Author + information
- Received December 11, 1996
- Revision received June 10, 1997
- Accepted June 21, 1997
- Published online October 1, 1997.
- Paul T. Vaitkus, MD, FACCABC,*,
- William T. Witmer, MDAC,
- Richard G. Brandenburg, PhDB,
- Susannah K. WellsC and
- Jonathan B. ZehnackerC
- ↵*Dr. Paul T. Vaitkus, Cardiology Division, University Hospitals of Cleveland, Case Western Reserve University, 11100 Euclid Avenue, Cleveland, Ohio 44122.
Objectives. We sought to broaden assessment of the economic impact of percutaneous transluminal coronary angioplasty (PTCA) revascularization salvage strategies by taking into account costs, revenues, the off-setting effects of prevented clinical complications and the effects of payer mix.
Background. Previous economic analyses of PTCA have focused on the direct costs of treatment but have not accounted either for associated revenues or for the ability of costly salvage techniques such as coronary stenting to reduce even costlier complications.
Methods. Procedural costs, revenues and contribution margins (i.e., “profit”) were measured for 765 consecutive PTCA cases to assess the economic impact of salvage techniques (prolonged heparin administration, thrombolysis, intracoronary stenting or use of perfusion balloon catheters) and clinical complications (myocardial infarction, coronary artery bypass graft surgery [CABG] or acute vessel closure with repeat PTCA). To assess the economic impact of various salvage techniques for failed PTCA, we used actual 1995 financial data as well as models of various mixes of fee-for-service, diagnosis-related group (DRG) and capitated payers.
Results. Under fee-for-service arrangements, most salvage techniques were profitable for the hospital. Stents were profitable at almost any level of clinical effectiveness. Under DRG-based systems, most salvage techniques such as stenting produced a financial loss to the hospital because one complication (CABG) remained profitable. Under capitated arrangements, stenting and other salvage modalities were profitable only if they were clinically effective in preventing complications in >50% of cases in which they were used.
Conclusions. The economic impact of PTCA salvage techniques depends on their clinical effectiveness, costs and revenues. In reimbursement systems dominated by DRG payers, salvage techniques are not rewarded, whereas complications are. Under capitated systems, the level of clinical effectiveness needed to achieve cost savings is probably not achievable in current practice. Further studies are needed to define equitable reimbursement schedules that will promote clinically effective practice.
The delivery of interventional cardiology services is expensive, involving both considerable fixed costs (cardiac catheterization laboratories and inpatient hospital facilities) and large variable costs (disposable equipment and concentrated personnel time). Because of the high prevalence of coronary disease and the expenses involved in providing the service, interventional cardiology has appropriately become the subject of considerable scrutiny [1–3]. Previous studies in this area [1, 2, 4–8]have focused exclusively on the single dimension of accounting for costs associated with percutaneous transluminal coronary angioplasty (PTCA). A recent American College of Cardiology consensus document emphasized that previous economic analyses did not account for the cost-savings potential of stenting in preventing emergency coronary artery bypass graft surgery (CABG).
Complications of PTCA such as myocardial infarction, CABG or repeat PTCA [10–14]add considerably to the costs of an interventional program . Various salvage techniques, including either technical tools (intracoronary stenting [16, 17]or perfusion balloon catheters [18, 19]) or pharmacologic strategies (prolonged postprocedure heparin administration or the intracoronary administration of thrombolytic agents [21, 22]) have been applied in an effort to treat or prevent vessel closure. Our aim was to expand the economic analysis by incorporating the additional dimensions of the clinical context within which PTCA is performed; the need to account for costs of multiple simultaneous salvage therapies, associated revenues, savings achieved from prevented complications and the payer mix. Previous economic analyses of PTCA [1–3, 8, 16]have included few or none of these other dimensions. Health care economic research must specify the perspective it adopts and must maintain this perspective consistently . All previous studies [1, 2, 4–8, 15]have implicitly adopted the perspective of the hospital by assessing hospital costs. We adopt the same perspective explicitly. This perspective is relevant to physicians because it is the economic platform on which the medical profession functions. Physicians are both responsible for and dependent on the economic well-being of the hospitals in which they work.
Of 817 consecutive interventional cases during a 13-month period (February 1994 to February 1995), 765 were included in our study. To achieve comparability of data across clinical strata and among salvage modalities, we focused on cases of conventional PTCA. Patients with separate planned PTCAs in different coronary arteries on different days (“staged PTCAs,” n = 22) were excluded because there was no satisfactory method of assigning inpatient hospital costs to each of the two or more PTCA procedures. Primary directional coronary atherectomies (n = 26) were excluded. There were no primary stent implantations during the study. Four patients (none of whom had received a stent or thrombolysis or had sustained any complication) were excluded because their hospital charts could not be located.
Complications included CABG (n = 16), myocardial infarction (n = 15) and sudden out-of-laboratory vessel closure involving successful repeat PTCA (n = 17). We tracked all in-hospital complications of PTCA cases, whether they occurred in the catheterization laboratory or later in the hospital stay. The salvage modalities considered included “bail-out” intracoronary stent implantation (n = 12), prolonged inflation with a perfusion balloon catheter (n = 14), intracoronary administration of thrombolytic agents (n = 15) and prolonged postprocedure heparin administration (n = 184). The latter was defined as heparin administration continued beyond the 1st 24 h after PTCA, which was the standard of care in our institution. Angiographic success was defined as achieving a final stenosis of <50% in at least one lesion with attempted PTCA. Clinical success was defined as angiographic success in the absence of any complications.
Revenue data were obtained from our patient billing department and were complete for all patients. Our mature cost accounting system systematically and comprehensively captures fixed and variable costs by using accepted accounting methods. Each patient care activity has assigned to it a cost that has been derived by summing all identified variable costs as well as an allocated component of fixed costs. The allocation of institutional overhead to cost centers is based on activity-based costing. Subsequent cost center overhead and fixed costs are assigned to specific patient care activities by a collaborative process involving the cost center manager and the cost-accounting staff. The cost accountants ensure that all costs are allocated comprehensively and in accordance with standard cost-accounting methods, whereas the managers assess that costs assigned to specific patient care activities represent a fair and equitable allocation. Cost information is linked to each patient charge code. An individual patient’s cost summary is thus derived from his or her itemized hospital bill.
The contribution margin for each case was calculated as revenues minus costs. However, the relevant figure in cases of complications and salvage treatments is not the total costs or the entire contribution margin but the increment (or decrement) above or below that which would have occurred had the case proceeded in an uncomplicated fashion.
Patients may undergo more than one salvage modality and costs in these cases have to be allocated to each modality. We first analyzed our data set using stepwise regression analysis with costs as the dependent variable and each of the clinical indications, complications and salvage techniques as an independent variable. An F value of 3.96 was accepted as evidence of a significant relation.
Standard statistical techniques are insufficient to correctly identify the incremental costs of complications and salvage techniques. Stepwise logistic regression will identify a statistically significant relation only if the independent variable accounts for a significant amount of the residual variation in the dependent variable in the entire cohort. Interventions that are used relatively infrequently are thus unlikely to appear as statistically significant independent variables. However, it would be incorrect to conclude that there is no incremental cost associated with their use. We thus sought to identify the incremental costs and contribution margin through iterative analysis, a method frequently employed in business to solve problems that can be expressed in terms of two or more equations whose results refer back to one another in a circular fashion . The incremental cost of a given salvage technique was calculated as the average cost of cases in which that technique is used minus the costs attributable to uncomplicated PTCA and costs attributable to any concomitantly used salvage technique. Therefore, the formula for each salvage modality necessarily refers to the formulas for every other modality. The incremental cost can be computed by multiple reiterations of the formulas. We programmed the calculation to proceed through a maximum of 10,000 iterations or until the change between iterations was less than 1 cent. The formulas used are detailed in the Appendix A.
To account for differences in case mix, PTCAs were stratified according to indication, and all analyses were carried out within a given clinical stratum. Patients underwent PTCA for stable angina, unstable angina, postmyocardial infarction, as primary therapy for acute myocardial infarction, cardiogenic shock and “other” indications. Once the incremental costs and contribution margin for each complication and salvage technique were calculated within the clinical strata, they were combined by summing them, weighted for the relative frequency of each clinical indication. We eliminated from analysis the few cases of cardiogenic shock and the PTCAs performed for “other” indications. Only one patient with cardiogenic shock underwent PTCA without complication and without salvage technique, an insufficient basis for determining the costs of uncomplicated, “unsalvaged” PTCA in this clinical stratum. To include patients with cardiogenic shock with those in the group with myocardial infarction would introduce an unfair bias against salvage techniques as the patients with shock invariably had very high costs that could not be construed as due to the salvage techniques. Similarly, we excluded the “other” indication cases because these patients were an extremely heterogeneous group with no clinically meaningful unifying theme. Most of the costs of these patient’s hospital stays were due to treatment of other dimensions of their cardiac disease, such as arrhythmias requiring electrophysiologic evaluations.
To model a purely fee-for-service world, we assigned to each case a reimbursement of a percent of actual charges. The percent used in this calculation was the average percent our institution actually received in cases covered by fee-for-service payers in 1995. To model a diagnosis-related group (DRG) payer environment, we assigned to each case reimbursement based on an average of the two largest DRG payers in our current mix (one being Medicare). Finally, to model a capitated environment, we assumed no revenue associated with any episode of care. We modeled a truly capitated arrangement with a fixed, absolute revenue to the provider organization based on the number of covered lives and not on alternative arrangements such as discounted, fixed payments for each percutaneous intervention. These alternative arrangements, although fairly common for managed care payers and providers, are merely variations on the DRG theme and do not represent true capitation.
Central to our analysis is the concept that salvage techniques are not adequately described by their direct costs and revenues alone. A successful salvage technique will prevent complications and their attendant costs. To model this, we can assume any level of clinical effectiveness. If a salvage modality is 0% effective in preventing complications, it will have no cost offset. A technique that is 100% effective (e.g., if each stent implanted for threatened vessel closure prevents one case of CABG) will have its costs offset by saving the costs of a complication.
Finally, we varied the degree to which the different types of payers had penetrated the market. Although it is possible for us to calculate the contribution margins of a salvage technique by using any proportion of different payers, for clarity and simplicity we restrict our reported results to varying the market mix between two pairs of payers, paralleling the historical developments in the marketplace. We first moved the market from a 100% fee-for-service/0% DRG-based to a 0% fee-for-service/100% DRG-based distribution. The second frontier we modeled was from a 100% DRG-based/0% capitated to a 0% DRG-based/100% capitated environment. We generated a family of curves to demonstrate the combined effects of payer mix and clinical efficacy on cost effectiveness.
The costs per case stratified by clinical indication are presented in Fig. 1. PTCA was performed in 206 patients for stable angina, in 337 for unstable angina, in 182 after myocardial infarction, in 25 as primary therapy for myocardial infarction, in 4 for cardiogenic shock and in 11 for other indications. Our payer mix is summarized in Table 1. Angiographic success was achieved in 668 cases (87.3%) and clinical success in 638 cases (83.4%).
Stepwise regression analysis revealed a significant relation between cost and the following independent variables (figures in parentheses are the regression coefficients): indications—stable angina (−2,783), cardiogenic shock (19,452), “other indication” (8,856); complications—CABG (27,412), myocardial infarction (6,253), repeat PTCA for out-of-laboratory closure (9,698); salvage techniques—prolonged heparin administration (1,544), stents (6,023), thrombolysis (3,693). The regression coefficients represent the dollar amount difference between a PTCA case including the independent variable and the “average” PTCA case.
The total costs of uncomplicated PTCA cases, complicated cases and cases involving salvage techniques are shown in Table 2. Table 2also lists the incremental cost of each complication and salvage technique and the incremental contribution margin of complications and salvage techniques under four payer environments: the 1995 University of Vermont payer mix, purely fee-for-service, DRG-based and capitated payment arrangements.
The finding that one salvage technique—perfusion balloon catheters—had an apparent negative associated incremental cost warranted closer scrutiny. The catheters were almost always used in association with other salvage techniques (23% of patients had stenting, 46% had prolonged heparin use, 23% had thrombolysis and only 31% had no other simultaneous salvage technique), and most of the excess costs were allocated to the other techniques.
Fig. 2summarizes the economic impact of one salvage technique, stenting, as a function of its clinical effectiveness and of the payer mix. Our model demonstrates that in a purely fee-for-service environment, the less clinically effective strategy is the most profitable. As DRG-based payers increase, this situation is rapidly reversed. In a DRG-dominated environment, stenting cannot achieve cost savings for the hospital regardless of its clinical effectiveness because the preventive measure brings no added revenues, whereas the economic impact of complications is driven by the fact that CABG is amply rewarded. Therefore, in a DRG-based environment, it is more profitable to have a mix of complications that includes some CABGs than it is to prevent complications with salvage techniques. The subsequent effects of an increase in capitated payers are summarized in Fig. 2. Salvage techniques become cost-reducing because there are no independent revenues associated with any event, including any of the complications. CABG is no longer profitable; it is purely a source of greater costs. Under a capitated system, the more clinically effective a salvage technique, the more cost saving it is for the hospital. The break-even point for clinical effectiveness for stenting to become cost savings under a capitated system is 51%; that is, if for approximately every two stents implanted, one complication is successfully prevented, the cost savings of prevented complications will pay for the excess costs of the stenting.
In the present study we have quantitated the costs, revenues and contribution margin of several commonly used interventional cardiology salvage strategies and the offsetting effects of prevented complications as a framework for evaluating the economic impact of salvage techniques. We have rendered these analyses more generalizable by creating a model to account for differences in payer mix and different degrees of clinical effectiveness. Previous work in this field raised awareness of the high costs involved with interventional techniques, but it did not account for the revenues, the concomitant use of multiple modalities or the off-setting effects of prevented complications. Whereas most previous discussions of stents has focused on their higher costs than those of conventional PTCA, we demonstrate that under certain conditions, stenting can be cost saving for the hospital by preventing more costly complications. However, applying this concept in clinical practice will prove challenging. As there are no controlled clinical trials of stents as salvage tools (nor of any of the other salvage modalities included in our study), we have only a limited basis for assessing their effectiveness as salvage tools. Two groups of investigators [16, 25]have reported temporal trends of CABG as a complication of PTCA since the advent of stents. Altmann et al. reported a decline in emergency CABG that was almost exactly equal to the number of stents implanted, a clinical effectiveness that approaches 100%. Lindsay et al. reported a similar decline in emergency CABG (∼3%) but a stenting rate that approached 50%, a figure that is closer to current practice at most centers. In clinical practice, it is nearly impossible to precisely gauge the clinical effectiveness of salvage techniques, but it is likely that a practice of stenting in 50% of all PTCAs will not have a sufficient offset of prevented complications to render stents cost saving for the hospitals.
Our results also illustrate how payment schemes can cause clinical and financial incentives to become misaligned. Although stenting is profitable for the hospital under fee-for-service arrangements, under the DRG-based prospective payment system extant during the time of our study, clinically effective behavior was economically detrimental. Whereas complications (i.e., CABG) were financially rewarded, salvage techniques were not. The physician is placed in the position whereby acting in the best interests of the patient is directly counter to the economic interests of the hospital, setting the stage for potential conflicts of interest and contentious relations between hospital administration and physician staff. One solution is to recalibrate the DRG system for stenting to reflect both its ability to prevent complications and its long-term benefits of reducing restenosis. In October 1996 the Health Care Finance Administration recalibrated reimbursement for DRG 112. However, because this higher reimbursement applied equally to conventional PTCA and to stenting, the findings of our analyses are not altered by this recalibration.
Concern has been expressed about the moral hazards that capitation payment systems create. Because under these systems providing less care can be financially rewarding, it is feared that capitation could encourage physicians to withhold care from patients inappropriately . However, in our model, for the specific clinical question we probed, capitation at least in part realigned financial and clinical incentives. Under a capitated system, the hospital that successfully averts CABG as a complication of PTCA when stents are implanted realizes a financial benefit as the patient realizes a clinical benefit. Furthermore, under a capitation system, the providers also realize the long-term benefits of stenting in that fewer repeat revascularizations are necessary , whereas under current fee-for-service and DRG-based arrangements, the loss of a subsequent revascularization represents a loss to the hospital.
3.1 Limitations of the study.
Our study has several limitations:
1. We examined the PTCA process after the point of entry of the patient into the catheterization laboratory. That is, we examined the economic effects of the interventional cardiologist’s behavior and decisions, but we did not examine the economics of the decision of whether or not to undertake percutaneous intervention. Furthermore, we limited our perspective to the initial episode of care. Although capitated contracts and integrated health care delivery systems may eventually erode the relevance of a single episode of care in favor of a longer-term perspective, this is not the case at present. Currently, capitation represents a small proportion of total cardiovascular services, and the episode-of-care perspective remains the dominant payment method and will maintain its relevance for physicians and hospitals for the foreseeable future.
2. We could not account for some indirect costs. If, for example, emergency CABG forces postponement of a scheduled surgical case, the cost of the added length of stay of the patient whose operation was canceled should be assigned to the emergency case. However, we had no means of identifying such indirect costs.
3. Our economic analyses are based on the unique cost structure of our institution, and the cost structure of other hospitals is likely to be different. However, our study provides a framework that can be adapted to local practices and local economic environments.
4. We did not evaluate every PTCA salvage modality (10–13), restricting ourselves to techniques used with a measurable frequency in our institution. Clinical practice will be different in other localities. Interventional cardiology is evolving rapidly and new modalities are being incorporated into clinical practice. For example, novel anticoagulant agents are emerging as an important therapeutic option . However, there are unavoidable temporal delays in gathering economic data, in particular, revenue data.
5. Finally, we did not model either future changes in DRG reimbursement for stenting (or other salvage techniques) or future changes in stent costs. As new stents enter the marketplace, it is likely that stent prices will fall and new break-even points will need to be calculated.
Our study has examined the economic impact of PTCA salvage techniques on provider organizations by considering multiple dimensions including costs, revenues, the offsetting effects of prevented complications, and payer mix. We have expanded economic analysis beyond the isolated consideration of the single dimension of cost and have thus developed a method that can be used to evaluate the costs and savings associated with different interventional strategies and techniques. Under fee-for-service arrangements, stenting is profitable to the hospital at almost any level of clinical effectiveness. Under current DRG systems, stenting will always have a negative economic impact on the hospital, suggesting a need to recalibrate the DRG system for stenting to realign clinical and financial incentives. Under capitated systems, stenting may be cost saving for the hospital at certain levels of effectiveness in preventing complications, but in practice it is nearly impossible to know the level of effectiveness being achieved. Furthermore, at current levels of usage, stenting is unlikely to pay for itself through prevented complications.
A.1 Formulas Used in Iterative Analysis
- K = No. of PTCAs without complications and with no salvage strategy used: K = K1+ K2+ K3+ K4+ K5+ K6, where 1 to 6 denote indications for PTCA (1 = stable angina; 2 = unstable angina; 3 = postmyocardial infarction; 4 = primary therapy for acute myocardial infarction; 5 = cardiogenic shock; 6 = other; These subscripts are similarly applied to all subsequent categories below).
- L = No. of PTCAs with subsequent CABG.
- M = No. of PTCAs with subsequent myocardial infarction.
- N = No. of PTCAs with subsequent out-of-laboratory acute vessel closure and repeat PTCA.
A.1.2 Salvage Strategies
- P = No. of PTCAs followed by prolonged heparin administration alone.
- Q = No. of PTCAs followed by use of perfusion balloon alone.
- R = No. of PTCAs followed by use of thrombolytic agent alone.
- S = No. of PTCAs followed by stenting alone.
- T = No. of PTCAs followed by prolonged heparin administration + perfusion balloon.
- U = No. of PTCAs followed by prolonged heparin administration + thrombolytic agents.
- V = No. of PTCAs followed by prolonged heparin administration + thrombolytic agents + perfusion balloon.
- W = No. of PTCAs followed by thrombolytic agents + perfusion balloon.
- X = No. of PTCAs followed with stenting + perfusion balloon.
- Y = No. of PTCAs followed with stenting + thrombolytic agents.
A.1.3 Cost Calculations
- ATCPTCA= Average total cost associated with PTCA without complications and without salvage strategy; thus,
- ATCPTCA−1= Average total cost of PTCA for stable angina without complications and without salvage strategy.
- TCPi= Total costs of the ith case of PTCA followed by strategy P (prolonged heparin administration) alone;
- TCQi= Total cost of the ith case of PTCA followed with strategy Q (perfusion balloon) alone; etc.
A.1.4 Incremental Costs of Salvage Strategies
Incremental costs (IC) of each salvage therapy are the average costs of all cases involving salvage therapy less the incremental costs attributable to any concomitant salvage therapy less the average cost of an uncomplicated PTCA. Each IC calculation was carried out separately for each of the clinical indications 1 to 6. A weighted average of incremental cost for each salvage therapy was calculated on the basis of relative frequency of PTCA for each indication. Thus, Incremental costs for other salvage therapies were calculated in an identical fashion.
A.1.5 Incremental Costs of Complications
Equations for the incremental costs of myocardial infarction (MI) and out of laboratory acute vessel closure with repeat PTCA paralleled the equations for incremental costs of salvage therapy; that is, the incremental cost of MI was the average total cost of all cases with MI less the incremental cost of any concomitant complication (i.e., repeat PTCA) less the average total cost of an uncomplicated PTCA.
A.1.6 Calculations for Contribution Margins
Contribution margins were calculated in the same manner as cost calculations.
- coronary artery bypass graft surgery
- diagnosis-related group
- percutaneous transluminal coronary angioplasty
- Received December 11, 1996.
- Revision received June 10, 1997.
- Accepted June 21, 1997.
- The American College of Cardiology
- Cohen DJ,
- Breall JA,
- Ho KKL,
- et al.
- Ellis SE,
- Miller DP,
- Brown KJ,
- et al.
- Wolfe MW,
- Roubin GS,
- Schweiger M,
- et al.
- Pilon C,
- Foley B,
- Penn IA,
- Brown RIG
- Weintraub WS,
- Mauldin P,
- Gilbert S,
- Hicks F,
- Canup D
- Lincoff AM,
- Popma JJ,
- Ellis SE,
- Hacker JA,
- Topol EJ
- Popma JJ,
- Redwood SR,
- Chuang YC,
- et al.
- George BS,
- Voorhees WD,
- Roubin GS,
- et al.
- van der Linden LP,
- Bakx ALM,
- Sedney MI,
- Buis B,
- Bruschke AVG
- Friedman HZ,
- Cragg DR,
- Glazier SM,
- et al.
- Vaitkus PT,
- Laskey WK
- ↵Microsoft Corporation. Using iteration to solve a circular reference formula. Microsoft Excel User’s Guide 2, 1993:13–4.
- Altmann DB,
- Racz M,
- Battleman DS,
- et al.
- Cohen DJ,
- Krumholz HM,
- Sukin CA,
- et al.