New York's higher CABG rate probably includes all of these patients but also includes many patients with limited coronary disease in whom the primary indication for surgery is presumably relief of angina rather than extension of life-expectancy. Determining whether New York's higher CABG surgery rate is more or less "right" than the rate in Ontario is difficult; this determination depends on the manner in which one values the surgical relief of angina in patients with limited coronary disease as opposed to alternative uses for the same resources.

The lower rate of surgery for left main disease in Ontario probably reflects underdetection of severe coronary artery disease, assuming that most patients in Ontario who have left main disease detected at angiography are offered and have CABG surgery. Deaths among patients on the waiting list for CABG surgery are relatively uncommon in Ontario, and only a small number of patients (<150 patients in 1993) have surgery outside the province [26]. The aforementioned RAND-Institute for Clinical Evaluative Sciences comparative study [8] showed that coronary angiography was more likely to show insignificant disease in New York (33%) than in Ontario (18%), reinforcing the trade-off paradigm seen on assessment of cardiac procedure rates. Thus, given that the ratio of angiography to CABG surgery is similar in New York and Ontario, an expansion of both cardiac catheterization and cardiac surgery facilities would be required in Ontario before more surgery for left main disease could be done. However, although the absolute number of persons found to have left main disease would increase with higher angiography rates, so too would the rate of angiography when no significant disease is found.

Our study highlights the need for further studies defining the appropriate indications for CABG surgery in the subgroups of patients in whom surgery is done in New York at a higher rate: the elderly, women, and patients who have recently had myocardial infarction. The major clinical trials comparing medical therapy with CABG surgery were started more than 15 to 20 years ago, and both types of treatment have improved significantly since that time. A recent meta-analysis of these studies [13] showed that CABG surgery offers significant survival benefits in patients with left main and three-vessel disease, but the generalizability of these results is less certain given that patients older than 65 years of age were excluded and only 3.2% of the participants were women. Recent studies also suggest that the higher rate of revascularization after myocardial infarction in the United States compared with Canada offers benefits in terms of quality of life, although substantial survival benefits have not been shown [19, 20].

Our study has some limitations. First, data on the long-term outcomes of patients in New York and Ontario who did and did not have the procedure were not available. Such data would help to determine the marginal benefits of the higher rate of CABG surgery in New York. Second, the case mix of patients having CABG surgery in Ontario and New York may not necessarily be representative of the situation in other parts of Canada or the United States. Rates of CABG surgery in Ontario are similar to the rates in most of the other parts of Canada, whereas rates of CABG surgery in New York are lower than those in most of the other parts of the United States [3, 8, 27].

In conclusion, our study provides new insights into differences in the rates of CABG surgery between New York State and Ontario and, more generally, provides a different perspective on the question, "Which rate is right?" [1]. At very high and very low rates of revascularization, a rate could be categorized as "wrong." For example, rates of revascularization after myocardial infarction that were 50% higher in Texas than in New York State have not been shown to improve survival or quality of life [28]. However, we suggest that there is probably neither one "right" rate nor a simple relation between service rates and appropriateness of case selection. Instead, policymakers must work within a paradigm of trade-offs, and there may be an intermediate range in which different CABG surgery rates can be considered acceptable. As CABG surgery rates increase, an increasing absolute number of citizens will receive important benefits from surgery; however, marginal returns on limited health dollars will probably diminish as proportionately more patients have surgery for weaker but still defensible indications. Policymakers in different countries will increasingly face these difficult trade-offs in the future as they seek to contain costs in their health care systems. In the U.S. market-based system, managed care companies will increasingly influence decisions about the use of CABG procedures for their enrollees; in the Canadian system, the government will balance increasing demand for CABG surgery against other uses for the same limited resources.

Appendix

The following are the members of the Steering Committee of the Cardiac Care Network of Ontario (formerly the Provincial Adult Cardiac Care Network of Ontario): Donald S. Beanlands, MD, and Lorna Bickerton, BScN (University of Ottawa Heart Institute, Ottawa, Ontario); Robert Chisholm, MD, and Jeffrey Lozon, MHA (St. Michael's Hospital, Toronto, Ontario); Martin Goldbach, MD (Victoria Hospital, London, Ontario); Vicki Kaminski, BScN (Sudbury Memorial Hospital, Sudbury, Ontario); Barry J. Monaghan, BComm, DHA (West Park Hospital, Toronto, Ontario); Neil McKenzie, MB, BCh (University Hospital, London, Ontario); Christopher D. Morgan, MD (Sunnybrook Health Science Centre, Toronto, Ontario); John Pym, MB, BCh (Kingston General Hospital, Kingston, Ontario); Hugh Scully, MD (Toronto Hospital, Toronto, Ontario); B. William Shragge, MD (Hamilton Civic Hospitals-General Division, Hamilton, Ontario); and James Swan, MD (Scarborough Centenary Health Centre, Scarborough, Ontario).

Dr. Hannan and Mr. Kumar: Department of Health Policy and Management, University at Albany School of Public Health, One University Place, Rensselaer, NY 12144-3456.

Dr. DeBuono: New York State Department of Health, Corning Tower, Empire State Plaza, Albany, NY 12237.

Dr. McNeil: Department of Health Care Policy, Harvard Medical School, 25 Shattuck Street, Parcel B, First Floor, Boston, MA 02115.