A Comparison of the Costs of and Quality of Life After Coronary Angioplasty or Coronary Surgery for Multivessel Coronary Artery Disease
Results From the Emory Angioplasty Versus Surgery Trial (EAST)
Background The Emory Angioplasty Versus Surgery Trial (EAST) is a randomized trial that compares, by intention to treat, the clinical outcome and costs of percutaneous transluminal coronary angioplasty (PTCA) and coronary surgery for multivessel coronary artery disease.
Methods and Results The primary end point was a composite of death, Q-wave myocardial infarction, and a large reversible thallium defect at 3 years. Multiple measures of quality of life also were made. Charges were assessed from the hospital UB-82 bills; professional charges were assessed from the Emory Clinic. Hospital charges were reduced to cost through step-down accounting methods. All costs and charges were deflated to 1987 dollars. Costs were assessed for the initial hospitalization and the cumulative costs of the initial hospitalization and additional revascularization procedures for up to 3 years. There was no difference in mortality or the primary end point. Mean initial hospital charges were $12 654 for the PTCA group and $20 214 for the surgery group (P<.0001). Professional charges were $4538 for PTCA and $9426 for surgery (P<.0001). Three-year hospital charges were $19 047 for PTCA and $21 174 for coronary surgery (P<.0001). Three-year professional charges were $6412 for PTCA and $9861 for surgery (P<.0001). Three-year total charges were $25 458 for PTCA and $31 033 for surgery (P<.0001). Total 3-year costs were $23 734 for PTCA and $25 310 for coronary surgery (P<.0001). There were more hospitalizations for angina and more antianginal medications used in the PTCA group, which would further narrow the differences in cost.
Conclusions There was no difference in the primary end point or its components at 3 years. Although the primary procedural costs of coronary surgery are more than for coronary angioplasty, this cost advantage is largely, although probably not completely, lost by 3 years because of more frequent additional procedures and other resource consumption after a first revascularization by PTCA.
As coronary angioplasty became as common as coronary surgery,1 2 many patients were found to be anatomically suitable for either procedure. Several trials currently are addressing the relative efficacies of these procedures. Efficacy may be measured by mortality, acute myocardial infarction, additional revascularization procedures, angina, and use of antianginal medications. There are additional aspects of efficacy generally included under the term “quality of life” that are more difficult to measure. These include returning to work, social functioning, and psychological state.
Another aspect of a comparison of the two procedures is cost. In principle, if both relative efficacy and relative cost could be measured, then determining which form of therapy was superior in terms of improved efficacy per dollar spent would be simple.3 The multiple measures of efficacy make this a difficult proposition. Similarly, it is not possible to arrive at one final figure for costs. Costs may be reflected in initial in-hospital costs and various measures of cost over different periods of time. Determining the boundaries of what to include in the cost of a procedure after discharge is difficult and arbitrary. Furthermore, there are multiple potential measures of cost.4
The purpose of this article is to present the initial in-hospital charges and costs of procedures in patients with multivessel disease randomized to percutaneous transluminal coronary angioplasty (PTCA) or coronary artery bypass graft surgery (CABG) in the Emory Angioplasty Versus Surgery Trial (EAST).5 6 In addition, the costs of additional revascularization procedures over 3 years and several indexes of quality of life are presented.
Overall Trial Description
EAST is a single-center, randomized clinical trial designed to evaluate whether initial revascularization with PTCA is as efficacious and safe as CABG for patients with multivessel coronary artery disease.5 6 Patients were randomized in EAST from 1987 through 1990 within strata of severity of disease (see definitions below). The primary end point was a composite of death, Q-wave myocardial infarction within 3 years, or a large reversible thallium defect at 3 years. A registry included all patients eligible but not randomized and a complete list of all patients screened for the trial.
Clinical variables were defined in a data dictionary in the EAST trial.5 6 Double-vessel disease meant the presence of ≥50% diameter luminal narrowing in two of the three major epicardial vessel systems. Triple-vessel disease was defined as the presence of ≥50% diameter luminal narrowing in all three major epicardial vessel systems or in the left anterior descending and proximal circumflex arteries in patients with nondominant right coronary arteries. Strata 1 included double-vessel disease with ≥50% diameter luminal narrowing at one site in each affected vessel. Strata 2 was double-vessel disease with ≥50% diameter luminal narrowing in more than one site in at least one affected vessel. Strata 3 included triple-vessel disease with ≥50% diameter luminal narrowing at one site in each affected vessel. Strata 4 was triple-vessel disease with ≥50% diameter luminal narrowing in more than one site in at least one affected vessel. A staged procedure was angioplasty during the initial hospitalization at a site different from any previously dilated site and within the same hospitalization or 14 days of the original procedure. Myocardial infarction was defined as Q-wave myocardial infarction determined by the ECG committee. Q waves were analyzed according to the Minnesota Code. ECGs were gathered during the index hospitalization and for all cardiac-related hospitalizations during follow-up. Thallium defect was defined as a large reversible thallium defect at the 3-year follow-up as judged by the thallium committee. The hospitalization for chest pain category included patients hospitalized for chest pain without myocardial infarction or additional procedures. Variables defined by patient history were hypertension, diabetes, severity of angina, and prior myocardial infarction. Angina was defined by the Canadian Cardiovascular Society Classification.7 Congestive heart failure was defined by the New York Heart Association criteria.
Procedure and Data Collection
All angioplasty and surgery procedures were performed with previously described standard techniques.8 9 Baseline and restudy demographic, clinical, angiographic, and procedural data, including complications, were recorded prospectively on standardized forms and entered into a computerized database. All fields were defined in a data dictionary.
Follow-up information was obtained from the patients or physicians. Follow-up was obtained by telephone every 6 months for 3 years. Follow-up status for each end point also was assessed at each subsequent hospital admission. Patients not readmitted were contacted by telephone or letter. Information obtained included occurrence of myocardial infarction since the initial angioplasty, subsequent need for an additional revascularization procedure (angioplasty or coronary surgery), death (cardiac and noncardiac), recurrent angina, and information on economies and quality of life. All follow-up information was recorded on standardized forms and entered into the computerized database. All repeated procedures performed at Emory University Hospital were confirmed by the database.
Economic Data and Analyses
All hospital charges were obtained from the UB-82 hospital billing form from the hospital finance offices. The UB-82 is a uniform billing statement used by all third-party insurance carriers. Although all available codes were used in this study, charges for catheterizations before and for days in the hospital before the original angioplasty or coronary surgery were not included. All professional charges for all hospitalizations at Emory were obtained electronically from the Emory Clinic Business Office, which performs all professional billing at Emory.
Costs were estimated from the charge data. From the actual resource use in each category, cost-to-charge ratios were determined by the Emory University Hospital Finance Department by use of variations of microcosting and vertical costing.10 These estimates were made by assessment of the use of space, manpower, equipment, and time of services for procedures. Total charges for specific services and centers were determined as follows: once direct costs were calculated for each center, the department-specific indirect cost was added on the basis of the step-down allocation methodology10 used in Medicare cost reports, and an additional amount based on the history of noncollectibles of the particular department was added to the cost figure. Finally, an amount was added for nonpayments. With knowledge of costs and charges for each department, the cost-to-charge ratios were calculated and applied to our sample of patients to determine actual hospital resource use (costs). An example of the calculation of cost-to-charge ratios and application of this procedure to a patient bill is presented in the “Appendix.” Physician charges were used as a proxy for professional costs. The mean charges and costs from Emory were used to estimate those for follow-up procedures performed at other institutions. Because treatment therapies should be compared at one point in time, the value of resources that are not used entirely in the present should be adjusted for annual inflation, the percentage rate at which the general level of prices is changing. Therefore, costs were deflated to 1987 dollars by use of the Medical Consumer Price Index. This is particularly important because the patterns of resource use are different for each therapy, with patients undergoing CABG consuming more resources during the original hospitalization and patients undergoing PTCA requiring additional resources over several years. In principle, costs may be inflated to 1993 by multiplying all dollar figures in this article by 1.56. Although 1993 costs might seem more relevant, costs inflated to 1993 may not fully reflect current costs because the efficiency of intervention procedures and cardiac surgery services probably has increased during the last couple of years. Thus, the costs may best be used for comparison between procedures.
Data are expressed as proportions or mean±SD and as median for the cost data. The two groups were compared according to intention to treat. Follow-up for mortality and repeated procedures was complete at 3 years or the time of death for all patients. Because the data for procedures were not censored, except for death, and all procedures, not just the first one in the follow-up period, were considered, these usually timed end points were considered proportions; actuarial statistics were not used. This was necessary because actuarial statistics will count only the first follow-up procedure; the error that resulted because actuarial statistics were not used was minimal and similar in the two groups because at 3 years the actuarial mortality was just 7% for PTCA patients and 6% for CABG patients.5 Information on myocardial infarctions was dependent on the availability of ECGs at 3 years; thus, some data were missing for myocardial infarctions and the primary variable. More data were missing for quality of life at 3 years. The denominator is presented for all data when any data are missing. Cost data are available on 384 patients. The 8 patients with missing data had procedures performed at the Atlanta Veterans Administration Hospital. Three-year cost data include data on those patients who died up to the time of their deaths. Differences in categorical variables were analyzed by χ2 (or Fisher’s exact test); differences in continuous variables were analyzed by Student’s t tests. When the assumption of normality was violated, largely for the cost data, the two arms were compared by the Mann-Whitney test and across strata by the Kruskal-Wallis test. Correlates of costs in-hospital and over 3 years were determined with multiple regression. Models were created by use of preprocedural variables only and preprocedural and postprocedural variables. All variables presented in the tables were evaluated for each set of regression analyses. Log-linear models also were tested. Statistical testing was performed with sas, s-plus, and bmdp statistical software packages.
Baseline and clinical end point data in the EAST trial have been published5 and are reviewed in Table 1⇓. There was no difference in the baseline values for age, sex, presence of diabetes, systolic arterial hypertension, severity of angina, or congestive heart failure. Most of the patients suffered from severe angina. Despite a significant incidence of prior myocardial infarction, few patients had congestive failure. In each arm, 60% of the patients had two-vessel disease and 40% had three-vessel disease. There was no difference in ejection fraction between the treatment arms. In-hospital deaths were infrequent at 1%. This was also the mortality at 30 days. There was a trend toward more in-hospital Q-wave myocardial infarctions in the surgical group. At 3 years, there was no difference noted in death, Q-wave myocardial infarction, or the primary end point.
Table 2⇓ gives the initial, additional in-hospital, and follow-up procedures. A complex pattern of original in-hospital resource use can be noted. To summarize, 196 of 198 patients in the PTCA arm initially had angioplasty. Of these 196 patients, 88 (45%) had additional angioplasties during the initial hospitalization: 74 (38%) were treated with staged procedures and 14 (7%) had repeated angioplasty because of an adverse outcome of the initial procedure. Of the PTCA patients, 10% required in-hospital coronary surgery, while only 1 CABG patient required an additional revascularization procedure. Over the next 3 years, many more PTCA than CABG patients required additional angioplasties and coronary surgical procedures. Of 162 additional procedures, only 8 (4.9%) were performed at institutions other than Emory.
Table 3⇓ shows additional hospitalizations for myocardial infarction, chest pain, congestive heart failure, and congestive heart failure plus chest pain. There were more hospitalizations for chest pain in the PTCA than the CABG group. Whereas no statistically significant difference was noted for any other category, there was a trend for more hospitalizations for myocardial infarction in the PTCA group.
Table 4⇓ shows quality of life parameters at 3 years. Angina was more prevalent in the PTCA patients. Consistent with this finding, there was a need for more antianginal medication in the PTCA group. Congestive heart failure remained infrequent. Almost two thirds of the patients reported good or very good health. There was a strong trend for CABG patients to report that they believed they had recovered completely compared with the PTCA group. Conversely, there was a slight trend for more angioplasty patients to feel optimistic about their health compared with CABG patients. Most patients in each group reported the same or improved economic status 3 years after their procedures as before the procedure. Only about half the patients were working at the time of entry into the study. Almost all the patients who were working before their procedures returned to work, at least part-time. Nonetheless, of the patients who were working before their original procedures, ≈40% in each group were retired by 3 years. This was consistent with a mean age of 62 at the time of randomization. No difference was noted between the treatment arms in job status before or after the original procedure or at follow-up.
Table 5⇓ gives initial in-hospital and 3-year procedural charges and costs. Cost data are displayed as mean±SD; the median value also is displayed. The median values were always less than the mean values, which is consistent with a tail of higher costs for some patients. For all cost estimates, the difference between median and mean was greater for angioplasty, and there were larger SDs for each measure of cost for angioplasty, consistent with more variability in the number of procedures performed. The initial hospital charges, professional charges, and hospital costs in the group randomized to surgery were all higher than for the angioplasty group. The sum of hospital costs and professional charges was $16 223±11 552 for the PTCA group and $24 005±6222 for the CABG group (P<.0001). If these numbers were inflated to 1993 dollars, the figures would be $24 821±17 675 for PTCA and $36 728±9520 for CABG. The in-hospital charges for angioplasty were profoundly affected by the number of procedures. Thus, for the 94 angioplasty patients with one angioplasty procedure and no coronary surgery, the original hospital charge was $8086±3904, the hospital cost was $7946±3975, and the professional charge was $3104±715.
Although total procedural charges and costs for the two groups became closer at 3 years, the professional charges remained higher for the CABG group. There was no difference in total hospital costs, with a trend toward a higher mean value for angioplasty, but a higher median for coronary surgery. Thus, in these patients with multivessel disease, the total hospital and professional charges remained higher for coronary surgery than for coronary angioplasty. More than half of the cost advantage of angioplasty was lost when the charges were reduced to cost. At 3 years, the sum of professional charges and hospital costs was $23 734±15 798 (median, $19 059) for coronary angioplasty and $25 310±7480 (median, $23 572) for coronary surgery, numbers that were close but statistically different (P<.0001). Inflated to 1993 dollars, the sum for angioplasty becomes $36 313±24 171 (median, $29 160); for surgery, $38 724±11 444 (median, $36 065).
Figs 1⇓ and 2⇓ show initial hospital and the total of initial hospital and 3-year costs, respectively, for the cumulative percent of patients. The costs presented are hospital costs and professional charges. In these types of figures, a normal distribution would have a symmetrical sigmoid shape. The initial hospital costs are clearly higher for CABG. In both groups but especially for PTCA patients, the lack of a normal distribution may be noted, with a tail of the last 10% to 15% of cases having costs that are several times the median. At 3 years, the separation between groups is less certain, with more PTCA than CABG patients having lower costs, but the overlap between groups is greater than initially. Thus, the curves actually cross. Although both groups now have a tail of patients with higher costs, this tail is more prominent in the PTCA group.
Fig 3⇓ shows the initial hospital costs plus physician charges by assigned treatment group and strata. For each stratum, the costs were higher for CABG than PTCA. Costs rose slightly with CABG from strata 1 to 4 (P=.013), and there was a slight increase for PTCA (P=.09). Fig 4⇓ shows similar data for the full 3 years. Once again, costs rose slightly for CABG from strata 1 to 4 (P=.028), and there was a slight trend in favor of PTCA (P=.14). At 3 years, the variance in costs for PTCA is much wider than initially and wider than for CABG. This was due to considerable variation in the number of procedures for PTCA patients; most CABG patients had only the original procedure. The costs remain higher with surgery, although they drew closer together by 3 years compared with after the initial hospitalization for each stratum. There remains a difference between angioplasty (mean, $20 875±13 533; median, $16 392) and surgery (mean, $23 639±6848; median, $22 063) for stratum 1, while for stratum 4 the difference between the mean costs of angioplasty (mean, $23 970±10 790; median, $22 944) and surgery (mean, $27 022±7495; median, $25 591) is smaller and of borderline statistical significance.
Multiple regression was used to determine correlates of cost. Table 6⇓ gives correlates of the total hospital costs and physician charges. The clinical variables and complications in Table 1⇑ were considered potential correlates. The most powerful correlate was assignment to the surgical group. The other correlates were a history of congestive heart failure, older age, male sex, a history of hypertension, and trends for angina class and strata. Note that the variables are weak correlates, except for surgical assignment. The multiple r value was .46 and r2 was .21, revealing that only 21% of the variability in cost could be accounted for by these variables. Adding length of stay to the model would increase the r value to .86 and the r2 value to .74. However, length of stay is a composite variable incorporating many other variables; thus, it destroys the ability to examine the individual clinical correlates of cost. Table 7⇓ gives correlates of the 3-year cumulative hospital costs and physician charges. The statistically significant correlates were ejection fraction, a history of hypertension, and male sex, and there were trends for angina class, older age, strata, surgical group, and congestive heart failure. The big difference from initial costs is that assignment to the CABG group changes from being the only powerful correlate to having just a bare trend. In addition, the r value fell to .27 and the r2 value to .07, meaning that only 7% of the variability in cost was accounted for. If the number of angioplasty and coronary surgical procedures was included, r would rise to .75 and r2 to .57. However, this would essentially be a tautology in which procedures predict procedures.
CABG and PTCA are both expensive; >600 000 of these procedures are performed annually11 at a cost of more than $17 billion. There has been uncertainty as to their relative efficacies and costs in the treatment of multivessel coronary artery disease. Results from the 3-year evaluation of EAST reveal no difference in the primary end point or the components of the primary end point—death, Q-wave myocardial infarction, or a large reversible thallium defect—between treatment arms.5 In this investigation, we report on the hospital and professional charges and costs of the two procedures at 3 years, all deflated to the starting year of the trial, 1987. Some costs inflated to 1993 are included for comparison. Direct conversion to expected contemporary costs is not possible because of recent efforts to increase efficiency, technological change, and reimbursement reductions from Medicare and private insurance carriers.
At 3 years, the EAST results present a clear and striking use pattern. During the initial hospitalization, 102 of 198 patients (52%) in the PTCA group were treated with more than one procedure. Thereafter, the high percentage of patients with repeated revascularizations in the PTCA group resulted in a substantial increase in resource use, with 106 additional angioplasties and 23 additional coronary surgeries in the PTCA group. In contrast, after 3 years barely a quarter as many follow-up angioplasty procedures (28) and only one additional coronary surgery were required in the CABG group. As a consequence, at 3 years there remained only a small, albeit statistically significant, difference between the CABG and PTCA groups in the overall composite measure of hospital costs and physician charges. However, procedural costs do not reflect total direct costs. Total direct costs are probably even closer than the cost data presented here because of the additional hospitalizations for angina and additional use of antianginal medications in the PTCA group. As an estimate, if the cost of hospitalization for an angina patient without myocardial infarction and without or before additional revascularization is between $3000 and $6000 and there were 44 more such hospitalizations in the PTCA group, this would raise the mean cost in the PTCA group overall by $666 to $1333. If the total mean direct cost in the angioplasty group increased by $1000 to $2000 more than in the CABG group over 3 years as a result of additional medication and hospitalization for chest pain, then angioplasty would cost from $568 less to $432 more than coronary surgery. On the other hand, the costs of loss of income or other opportunities resulting from more prolonged hospitalization, time lost from work, or medical or nursing care after coronary surgery may be greater. Thus, the uncertainties of measurement are great, and the differences in cost over several years are small.
In terms of initial hospitalization, PTCA patients’ hospital costs were significantly lower than CABG patients’ costs, $11 684 versus $14 579. At 3 years, however, cumulative hospital costs for the two groups were similar. This pattern was seen with hospital costs specifically because the catheterization laboratory at Emory uses a higher cost-to-charge ratio than the operating room or surgical intensive care units and because of the larger number of repeated angioplasties in the PTCA group.
Quality of life measures also suggest some interesting patterns. The less complete revascularization with angioplasty appears to result in more frequent continuing angina in the PTCA group. In addition, fewer patients in this group felt that their recoveries were complete. However, more PTCA than CABG patients were optimistic about their health. This difference between PTCA patients’ perceptions of their current and future health may relate to the more invasive nature of CABG and the resulting psychological impact. Future results from the ongoing EAST should shed more light on the relation between patients’ perceptions of their health and their ultimate functional capabilities.
The EAST results are consistent with those from other trials reported to date,12 13 14 which have shown few clinical differences between the two forms of revascularization in acute myocardial infarction or death, except for more revascularization procedures in the angioplasty group. Although preliminary results of one Argentinean trial15 have found PTCA to be less expensive than CABG and economic outcome data are still expected from some European trials, the comparability of results from the European and South American trials and US trials, with varying clinical practices and reimbursement systems, is unclear and awaits further study. Neither clinical nor economic results from the American multicenter trial16 are available.
Considering the cost of medical procedures, the perspective that would be most meaningful would be overall resources consumed by society as a consequence of the procedure. This will always require some assumptions, and it probably is impossible to assess this in an accurate, thorough manner for complex procedures with consequences extending over several years. The limitations of such a complete cost analysis are detailed below. In lieu of a complete picture from a societal standpoint, we have developed a perspective from the point of view of the providers (hospitals and professionals), with some information about additional resource use from a patient perspective.
In this study, we have presented hospital charges and costs as derived from the UB-82 hospital billing form. Hospital charges are widely recognized as weak proxies for hospital costs, which depend more on institutional accounting peculiarities than uniform standards.17 However, to the degree that hospital charges reflect the relative relation between variables, they are informative and revealing. Multiple accounting methods have been proposed to try to extract costs from charges, although some might argue with limited success.18 The top-down method used in this study assumes a constant cost-to-charge ratio across a department.10 If the cost-to-charge ratio was lower for the catheterization laboratory, then the in-hospital and follow-up hospitalization costs for the angioplasty arm would be lower, while the costs for coronary surgery would be largely unchanged.
In principle, a better method to calculate cost than the top down would be to do accurate cost accounting for each service. Although cost accounting systems are beginning to be used in hospitals, these methods are only as good as the accuracy of the measurements, which may be both complicated and of questionable accuracy and validity. The costs presented in this study are average costs. Another approach would be to calculate marginal cost, the cost of doing one more procedure of a specific type.4 This might or might not be of more interest but would involve yet another set of assumptions about what constitutes marginal cost. Marginal costs are of more interest only if at least some fixed costs could be eliminated. There are also difficulties in determining physician costs. In the present study, we have presented physician charges as a proxy for costs. These charges do not equate to reimbursement. Collections may be expected to be a smaller fraction of charges in the future as Medicare and managed care programs reduce professional reimbursements. Future analyses may examine the Resource Based Relative Value Scale as a measure of professional costs.19 20
Follow-up direct and indirect costs are more difficult to measure than in-hospital costs. Direct costs are all direct medical expenses, including additional procedures, related hospitalizations, office visits, and medications. Indirect costs include additional expenses such as time lost from work. It is difficult to know which costs over several years of follow-up to attribute to a procedure. For instance, is a hospitalization for pneumonia 2 weeks after a myocardial infarction experienced 1 year after angioplasty attributable to the angioplasty? Indirect costs suffer similar difficulties. Is time lost from work merely hours lost times wage? The answer is “yes” if the patient is not paid, but if the patient continues to be paid, the patient suffers no loss of income. Can we attribute the cost to a loss to the employer? The answer is “yes” if the wage is the real loss to the employer. However, the loss to the employer may be either much less or much greater and is unmeasurable in a study such as this one. Thus, in a clinical cost study there must inevitably be a recognition that all follow-up costs cannot be accounted for. What is reported, however, should be complete or at least accounted for. A simple, straightforward approach was taken in the present study: the cost data presented are only for additional revascularization procedures. Collection of these data was straightforward because all but 8 procedures (5.0%) were performed at Emory and could easily be attributed to the original procedure. However, as noted above, the differences in the procedures make it likely that there are differences in follow-up direct and indirect costs. Additional analyses may explore other costs, including costs for lost wages or other opportunities, albeit with more assumptions and uncertainty about the data.
In an era of great changes in medicine, all areas examined in EAST are “moving targets.” Coronary surgery is changing the least, while angioplasty continues to undergo methodological change. Although intervention today includes new devices,21 the impact on overall results has been variable.22 23 Perhaps the greatest changes are in cost. In EAST, 38% of the angioplasties were performed as staged procedures; today, this figure would be much lower. In addition, professional fees are falling. Hospitals are currently making great efforts to become more efficient, and length of stay is decreasing. Thus, interpretation of the EAST economic data over the next several years will have to be done from the perspective of changes in hospital and professional economic trends and secular inflationary trends.
It should be noted that while these analyses are for only 3 years, the EAST patients will be followed for 8 to 11 years. A recent study from Emory24 has shown that over a longer period, up to 12 years, coronary surgery patients also experience increasing rates of repeated revascularizations, which may accelerate the late cost of coronary surgery. Thus, it remains to be seen from an economic perspective which procedure will ultimately show the larger long-term cost increase.
Finally, the results are from a randomized cohort of patients with multivessel disease in a single institution and must be generalized to broader populations with caution. Other institutions may have different cost bases for both hospitalizations and professional fees. Other institutions also may have different practice patterns and use resources, especially for procedures, in somewhat different ways. Nonetheless, in such an important area where virtually no comprehensive data from randomized trials exist, investigations like the present one represent important starting points.
What do these data suggest for medical decision making? At 3 years, it appears that in patients with multivessel disease who are anatomically suitable for either procedure, no significant differences were found in death rates, myocardial infarction, or ischemic potential, while there was little difference in cost. Quality of life measures, especially angina, favored coronary surgery, although patients have to undergo the more difficult procedure to achieve this improved quality of life. Thus, when revascularization is needed in patients for whom either procedure is appropriate, it is essential for physicians and medical personnel to work with patients and families to make the best choice. Some patients may seek to avoid surgery and would prefer the less invasive approach of angioplasty, knowing that they have a greater probability of recurrent symptoms. Conversely, other patients may prefer coronary surgery over angioplasty because of the lower probability of additional revascularization and recurrent symptoms over the first several years after the original procedure.
This investigation underscores the importance of patient satisfaction and quality of life in evaluating the outcomes of these procedures. In light of the growing pressure to contain costs and provide patient satisfaction, data on patients’ subjective evaluations of outcome, symptomatic status, rates of death, myocardial infarction, and additional revascularization and on cost must be considered with each patient’s specific clinical status and point of view to select the appropriate therapy. Because the randomization process eliminates the inevitable selection bias when patients and physicians choose the form of therapy, the clinical trials comparing coronary angioplasty and coronary surgery are providing important data for making sound clinical decisions.
Calculation of Hospital Costs From Charges
The primary sources of information to analyze the UB-82 statements were the cost reports prepared annually by the hospital’s Finance Department. Total charges for specific services and centers are developed by the Department of Financial Planning of each hospital. The method for charge determination is as follows. Once direct costs have been calculated for each procedure, the department-specific indirect cost is added on the basis of the step-down allocation methodology10 (as used in Medicare cost reports), and additional amounts for uncollectibles and profit are added to ensure that the hospital remains solvent.
Table 8⇓ contains a hypothetical translation of a patient’s UB-92 statement to economic cost. The operating room charge of $7987.24 multiplied by the cost-to-charge ratio of 0.627852 yields a cost of $5014.80. The other services and centers on the UB-92 form, except for the patient room charges, are calculated by the same method. General room cost and the intensive care unit are presented as cost per diem rather than left as cost-to-charge ratios. This methodology conforms to microcosting principles and is very similar to the calculations for the other cost centers. As shown in Table 8⇓, the calculations of the costs of the private room and intensive care unit encompass the same cost items (ie, laundry, nursing, and administration) as all other cost centers and services (ie, operating room and anesthesia). However, the total cost for the room units is divided by the total inpatient days attributed to those units, calculating a cost per diem. Thus, when applied to the UB-82 form (Table 8⇓), this cost is multiplied by the total units (days) used by the patient, resulting in the total cost for the use of the unit by the patient.
This work was supported in part by grant R01-HL-33965 from the NHLBI, Bethesda, Md. We thank Chuck Ackers for his tireless efforts with the EAST data set.
Presented in part at the 66th Scientific Sessions of the American Heart Association, Atlanta, Ga, November 8-11, 1993.
- Received December 20, 1994.
- Revision received May 24, 1995.
- Accepted June 23, 1995.
- Copyright © 1995 by American Heart Association
Hlatky MA, Lipscomb J, Nelson C, Califf RM, Pryor D, Wallace AG, Mark DB. Resource use and cost of initial coronary revascularization: coronary angioplasty versus coronary bypass surgery. Circulation. 1990;82(suppl IV):IV-208-IV-213.
King SB III, Lembo NJ, Weintraub WS, Barnhart HX, Kosinski AS, Kutner MH. The Emory Angioplasty Versus Surgery Trial: design, recruitment and baseline description of patients. Am J Cardiol. 1995;75:426-596.
Campeau L. Grading of angina pectoris. Circulation. 1975;54:522-523. Letter.
Douglas JS Jr, King SB III, Roubin GS. Technique of percutaneous transluminal angioplasty of the coronary, renal, mesenteric, and peripheral arteries. In: Hurst JW, Schlant RC, Rackley CE, Sonnenblick EH, Wenger NK, eds. The Heart. 7th ed. New York, NY: McGraw-Hill Publishing Co; 1990:2131-2156.
Deakin EB, Waher MW. Allocating service department costs. In: Cost Accounting. Homewood, Ill: Richard D Irwin Inc; 1987:251-294.
Detailed diagnosis and procedures for patients discharged from short-stay hospitals: United States, 1987. Vital Health Stat 13. 1989;100:181.
Rodriguez A, Boullon F, Perez-Balino N, Paviotti C, Lipandi MIS, Palacios IF. Argentine randomized trial of percutaneous transluminal coronary angioplasty versus coronary artery bypass surgery in multivessel disease (ERACI): in-hospital results and 1-year follow-up. J Am Coll Cardiol. 1993;22:1060-1067.
Hamm CW, Reimers J, Ischinger T, Rupprecht HJ, Berger J, Bleifeld W, for the German Angioplasty Bypass Surgery Investigation. A randomized study of coronary angioplasty compared with bypass surgery in patients with symptomatic multivessel coronary disease. N Engl J Med. 1994;331:1037-1043.
Rodriquez A, Ahualli P, Balino NP, Meie E, Paviotti C, Liprandi MS, Palacios I. Argentine randomized trial of percutaneous transluminal coronary angioplasty versus coronary artery bypass surgery in multivessel disease (ERACI): late costs and three years follow up results. J Am Coll Cardiol. 1994;23:469A. Abstract.
Protocol for the Bypass Angioplasty Revascularization Investigation. Circulation. 1991;84(suppl V):V-1-V-27.
Finkler SA. The distinction between cost and charges. Ann Intern Med. 1982;96:102-109.
Baim DS, Kent KM, King SB III, Safian RD, Cowley MJ, Holmes DR, Roubin GS, Gallup D, Steenkiste AR, Detre K. Evaluating new devices: acute (in-hospital) results from the New Approaches to Coronary Intervention Registry. Circulation. 1994;89:471-481.
Topol EJ, Leya F, Pinkerton CA, Whitlow PL, Hofling B, Simonton CA, Masden RR, Serruys RW, Leon MB, Williams DO, King SB III, Mark KB, Isner JM, Holmes DR Jr, Ellis SG, Lee KL, Keeler GP, Berdan LG, Hinohara T, Califf RM, for the CAVEAT Study Group. A comparison of directional atherectomy with coronary angioplasty in patients with coronary artery disease. N Engl J Med. 1993;329:221-227.
Fischman DL, Leon MB, Baim DS, Schatz RA, Savage MP, Penn I, Detre K, Veltri L, Ricci D, Nobuyoshi M, Cleman M, Heuser R, Almond D, Teirstein PS, Fish RD, Colombo A, Brinker J, Moses J, Shaknovich A, Hirshfeld J, Bailey S, Ellis S, Rake R, Goldberg S, for the Stent Restenosis Study Investigators. A randomized comparison of coronary-stent placement and balloon angioplasty in the treatment of coronary artery disease. N Engl J Med. 1994;331:496-501.