Eur J Health Econ DOI 10.1007/s10198-015-0670-4
ORIGINAL PAPER
Understanding the economic impact of intravascular ultrasound (IVUS) Alessandro Alberti • Pietro Giudice • Alessandra Gelera Luca Stefanini • Virginia Priest • Michael Simmonds • Christa Lee • Matthew Wasserman
•
Received: 17 June 2014 / Accepted: 26 January 2015 Springer-Verlag Berlin Heidelberg 2015
Abstract Aims To examine the cost-effectiveness of intravascular ultrasound (IVUS) use during percutaneous coronary intervention (PCI) with drug-eluting stents (DES) in treating coronary artery disease (CAD). Methods and results A Markov model was constructed with a lifetime horizon to compare costs and health outcomes between IVUS-guided PCI and PCI guided solely by angiography from an Italian healthcare payer perspective. The population examined included CAD patients undergoing PCI with DES. From a healthcare payer perspective, the resulting incremental cost-effectiveness ratio (ICER) per quality-adjusted life-year was negative in the base-case scenario (i.e., IVUS benefit assumed to persist beyond the first year). When IVUS benefit was assumed to be limited Electronic supplementary material The online version of this article (doi:10.1007/s10198-015-0670-4) contains supplementary material, which is available to authorized users.
to the first year, the ICER increased to €9,624. This conclusion remained consistent even when scenarios varied regarding the duration of the device’s effect. Furthermore, benefits of using IVUS were greater for patients with acute coronary syndrome, renal insufficiency, and diabetes. Conclusions Using IVUS with angiography is a dominant strategy in Italy, and results demonstrate that it is desirable to target those at a greater risk of restenosis (i.e., patients with diabetes, chronic kidney disease, and acute coronary syndrome), who tend to benefit more from accurate stent implantation. Further information is necessary regarding the long-term benefits of IVUS, however sensitivity analysis presented in this research demonstrates a strong argument supporting the cost-effectiveness of IVUS. Keywords Angiography Cost-effectiveness analyses Drug-eluting stent Economic evaluations Intravascular ultrasound Percutaneous coronary intervention
A. Alberti Istituto Clinico Sant’Ambrogio, Milan, Italy
Introduction
P. Giudice San Giovanni Di Dio e Ruggi D’Aragona University Hospital, Salerno, Italy
Background
A. Gelera L. Stefanini Boston Scientific, Milan, Italy V. Priest Boston Scientific, Singapore, Singapore M. Simmonds Boston Scientific, Sydney, Australasia C. Lee M. Wasserman (&) Double Helix Consulting, London, UK e-mail: [email protected]
Coronary artery disease (CAD) is a medical condition associated with narrowing of the coronary arteries. Often led by atherosclerosis, the narrowing consequently reduces blood flow to the heart muscle. CAD is frequently observed in population groups with smoking history, diabetes, hypertension, or hypercholesterolemia, and can result in major adverse cardiac events and even death [1]. CAD represents a high cost to the healthcare system as well as society due to its high prevalence and associated poor health outcomes. For instance, estimated direct and indirect costs of CAD amounted to €45 billion in the European
123
A. Alberti et al.
Union (EU) in 2003, which was a quarter of the entire expenditure for cardiovascular disease [2, 3]. Treatment for CAD For CAD patients, especially those categorized as high risk, angiography would be used if no contraindications exist. According to angiographic results, cardiologists would then choose to treat the patients with medication only, or to perform either a coronary artery bypass graft (CABG) or percutaneous coronary intervention (PCI) [4]. PCI is usually undertaken with either a bare metal stent (BMS) or a drug-eluting stent (DES), to reduce the chance of restenosis. Comparing the efficacy between BMS and DES, a number of studies showed superiority of DES in reducing patients’ need for revascularization [5], especially for patient groups at a greater risk of post-PCI adverse cardiac events [6]. Although DES contributes to reduced restenosis, further reduction is possible with a more accurate implantation of the stent in the artery [7]. The current practice of PCI is often undertaken with angiography alone, which only displays the lumen [5]. The use of intravascular ultrasound (IVUS) along with angiography enables cardiologists to assess the volume, extension, composition, morphology of the plaque and severity of stenosis within the vessel more thoroughly than angiography alone. Thus, this better imaging can assist cardiologists in selecting the appropriate stent for patients’ arterial conditions and aid in optimal stent placement [8–10]. Objectives of this study Acknowledging both potential improvements in health outcomes and additional costs incurred through IVUS use, this study attempts to compare the costs and outcomes between IVUS-guided and angiography-guided PCI in the context of Italy using a deterministic Markov model. It accounts for some of the critical concerns emerging in previous studies including sample size, presentations of patients, the number of endpoints, and time horizon. Also, this study assesses the effect of IVUS only when PCI was completed with DES. An approach to focus solely on DES rather than all stents was chosen as patients implanted with DES are more likely to need IVUS during a PCI procedure than those with BMS as both IVUS and DES target high-risk patients.
Description of the model Type of economic evaluation A deterministic Markov model was constructed to examine the long-term costs and health outcomes of IVUS use in
123
DES placement, comparing between IVUS-guided PCI and PCI guided solely by angiography. With the selected model, this study conducted cost-effectiveness analysis (CEA) and cost-utility analysis (CUA), in which outcomes were presented in both life years gained (LYG) and quality-adjusted life year (QALY) gained, respectively. Patient population The treatment population in the model is based on the patient population examined in a recent meta-analysis by Ahn et al. [11]. Ahn et al. [11] is the most recent metaanalysis on IVUS use in DES implantation to date. The authors investigated both acute and chronic CAD patients going through either IVUS-guided or angiography-guided PCI with DES. Ahn et al. also included patients with diabetes mellitus, renal insufficiency (creatinine clearance \60 ml/min), acute coronary syndrome (ACS), or smoking history, who represent patients at a high risk of post-PCI adverse cardiac events [11]. Perspectives This study took the Italian healthcare payer perspective. Due to the presence of hospital staff, informal care is not a critical factor in costing for IVUS, which is utilized in inpatient settings. Also, productivity loss is often small for elderly patients. Given the constraints and circumstances in which IVUS would be employed, this study considered accounting for direct costs while excluding indirect costs to be a reasonable approach. Model design, health states, and time horizon The model began with a hypothetical cohort of patients with an average age of 63, the average age reported by Ahn et al. [11]. All patients underwent PCI with DES, guided either by IVUS and angiography, or angiography alone. However, patients who developed restenosis were not considered for IVUS during their additional treatments. Such assumption was made because none of the relevant clinical trials indicated that IVUS accompanied subsequent treatments for post-PCI adverse events [11]. The Markov transition model included four health states: event-free, myocardial infarction (MI), revascularization [defined as a composite between target lesion revascularization (TLR) and target vessel revascularization (TVR)], and death. Since both medical therapy and interventions (PCI or CABG) can treat MI, this model made a distinction between the two treatment methods (Fig. 1). The baseline analysis assessed the use of IVUS with a lifetime horizon in order to evaluate outcomes over the full life expectancy of the patients, assuming the benefits of
Understanding the economic impact of intravascular ultrasound MI Revasc Event Free MI (Medical Therapy)
Dead
MI MI + Revasc IVUS+Angio
MI Re va sc
De a d
MI
Re va sc
Re va sc
De a d
De a d
CAD_DES MI Revasc Event Free MI (Medical Therapy)
Dead
MI MI + Revasc Angio
MI Re va sc
De a d
MI
Re va sc
Re va sc
De a d
De a d
Fig. 1 Deterministic Markov model for IVUS use during PCI with DES implantation. Note: the model ends after 38 years. Markov transition shown in this figure continues until the model ends
IVUS would persist throughout their lifetime [12]. The model used monthly cycles for the first year and full-year cycles for the remaining years in order to keep consistency with the BMS model, where there was clinical evidence to support a higher event rate in the initial months of post-PCI (see Appendix A for BMS results). Discount rates of 3 % per annum were applied to both costs and outcomes [12].
simulations were run based on probability distributions described in Table 3. These results were then used to create an incremental cost-effectiveness plane.
Model inputs Treatment effect
Scenario and sensitivity analyses Although the benefits of IVUS continue throughout a patient’s lifetime in the base-case analysis, the duration of IVUS effect is uncertain because of the short follow-up of the studies included in Ahn et al. [11]. Therefore, a scenario analysis was conducted, assuming that the benefits of IVUS would be limited to the first year (for further information, see ‘‘Model Input’’ section). For all parameters, one-way sensitivity analyses (OWSA) were completed using the 95 % confidence intervals reported in the literature (Table 3). Where 95 % confidence interval was not available, a 10 % variation from the mean was used. For those inputs that greatly influenced the incremental cost-effectiveness ratio (ICER) in their OWSA, the study conducted two-way sensitivity analyses (TWSA). Probabilistic sensitivity analysis (PSA) was also undertaken to present overall uncertainty of the model on the cost-effectiveness of IVUS. PSA allows all parameter values to vary simultaneously when sampled from a prioridefined probability distributions [13]. One-thousand
The probabilities of revascularization and MI for the angiography arm were estimated based on the results of COMPARE trial as paclitaxel-eluting and everolimus-eluting stents remain the most widely used forms of DES worldwide and the Ahn et al. [14] meta-analyses include trials and studies investigating both types of stents (Table 1). In order to obtain the estimates for the same variables in the IVUS arm, risk ratios (RRs) for experiencing MI and revascularization between IVUS and angiography groups were multiplied to the corresponding pooled probabilities produced from the COMPARE trial [11, 14] (for all model inputs, see Table 3). A supplementary analysis was also conducted evaluating only everolimus-eluting stents given their prevalent use in the Italian market. The baseline analysis assumed the benefits of IVUS would persist because precise DES implantation may lower rates of restenosis and cardiac events in the longer term [8] (Table 2). However, for the short duration of the trials included in Ahn et al., additional analysis was undertaken, assuming that the benefits of IVUS for MI and revascularization would be restrained to the first year.
123
A. Alberti et al. Table 1 Pooled base-case probability estimates [14] Clinical endpoint
Everolimus
Paclitaxel
Pooled probability (%)
1-month probability
MI
25/897
49/903
74/1,800 (4.11 %)
0.35 %
Revascularization
39/897
97/903
136/1,800 (7.56 %)
0.65 %
Using the results from the first-year follow-up, this table illustrates the pooling process. These estimates were applied to the first year of the current model in the angiography arm. The same method was applied to the data obtained from the second-year follow-up
Table 2 Extrapolated post-year-one probabilities [11, 14] Probability of having MI
Probability of having revascularization
Base-case analysis: continuation of IVUS benefits beyond the first year IVUS arm
0.92 %
2.23 %
Angiography arm 1.61 % 2.72 % Scenario analysis: benefits of IVUS limited to the first year Both arms
1.61 %
2.72 %
All tables present values rounded up to the second decimal place. Only when necessary, they were reported up to the third decimal. However, for costs and ICERs, values were rounded up to a whole number. When these measures were put into the model, those figures before rounding were inputted. Numbers surrounded by round parentheses are negative values The extrapolated probabilities were applied to the all cycles after the first year
Costs Costs were derived by weighting DRG tariffs with the distribution of the number of PCI reported in each DRG code. The government publications such as the national formulary provided information on medication costs for managing post-MI treatment [24, 25]. All cost values were adjusted and presented in 2013 euros. Summary of model assumptions This model made assumptions throughout the analyses because of data limitations or uncertainty. The ‘‘Results’’ and ‘‘Discussion’’ sections discuss the implications of the major assumptions. I.
•
Treatment effect: subgroups A number of studies concluded that clinical experience of patients with particular comorbid conditions would differ from that of the general population [15, 16]. Acknowledging those risk factors, this study conducted subgroup analyses for patients with diabetes, renal insufficiency, and ACS. The hazard ratios (HRs) derived from ADAPT-DES [6] were applied to the probabilities of having MI in both arms. The ratios concerning TLR/TVR were not available for these subgroups; thus, they were set at one in the baseline analysis (Table 3).
Assumptions about the model:
• II.
Assumptions about efficacy inputs: •
• III.
The average age of the patients who undergo PCI with DES implantation is 63 years. The model ends after about 38 years. PCI does not employ IVUS after the first cycle of the model.
The probabilities of having MI and undergoing revascularization remain consistent from the second year after the patients’ initial PCI. Designs of DES do not influence the efficacy of the procedure.
Assumptions about health outcomes:
Mortality
•
The background (natural) death probability was calculated using lifetime tables for Italy from World Health Organization [17]. Cardiac death probabilities were mainly derived from Gan et al. [18] and Serruys et al. [19].
•
The utilities associated with each health state are the same between the two arms. The disutility associated with MI and revascularization persists for 1 year.
Outcomes
Results
Overall LYG related to IVUS use was calculated using the difference in life expectancy between patients in each arm [18, 19]. Utility weights from the literature were used to calculate the corresponding outcomes in QALYs (Table 3) [20–23].
Baseline analyses
123
From the healthcare payer perspective, IVUS is a dominant treatment option, providing improved outcomes at lower costs when the benefits of IVUS continue beyond 1 year
Understanding the economic impact of intravascular ultrasound Table 3 Model parameters, data sources, and values used in the model Variable
Mean
Lower value
Upper value
PSA distribution
Source
0.57
0.44b
0.75b
Gamma
[11]
0.82
b
0.70
0.97b
1.48
1.15b
1.89b
Gamma
[6]
1.64
b
1.19
2.27b
1.41
1.10b
1.80b
With IVUS
0.20 %
0.18 %
0.22 %
Beta
[14]
With angiography
0.35 %
0.31 %
0.38 %
With IVUS
0.54 %
0.48 %
0.59 %
Beta
[14]
With angiography
0.65 %
0.59 %
0.72 %
With IVUS
0.92 %
0.83 %
1.01 %
Beta
[14]
With angiography
1.61 %
1.45 %
1.77 %
With IVUS
2.23 %
2.01 %
2.46 %
Beta
[14]
With angiography
2.72 %
2.45 %
2.99 %
Relative risk (RR) for MI and revascularization MI RRa a
Revascularization RR
HR for subgroup analyses Diabetes HRc Renal insufficiency HR
c
ACS HRc Probabilities Probability of having MI: year one (per month)
Probability of having revascularization: year one (per month)
Probability of having MI: post year oned
Probability of having revascularization: post year oned
Probability of death after MI
17.96 %
16.16 %
19.76 %
Beta
[18]
Probability of death after CABG or PCI
3.54 %
3.19 %
3.89 %
Beta
[19]
Probability of MI being treated with revascularizatione,f
50 %
45 %
55 %
Beta
Assumed
Probability of revascularization being treated with CABGe,f
10 %
9%
11 %
Beta
Assumed
Discounting rates for costs and outcomes
3%
0%
8%
–
[12]
Utilities No adverse events occurred after the procedure
0.85
0.77
0.94
Beta
[20]
Treated for MI
0.68
0.61
0.75
Beta
[21, 22]
Treated for revascularization with PCI
0.77
0.69
0.85
Beta
[20]
Treated for revascularization with CABG
0.73
0.66
0.80
Beta
[23]
Death
0
0
0
–
IVUS
650
585
715
Gamma
Initial PCI with angiography and IVUS
7,084
6,376
7,792
Initial PCI with angiography alone
6,434
5,791
7,077
MI treated with medications
3,875
3,487
4,262
MI treated with PCI
8,128
7,315
8,941
MI treated with CABG
22,125
19,913
24,338
Revascularization treated with PCI
6,434
5,791
7,077
Costs (€)
a
Revascularization treated with CABG
19,112
17,201
21,023
Medication to manage post-MI
511
460
562
[24, 25]
p values of RRs for revascularization and MI were 0.022 and \0.001 individually
b
Lower and upper bound of 95 % CI
c
p values of HRs for subgroups of patients with diabetes, renal insufficiency, and ACS were 0.0021, 0.0025 and 0.0064, respectively
d
The post-year-one probabilities shown in the table were based on the assumption that benefits of IVUS would persist throughout the cycles (Table 2)
e
Only two out of the four country-adjusted probabilities were displayed. The others, the probability of MI treated with medications and that of having revascularization with PCI, were calculated as follows: 100 less each of the two variables included in the table (%)
f
These clinical probabilities were obtained and confirmed through consultations with local clinicians
123
A. Alberti et al. Table 4 Summary lifetime ICERs in baseline and subgroup analyses IVUS
Angiography
Incremental
12.200
0.217
Cost per QALY gained
(€3,430)a
IVUS was more cost-effective in patients with co-morbid conditions who are at a higher risk of cardiac events: the ICERs were significantly lower for these subgroups than those derived from the general patient population. These results further support the reimbursement of IVUS especially for the patients with diabetes, renal insufficiency and ACS (Table 4). Supplementary analyses on BMS
General patient population (baseline) Cost
€11,182
€11,928
(€746)
LYs
15.180
14.941
0.240
Cost per LYG
(€3,112)a
QALYs
12.417
Subgroup analyses
Diabetes mellitus Cost
€11,710
€12,759
(€1,049)
LYs
15.046
14.714
0.332
Cost per LYG
(€3,160)a
11.997 Cost per QALY gained
0.300 (€3,493)a
€13,024
(€1,142)
QALYs
12.297
Renal insufficiency Cost
€11,881
LYs
15.002
QALYs
12.257
14.640
0.362
Cost per LYG
(€3,156)a
11.930
0.327
Cost per QALY gained
(€3,491)a
€12,642
(€1,007)
ACS Cost
€11,635
LYs
15.066
QALYs
12.314
14.747
0.319
Cost per LYG
(€3,160)a
12.026
0.288
Cost per QALY gained
(€3,492)a
General patient population, using only everolimus-eluting stents Cost
€9,825
€10,170
(€345)
LYs
15.305
15.139 Cost per LYG
0.166 (€2,077)a
QALYs
12.534
12.383
0.151
Cost per QALY gained
(€2,288)a
a Negative ICERs imply that IVUS-guided PCI with DES is a dominant treatment option compared to angiography-guided PCI with DES
Recognizing that IVUS and DES both aim to treat high-risk patients, this current study focused on DES implantation only. Nonetheless, given that patients who are treated with BMS may still benefit from IVUS-guided stent placement, we also assessed the cost-effectiveness of IVUS use in BMS placement. The relevant inputs and results can be found in Appendix A. Supplementary analyses on everolimus-eluting stents Given that in Italy everolimus-eluting stents are predominantly used, a supplementary analysis was undertaken to evaluate the cost-effectiveness when only these stents are used (Table 4). Given that the final ICER for the general patient population is negative, IVUS use would remain as a dominant option for the chosen subgroups. Results of paclitaxel-eluting stents can be found in Appendix B. Results for scenario and sensitivity analyses This study conducted scenario and sensitivity analyses to determine the robustness of the presented results. This section only refers to the ICER per QALY gained for Italy because the WTP threshold for cost per LYG is unknown. Benefits of IVUS
(Table 4). The negative ICER value implies that IVUS use during PCI with DES is a superior strategy regardless of the ICER threshold; the €25,000 per QALY is often employed in Italy as its implicit willingness-to-pay (WTP) threshold [26, 27]. Since the data were extrapolated beyond the observation period, the study assessed final results at shorter time intervals in addition to the full lifetime horizon. A longer time frame contributed to lower ICERs because the costdifference between the two arms diminished over time while the positive impact of IVUS guidance continued. The ICER was estimated at €166,641 at a 1-year time interval; however, it fell below the threshold between 3- and 5-year time intervals, at which the ICERs were €26,967 and €7,409, respectively.
123
Even with a conservative assumption that the benefits of IVUS are restrained to the first year, the ICER remained below the WTP threshold (Table 5).
Table 5 The resulting ICERs per QALY with limited benefits of IVUS IVUS
Angiography
Incremental
The benefits of IVUS limited to the first year Cost
€12,354
€11,928
€426
QALYs
12.244
12.200
0.044
Cost per QALY gained
€9,624
Understanding the economic impact of intravascular ultrasound €-
€ 10,000.0 € 20,000.0 € 30,000.0 € 40,000.0 € 50,000.0 € 60,000.0 € 70,000.0
Probability of MI (angiography) (post year 1) Probability of MI (IVUS) (post year 1) Probability of revasc. (IVUS) (post year 1) Probability of revasc. (angiography) (post year 1)
Variable
Fig. 2 Tornado diagram of the scenario analysis (i.e., the benefits of IVUS limited to the first year). Note: The light and dark blue bars show the direction of ICER change for the upper and lower input bounds, respectively. The vertical axis that crosses the tornado diagram above indicates the base-case ICER of €9,624 in the scenario in which the benefits of IVUS are limited to the first year (color figure online)
MI RR Discounng rate for outcomes Revasc. RR No event ulity Probability of death aer MI if a Man Discounng rate for costs
Parameters None of the parameters significantly affected the final results. Even in the scenario analysis, where the benefits of IVUS were limited to the first year, the ICERs were sensitive to the following parameters only: the probabilities of having MI (post year one) and the cost of initial PCI with IVUS and angiography (Fig. 2). Variations in the two cost parameters made ICERs negative, further supporting the base-case outcome that IVUS is a dominant strategy: the cost of initial PCI with IVUS and angiography and the cost of initial PCI with angiography. The results from OWSA clearly identified the drivers of this model. While TWSA was preformed, results do not provide any further granularity or deviation from OWSA. PSA PSA identifies the optimal clinical strategy and the associated probability the strategy is cost effective given relevant WTP thresholds. As demonstrated in the incremental
3,000
ICERs generated from PSA
2,000
Mean ICER
Incremental Costs
4,000
1,000 0 -0.1
0
0.1
0.2
-1,000 -2,000 -3,000 -4,000 -5,000 -6,000
Incremental QALYs
Fig. 3 Incremental cost-effectiveness plane
0.3
0.4
0.5
cost-effectiveness plane in Fig. 3, about 71 % of the simulations fell into the south–east quadrant, implying that IVUS use is a dominant option. The remaining 29 % fell into the north–east quadrant; however, the resulting ICERs were always below €18,000. From an Italian healthcare payer perspective, the choice PCI guided by both IVUS and angiography is always the optimal strategy given the €25,000 WTP threshold.
Discussion This study demonstrates that, in Italy, using IVUS in addition to angiography is a favorable strategy over employing angiography alone during PCI with DES. Covering the full life expectancy of the patients, this study produced negative ICER. The final ICER remained lower than the implicit WTP threshold even when the duration of IVUS benefit was assumed to be limited to the first year. IVUS use in patients at a higher risk of restenosis was more costeffective than use in the general patient population. Furthermore, positive results of the ADAPT-DES study at 1 year appear to prolong to 2-year follow-up and therefore the benefits of IVUS continuing beyond the first year may be a more plausible assumption [6, 28]. Incorporating a number of potential scenarios and mitigating the common problems emerging in the earlier literature, this study improves the evidence on the impact of IVUS in terms of both costs and health outcomes. Despite these strengths, the results were sensitive to the length of the assessed time horizon and the key parameters: the probabilities of having MI (post year-one) and the cost of initial PCI with IVUS and angiography. In addition to the uncertainty with the results, this study suffers from several limitations due to data unavailability, modeling, and estimation assumptions. First, in this study, it was assumed using the 2-year follow-up data from the COMPARE trial that the
123
A. Alberti et al.
probabilities of having MI and TLR/TVR in both arms would remain constant after 2 years following the event. This assumes that the chance of having subsequent events remains the same as the number of successive treatments increases [14]. Nevertheless, the error on the final ICERs should be minimal in that a majority of the patients would not need numerous subsequent treatments for restenosis. Moreover, this study supposed that types of the drugs coated on the stents would not affect efficacy of the procedure, since Smits et al. [14] only incorporated two designs of DES in their examinations, while cardiologists employ other designs in their clinical practice. However, although the validity of their argument is still controversial, some researchers advocate for the superiority of everolimus-eluting stents over paclitaxel-eluting stents [29, 30]. Given the near-ubiquitous use in Italy, the model was tested for sensitivity in limiting usage to everolimus-eluting stents, and for high-risk patient groups, lifetime ICERs were found to be consistent with base-case estimates. These stratified results should only be considered secondarily, as the clinical inputs from Ahn et al. [11] included a mixture of both stent types. Owing to the paucity of IVUS-specific information, the analysis assumed the same utilities for patients in each arm. Nevertheless, the users of IVUS insist that the patients with IVUS-driven PCI have a better experience than those using angiography alone because better imaging serves improved stent implantation [7]. If this claim is supported with more clinical evidence, IVUS would become more costeffective. Furthermore, the study only incorporated direct costs incurred to the healthcare payers because of the data constraints and the settings in which the device would be utilized. For the examined population of this study, productivity loss of caregivers can be considerable given the patients’ average age. Nonetheless, the underestimation of costs should not influence the ICERs if the degrees of inaccuracy in both arms were similar. Also, because none of the clinical trials specified that IVUS was used for subsequent treatments, this device only accompanied PCI of the first cycle [11]. If cardiologists frequently employ IVUS for post-PCI cardiac issues in their daily practices, the estimated costs in the IVUS arm, and thus, the final ICERs might be underestimated.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
Conflict of interest Dr. Alessandro Alberti received consulting fees from Boston Scientific. Dr. Pietro Giudice: none to declare. 12.
References 1. Hill, RA., Boland, A., Dickson, R., Dundar, Y., Haycox, A., McLeod, C., Mujica Mota, R., Walley, T., Bagust, A.: Drug-
123
13.
eluting stents: a systematic review and economic evaluation. Health Technol. Assessment November. Vol. 11, No.46 (2007) Cleveland Clinic Center for Continuing Education. Coronary Artery Disease. http://www.clevelandclinicmeded.com/medi calpubs/diseasemanagement/cardiology/coronary-artery-disease. Accessed 14 July 2013 Leal, J., Luengo-Fernandez, R., Gray, A., Petersen, S., Rayner, M.: Economic burden of cardiovascular diseases in the enlarged European Union. Eur. Soc. Cardiol. 27, 1610–1619 (2006) National Institute for Health and Care Excellence Unstable angina and NSTEMI: the early management of unstable angina and non-ST-segment-elevation myocardial infarction. [CG94]. London: National Institute for Health and Care Excellence (2010) De LUCA, G., Dirksen, M.T., Spaulding, C., Kelbaek, H., Schalij, M., Thuesen, L., van der Hoeven, B., Vink, M.A., Kaiser, C., Musto, C., Chechi, T., Spaziani, G., Diaz de la Llera, L.S., Pasceri, V., Di Lorenzo, E., Violini, R., Cortese, G., Suryapranata, H., Stone, G.W.: Drug-eluting vs. bare-metal stents in primary angioplasty: a pooled patient-level meta-analysis of randomized trials. Br. Med. J. 172(8), 611–621 (2012) Witzenbichler, B., Maehara, A., Weisz, G., Neumann, F.J., Rinaldi, M., Metzger, D.C., Henry, T., Cox, D., Duffy, P., Brodie, B., Stuckey, T., Mazzaferri, E., Xu, K., Mehran, R., Praise, H., Mintz, G., Stone, G.: Relationship between intravascular ultrasound and guidance and clinical outcomes after drug-eluting stents: the assessment of dual antiplatelet therapy with drug-eluting stents (ADAPT-DES) study. Circulation 129(4), 463–470 (2014) Rogacka, R., Latib, A., Colombo, A.: IVUS-guided stent implantation to improve outcome: a promise to be fulfilled. Curr. Cardiol. Rev. 5(2), 78–86 (2009) Gaster, A.L., Slothuus Skjoldborg, U., Larsen, J., Korsholm, L., von Birgelen, C., Jensen, S., Thayssen, P., Pedersen, K.E., Haghfelt, T.H.: Continued improvement of clinical outcome and cost effectiveness following intravascular ultrasound guided PCI: insights from a prospective, randomised study. Heart 89, 1043–1049 (2003) Biondi-Zoccai, G., Sheiban, I., Romagnoli, E., De Servi, S., Tamburino, C., Colombo, A., Burzotta, F., Presbitero, P., Bolognese, L., Paloscia, L., Rubino, P., Sardella, G., Briguori, C., Niccoli, L., Franco, G., Di Girolamo, D., Piatti, L., Greco, C., Capodanno, D., Sangiorgi, G.: Is intravascular ultrasound beneficial for percutaneous coronary intervention of bifurcation lesions? Evidence from a 4,314-patient registry. Clin. Res. Cardiol. 100(11), 1021–1028 (2011) Claessen, B.E., Mehran, R., Mintz, G.S., Weisz, G., Leon, M.B., Dogan, O., de Ribamar Costa Jr, J., Stone, G.W., Apostolidou, I., Morales, A., Chantziara, V., Syros, G., Sanidas, E., Xu, K., Tijsse, J.G., Henriques, J.P., Piek, J.J., Moses, J.W., Maehara, A., Dangas, G.D.: Impact of intravascular ultrasound imaging on early and late clinical outcomes following percutaneous coronary intervention with drug-eluting stents. J. Am. Coll. Cardiol. 4, 974–981 (2011) Ahn, J.M., Kang, S.J., Yoon, S.H., Park, H.W., Kang, S.M., Lee, J.Y., Lee, S.W., Kim, Y.H., Lee, C.W., Park, S.W., Mintz, G.S., Park, S.J.: Meta-analysis of outcomes after intravascular ultrasound-guided versus angiography-guided drug-eluting stent implantation in 26,503 patients enrolled in three randomized trials and 14 observational studies. Am. J. Cardiol. 113(8), 1338–1347 (2014) International Society for Pharmacoeconomics and Outcomes Research. Pharmacoeconomic Guidelines Around the World. http://www.ispor.org/PEguidelines/index.asp. Accessed 20 July 2013 Briggs, A., Weinstein, M., Fenwick, E., Karnon, J., Sculpher, M., Paltiel, D., ISPOR-SMDM Modeling Good Research Practices
Understanding the economic impact of intravascular ultrasound
14.
15.
16.
17.
18.
19.
20.
21.
Task Force: Model parameter estimation and uncertainty: a report of the ISPOR-SMDM modeling good research practices task froce-6. Value Health 15, 835–842 (2012) Smits, P., Kedhi, E., Royaards, K.J., Joesoef, K., Wassing, J., Rademaker-Havinga, T.A., McFadden, E.: 2-Year follow-up of a randomized controlled trial of everolimus- and paclitaxel-eluting stents for coronary revascularisation in daily practice. J. Am. Coll. Cardiol. 58(1), 11–18 (2011) Mathew, V., Gersh, B., Williams, B., Laskey, W., Willerson, J., Tilbury, T., Davis, B., Holmes, D.: Outcomes in patients with diabetes mellitus undergoing percutaneous coronary intervention in the current era. A report from the prevention of restenosis with tranilast and its outcomes (PRESTO) trial. J. Am. Heart Assoc. 109, 476–480 (2004) Garg, P., Normand, S., Silbaugh, T., Wolf, R., Zelevinsky, K., Lovett, A., Varma, M., Zhou, Z., Mauri, L.: Drug-eluting or baremetal stenting in patients with diabetes mellitus: results from the Massachusetts Data Analysis Center Registry. J. Am. Heart Assoc. 118, 2277–2285 (2008) World Health Organization (WHO). Global Health Observatory Data Repository: Life Tables. http://www.who.int/gho/mortality_ burden_disease/life_tables/life _tables/en. Accessed 1 Aug 2013 Gan, S.C., Beaver, S.K., Houck, P.M., MacLehose, R.F., Lawson, H.W., Chan, L.: Treatment of acute myocardial infarction and 30-day mortality among women and men. N. Engl. J. Med. 343, 8–15 (2000) Serruys, P.W., Morice, M.C., Kappetein, A.P., Colombo, A., Holmes, Mack, M.J., Stahle, E., Feldman, T.E., van den Brand, M., Bass, E.J., Van Dyck, N., Leadley, K., Dawkins, K.D., Mohr, F.W., SYNTAX Investigators: Percutaneous coronary intervention versus coronary-artery bypass grafting for severe coronary artery disease. N. Engl. J. Med. 360(10), 961–972 (2009) Shrive, F., Manns, B., Galbraith, D., Knudtson, M., Ghali, W.: Economic evaluation of sirolimus-eluting stents. Can. Med. Assoc. J. 172, 345–351 (2005) Dhalla, I., Smith, M., Choudhry, N., Denburg, A.: Costs and benefits of free medications after myocardial infarction. Health Policy 5(2), 68–86 (2009)
22. Clarke, P., Gray, A., Holman, R.: Estimating Utility Values for Health States of Type 2 Diabetic Patients Using the EQ-5D (UKPDS 62). Med. Decis. Mak. 22(4), 340–349 (2000) 23. Puskas, J.D., Williams, W.H., Mahoney, E.M., Huber, P.R., Block, P.C., Duke, P.G., Staples, J.R., Glas, K.E., Marshall, J.J., Leimbach, M.E., McCall, S.A., Petersen, R.J., Bailey, D.E., Weintraub, W.S., Guyton, R.A.: Off pump vs. conventional coronary artery bypass grafting: early and 1 year graft patency, cost, and quality of life outcomes. J. Am. Med. Assoc. 291(15), 1841–1849 (2004) 24. Nello Martini.: The New National Pharmaceutical Formulary 2003. Ministry of Health. 2003 25. AIFA Agenzia Italiana del Farmaco. http://www.agenziafarmaco. gov.it. Accessed 20 July 2013 26. Mencacci, C., Di, Sciascio G., Katz, P., Ripellino, C.: Cost-effectiveness evaluation of escitalopram in major depressive disorder in Italy. J. ClinicoEcon. Outcomes Res. 5, 87–99 (2013) 27. Lazzaro, C., Bordonaro, R., Cognetti, F., Fabi, A., De Placido, S., Arpino, G., Marchetti, P., Botticelli, A., Pronzato, P., Martelli, E.: An Italian cost-effectiveness analysis of paclitaxel albumin (nabpaclitaxel) versus conventional paclitaxel for metastatic breast cancer patients: the COSTANza study. J. ClinicoEcon. Outcomes Res. 5, 125–135 (2013) 28. Maehara, A., Weisz, G., Neumann, F., Metzger, C., Henry, T., Cox, D. et. al.: TCT-68 does IVUS reduce stent thrombosis with DES? Two-year results from the prospective multicentre ADAPT-DES study. J. Am. Coll. Cardiol. B21–B22 (2013) 29. Stone, G.W., Midei, M., Newman, W., Sanz, M., Hermiller, J.B., Williams, J., Farhat, N., Mahaffey, K.W., Cutlip, D.E., Fitzgerald, P.J., Sood, P., Su, X., Lansky, A.J.: SPIRIT III investigators.: comparison of an everolimus-eluting stent and a paclitaxel-eluting stent in patients with coronary artery disease: a randomised trial. J. Am. Med. Assoc. 299(16), 1903–1913 (2008) 30. Joner, M., Nakazawa, G., Finn, A.V., Quee, S.C., Coleman, L., Acampado, E., Wilson, P.S., Skorija, K., Cheng, Q., Xu, X., Gold, H.K., Kolodgie, F.D., Virmani, R.: Endothelial cell recovery between comparator polymer-based drug-eluting stents. J. Am. Coll. Cardiol. 52, 333–342 (2008)
123
ORIGINAL PAPER
Understanding the economic impact of intravascular ultrasound (IVUS) Alessandro Alberti • Pietro Giudice • Alessandra Gelera Luca Stefanini • Virginia Priest • Michael Simmonds • Christa Lee • Matthew Wasserman
•
Received: 17 June 2014 / Accepted: 26 January 2015 Springer-Verlag Berlin Heidelberg 2015
Abstract Aims To examine the cost-effectiveness of intravascular ultrasound (IVUS) use during percutaneous coronary intervention (PCI) with drug-eluting stents (DES) in treating coronary artery disease (CAD). Methods and results A Markov model was constructed with a lifetime horizon to compare costs and health outcomes between IVUS-guided PCI and PCI guided solely by angiography from an Italian healthcare payer perspective. The population examined included CAD patients undergoing PCI with DES. From a healthcare payer perspective, the resulting incremental cost-effectiveness ratio (ICER) per quality-adjusted life-year was negative in the base-case scenario (i.e., IVUS benefit assumed to persist beyond the first year). When IVUS benefit was assumed to be limited Electronic supplementary material The online version of this article (doi:10.1007/s10198-015-0670-4) contains supplementary material, which is available to authorized users.
to the first year, the ICER increased to €9,624. This conclusion remained consistent even when scenarios varied regarding the duration of the device’s effect. Furthermore, benefits of using IVUS were greater for patients with acute coronary syndrome, renal insufficiency, and diabetes. Conclusions Using IVUS with angiography is a dominant strategy in Italy, and results demonstrate that it is desirable to target those at a greater risk of restenosis (i.e., patients with diabetes, chronic kidney disease, and acute coronary syndrome), who tend to benefit more from accurate stent implantation. Further information is necessary regarding the long-term benefits of IVUS, however sensitivity analysis presented in this research demonstrates a strong argument supporting the cost-effectiveness of IVUS. Keywords Angiography Cost-effectiveness analyses Drug-eluting stent Economic evaluations Intravascular ultrasound Percutaneous coronary intervention
A. Alberti Istituto Clinico Sant’Ambrogio, Milan, Italy
Introduction
P. Giudice San Giovanni Di Dio e Ruggi D’Aragona University Hospital, Salerno, Italy
Background
A. Gelera L. Stefanini Boston Scientific, Milan, Italy V. Priest Boston Scientific, Singapore, Singapore M. Simmonds Boston Scientific, Sydney, Australasia C. Lee M. Wasserman (&) Double Helix Consulting, London, UK e-mail: [email protected]
Coronary artery disease (CAD) is a medical condition associated with narrowing of the coronary arteries. Often led by atherosclerosis, the narrowing consequently reduces blood flow to the heart muscle. CAD is frequently observed in population groups with smoking history, diabetes, hypertension, or hypercholesterolemia, and can result in major adverse cardiac events and even death [1]. CAD represents a high cost to the healthcare system as well as society due to its high prevalence and associated poor health outcomes. For instance, estimated direct and indirect costs of CAD amounted to €45 billion in the European
123
A. Alberti et al.
Union (EU) in 2003, which was a quarter of the entire expenditure for cardiovascular disease [2, 3]. Treatment for CAD For CAD patients, especially those categorized as high risk, angiography would be used if no contraindications exist. According to angiographic results, cardiologists would then choose to treat the patients with medication only, or to perform either a coronary artery bypass graft (CABG) or percutaneous coronary intervention (PCI) [4]. PCI is usually undertaken with either a bare metal stent (BMS) or a drug-eluting stent (DES), to reduce the chance of restenosis. Comparing the efficacy between BMS and DES, a number of studies showed superiority of DES in reducing patients’ need for revascularization [5], especially for patient groups at a greater risk of post-PCI adverse cardiac events [6]. Although DES contributes to reduced restenosis, further reduction is possible with a more accurate implantation of the stent in the artery [7]. The current practice of PCI is often undertaken with angiography alone, which only displays the lumen [5]. The use of intravascular ultrasound (IVUS) along with angiography enables cardiologists to assess the volume, extension, composition, morphology of the plaque and severity of stenosis within the vessel more thoroughly than angiography alone. Thus, this better imaging can assist cardiologists in selecting the appropriate stent for patients’ arterial conditions and aid in optimal stent placement [8–10]. Objectives of this study Acknowledging both potential improvements in health outcomes and additional costs incurred through IVUS use, this study attempts to compare the costs and outcomes between IVUS-guided and angiography-guided PCI in the context of Italy using a deterministic Markov model. It accounts for some of the critical concerns emerging in previous studies including sample size, presentations of patients, the number of endpoints, and time horizon. Also, this study assesses the effect of IVUS only when PCI was completed with DES. An approach to focus solely on DES rather than all stents was chosen as patients implanted with DES are more likely to need IVUS during a PCI procedure than those with BMS as both IVUS and DES target high-risk patients.
Description of the model Type of economic evaluation A deterministic Markov model was constructed to examine the long-term costs and health outcomes of IVUS use in
123
DES placement, comparing between IVUS-guided PCI and PCI guided solely by angiography. With the selected model, this study conducted cost-effectiveness analysis (CEA) and cost-utility analysis (CUA), in which outcomes were presented in both life years gained (LYG) and quality-adjusted life year (QALY) gained, respectively. Patient population The treatment population in the model is based on the patient population examined in a recent meta-analysis by Ahn et al. [11]. Ahn et al. [11] is the most recent metaanalysis on IVUS use in DES implantation to date. The authors investigated both acute and chronic CAD patients going through either IVUS-guided or angiography-guided PCI with DES. Ahn et al. also included patients with diabetes mellitus, renal insufficiency (creatinine clearance \60 ml/min), acute coronary syndrome (ACS), or smoking history, who represent patients at a high risk of post-PCI adverse cardiac events [11]. Perspectives This study took the Italian healthcare payer perspective. Due to the presence of hospital staff, informal care is not a critical factor in costing for IVUS, which is utilized in inpatient settings. Also, productivity loss is often small for elderly patients. Given the constraints and circumstances in which IVUS would be employed, this study considered accounting for direct costs while excluding indirect costs to be a reasonable approach. Model design, health states, and time horizon The model began with a hypothetical cohort of patients with an average age of 63, the average age reported by Ahn et al. [11]. All patients underwent PCI with DES, guided either by IVUS and angiography, or angiography alone. However, patients who developed restenosis were not considered for IVUS during their additional treatments. Such assumption was made because none of the relevant clinical trials indicated that IVUS accompanied subsequent treatments for post-PCI adverse events [11]. The Markov transition model included four health states: event-free, myocardial infarction (MI), revascularization [defined as a composite between target lesion revascularization (TLR) and target vessel revascularization (TVR)], and death. Since both medical therapy and interventions (PCI or CABG) can treat MI, this model made a distinction between the two treatment methods (Fig. 1). The baseline analysis assessed the use of IVUS with a lifetime horizon in order to evaluate outcomes over the full life expectancy of the patients, assuming the benefits of
Understanding the economic impact of intravascular ultrasound MI Revasc Event Free MI (Medical Therapy)
Dead
MI MI + Revasc IVUS+Angio
MI Re va sc
De a d
MI
Re va sc
Re va sc
De a d
De a d
CAD_DES MI Revasc Event Free MI (Medical Therapy)
Dead
MI MI + Revasc Angio
MI Re va sc
De a d
MI
Re va sc
Re va sc
De a d
De a d
Fig. 1 Deterministic Markov model for IVUS use during PCI with DES implantation. Note: the model ends after 38 years. Markov transition shown in this figure continues until the model ends
IVUS would persist throughout their lifetime [12]. The model used monthly cycles for the first year and full-year cycles for the remaining years in order to keep consistency with the BMS model, where there was clinical evidence to support a higher event rate in the initial months of post-PCI (see Appendix A for BMS results). Discount rates of 3 % per annum were applied to both costs and outcomes [12].
simulations were run based on probability distributions described in Table 3. These results were then used to create an incremental cost-effectiveness plane.
Model inputs Treatment effect
Scenario and sensitivity analyses Although the benefits of IVUS continue throughout a patient’s lifetime in the base-case analysis, the duration of IVUS effect is uncertain because of the short follow-up of the studies included in Ahn et al. [11]. Therefore, a scenario analysis was conducted, assuming that the benefits of IVUS would be limited to the first year (for further information, see ‘‘Model Input’’ section). For all parameters, one-way sensitivity analyses (OWSA) were completed using the 95 % confidence intervals reported in the literature (Table 3). Where 95 % confidence interval was not available, a 10 % variation from the mean was used. For those inputs that greatly influenced the incremental cost-effectiveness ratio (ICER) in their OWSA, the study conducted two-way sensitivity analyses (TWSA). Probabilistic sensitivity analysis (PSA) was also undertaken to present overall uncertainty of the model on the cost-effectiveness of IVUS. PSA allows all parameter values to vary simultaneously when sampled from a prioridefined probability distributions [13]. One-thousand
The probabilities of revascularization and MI for the angiography arm were estimated based on the results of COMPARE trial as paclitaxel-eluting and everolimus-eluting stents remain the most widely used forms of DES worldwide and the Ahn et al. [14] meta-analyses include trials and studies investigating both types of stents (Table 1). In order to obtain the estimates for the same variables in the IVUS arm, risk ratios (RRs) for experiencing MI and revascularization between IVUS and angiography groups were multiplied to the corresponding pooled probabilities produced from the COMPARE trial [11, 14] (for all model inputs, see Table 3). A supplementary analysis was also conducted evaluating only everolimus-eluting stents given their prevalent use in the Italian market. The baseline analysis assumed the benefits of IVUS would persist because precise DES implantation may lower rates of restenosis and cardiac events in the longer term [8] (Table 2). However, for the short duration of the trials included in Ahn et al., additional analysis was undertaken, assuming that the benefits of IVUS for MI and revascularization would be restrained to the first year.
123
A. Alberti et al. Table 1 Pooled base-case probability estimates [14] Clinical endpoint
Everolimus
Paclitaxel
Pooled probability (%)
1-month probability
MI
25/897
49/903
74/1,800 (4.11 %)
0.35 %
Revascularization
39/897
97/903
136/1,800 (7.56 %)
0.65 %
Using the results from the first-year follow-up, this table illustrates the pooling process. These estimates were applied to the first year of the current model in the angiography arm. The same method was applied to the data obtained from the second-year follow-up
Table 2 Extrapolated post-year-one probabilities [11, 14] Probability of having MI
Probability of having revascularization
Base-case analysis: continuation of IVUS benefits beyond the first year IVUS arm
0.92 %
2.23 %
Angiography arm 1.61 % 2.72 % Scenario analysis: benefits of IVUS limited to the first year Both arms
1.61 %
2.72 %
All tables present values rounded up to the second decimal place. Only when necessary, they were reported up to the third decimal. However, for costs and ICERs, values were rounded up to a whole number. When these measures were put into the model, those figures before rounding were inputted. Numbers surrounded by round parentheses are negative values The extrapolated probabilities were applied to the all cycles after the first year
Costs Costs were derived by weighting DRG tariffs with the distribution of the number of PCI reported in each DRG code. The government publications such as the national formulary provided information on medication costs for managing post-MI treatment [24, 25]. All cost values were adjusted and presented in 2013 euros. Summary of model assumptions This model made assumptions throughout the analyses because of data limitations or uncertainty. The ‘‘Results’’ and ‘‘Discussion’’ sections discuss the implications of the major assumptions. I.
•
Treatment effect: subgroups A number of studies concluded that clinical experience of patients with particular comorbid conditions would differ from that of the general population [15, 16]. Acknowledging those risk factors, this study conducted subgroup analyses for patients with diabetes, renal insufficiency, and ACS. The hazard ratios (HRs) derived from ADAPT-DES [6] were applied to the probabilities of having MI in both arms. The ratios concerning TLR/TVR were not available for these subgroups; thus, they were set at one in the baseline analysis (Table 3).
Assumptions about the model:
• II.
Assumptions about efficacy inputs: •
• III.
The average age of the patients who undergo PCI with DES implantation is 63 years. The model ends after about 38 years. PCI does not employ IVUS after the first cycle of the model.
The probabilities of having MI and undergoing revascularization remain consistent from the second year after the patients’ initial PCI. Designs of DES do not influence the efficacy of the procedure.
Assumptions about health outcomes:
Mortality
•
The background (natural) death probability was calculated using lifetime tables for Italy from World Health Organization [17]. Cardiac death probabilities were mainly derived from Gan et al. [18] and Serruys et al. [19].
•
The utilities associated with each health state are the same between the two arms. The disutility associated with MI and revascularization persists for 1 year.
Outcomes
Results
Overall LYG related to IVUS use was calculated using the difference in life expectancy between patients in each arm [18, 19]. Utility weights from the literature were used to calculate the corresponding outcomes in QALYs (Table 3) [20–23].
Baseline analyses
123
From the healthcare payer perspective, IVUS is a dominant treatment option, providing improved outcomes at lower costs when the benefits of IVUS continue beyond 1 year
Understanding the economic impact of intravascular ultrasound Table 3 Model parameters, data sources, and values used in the model Variable
Mean
Lower value
Upper value
PSA distribution
Source
0.57
0.44b
0.75b
Gamma
[11]
0.82
b
0.70
0.97b
1.48
1.15b
1.89b
Gamma
[6]
1.64
b
1.19
2.27b
1.41
1.10b
1.80b
With IVUS
0.20 %
0.18 %
0.22 %
Beta
[14]
With angiography
0.35 %
0.31 %
0.38 %
With IVUS
0.54 %
0.48 %
0.59 %
Beta
[14]
With angiography
0.65 %
0.59 %
0.72 %
With IVUS
0.92 %
0.83 %
1.01 %
Beta
[14]
With angiography
1.61 %
1.45 %
1.77 %
With IVUS
2.23 %
2.01 %
2.46 %
Beta
[14]
With angiography
2.72 %
2.45 %
2.99 %
Relative risk (RR) for MI and revascularization MI RRa a
Revascularization RR
HR for subgroup analyses Diabetes HRc Renal insufficiency HR
c
ACS HRc Probabilities Probability of having MI: year one (per month)
Probability of having revascularization: year one (per month)
Probability of having MI: post year oned
Probability of having revascularization: post year oned
Probability of death after MI
17.96 %
16.16 %
19.76 %
Beta
[18]
Probability of death after CABG or PCI
3.54 %
3.19 %
3.89 %
Beta
[19]
Probability of MI being treated with revascularizatione,f
50 %
45 %
55 %
Beta
Assumed
Probability of revascularization being treated with CABGe,f
10 %
9%
11 %
Beta
Assumed
Discounting rates for costs and outcomes
3%
0%
8%
–
[12]
Utilities No adverse events occurred after the procedure
0.85
0.77
0.94
Beta
[20]
Treated for MI
0.68
0.61
0.75
Beta
[21, 22]
Treated for revascularization with PCI
0.77
0.69
0.85
Beta
[20]
Treated for revascularization with CABG
0.73
0.66
0.80
Beta
[23]
Death
0
0
0
–
IVUS
650
585
715
Gamma
Initial PCI with angiography and IVUS
7,084
6,376
7,792
Initial PCI with angiography alone
6,434
5,791
7,077
MI treated with medications
3,875
3,487
4,262
MI treated with PCI
8,128
7,315
8,941
MI treated with CABG
22,125
19,913
24,338
Revascularization treated with PCI
6,434
5,791
7,077
Costs (€)
a
Revascularization treated with CABG
19,112
17,201
21,023
Medication to manage post-MI
511
460
562
[24, 25]
p values of RRs for revascularization and MI were 0.022 and \0.001 individually
b
Lower and upper bound of 95 % CI
c
p values of HRs for subgroups of patients with diabetes, renal insufficiency, and ACS were 0.0021, 0.0025 and 0.0064, respectively
d
The post-year-one probabilities shown in the table were based on the assumption that benefits of IVUS would persist throughout the cycles (Table 2)
e
Only two out of the four country-adjusted probabilities were displayed. The others, the probability of MI treated with medications and that of having revascularization with PCI, were calculated as follows: 100 less each of the two variables included in the table (%)
f
These clinical probabilities were obtained and confirmed through consultations with local clinicians
123
A. Alberti et al. Table 4 Summary lifetime ICERs in baseline and subgroup analyses IVUS
Angiography
Incremental
12.200
0.217
Cost per QALY gained
(€3,430)a
IVUS was more cost-effective in patients with co-morbid conditions who are at a higher risk of cardiac events: the ICERs were significantly lower for these subgroups than those derived from the general patient population. These results further support the reimbursement of IVUS especially for the patients with diabetes, renal insufficiency and ACS (Table 4). Supplementary analyses on BMS
General patient population (baseline) Cost
€11,182
€11,928
(€746)
LYs
15.180
14.941
0.240
Cost per LYG
(€3,112)a
QALYs
12.417
Subgroup analyses
Diabetes mellitus Cost
€11,710
€12,759
(€1,049)
LYs
15.046
14.714
0.332
Cost per LYG
(€3,160)a
11.997 Cost per QALY gained
0.300 (€3,493)a
€13,024
(€1,142)
QALYs
12.297
Renal insufficiency Cost
€11,881
LYs
15.002
QALYs
12.257
14.640
0.362
Cost per LYG
(€3,156)a
11.930
0.327
Cost per QALY gained
(€3,491)a
€12,642
(€1,007)
ACS Cost
€11,635
LYs
15.066
QALYs
12.314
14.747
0.319
Cost per LYG
(€3,160)a
12.026
0.288
Cost per QALY gained
(€3,492)a
General patient population, using only everolimus-eluting stents Cost
€9,825
€10,170
(€345)
LYs
15.305
15.139 Cost per LYG
0.166 (€2,077)a
QALYs
12.534
12.383
0.151
Cost per QALY gained
(€2,288)a
a Negative ICERs imply that IVUS-guided PCI with DES is a dominant treatment option compared to angiography-guided PCI with DES
Recognizing that IVUS and DES both aim to treat high-risk patients, this current study focused on DES implantation only. Nonetheless, given that patients who are treated with BMS may still benefit from IVUS-guided stent placement, we also assessed the cost-effectiveness of IVUS use in BMS placement. The relevant inputs and results can be found in Appendix A. Supplementary analyses on everolimus-eluting stents Given that in Italy everolimus-eluting stents are predominantly used, a supplementary analysis was undertaken to evaluate the cost-effectiveness when only these stents are used (Table 4). Given that the final ICER for the general patient population is negative, IVUS use would remain as a dominant option for the chosen subgroups. Results of paclitaxel-eluting stents can be found in Appendix B. Results for scenario and sensitivity analyses This study conducted scenario and sensitivity analyses to determine the robustness of the presented results. This section only refers to the ICER per QALY gained for Italy because the WTP threshold for cost per LYG is unknown. Benefits of IVUS
(Table 4). The negative ICER value implies that IVUS use during PCI with DES is a superior strategy regardless of the ICER threshold; the €25,000 per QALY is often employed in Italy as its implicit willingness-to-pay (WTP) threshold [26, 27]. Since the data were extrapolated beyond the observation period, the study assessed final results at shorter time intervals in addition to the full lifetime horizon. A longer time frame contributed to lower ICERs because the costdifference between the two arms diminished over time while the positive impact of IVUS guidance continued. The ICER was estimated at €166,641 at a 1-year time interval; however, it fell below the threshold between 3- and 5-year time intervals, at which the ICERs were €26,967 and €7,409, respectively.
123
Even with a conservative assumption that the benefits of IVUS are restrained to the first year, the ICER remained below the WTP threshold (Table 5).
Table 5 The resulting ICERs per QALY with limited benefits of IVUS IVUS
Angiography
Incremental
The benefits of IVUS limited to the first year Cost
€12,354
€11,928
€426
QALYs
12.244
12.200
0.044
Cost per QALY gained
€9,624
Understanding the economic impact of intravascular ultrasound €-
€ 10,000.0 € 20,000.0 € 30,000.0 € 40,000.0 € 50,000.0 € 60,000.0 € 70,000.0
Probability of MI (angiography) (post year 1) Probability of MI (IVUS) (post year 1) Probability of revasc. (IVUS) (post year 1) Probability of revasc. (angiography) (post year 1)
Variable
Fig. 2 Tornado diagram of the scenario analysis (i.e., the benefits of IVUS limited to the first year). Note: The light and dark blue bars show the direction of ICER change for the upper and lower input bounds, respectively. The vertical axis that crosses the tornado diagram above indicates the base-case ICER of €9,624 in the scenario in which the benefits of IVUS are limited to the first year (color figure online)
MI RR Discounng rate for outcomes Revasc. RR No event ulity Probability of death aer MI if a Man Discounng rate for costs
Parameters None of the parameters significantly affected the final results. Even in the scenario analysis, where the benefits of IVUS were limited to the first year, the ICERs were sensitive to the following parameters only: the probabilities of having MI (post year one) and the cost of initial PCI with IVUS and angiography (Fig. 2). Variations in the two cost parameters made ICERs negative, further supporting the base-case outcome that IVUS is a dominant strategy: the cost of initial PCI with IVUS and angiography and the cost of initial PCI with angiography. The results from OWSA clearly identified the drivers of this model. While TWSA was preformed, results do not provide any further granularity or deviation from OWSA. PSA PSA identifies the optimal clinical strategy and the associated probability the strategy is cost effective given relevant WTP thresholds. As demonstrated in the incremental
3,000
ICERs generated from PSA
2,000
Mean ICER
Incremental Costs
4,000
1,000 0 -0.1
0
0.1
0.2
-1,000 -2,000 -3,000 -4,000 -5,000 -6,000
Incremental QALYs
Fig. 3 Incremental cost-effectiveness plane
0.3
0.4
0.5
cost-effectiveness plane in Fig. 3, about 71 % of the simulations fell into the south–east quadrant, implying that IVUS use is a dominant option. The remaining 29 % fell into the north–east quadrant; however, the resulting ICERs were always below €18,000. From an Italian healthcare payer perspective, the choice PCI guided by both IVUS and angiography is always the optimal strategy given the €25,000 WTP threshold.
Discussion This study demonstrates that, in Italy, using IVUS in addition to angiography is a favorable strategy over employing angiography alone during PCI with DES. Covering the full life expectancy of the patients, this study produced negative ICER. The final ICER remained lower than the implicit WTP threshold even when the duration of IVUS benefit was assumed to be limited to the first year. IVUS use in patients at a higher risk of restenosis was more costeffective than use in the general patient population. Furthermore, positive results of the ADAPT-DES study at 1 year appear to prolong to 2-year follow-up and therefore the benefits of IVUS continuing beyond the first year may be a more plausible assumption [6, 28]. Incorporating a number of potential scenarios and mitigating the common problems emerging in the earlier literature, this study improves the evidence on the impact of IVUS in terms of both costs and health outcomes. Despite these strengths, the results were sensitive to the length of the assessed time horizon and the key parameters: the probabilities of having MI (post year-one) and the cost of initial PCI with IVUS and angiography. In addition to the uncertainty with the results, this study suffers from several limitations due to data unavailability, modeling, and estimation assumptions. First, in this study, it was assumed using the 2-year follow-up data from the COMPARE trial that the
123
A. Alberti et al.
probabilities of having MI and TLR/TVR in both arms would remain constant after 2 years following the event. This assumes that the chance of having subsequent events remains the same as the number of successive treatments increases [14]. Nevertheless, the error on the final ICERs should be minimal in that a majority of the patients would not need numerous subsequent treatments for restenosis. Moreover, this study supposed that types of the drugs coated on the stents would not affect efficacy of the procedure, since Smits et al. [14] only incorporated two designs of DES in their examinations, while cardiologists employ other designs in their clinical practice. However, although the validity of their argument is still controversial, some researchers advocate for the superiority of everolimus-eluting stents over paclitaxel-eluting stents [29, 30]. Given the near-ubiquitous use in Italy, the model was tested for sensitivity in limiting usage to everolimus-eluting stents, and for high-risk patient groups, lifetime ICERs were found to be consistent with base-case estimates. These stratified results should only be considered secondarily, as the clinical inputs from Ahn et al. [11] included a mixture of both stent types. Owing to the paucity of IVUS-specific information, the analysis assumed the same utilities for patients in each arm. Nevertheless, the users of IVUS insist that the patients with IVUS-driven PCI have a better experience than those using angiography alone because better imaging serves improved stent implantation [7]. If this claim is supported with more clinical evidence, IVUS would become more costeffective. Furthermore, the study only incorporated direct costs incurred to the healthcare payers because of the data constraints and the settings in which the device would be utilized. For the examined population of this study, productivity loss of caregivers can be considerable given the patients’ average age. Nonetheless, the underestimation of costs should not influence the ICERs if the degrees of inaccuracy in both arms were similar. Also, because none of the clinical trials specified that IVUS was used for subsequent treatments, this device only accompanied PCI of the first cycle [11]. If cardiologists frequently employ IVUS for post-PCI cardiac issues in their daily practices, the estimated costs in the IVUS arm, and thus, the final ICERs might be underestimated.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
Conflict of interest Dr. Alessandro Alberti received consulting fees from Boston Scientific. Dr. Pietro Giudice: none to declare. 12.
References 1. Hill, RA., Boland, A., Dickson, R., Dundar, Y., Haycox, A., McLeod, C., Mujica Mota, R., Walley, T., Bagust, A.: Drug-
123
13.
eluting stents: a systematic review and economic evaluation. Health Technol. Assessment November. Vol. 11, No.46 (2007) Cleveland Clinic Center for Continuing Education. Coronary Artery Disease. http://www.clevelandclinicmeded.com/medi calpubs/diseasemanagement/cardiology/coronary-artery-disease. Accessed 14 July 2013 Leal, J., Luengo-Fernandez, R., Gray, A., Petersen, S., Rayner, M.: Economic burden of cardiovascular diseases in the enlarged European Union. Eur. Soc. Cardiol. 27, 1610–1619 (2006) National Institute for Health and Care Excellence Unstable angina and NSTEMI: the early management of unstable angina and non-ST-segment-elevation myocardial infarction. [CG94]. London: National Institute for Health and Care Excellence (2010) De LUCA, G., Dirksen, M.T., Spaulding, C., Kelbaek, H., Schalij, M., Thuesen, L., van der Hoeven, B., Vink, M.A., Kaiser, C., Musto, C., Chechi, T., Spaziani, G., Diaz de la Llera, L.S., Pasceri, V., Di Lorenzo, E., Violini, R., Cortese, G., Suryapranata, H., Stone, G.W.: Drug-eluting vs. bare-metal stents in primary angioplasty: a pooled patient-level meta-analysis of randomized trials. Br. Med. J. 172(8), 611–621 (2012) Witzenbichler, B., Maehara, A., Weisz, G., Neumann, F.J., Rinaldi, M., Metzger, D.C., Henry, T., Cox, D., Duffy, P., Brodie, B., Stuckey, T., Mazzaferri, E., Xu, K., Mehran, R., Praise, H., Mintz, G., Stone, G.: Relationship between intravascular ultrasound and guidance and clinical outcomes after drug-eluting stents: the assessment of dual antiplatelet therapy with drug-eluting stents (ADAPT-DES) study. Circulation 129(4), 463–470 (2014) Rogacka, R., Latib, A., Colombo, A.: IVUS-guided stent implantation to improve outcome: a promise to be fulfilled. Curr. Cardiol. Rev. 5(2), 78–86 (2009) Gaster, A.L., Slothuus Skjoldborg, U., Larsen, J., Korsholm, L., von Birgelen, C., Jensen, S., Thayssen, P., Pedersen, K.E., Haghfelt, T.H.: Continued improvement of clinical outcome and cost effectiveness following intravascular ultrasound guided PCI: insights from a prospective, randomised study. Heart 89, 1043–1049 (2003) Biondi-Zoccai, G., Sheiban, I., Romagnoli, E., De Servi, S., Tamburino, C., Colombo, A., Burzotta, F., Presbitero, P., Bolognese, L., Paloscia, L., Rubino, P., Sardella, G., Briguori, C., Niccoli, L., Franco, G., Di Girolamo, D., Piatti, L., Greco, C., Capodanno, D., Sangiorgi, G.: Is intravascular ultrasound beneficial for percutaneous coronary intervention of bifurcation lesions? Evidence from a 4,314-patient registry. Clin. Res. Cardiol. 100(11), 1021–1028 (2011) Claessen, B.E., Mehran, R., Mintz, G.S., Weisz, G., Leon, M.B., Dogan, O., de Ribamar Costa Jr, J., Stone, G.W., Apostolidou, I., Morales, A., Chantziara, V., Syros, G., Sanidas, E., Xu, K., Tijsse, J.G., Henriques, J.P., Piek, J.J., Moses, J.W., Maehara, A., Dangas, G.D.: Impact of intravascular ultrasound imaging on early and late clinical outcomes following percutaneous coronary intervention with drug-eluting stents. J. Am. Coll. Cardiol. 4, 974–981 (2011) Ahn, J.M., Kang, S.J., Yoon, S.H., Park, H.W., Kang, S.M., Lee, J.Y., Lee, S.W., Kim, Y.H., Lee, C.W., Park, S.W., Mintz, G.S., Park, S.J.: Meta-analysis of outcomes after intravascular ultrasound-guided versus angiography-guided drug-eluting stent implantation in 26,503 patients enrolled in three randomized trials and 14 observational studies. Am. J. Cardiol. 113(8), 1338–1347 (2014) International Society for Pharmacoeconomics and Outcomes Research. Pharmacoeconomic Guidelines Around the World. http://www.ispor.org/PEguidelines/index.asp. Accessed 20 July 2013 Briggs, A., Weinstein, M., Fenwick, E., Karnon, J., Sculpher, M., Paltiel, D., ISPOR-SMDM Modeling Good Research Practices
Understanding the economic impact of intravascular ultrasound
14.
15.
16.
17.
18.
19.
20.
21.
Task Force: Model parameter estimation and uncertainty: a report of the ISPOR-SMDM modeling good research practices task froce-6. Value Health 15, 835–842 (2012) Smits, P., Kedhi, E., Royaards, K.J., Joesoef, K., Wassing, J., Rademaker-Havinga, T.A., McFadden, E.: 2-Year follow-up of a randomized controlled trial of everolimus- and paclitaxel-eluting stents for coronary revascularisation in daily practice. J. Am. Coll. Cardiol. 58(1), 11–18 (2011) Mathew, V., Gersh, B., Williams, B., Laskey, W., Willerson, J., Tilbury, T., Davis, B., Holmes, D.: Outcomes in patients with diabetes mellitus undergoing percutaneous coronary intervention in the current era. A report from the prevention of restenosis with tranilast and its outcomes (PRESTO) trial. J. Am. Heart Assoc. 109, 476–480 (2004) Garg, P., Normand, S., Silbaugh, T., Wolf, R., Zelevinsky, K., Lovett, A., Varma, M., Zhou, Z., Mauri, L.: Drug-eluting or baremetal stenting in patients with diabetes mellitus: results from the Massachusetts Data Analysis Center Registry. J. Am. Heart Assoc. 118, 2277–2285 (2008) World Health Organization (WHO). Global Health Observatory Data Repository: Life Tables. http://www.who.int/gho/mortality_ burden_disease/life_tables/life _tables/en. Accessed 1 Aug 2013 Gan, S.C., Beaver, S.K., Houck, P.M., MacLehose, R.F., Lawson, H.W., Chan, L.: Treatment of acute myocardial infarction and 30-day mortality among women and men. N. Engl. J. Med. 343, 8–15 (2000) Serruys, P.W., Morice, M.C., Kappetein, A.P., Colombo, A., Holmes, Mack, M.J., Stahle, E., Feldman, T.E., van den Brand, M., Bass, E.J., Van Dyck, N., Leadley, K., Dawkins, K.D., Mohr, F.W., SYNTAX Investigators: Percutaneous coronary intervention versus coronary-artery bypass grafting for severe coronary artery disease. N. Engl. J. Med. 360(10), 961–972 (2009) Shrive, F., Manns, B., Galbraith, D., Knudtson, M., Ghali, W.: Economic evaluation of sirolimus-eluting stents. Can. Med. Assoc. J. 172, 345–351 (2005) Dhalla, I., Smith, M., Choudhry, N., Denburg, A.: Costs and benefits of free medications after myocardial infarction. Health Policy 5(2), 68–86 (2009)
22. Clarke, P., Gray, A., Holman, R.: Estimating Utility Values for Health States of Type 2 Diabetic Patients Using the EQ-5D (UKPDS 62). Med. Decis. Mak. 22(4), 340–349 (2000) 23. Puskas, J.D., Williams, W.H., Mahoney, E.M., Huber, P.R., Block, P.C., Duke, P.G., Staples, J.R., Glas, K.E., Marshall, J.J., Leimbach, M.E., McCall, S.A., Petersen, R.J., Bailey, D.E., Weintraub, W.S., Guyton, R.A.: Off pump vs. conventional coronary artery bypass grafting: early and 1 year graft patency, cost, and quality of life outcomes. J. Am. Med. Assoc. 291(15), 1841–1849 (2004) 24. Nello Martini.: The New National Pharmaceutical Formulary 2003. Ministry of Health. 2003 25. AIFA Agenzia Italiana del Farmaco. http://www.agenziafarmaco. gov.it. Accessed 20 July 2013 26. Mencacci, C., Di, Sciascio G., Katz, P., Ripellino, C.: Cost-effectiveness evaluation of escitalopram in major depressive disorder in Italy. J. ClinicoEcon. Outcomes Res. 5, 87–99 (2013) 27. Lazzaro, C., Bordonaro, R., Cognetti, F., Fabi, A., De Placido, S., Arpino, G., Marchetti, P., Botticelli, A., Pronzato, P., Martelli, E.: An Italian cost-effectiveness analysis of paclitaxel albumin (nabpaclitaxel) versus conventional paclitaxel for metastatic breast cancer patients: the COSTANza study. J. ClinicoEcon. Outcomes Res. 5, 125–135 (2013) 28. Maehara, A., Weisz, G., Neumann, F., Metzger, C., Henry, T., Cox, D. et. al.: TCT-68 does IVUS reduce stent thrombosis with DES? Two-year results from the prospective multicentre ADAPT-DES study. J. Am. Coll. Cardiol. B21–B22 (2013) 29. Stone, G.W., Midei, M., Newman, W., Sanz, M., Hermiller, J.B., Williams, J., Farhat, N., Mahaffey, K.W., Cutlip, D.E., Fitzgerald, P.J., Sood, P., Su, X., Lansky, A.J.: SPIRIT III investigators.: comparison of an everolimus-eluting stent and a paclitaxel-eluting stent in patients with coronary artery disease: a randomised trial. J. Am. Med. Assoc. 299(16), 1903–1913 (2008) 30. Joner, M., Nakazawa, G., Finn, A.V., Quee, S.C., Coleman, L., Acampado, E., Wilson, P.S., Skorija, K., Cheng, Q., Xu, X., Gold, H.K., Kolodgie, F.D., Virmani, R.: Endothelial cell recovery between comparator polymer-based drug-eluting stents. J. Am. Coll. Cardiol. 52, 333–342 (2008)
123
