International Journal of Cardiology 172 (2014) 109–114
Contents lists available at ScienceDirect
International Journal of Cardiology journal homepage: www.elsevier.com/locate/ijcard
Long-term cost-effectiveness of transcatheter versus surgical closure of secundum atrial septal defect in adults☆ Darren Mylotte a,1, Stéphane P. Quenneville a,1, Mark A. Kotowycz a,1, Xuanqian Xie b,1, James M. Brophy c,d,1, Raluca Ionescu-Ittu a,1, Giuseppe Martucci a,1, Louise Pilote e,1, Judith Therrien a,1, Ariane J. Marelli a,⁎,1 a
McGill Adult Unit for Congenital Heart Disease Excellence (MAUDE Unit), McGill University Health Centre, Canada Technology Assessment Unit, McGill University Health Centre, Canada Department of Medicine, McGill University, Canada d Department of Epidemiology & Biostatistics, McGill University, Canada e Division of Clinical Epidemiology, McGill University Health Centre, Canada b c
a r t i c l e
i n f o
Article history: Received 18 July 2013 Received in revised form 2 December 2013 Accepted 26 December 2013 Available online 8 January 2014 Keywords: Atrial septal defect Surgery Transcatheter Adult congenital heart disease Cost-effectiveness
a b s t r a c t Background: The most common congenital anomaly in adults is secundum atrial septal defect (ASD), which can be closed using a surgical or transcatheter approach. Despite the growing use of transcatheter ASD closure, few studies have examined the cost-effectiveness of this strategy. We sought to compare the long-term cost effectiveness of transcatheter and surgical closure of secundum ASD in adults. Methods: A decision-analytic model was used with all clinical outcome parameter estimates obtained from the province-wide Québec Congenital Heart Disease Database. Costs were obtained from a single academic centre (Canadian dollars). A cost-effectiveness analysis using a discrete event Monte Carlo simulation model from the perspective of a single third party payer and multiple sensitivity analyses were performed. Patients were followed for a maximum of 5 years after ASD closure. Results: Between l998 and 2005, we identified 718 adults (n = 335 transcatheter; n = 383 surgical) who underwent ASD closure in Quebec. The 5-year cost of surgical closure was $15,304 SD $4581 versus $11,060 SD $5169 for the transcatheter alternative. At 5 years, transcatheter closure was marginally more effective than surgery (4.683 SD 0.379 life-years versus 4.618 SD 0.638 life-years). Probabilistic sensitivity analyses demonstrated that transcatheter ASD closure was a dominant strategy with an 80% probability of cost savings and equal or greater efficacy compared to surgical treatment. Conclusion: Although definitive conclusions are limited given the observational nature of the primary data sources, transcatheter ASD closure appeared to be a cost-effective strategy associated with slightly improved clinical outcomes and reduced costs compared to surgical closure at 5-years follow-up. © 2014 Published by Elsevier Ireland Ltd.
1. Introduction The population of adults living with congenital heart disease (CHD) has grown significantly over the last decades [1], with atrial septal defect (ASD) being the most common lesion. While ASD can be closed either surgically or with a transcatheter device, the less invasive transcatheter strategy has become the de facto accepted standard of care
☆ Sources of funding: Dr. Mylotte is funded by the Beth Raby Fellowship. Dr. Brophy is a research scholar, Les Fonds de la Recherche en Santé du Québec. Dr. Ionescu-Ittu was funded by Les Fonds de la Recherche du Québec (postdoctoral scholarship) at the time that the study was conducted. Dr. Marelli is a research scholar, Les Fonds de la Recherche en Santé du Québec and is funded by the Heart and Stroke Foundation of Québec and the Canadian Institutes of Health Research.There are no relationships with industry. ⁎ Corresponding author at: McGill Adult Unit for Congenital Heart Disease (MAUDE Unit), 687 Pine Avenue West, Room H4.33, Montréal, Québec H3A 1A1, Canada. E-mail address: [email protected] (A.J. Marelli). 1 This author takes responsibility for all aspects of the reliability and freedom from bias of the data presented and their discussed interpretation. 0167-5273/$ – see front matter © 2014 Published by Elsevier Ireland Ltd. http://dx.doi.org/10.1016/j.ijcard.2013.12.144
[2–4] despite a paucity of randomized clinical trials and long-term cost-effectiveness studies. There is an increasing interest in comparative effectiveness studies to examine interventional outcomes in a setting that reflects clinical practice [5]. In 2010, The Patient Protection and Affordable Care Act established comparative effectiveness research as central to health care improvement [6]. Although cost-effectiveness is not an outcome as universally accepted as clinical effectiveness, it provides an important measure of the value of health care particularly for costly procedures [7,8]. Because the use of transcatheter closure in ASD has become so wide-spread, it is relevant to examine its cost-effectiveness as an important dimension of comparative effectiveness. Several small studies have previously reported reduced in-hospital and procedural costs with transcatheter compared to surgical ASD closure [9–14]. However, the cost-effectiveness of novel medical technologies can only be sensibly interpreted when applied to the whole population susceptible to receiving the intervention and when both medium- and long-term clinical outcomes and costs are
110
D. Mylotte et al. / International Journal of Cardiology 172 (2014) 109–114
systematically evaluated. To the best of our knowledge, a comparative cost-effectiveness analysis at the population level has not been performed for ASD closure and this research aims to address this knowledge gap.
a procedural code for transcatheter ASD closure or a surgical code for ASD closure, billed during an observation period that started with the day following the index procedure and continued for the duration of follow-up.
2.3. Statistical analysis 2. Methods We used population-based health service data sources from Québec, Canada and relied on a discrete event Monte Carlo simulation model inputting effectiveness parameters from our large and previously validated administrative databases [1]. Cost parameters were derived from the McGill University Health Centre in Montreal, Quebec and from the literature. 2.1. Data input for effectiveness parameters 2.1.1. Data sources In Quebec, Canada there is universal access to health care and a unique health care number is assigned to every individual at birth. This number is used to record all diagnoses, hospitalisations and health service encounters throughout an individual's lifetime. This information is recorded into two databases: the medical claims database of the provincial authority, the Regie de l'Assurance Maladie du Quebec (RAMQ), and the hospital provincial discharge summary database (Med-Echo). The RAMQ database also includes patient demographic information, including sex, and dates of birth and death. The Quebec CHD database was created by merging these two databases and developing algorithms to correctly identify CHD patients and extract CHD diagnoses [1]. Diagnoses recorded in the databases adhered to the International Classification of Diseases, Ninth Revision (ICD-9). Patients that had at least one diagnostic code for CHD and/or a CHD-specific surgical procedure were identified. Provider codes were used to select those CHD billings made by cardiovascular disease specialists, imaging specialists, and primary care physicians. Patients were assigned one or two CHD diagnoses via the use of a previously described hierarchical algorithm [1]. All information was cross-referenced between outpatient and inpatient data sources. Our CHD database therefore contains comprehensive longitudinal, demographic, diagnostic and therapeutic records of all patient encounters with the provincial health care system between January 1, 1983 and December 31, 2005 inclusively. In Québec death attestation is sent to the Québec Health Insurance Board, which then updates the medical claims database with this information, thereby making the documentation of death complete in the database, regardless whether death occurred in or out of hospital. 2.2. Study population Our study population, derived from the Québec CHD database, consisted of all adult patients aged N18 and ≤75 years with both ASD diagnosis and either (i) a procedural code for transcatheter ASD closure, billed by an interventional cardiologist in the period 1998–2005, or (ii) a procedural code for surgical ASD closure, billed by a cardiac surgeon or general surgeon in the same period. To ensure that the study population included only patients with secundum ASD, the ASD patients who also had codes for other congenital anomalies were excluded. We also excluded patients who had surgical ASD closure billed in the same day with other cardiac surgeries, such as coronary artery bypass grafting surgery, aortic arch surgery, or valvular surgery. The ICD-9 code for secundum ASD (745.5) does not differentiate between secundum ASD and patent foramen ovale (PFO). Given that the majority of PFO closures occur in patients with cryptogenic stroke, we also excluded patients with stroke in the year prior to septal closure. Baseline characteristics included age, sex, and the following comorbidities, measured in the 5 years before intervention: coronary artery disease, hypertension, diabetes mellitus, congestive heart failure (CHF), pulmonary hypertension, and atrial fibrillation (A-Fib). Comorbidities were defined using codes from the ICD-9. We also included the Charlson comorbidity index, an aggregate measure of patients' overall burden of disease [15]. We followed patients for a maximum of 5 years following ASD closure, until the earliest of one of the following: death, end of the 5-year follow-up, or the end of the study on Dec 31, 2005. Procedure-related mortality was defined as death within 30 days of ASD closure. Long-term mortality was defined as death within 5 years of the index ASD closure. Other clinical outcomes that were assessed during the 5-year follow-up included ASD closure success, stroke, CHF and A-Fib. ASD closure success was defined as freedom from a further procedural code for ASD closure. Re-intervention was defined as
Baseline descriptive statistics of the two intervention groups are presented with categorical and continuous variables presented as percentages and medians (25th percentile, 75th percentile), respectively. Clinical outcomes within 5 years of the index ASD closure are described as unadjusted (i) percentages over the 5 years of follow-up and (ii) means and standard errors of annual event rates. The annual event rates were obtained by transforming the cumulative incidence rates for mortality, stroke, CHF and A-Fib in the two groups into annual incidence. We assumed that the risks of all events were constant during the entire follow-up period and random censoring was non-informative in both arms. In the case of low cumulative incidence as the ones observed in this study, these assumptions imply that the annual event rate approximates the cumulative incidence rate divided by the median years of follow-up. The standard errors of the annual event rates were calculated using approximate approach to binomial distribution. Data were analysed on the intention-to-treat principle, with patients assigned to the surgical or catheterisation groups at the time of their first ASD closure intervention recorded in the database. Cox regression analysis was performed to determine if any baseline characteristic predicted 5-year mortality. All statistical tests were 2-sided, and p values b0.05 were considered to be statistically significant. Data was analysed using SAS 9.2.
2.4. Data input for cost parameters Costs were evaluated from the perspective of the Canadian health care system and thus, focused on direct health care costs, and physician fees associated with ASD closure. These cost data were obtained directly from the accounting and resource departments of the McGill University Health Centre (MUHC) for the year 2009. In a province with universal health care such as Quebec, costs are not expected to vary across centres and costs estimated at one centre are expected to generalize to other centres in the province. We sought to estimate the incremental cost of both treatment strategies. Since patients are managed identically after being discharged from hospital following the index transcatheter or surgical ASD closure, the usual care costs incurred during these post procedural health states are balanced and were ignored. We also included the costs of secondary outcomes: stroke, CHF and A-Fib during follow-up. These costs were derived from published literature in Canadian populations [16–18]. All costs were expressed in 2010 Canadian dollars ($).
2.5. Cost-effectiveness model A discrete-event simulation (DES) model was developed to evaluate the clinical and economic consequences of surgical and transcatheter ASD closure strategies [19]. The DES model is governed by a chronologically ordered event list, and a finite number of discrete changes in a state (versus continuous model), and follow events that have occurred. DES is commonly used in physical systems, but has also been applied recently in the heath care field, as it allows complex modelling including that for waiting lists [20], competing resources [20], patients' baseline characteristics [21], and/or disease progression [22]. Our model starts with patients who meet clinical criteria for surgical or transcatheter ASD closure (Fig. 1). Simulated paired individuals (identical clones) were assigned to undergo surgical or transcatheter ASD closure. Our model presents the patient's course naturally, according to a scheduled order in which the potential events (e.g. operation, procedure-related mortality, re-operation, stroke, CHF, A-Fib, and death) can occur [19]. Each event was associated with resource consumption. For simplicity, minor procedural complications were not included in the model. We assumed, conservatively, that if the first transcatheter intervention was not successful, patients automatically underwent surgical closure with a 100% success rate and no procedure-related mortality. After successful ASD closure, patients may go on to develop stroke, CHF, A-Fib and to succumb to death with defined probabilities, based on our analyses of clinical outcomes observed in the Quebec CHD database. Simulated patients that did not succumb to death during follow-up were censored at 5 years.
Fig. 1. Discrete-event simulation model. Flow diagram of patients in the study. ASD = atrial septal defect; CHF = congestive heart failure; A-Fib = atrial fibrillation.
D. Mylotte et al. / International Journal of Cardiology 172 (2014) 109–114 2.6. Cost-effectiveness analysis
111
Table 2 Clinical outcomes.
Using our DES model, we compared the cost and effectiveness of the transcatheter and surgical strategies for ASD closure. The principal outcome measured was the cost of each procedure. We also present other outcomes, such as average life years, number of fatal and non-fatal events per 1000 patients, incremental cost and incremental effectiveness within 5 years. We applied the deterministic model as base case, averaging estimation for all parameters. We examined the robustness of our base case estimates by varying different inputs within our model using a univariate sensitivity analysis. The following inputs were varied: duration of follow-up; annual discount rate; success rate of transcatheter ASD closure; procedural cost; procedural mortality; and annual mortality. To explore patient level variability, we simulated 10,000-paired individuals using the stochastic DES model by Monte Carlo simulation. The Monte Carlo simulation is a technique that uses repeated random sampling to approximate the probability of certain outcomes. Since this is a stochastic method based on chance, it captures the effects of random uncertainty. All stochastic parameters are presented as distributions. Cost data must be positive and is usually skewed, and therefore was assigned to Gamma distributions [23]. Any clinical input (i.e. incidence rate) bounded between 0 and 1 was assigned to Beta distributions [23]. We arbitrarily defined the standard deviation at 25% of the mean, when variances of costs were not available. Because the device used in MUHC during the study period was the Amplatzer septal occluder (St. Jude Medical, Inc. St. Paul, MN, USA)., we also performed a sensitivity analysis in order to test the robustness of our results with different device costs. In this analysis, the cost input for surgery included: the operating room; intensive care unit stay; cardiac surgery ward stay; and physician fees for the procedure and the index hospitalisation. The cost inputs for the transcatheter procedure were: the Amplatzer septal occluder; delivery catheters; catheterisation suite; ancillary equipment; coronary care unit stay; and physician fees. Both costs and life-years were accrued over a 5-year period. This period was selected because it captures all re-interventions and also because the number of patients undergoing catheterisation who have more than 5 years is too small to draw accurate inferences. An annual discount rate of 3% was applied for costs as well as life years. Analyses were conducted using Treeage 2010 and Excel 2003.
Death Procedural Long-term Successful ASD closure Stroke Congestive heart failure Atrial fibrillation Annual event rate, %, SEMa Death Stroke Congestive heart failure Atrial fibrillation
Transcatheter closure (N = 335)
Surgical closure (N = 383)
4 (1.2) 1 (0.3) 3 (0.9) 317 (94.9) 1 (0.3) 1 (0.3) 14 (4.2)
23 (6.0) 4 (1.0) 19 (5.0) 383 (100) 3 (0.8) 5 (1.3) 28 (7.4)
0.4 SEM 0.3 0.1 SEM 0.2 0.1 SEM 0.2 1.7 SEM 0.7
1.0 SEM 0.1 0.2 SEM 0.2 0.3 SEM 0.3 1.5 SEM 0.6
Data are presented as number and percentage. ASD = atrial septal defect; SEM = standard error of the mean. a SEM was calculated using the approximate approach to binomial distribution.
Clinical outcomes are summarized in Table 2. Compared with surgical treatment, the transcatheter approach was associated with a lower risk of procedure-related mortality (0.3% versus 1.0%), lower annual mortality (0.4% versus 1.0%), and lower procedural success rates (94.9% versus 100%).
3.2. Costs 3. Results 3.1. Clinical outcomes Between 1988 and 2005 we identified 335 adult patients who had transcatheter ASD closure and 383 who had surgical closure. The clinical characteristics and outcomes of these patient groups have previously been described [24]. The number of transcatheter closures grew rapidly after the technique was introduced, with a corresponding decrease in the number of surgical closures. The patients who underwent transcatheter closure were older, had a higher incidence of hypertension, and a higher Charlson comorbidity index than those who underwent surgical closure (Table 1). One exception was the history of pulmonary hypertension, which was more common in the surgical group. Cox regression analysis identified increasing patient age (Hazard Ratio [HR] 1.057; 95% confidence intervals [CI]: 1.026–1.088, p = 0.0002) and the Charlson comorbidity index (HR 1.241; 95% CI 1.128–1.364, p b 0.0001) as the only baseline characteristics to independently predict 5-year mortality. The median follow-up time was significantly shorter in the transcatheter group compared to the surgical group, 3.0 (1.2, 3.9) and 10.0 (7.0, 13.0) years, respectively.
Table 1 Baseline characteristics of study population.
Age, years Male gender Coronary artery disease Hypertension Diabetes mellitus Congestive heart failure Pulmonary hypertension Atrial fibrillation Charlson comorbidity index Median follow-up, years
Transcatheter closure (N = 335)
Surgical closure (N = 383)
p-Value
49 (39–61) 102 (30.5) 4 (1.3) 97 (32.0) 20 (6.6) 37 (12.2) 23 (7.6) 64 (21.1) 2.3 (0.2–4.0) 3 (1.2, 3.9)
43 (32–54) 110 (28.7) 15 (4.0) 94 (24.8) 15 (4.0) 63 (16.6) 62 (16.4) 76 (20.0) 2.0 (0.1–3.0) 10.0 (7.0, 13.0)
b0.001 0.61 0.06 0.04 0.12 0.11 b0.001 0.73 b0.001 b0.001
Values are presented as N (%) or median (interquartile range). ASD = atrial septal defect.
Costs of transcatheter ASD and surgical ASD closure were derived from the MUHC costing centre (Table 3). The unit costs of surgical ASD closure included the operating room ($5800), intensive care unit stay ($1387), cardiac surgery ward stay ($3840) and physician fees for the surgical procedure and the index hospitalisation ($3421). The unit costs of transcatheter ASD closure included the Amplatzer septal occluder ($5500), delivery catheters ($480), the catheterisation suite ($484), ancillary equipment ($700), coronary care unit stay ($631), and physician fees ($900).
3.3. Base case analysis of economic model Based on the model proposed in Fig. 1 and using the parameter estimates given in Tables 2 and 3, the cost and effectiveness of transcatheter and surgical ASD closure strategies for adults over 5 years are $11,029 and 4.664 life years and $15,467 and 4.557 life years, respectively (Table 4). The transcatheter approach is superior to the surgical treatment with marginally greater effectiveness and a lower cost. Compared with surgical ASD closure, transcatheter ASD closure had lower risk of mortality and CHF, whereas the risks of stroke and A-Fib were similar in both groups. The numbers of events per 1000 hypothetical patients over 5 years in transcatheter and surgical arms were: death: 21 versus 59; stroke: 6 versus 8; CHF: 6 versus 13; and A-Fib: 83 versus 71.
Table 3 Cost inputs.
In-hospital Transcatheter ASD closure Surgical ASD closure Cost of adverse outcomes per annum Stroke Atrial fibrillation Congestive heart failure MUHC; McGill University Health Centre.
Canadian dollars
Source
$9440 $14,648
MUHC MUHC
$10,186 $3978 $3742
Teng et al. [16] Marshall et al. [17] Finlayson et al. [18]
112
D. Mylotte et al. / International Journal of Cardiology 172 (2014) 109–114
Table 4 Base case analysis. Strategy
Total cost (CAD $)
Surgical ASD closure Transcatheter ASD closure
15,467 11,029
Incremental cost (CAD $)
Life-years
Incremental life-years
ICER
−4438
4.557 4.664
0.107
–
ASD: atrial septal defect; CAD $: Canadian dollars; ICER: incremental cost-effectiveness ratio. --: negative ICER not reported.
3.4. Univariate sensitivity analyses When the model inputs were varied, the transcatheter approach remained superior to the surgical approach in almost all scenarios (Table 5). The main parameters influencing incremental effectiveness were the procedural and annual mortality rates. The main parameters impacting the incremental cost were the initial costs of the surgical and transcatheter ASD closure procedure. 3.5. Monte Carlo simulation The results of the first order Monte Carlo simulations are consistent with those in the base case. Specifically, transcatheter closure led to cost savings and equal or greater life-time in comparison to surgical treatment in 79.5% of simulated individuals. The cost and effectiveness of the transcatheter and surgical strategies were $11,060 SD $5169 and 4.683 SD 0.379 life years and $15,304 SD $4581 and 4.618 SD 0.638 life-years, respectively. The incremental cost and the incremental effectiveness were $4243 SD $6765 and 0.065 SD 0.653 life years respectively. The distribution of incremental cost of transcatheter compared to surgical ASD closure is shown in Fig. 2. For almost 80% of simulated individuals, the transcatheter approach is cheaper than the surgical one. 4. Discussion In this population-based study of ASD closure, our base case analysis demonstrated that transcatheter and surgical ASD closure were associated with 5-year costs of $11,029 and $15,467, respectively. In addition, the relative effectiveness also favoured the transcatheter approach with equal or greater life-time compared to surgical treatment. The first order Monte Carlo simulation substantiated the lower costs and improved clinical effectiveness associated with percutaneous ASD closure. Since the first transcatheter closure of a secundum ASD was reported in 1976 [25], several studies have attempted to describe and compare the costs of transcatheter and surgical ASD closure [9–14]. However, all previous analyses have been limited to the costs associated with the initial hospitalisation. A true comparison of cost-effectiveness of
surgical and transcatheter ASD closure requires consideration of both in-hospital and long-term costs. In our base case analysis, we found that transcatheter ASD closure was less expensive than surgical ASD closure. The cost of the initial hospitalisation was reduced by $5208 or by 35.6% with transcatheter closure. When the cost of adverse events over the 5-year follow-up period was considered, transcatheter ASD closure remained $4438 or 28.7% less expensive than the surgical alternative. In addition, transcatheter closure increased life expectancy over 5 years of follow-up. The Monte Carlo simulation corroborated these results, with the transcatheter approach yielding lower costs than the surgical approach in 79.5% of the simulated individuals. Furthermore, sensitivity analyses confirmed that our findings are robust to changes in the base case model input parameters. A number of variables were entered into the model to determine the cost of both interventions at 5 years. Not surprisingly, the cost of the initial procedure had the greatest impact on the cost differential of the two approaches. In particular, the cost of the Amplatzer septal occluder was the greatest contributor to the transcatheter costs and the inpatient costs the greatest contributor in the surgical group. While the cost of the Amplatzer device is consistent across all implanting sites in Quebec, one must consider that surgical expertise can vary considerably between institutions. Indeed, a recent study found that minimally invasive surgical techniques for ASD closure reduced both hospital stay and cost to levels seen with transcatheter closure [10]. Thus, there remains potential for novel surgical techniques to enhance the cost-effectiveness of ASD closure. There were some differences in baseline characteristics between groups. The transcatheter group was older and had a higher incidence of comorbidity, with the exception of pulmonary hypertension, which was more common in the surgical group. We have previously demonstrated however, that the excess mortality in the surgical cohort remained consistent after adjusting for pulmonary hypertension and after excluding patients with pulmonary hypertension [24]. When comparing the cost-effectiveness of transcatheter and surgical procedures, the cost of significant morbidities associated with either intervention must be considered. In many instances, the cost of major morbidity, such as stroke, is far more costly than the procedure itself. We found that the cumulative incidence of stroke, CHF and A-Fib was
Table 5 Univariate sensitivity analysis. Variable, range
Incremental costs (CAD $)
Incremental effectiveness (life-year)
ICER (CAD $/life-year)
Follow up duration (3 to 10 years) Annual discount rate (0 to 5%) Success rate of transcatheter ASD closure (80 to 99%) Procedural mortality Transcatheter (0 to 1%) Surgical (0.5 to 2%) Annual mortality Transcatheter (0.36 to 1%) Surgical (0.5 to 1.5%) Procedural cost Transcatheter ($7000$ to $15,000) Surgical ($10,000 to $20,000)
−4462 to −4297 −4434 to −4441 −2259 to −5034
0.049 to 0.319 0.115 to 0.102 0.107 to 0.107
– – –
−4434 to −4450 −4443 to −4431
0.121 to 0.074 0.080 to 0.151
– –
−4438 to −4456 −4452 to −4425
0.107 to 0.035 0.051 to 0.161
– –
−6878 to 1122 −26 to −9520
0.107 to 0.107 0.107 to 0.107
– to 10,523 –
Incremental cost = accrued cost in transcatheter group − accrued cost in surgery group. Incremental effectiveness = accrued life-year in transcatheter group − accrued life-year in surgery group. CAD $: Canadian dollars; ICER: incremental cost-effectiveness ratio. --: ICER not reported when negative.
D. Mylotte et al. / International Journal of Cardiology 172 (2014) 109–114
113
Fig. 2. Distribution of incremental cost of transcatheter versus surgical ASD closure. The grey line represents the zero of the incremental cost where the two approaches would hypothetically cost the same. The left side of the dashed line indicates cost saving by a transcatheter approach.
higher in the surgical group compared to the transcatheter group. Although the duration of follow-up was significantly longer in the surgical group, the follow-up was restricted to the median follow-up among groups (5 years). The different time frames of follow-up are unlikely to have significantly influenced the results since the hazard ratios are defined per unit of time and because any differential events relating to the index intervention decrease with increasing follow-up. The cost to society of these adverse events is substantial over the lifetime of a patient [16–18]. In the current study, the higher incidence of adverse events in the surgical group at 5 years ensured that the cost of surgical ASD closure remained higher than that of the transcatheter approach. Put simply, our results demonstrate that in routine clinical practice, transcatheter ASD closure is a less expensive and more effective treatment than surgical ASD closure. When combined with evidence of comparable clinical effectiveness [24], these data support current clinical trends of adopting less invasive, innovative technology to treat patients with CHD. In light of escalating health care costs, physicians, the medical device industry, and other stakeholders have a responsibility to develop a coherent strategy for the judicious use of such technology. Thus, we provide important information, in keeping with the principles of the Patient Protection and Affordable Care Act [6], which confirms the clinical and economic efficacy of transcatheter ASD closure. There remain important anatomical criteria that determine the feasibility and safety of transcatheter ASD closure: large ASDs (N 30 mm) and those with insufficient rims may be more suitable for surgical intervention. Thus, surgical ASD closure remains a vital therapeutic option for a significant proportion of patients.
4.1. Limitations Our study has limitations. As effectiveness was determined using observational rather than randomized data, there were differences in baseline characteristics between the study groups. These differences could have impacted on in-hospital and long-term outcomes. However, the Cox regression only demonstrated increasing patient age and the Charlson comorbidity index to be predictive of 5-year mortality, and as both of these indices were significantly higher in the transcatheter group, the study conclusions should remain unchanged. The unequal duration of follow-up also introduces the possibility of secular changes in general hospital care that could also be a confounding factor. We restricted the duration of patient follow-up to 5 years as this was the median time of follow-up between groups and because this period was likely to capture all events related to the index procedure. However, a longer observation period may have been more informative. Transcatheter therapies have been associated with enhanced short-term quality of life compared with their surgical equivalents, though these
differences may be reversed in the longer term [26]. Calculating quality-adjusted life years (QALY) may be an appropriate method to compare cost effectiveness between these strategies, though there are limitations to this approach, particularly when using observational or population data [27,28]. Given that well measured QALYs for our study population do not exist as well as the controversy over their use, we elected not to calculate them in this study. We did not have access to patient level data that would have afforded more detailed cost analysis for individual patients. Specifically, information on medications used in hospital or in the follow-up period was not available. Differences in the use of medications, such as antiplatelet therapies, between groups could have impacted on the rates of adverse events. However, the higher rate of major morbidity observed in the surgical group implies a greater requirement for additional therapies and thus, the conclusions of the study would not have been altered. The anatomical characteristics of the ASDs treated in each group were not available. It is likely that larger and more complex lesions were treated in the surgical group and this may have impacted on clinical outcomes. Although we used 2009 costs and applied these costs retrospectively to the onset of our observation period in 1998, we have no evidence that the cost variations for the two techniques varied differentially over time. As we assumed that the costs associated with patient management following ASD closure were identical between the two groups, our estimates should not be interpreted as the total direct costs of these interventions in Quebec. Finally, the results of this analysis may not necessarily be transferrable to other health care systems, particularly in developing nations where the cost of medical device technology can have a greater influence on the final cost analysis [29]. 5. Conclusion Transcatheter ASD closure appears to be a more cost effective and potentially clinically effective treatment compared to surgical ASD closure. Acknowledgements The authors would like to thank Ms. Liming Guo for her help in statistical analysis and figure preparation. References [1] Marelli AJ, Mackie AS, Ionescu-Ittu R, Rahme E, Pilote L. Congenital heart disease in the general population: changing prevalence and age distribution. Circulation 2007;115:163–72. [2] Krasuski RA, Bashore TM. The emerging role of percutaneous intervention in adults with congenital heart disease. Rev Cardiovasc Med 2005;6:11–22. [3] Chan KC, Godman MJ, Walsh K, Wilson N, Redington A, Gibbs JL. Transcatheter closure of atrial septal defect and interatrial communications with a new self
114
[4]
[5] [6] [7] [8] [9] [10]
[11]
[12]
[13]
[14]
[15]
D. Mylotte et al. / International Journal of Cardiology 172 (2014) 109–114 expanding nitinol double disc device (Amplatzer septal occluder): multicentre UK experience. Heart 1999;82:300–6. Warnes CA, Williams RG, Bashore TM, et al. ACC/AHA 2008 guidelines for the management of adults with congenital heart disease: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (Writing Committee to Develop Guidelines on the Management of Adults With Congenital Heart Disease). Developed in Collaboration With the American Society of Echocardiography, Heart Rhythm Society, International Society for Adult Congenital Heart Disease, Society for Cardiovascular Angiography and Interventions, and Society of Thoracic Surgeons. J Am Coll Cardiol 2008;52:e143–263. VanLare JM, Conway PH, Sox HC. Five next steps for a new national program for comparative-effectiveness research. N Engl J Med 2010;362:970–3. Mushlin AI, Ghomrawi H. Health care reform and the need for comparativeeffectiveness research. N Engl J Med 2010;362:e6. Garber AM, Sox HC. The role of costs in comparative effectiveness research. Health Aff (Millwood) 2010;29:1805–11. Peterson ED. Innovation and comparative-effectiveness research in cardiac surgery. N Engl J Med 2009;361:1897–9. Baker SS, O'Laughlin MP, Jollis JG, Harrison JK, Sanders SP, Li JS. Cost implications of closure of atrial septal defect. Catheter Cardiovasc Interv 2002;55:83–7. Formigari R, Di Donato RM, Mazzera E, et al. Minimally invasive or interventional repair of atrial septal defects in children: experience in 171 cases and comparison with conventional strategies. J Am Coll Cardiol 2001;37:1707–12. Thomson JD, Aburawi EH, Watterson KG, Van Doorn C, Gibbs JL. Surgical and transcatheter (Amplatzer) closure of atrial septal defects: a prospective comparison of results and cost. Heart 2002;87:466–9. Hughes ML, Maskell G, Goh TH, Wilkinson JL. Prospective comparison of costs and short term health outcomes of surgical versus device closure of atrial septal defect in children. Heart 2002;88:67–70. Du ZD, Hijazi ZM, Kleinman CS, Silverman NH, Larntz K. Comparison between transcatheter and surgical closure of secundum atrial septal defect in children and adults: results of a multicenter nonrandomized trial. J Am Coll Cardiol 2002;39:1836–44. Kim JJ, Hijazi ZM. Clinical outcomes and costs of Amplatzer transcatheter closure as compared with surgical closure of ostium secundum atrial septal defects. Med Sci Monit 2002;8:CR787–91. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis 1987;40:373–83.
[16] Teng J, Mayo NE, Latimer E, et al. Costs and caregiver consequences of early supported discharge for stroke patients. Stroke 2003;34:528–36. [17] Marshall DA, Levy AR, Vidaillet H, et al. Cost-effectiveness of rhythm versus rate control in atrial fibrillation. Ann Intern Med 2004;141:653–61. [18] Finlayson G, Reimer J, Dahl M, Stargardter M, McGowan K. The direct cost of hospitalizations in Manitoba, 2005/06. Manitoba Centre for Health Policy. Available at: http://mchp-appserv.cpe.umanitoba.ca; 2009. [19] Caro JJ. Pharmacoeconomic analyses using discrete event simulation. Pharmacoeconomics 2005;23:323–32. [20] Jahn B, Pfeiffer KP, Theurl E, Tarride JE, Goeree R. Capacity constraints and costeffectiveness: a discrete event simulation for drug-eluting stents. Med Decis Making 2010;30:16–28. [21] Deniz HB, Ward A, Jaime Caro J, Alvarez P, Sadri H. Cost–benefit analysis of primary prevention of sudden cardiac death with an implantable cardioverter defibrillator versus amiodarone in Canada. Curr Med Res Opin 2009;25: 617–26. [22] Najafzadeh M, Marra CA, Galanis E, Patrick DM. Cost effectiveness of herpes zoster vaccine in Canada. Pharmacoeconomics 2009;27:991–1004. [23] Briggs AH. Handling uncertainty in cost-effectiveness models. Pharmacoeconomics 2000;17:479–500. [24] Kotowycz MA, Therrien J, Ionescu-Ittu R, et al. Long-term outcomes after surgical versus transcatheter closure of atrial septal defects in adults. JACC Cardiovasc Interv 2013;6:497–503. [25] King TD, Thompson SL, Steiner C, Mills NL. Secundum atrial septal defect. Nonoperative closure during cardiac catheterization. JAMA 1976;235:2506–9. [26] Cohen DJ, Van Hout B, Serruys PW, et al. Quality of life after PCI with drugeluting stents or coronary-artery bypass surgery. N Engl J Med 2011; 364:1016–26. [27] Neyt M, Cleemput I, Thiry N, De Laet C. Calculating an intervention's (cost-) effectiveness for the real-world target population: the potential of combining strengths of both RCTs and observational data. Health Policy 2012;106: 207–10. [28] Prieto L, Sacristan JA. Problems and solutions in calculating quality-adjusted life years (QALYs). Health Qual Life Outcomes 2003;1:80. [29] Vida VL, Barnoya J, O'Connell M, Leon-Wyss J, Larrazabal LA, Castaneda AR. Surgical versus percutaneous occlusion of ostium secundum atrial septal defects: results and cost-effective considerations in a low-income country. J Am Coll Cardiol 2006; 47:326–31.
Contents lists available at ScienceDirect
International Journal of Cardiology journal homepage: www.elsevier.com/locate/ijcard
Long-term cost-effectiveness of transcatheter versus surgical closure of secundum atrial septal defect in adults☆ Darren Mylotte a,1, Stéphane P. Quenneville a,1, Mark A. Kotowycz a,1, Xuanqian Xie b,1, James M. Brophy c,d,1, Raluca Ionescu-Ittu a,1, Giuseppe Martucci a,1, Louise Pilote e,1, Judith Therrien a,1, Ariane J. Marelli a,⁎,1 a
McGill Adult Unit for Congenital Heart Disease Excellence (MAUDE Unit), McGill University Health Centre, Canada Technology Assessment Unit, McGill University Health Centre, Canada Department of Medicine, McGill University, Canada d Department of Epidemiology & Biostatistics, McGill University, Canada e Division of Clinical Epidemiology, McGill University Health Centre, Canada b c
a r t i c l e
i n f o
Article history: Received 18 July 2013 Received in revised form 2 December 2013 Accepted 26 December 2013 Available online 8 January 2014 Keywords: Atrial septal defect Surgery Transcatheter Adult congenital heart disease Cost-effectiveness
a b s t r a c t Background: The most common congenital anomaly in adults is secundum atrial septal defect (ASD), which can be closed using a surgical or transcatheter approach. Despite the growing use of transcatheter ASD closure, few studies have examined the cost-effectiveness of this strategy. We sought to compare the long-term cost effectiveness of transcatheter and surgical closure of secundum ASD in adults. Methods: A decision-analytic model was used with all clinical outcome parameter estimates obtained from the province-wide Québec Congenital Heart Disease Database. Costs were obtained from a single academic centre (Canadian dollars). A cost-effectiveness analysis using a discrete event Monte Carlo simulation model from the perspective of a single third party payer and multiple sensitivity analyses were performed. Patients were followed for a maximum of 5 years after ASD closure. Results: Between l998 and 2005, we identified 718 adults (n = 335 transcatheter; n = 383 surgical) who underwent ASD closure in Quebec. The 5-year cost of surgical closure was $15,304 SD $4581 versus $11,060 SD $5169 for the transcatheter alternative. At 5 years, transcatheter closure was marginally more effective than surgery (4.683 SD 0.379 life-years versus 4.618 SD 0.638 life-years). Probabilistic sensitivity analyses demonstrated that transcatheter ASD closure was a dominant strategy with an 80% probability of cost savings and equal or greater efficacy compared to surgical treatment. Conclusion: Although definitive conclusions are limited given the observational nature of the primary data sources, transcatheter ASD closure appeared to be a cost-effective strategy associated with slightly improved clinical outcomes and reduced costs compared to surgical closure at 5-years follow-up. © 2014 Published by Elsevier Ireland Ltd.
1. Introduction The population of adults living with congenital heart disease (CHD) has grown significantly over the last decades [1], with atrial septal defect (ASD) being the most common lesion. While ASD can be closed either surgically or with a transcatheter device, the less invasive transcatheter strategy has become the de facto accepted standard of care
☆ Sources of funding: Dr. Mylotte is funded by the Beth Raby Fellowship. Dr. Brophy is a research scholar, Les Fonds de la Recherche en Santé du Québec. Dr. Ionescu-Ittu was funded by Les Fonds de la Recherche du Québec (postdoctoral scholarship) at the time that the study was conducted. Dr. Marelli is a research scholar, Les Fonds de la Recherche en Santé du Québec and is funded by the Heart and Stroke Foundation of Québec and the Canadian Institutes of Health Research.There are no relationships with industry. ⁎ Corresponding author at: McGill Adult Unit for Congenital Heart Disease (MAUDE Unit), 687 Pine Avenue West, Room H4.33, Montréal, Québec H3A 1A1, Canada. E-mail address: [email protected] (A.J. Marelli). 1 This author takes responsibility for all aspects of the reliability and freedom from bias of the data presented and their discussed interpretation. 0167-5273/$ – see front matter © 2014 Published by Elsevier Ireland Ltd. http://dx.doi.org/10.1016/j.ijcard.2013.12.144
[2–4] despite a paucity of randomized clinical trials and long-term cost-effectiveness studies. There is an increasing interest in comparative effectiveness studies to examine interventional outcomes in a setting that reflects clinical practice [5]. In 2010, The Patient Protection and Affordable Care Act established comparative effectiveness research as central to health care improvement [6]. Although cost-effectiveness is not an outcome as universally accepted as clinical effectiveness, it provides an important measure of the value of health care particularly for costly procedures [7,8]. Because the use of transcatheter closure in ASD has become so wide-spread, it is relevant to examine its cost-effectiveness as an important dimension of comparative effectiveness. Several small studies have previously reported reduced in-hospital and procedural costs with transcatheter compared to surgical ASD closure [9–14]. However, the cost-effectiveness of novel medical technologies can only be sensibly interpreted when applied to the whole population susceptible to receiving the intervention and when both medium- and long-term clinical outcomes and costs are
110
D. Mylotte et al. / International Journal of Cardiology 172 (2014) 109–114
systematically evaluated. To the best of our knowledge, a comparative cost-effectiveness analysis at the population level has not been performed for ASD closure and this research aims to address this knowledge gap.
a procedural code for transcatheter ASD closure or a surgical code for ASD closure, billed during an observation period that started with the day following the index procedure and continued for the duration of follow-up.
2.3. Statistical analysis 2. Methods We used population-based health service data sources from Québec, Canada and relied on a discrete event Monte Carlo simulation model inputting effectiveness parameters from our large and previously validated administrative databases [1]. Cost parameters were derived from the McGill University Health Centre in Montreal, Quebec and from the literature. 2.1. Data input for effectiveness parameters 2.1.1. Data sources In Quebec, Canada there is universal access to health care and a unique health care number is assigned to every individual at birth. This number is used to record all diagnoses, hospitalisations and health service encounters throughout an individual's lifetime. This information is recorded into two databases: the medical claims database of the provincial authority, the Regie de l'Assurance Maladie du Quebec (RAMQ), and the hospital provincial discharge summary database (Med-Echo). The RAMQ database also includes patient demographic information, including sex, and dates of birth and death. The Quebec CHD database was created by merging these two databases and developing algorithms to correctly identify CHD patients and extract CHD diagnoses [1]. Diagnoses recorded in the databases adhered to the International Classification of Diseases, Ninth Revision (ICD-9). Patients that had at least one diagnostic code for CHD and/or a CHD-specific surgical procedure were identified. Provider codes were used to select those CHD billings made by cardiovascular disease specialists, imaging specialists, and primary care physicians. Patients were assigned one or two CHD diagnoses via the use of a previously described hierarchical algorithm [1]. All information was cross-referenced between outpatient and inpatient data sources. Our CHD database therefore contains comprehensive longitudinal, demographic, diagnostic and therapeutic records of all patient encounters with the provincial health care system between January 1, 1983 and December 31, 2005 inclusively. In Québec death attestation is sent to the Québec Health Insurance Board, which then updates the medical claims database with this information, thereby making the documentation of death complete in the database, regardless whether death occurred in or out of hospital. 2.2. Study population Our study population, derived from the Québec CHD database, consisted of all adult patients aged N18 and ≤75 years with both ASD diagnosis and either (i) a procedural code for transcatheter ASD closure, billed by an interventional cardiologist in the period 1998–2005, or (ii) a procedural code for surgical ASD closure, billed by a cardiac surgeon or general surgeon in the same period. To ensure that the study population included only patients with secundum ASD, the ASD patients who also had codes for other congenital anomalies were excluded. We also excluded patients who had surgical ASD closure billed in the same day with other cardiac surgeries, such as coronary artery bypass grafting surgery, aortic arch surgery, or valvular surgery. The ICD-9 code for secundum ASD (745.5) does not differentiate between secundum ASD and patent foramen ovale (PFO). Given that the majority of PFO closures occur in patients with cryptogenic stroke, we also excluded patients with stroke in the year prior to septal closure. Baseline characteristics included age, sex, and the following comorbidities, measured in the 5 years before intervention: coronary artery disease, hypertension, diabetes mellitus, congestive heart failure (CHF), pulmonary hypertension, and atrial fibrillation (A-Fib). Comorbidities were defined using codes from the ICD-9. We also included the Charlson comorbidity index, an aggregate measure of patients' overall burden of disease [15]. We followed patients for a maximum of 5 years following ASD closure, until the earliest of one of the following: death, end of the 5-year follow-up, or the end of the study on Dec 31, 2005. Procedure-related mortality was defined as death within 30 days of ASD closure. Long-term mortality was defined as death within 5 years of the index ASD closure. Other clinical outcomes that were assessed during the 5-year follow-up included ASD closure success, stroke, CHF and A-Fib. ASD closure success was defined as freedom from a further procedural code for ASD closure. Re-intervention was defined as
Baseline descriptive statistics of the two intervention groups are presented with categorical and continuous variables presented as percentages and medians (25th percentile, 75th percentile), respectively. Clinical outcomes within 5 years of the index ASD closure are described as unadjusted (i) percentages over the 5 years of follow-up and (ii) means and standard errors of annual event rates. The annual event rates were obtained by transforming the cumulative incidence rates for mortality, stroke, CHF and A-Fib in the two groups into annual incidence. We assumed that the risks of all events were constant during the entire follow-up period and random censoring was non-informative in both arms. In the case of low cumulative incidence as the ones observed in this study, these assumptions imply that the annual event rate approximates the cumulative incidence rate divided by the median years of follow-up. The standard errors of the annual event rates were calculated using approximate approach to binomial distribution. Data were analysed on the intention-to-treat principle, with patients assigned to the surgical or catheterisation groups at the time of their first ASD closure intervention recorded in the database. Cox regression analysis was performed to determine if any baseline characteristic predicted 5-year mortality. All statistical tests were 2-sided, and p values b0.05 were considered to be statistically significant. Data was analysed using SAS 9.2.
2.4. Data input for cost parameters Costs were evaluated from the perspective of the Canadian health care system and thus, focused on direct health care costs, and physician fees associated with ASD closure. These cost data were obtained directly from the accounting and resource departments of the McGill University Health Centre (MUHC) for the year 2009. In a province with universal health care such as Quebec, costs are not expected to vary across centres and costs estimated at one centre are expected to generalize to other centres in the province. We sought to estimate the incremental cost of both treatment strategies. Since patients are managed identically after being discharged from hospital following the index transcatheter or surgical ASD closure, the usual care costs incurred during these post procedural health states are balanced and were ignored. We also included the costs of secondary outcomes: stroke, CHF and A-Fib during follow-up. These costs were derived from published literature in Canadian populations [16–18]. All costs were expressed in 2010 Canadian dollars ($).
2.5. Cost-effectiveness model A discrete-event simulation (DES) model was developed to evaluate the clinical and economic consequences of surgical and transcatheter ASD closure strategies [19]. The DES model is governed by a chronologically ordered event list, and a finite number of discrete changes in a state (versus continuous model), and follow events that have occurred. DES is commonly used in physical systems, but has also been applied recently in the heath care field, as it allows complex modelling including that for waiting lists [20], competing resources [20], patients' baseline characteristics [21], and/or disease progression [22]. Our model starts with patients who meet clinical criteria for surgical or transcatheter ASD closure (Fig. 1). Simulated paired individuals (identical clones) were assigned to undergo surgical or transcatheter ASD closure. Our model presents the patient's course naturally, according to a scheduled order in which the potential events (e.g. operation, procedure-related mortality, re-operation, stroke, CHF, A-Fib, and death) can occur [19]. Each event was associated with resource consumption. For simplicity, minor procedural complications were not included in the model. We assumed, conservatively, that if the first transcatheter intervention was not successful, patients automatically underwent surgical closure with a 100% success rate and no procedure-related mortality. After successful ASD closure, patients may go on to develop stroke, CHF, A-Fib and to succumb to death with defined probabilities, based on our analyses of clinical outcomes observed in the Quebec CHD database. Simulated patients that did not succumb to death during follow-up were censored at 5 years.
Fig. 1. Discrete-event simulation model. Flow diagram of patients in the study. ASD = atrial septal defect; CHF = congestive heart failure; A-Fib = atrial fibrillation.
D. Mylotte et al. / International Journal of Cardiology 172 (2014) 109–114 2.6. Cost-effectiveness analysis
111
Table 2 Clinical outcomes.
Using our DES model, we compared the cost and effectiveness of the transcatheter and surgical strategies for ASD closure. The principal outcome measured was the cost of each procedure. We also present other outcomes, such as average life years, number of fatal and non-fatal events per 1000 patients, incremental cost and incremental effectiveness within 5 years. We applied the deterministic model as base case, averaging estimation for all parameters. We examined the robustness of our base case estimates by varying different inputs within our model using a univariate sensitivity analysis. The following inputs were varied: duration of follow-up; annual discount rate; success rate of transcatheter ASD closure; procedural cost; procedural mortality; and annual mortality. To explore patient level variability, we simulated 10,000-paired individuals using the stochastic DES model by Monte Carlo simulation. The Monte Carlo simulation is a technique that uses repeated random sampling to approximate the probability of certain outcomes. Since this is a stochastic method based on chance, it captures the effects of random uncertainty. All stochastic parameters are presented as distributions. Cost data must be positive and is usually skewed, and therefore was assigned to Gamma distributions [23]. Any clinical input (i.e. incidence rate) bounded between 0 and 1 was assigned to Beta distributions [23]. We arbitrarily defined the standard deviation at 25% of the mean, when variances of costs were not available. Because the device used in MUHC during the study period was the Amplatzer septal occluder (St. Jude Medical, Inc. St. Paul, MN, USA)., we also performed a sensitivity analysis in order to test the robustness of our results with different device costs. In this analysis, the cost input for surgery included: the operating room; intensive care unit stay; cardiac surgery ward stay; and physician fees for the procedure and the index hospitalisation. The cost inputs for the transcatheter procedure were: the Amplatzer septal occluder; delivery catheters; catheterisation suite; ancillary equipment; coronary care unit stay; and physician fees. Both costs and life-years were accrued over a 5-year period. This period was selected because it captures all re-interventions and also because the number of patients undergoing catheterisation who have more than 5 years is too small to draw accurate inferences. An annual discount rate of 3% was applied for costs as well as life years. Analyses were conducted using Treeage 2010 and Excel 2003.
Death Procedural Long-term Successful ASD closure Stroke Congestive heart failure Atrial fibrillation Annual event rate, %, SEMa Death Stroke Congestive heart failure Atrial fibrillation
Transcatheter closure (N = 335)
Surgical closure (N = 383)
4 (1.2) 1 (0.3) 3 (0.9) 317 (94.9) 1 (0.3) 1 (0.3) 14 (4.2)
23 (6.0) 4 (1.0) 19 (5.0) 383 (100) 3 (0.8) 5 (1.3) 28 (7.4)
0.4 SEM 0.3 0.1 SEM 0.2 0.1 SEM 0.2 1.7 SEM 0.7
1.0 SEM 0.1 0.2 SEM 0.2 0.3 SEM 0.3 1.5 SEM 0.6
Data are presented as number and percentage. ASD = atrial septal defect; SEM = standard error of the mean. a SEM was calculated using the approximate approach to binomial distribution.
Clinical outcomes are summarized in Table 2. Compared with surgical treatment, the transcatheter approach was associated with a lower risk of procedure-related mortality (0.3% versus 1.0%), lower annual mortality (0.4% versus 1.0%), and lower procedural success rates (94.9% versus 100%).
3.2. Costs 3. Results 3.1. Clinical outcomes Between 1988 and 2005 we identified 335 adult patients who had transcatheter ASD closure and 383 who had surgical closure. The clinical characteristics and outcomes of these patient groups have previously been described [24]. The number of transcatheter closures grew rapidly after the technique was introduced, with a corresponding decrease in the number of surgical closures. The patients who underwent transcatheter closure were older, had a higher incidence of hypertension, and a higher Charlson comorbidity index than those who underwent surgical closure (Table 1). One exception was the history of pulmonary hypertension, which was more common in the surgical group. Cox regression analysis identified increasing patient age (Hazard Ratio [HR] 1.057; 95% confidence intervals [CI]: 1.026–1.088, p = 0.0002) and the Charlson comorbidity index (HR 1.241; 95% CI 1.128–1.364, p b 0.0001) as the only baseline characteristics to independently predict 5-year mortality. The median follow-up time was significantly shorter in the transcatheter group compared to the surgical group, 3.0 (1.2, 3.9) and 10.0 (7.0, 13.0) years, respectively.
Table 1 Baseline characteristics of study population.
Age, years Male gender Coronary artery disease Hypertension Diabetes mellitus Congestive heart failure Pulmonary hypertension Atrial fibrillation Charlson comorbidity index Median follow-up, years
Transcatheter closure (N = 335)
Surgical closure (N = 383)
p-Value
49 (39–61) 102 (30.5) 4 (1.3) 97 (32.0) 20 (6.6) 37 (12.2) 23 (7.6) 64 (21.1) 2.3 (0.2–4.0) 3 (1.2, 3.9)
43 (32–54) 110 (28.7) 15 (4.0) 94 (24.8) 15 (4.0) 63 (16.6) 62 (16.4) 76 (20.0) 2.0 (0.1–3.0) 10.0 (7.0, 13.0)
b0.001 0.61 0.06 0.04 0.12 0.11 b0.001 0.73 b0.001 b0.001
Values are presented as N (%) or median (interquartile range). ASD = atrial septal defect.
Costs of transcatheter ASD and surgical ASD closure were derived from the MUHC costing centre (Table 3). The unit costs of surgical ASD closure included the operating room ($5800), intensive care unit stay ($1387), cardiac surgery ward stay ($3840) and physician fees for the surgical procedure and the index hospitalisation ($3421). The unit costs of transcatheter ASD closure included the Amplatzer septal occluder ($5500), delivery catheters ($480), the catheterisation suite ($484), ancillary equipment ($700), coronary care unit stay ($631), and physician fees ($900).
3.3. Base case analysis of economic model Based on the model proposed in Fig. 1 and using the parameter estimates given in Tables 2 and 3, the cost and effectiveness of transcatheter and surgical ASD closure strategies for adults over 5 years are $11,029 and 4.664 life years and $15,467 and 4.557 life years, respectively (Table 4). The transcatheter approach is superior to the surgical treatment with marginally greater effectiveness and a lower cost. Compared with surgical ASD closure, transcatheter ASD closure had lower risk of mortality and CHF, whereas the risks of stroke and A-Fib were similar in both groups. The numbers of events per 1000 hypothetical patients over 5 years in transcatheter and surgical arms were: death: 21 versus 59; stroke: 6 versus 8; CHF: 6 versus 13; and A-Fib: 83 versus 71.
Table 3 Cost inputs.
In-hospital Transcatheter ASD closure Surgical ASD closure Cost of adverse outcomes per annum Stroke Atrial fibrillation Congestive heart failure MUHC; McGill University Health Centre.
Canadian dollars
Source
$9440 $14,648
MUHC MUHC
$10,186 $3978 $3742
Teng et al. [16] Marshall et al. [17] Finlayson et al. [18]
112
D. Mylotte et al. / International Journal of Cardiology 172 (2014) 109–114
Table 4 Base case analysis. Strategy
Total cost (CAD $)
Surgical ASD closure Transcatheter ASD closure
15,467 11,029
Incremental cost (CAD $)
Life-years
Incremental life-years
ICER
−4438
4.557 4.664
0.107
–
ASD: atrial septal defect; CAD $: Canadian dollars; ICER: incremental cost-effectiveness ratio. --: negative ICER not reported.
3.4. Univariate sensitivity analyses When the model inputs were varied, the transcatheter approach remained superior to the surgical approach in almost all scenarios (Table 5). The main parameters influencing incremental effectiveness were the procedural and annual mortality rates. The main parameters impacting the incremental cost were the initial costs of the surgical and transcatheter ASD closure procedure. 3.5. Monte Carlo simulation The results of the first order Monte Carlo simulations are consistent with those in the base case. Specifically, transcatheter closure led to cost savings and equal or greater life-time in comparison to surgical treatment in 79.5% of simulated individuals. The cost and effectiveness of the transcatheter and surgical strategies were $11,060 SD $5169 and 4.683 SD 0.379 life years and $15,304 SD $4581 and 4.618 SD 0.638 life-years, respectively. The incremental cost and the incremental effectiveness were $4243 SD $6765 and 0.065 SD 0.653 life years respectively. The distribution of incremental cost of transcatheter compared to surgical ASD closure is shown in Fig. 2. For almost 80% of simulated individuals, the transcatheter approach is cheaper than the surgical one. 4. Discussion In this population-based study of ASD closure, our base case analysis demonstrated that transcatheter and surgical ASD closure were associated with 5-year costs of $11,029 and $15,467, respectively. In addition, the relative effectiveness also favoured the transcatheter approach with equal or greater life-time compared to surgical treatment. The first order Monte Carlo simulation substantiated the lower costs and improved clinical effectiveness associated with percutaneous ASD closure. Since the first transcatheter closure of a secundum ASD was reported in 1976 [25], several studies have attempted to describe and compare the costs of transcatheter and surgical ASD closure [9–14]. However, all previous analyses have been limited to the costs associated with the initial hospitalisation. A true comparison of cost-effectiveness of
surgical and transcatheter ASD closure requires consideration of both in-hospital and long-term costs. In our base case analysis, we found that transcatheter ASD closure was less expensive than surgical ASD closure. The cost of the initial hospitalisation was reduced by $5208 or by 35.6% with transcatheter closure. When the cost of adverse events over the 5-year follow-up period was considered, transcatheter ASD closure remained $4438 or 28.7% less expensive than the surgical alternative. In addition, transcatheter closure increased life expectancy over 5 years of follow-up. The Monte Carlo simulation corroborated these results, with the transcatheter approach yielding lower costs than the surgical approach in 79.5% of the simulated individuals. Furthermore, sensitivity analyses confirmed that our findings are robust to changes in the base case model input parameters. A number of variables were entered into the model to determine the cost of both interventions at 5 years. Not surprisingly, the cost of the initial procedure had the greatest impact on the cost differential of the two approaches. In particular, the cost of the Amplatzer septal occluder was the greatest contributor to the transcatheter costs and the inpatient costs the greatest contributor in the surgical group. While the cost of the Amplatzer device is consistent across all implanting sites in Quebec, one must consider that surgical expertise can vary considerably between institutions. Indeed, a recent study found that minimally invasive surgical techniques for ASD closure reduced both hospital stay and cost to levels seen with transcatheter closure [10]. Thus, there remains potential for novel surgical techniques to enhance the cost-effectiveness of ASD closure. There were some differences in baseline characteristics between groups. The transcatheter group was older and had a higher incidence of comorbidity, with the exception of pulmonary hypertension, which was more common in the surgical group. We have previously demonstrated however, that the excess mortality in the surgical cohort remained consistent after adjusting for pulmonary hypertension and after excluding patients with pulmonary hypertension [24]. When comparing the cost-effectiveness of transcatheter and surgical procedures, the cost of significant morbidities associated with either intervention must be considered. In many instances, the cost of major morbidity, such as stroke, is far more costly than the procedure itself. We found that the cumulative incidence of stroke, CHF and A-Fib was
Table 5 Univariate sensitivity analysis. Variable, range
Incremental costs (CAD $)
Incremental effectiveness (life-year)
ICER (CAD $/life-year)
Follow up duration (3 to 10 years) Annual discount rate (0 to 5%) Success rate of transcatheter ASD closure (80 to 99%) Procedural mortality Transcatheter (0 to 1%) Surgical (0.5 to 2%) Annual mortality Transcatheter (0.36 to 1%) Surgical (0.5 to 1.5%) Procedural cost Transcatheter ($7000$ to $15,000) Surgical ($10,000 to $20,000)
−4462 to −4297 −4434 to −4441 −2259 to −5034
0.049 to 0.319 0.115 to 0.102 0.107 to 0.107
– – –
−4434 to −4450 −4443 to −4431
0.121 to 0.074 0.080 to 0.151
– –
−4438 to −4456 −4452 to −4425
0.107 to 0.035 0.051 to 0.161
– –
−6878 to 1122 −26 to −9520
0.107 to 0.107 0.107 to 0.107
– to 10,523 –
Incremental cost = accrued cost in transcatheter group − accrued cost in surgery group. Incremental effectiveness = accrued life-year in transcatheter group − accrued life-year in surgery group. CAD $: Canadian dollars; ICER: incremental cost-effectiveness ratio. --: ICER not reported when negative.
D. Mylotte et al. / International Journal of Cardiology 172 (2014) 109–114
113
Fig. 2. Distribution of incremental cost of transcatheter versus surgical ASD closure. The grey line represents the zero of the incremental cost where the two approaches would hypothetically cost the same. The left side of the dashed line indicates cost saving by a transcatheter approach.
higher in the surgical group compared to the transcatheter group. Although the duration of follow-up was significantly longer in the surgical group, the follow-up was restricted to the median follow-up among groups (5 years). The different time frames of follow-up are unlikely to have significantly influenced the results since the hazard ratios are defined per unit of time and because any differential events relating to the index intervention decrease with increasing follow-up. The cost to society of these adverse events is substantial over the lifetime of a patient [16–18]. In the current study, the higher incidence of adverse events in the surgical group at 5 years ensured that the cost of surgical ASD closure remained higher than that of the transcatheter approach. Put simply, our results demonstrate that in routine clinical practice, transcatheter ASD closure is a less expensive and more effective treatment than surgical ASD closure. When combined with evidence of comparable clinical effectiveness [24], these data support current clinical trends of adopting less invasive, innovative technology to treat patients with CHD. In light of escalating health care costs, physicians, the medical device industry, and other stakeholders have a responsibility to develop a coherent strategy for the judicious use of such technology. Thus, we provide important information, in keeping with the principles of the Patient Protection and Affordable Care Act [6], which confirms the clinical and economic efficacy of transcatheter ASD closure. There remain important anatomical criteria that determine the feasibility and safety of transcatheter ASD closure: large ASDs (N 30 mm) and those with insufficient rims may be more suitable for surgical intervention. Thus, surgical ASD closure remains a vital therapeutic option for a significant proportion of patients.
4.1. Limitations Our study has limitations. As effectiveness was determined using observational rather than randomized data, there were differences in baseline characteristics between the study groups. These differences could have impacted on in-hospital and long-term outcomes. However, the Cox regression only demonstrated increasing patient age and the Charlson comorbidity index to be predictive of 5-year mortality, and as both of these indices were significantly higher in the transcatheter group, the study conclusions should remain unchanged. The unequal duration of follow-up also introduces the possibility of secular changes in general hospital care that could also be a confounding factor. We restricted the duration of patient follow-up to 5 years as this was the median time of follow-up between groups and because this period was likely to capture all events related to the index procedure. However, a longer observation period may have been more informative. Transcatheter therapies have been associated with enhanced short-term quality of life compared with their surgical equivalents, though these
differences may be reversed in the longer term [26]. Calculating quality-adjusted life years (QALY) may be an appropriate method to compare cost effectiveness between these strategies, though there are limitations to this approach, particularly when using observational or population data [27,28]. Given that well measured QALYs for our study population do not exist as well as the controversy over their use, we elected not to calculate them in this study. We did not have access to patient level data that would have afforded more detailed cost analysis for individual patients. Specifically, information on medications used in hospital or in the follow-up period was not available. Differences in the use of medications, such as antiplatelet therapies, between groups could have impacted on the rates of adverse events. However, the higher rate of major morbidity observed in the surgical group implies a greater requirement for additional therapies and thus, the conclusions of the study would not have been altered. The anatomical characteristics of the ASDs treated in each group were not available. It is likely that larger and more complex lesions were treated in the surgical group and this may have impacted on clinical outcomes. Although we used 2009 costs and applied these costs retrospectively to the onset of our observation period in 1998, we have no evidence that the cost variations for the two techniques varied differentially over time. As we assumed that the costs associated with patient management following ASD closure were identical between the two groups, our estimates should not be interpreted as the total direct costs of these interventions in Quebec. Finally, the results of this analysis may not necessarily be transferrable to other health care systems, particularly in developing nations where the cost of medical device technology can have a greater influence on the final cost analysis [29]. 5. Conclusion Transcatheter ASD closure appears to be a more cost effective and potentially clinically effective treatment compared to surgical ASD closure. Acknowledgements The authors would like to thank Ms. Liming Guo for her help in statistical analysis and figure preparation. References [1] Marelli AJ, Mackie AS, Ionescu-Ittu R, Rahme E, Pilote L. Congenital heart disease in the general population: changing prevalence and age distribution. Circulation 2007;115:163–72. [2] Krasuski RA, Bashore TM. The emerging role of percutaneous intervention in adults with congenital heart disease. Rev Cardiovasc Med 2005;6:11–22. [3] Chan KC, Godman MJ, Walsh K, Wilson N, Redington A, Gibbs JL. Transcatheter closure of atrial septal defect and interatrial communications with a new self
114
[4]
[5] [6] [7] [8] [9] [10]
[11]
[12]
[13]
[14]
[15]
D. Mylotte et al. / International Journal of Cardiology 172 (2014) 109–114 expanding nitinol double disc device (Amplatzer septal occluder): multicentre UK experience. Heart 1999;82:300–6. Warnes CA, Williams RG, Bashore TM, et al. ACC/AHA 2008 guidelines for the management of adults with congenital heart disease: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (Writing Committee to Develop Guidelines on the Management of Adults With Congenital Heart Disease). Developed in Collaboration With the American Society of Echocardiography, Heart Rhythm Society, International Society for Adult Congenital Heart Disease, Society for Cardiovascular Angiography and Interventions, and Society of Thoracic Surgeons. J Am Coll Cardiol 2008;52:e143–263. VanLare JM, Conway PH, Sox HC. Five next steps for a new national program for comparative-effectiveness research. N Engl J Med 2010;362:970–3. Mushlin AI, Ghomrawi H. Health care reform and the need for comparativeeffectiveness research. N Engl J Med 2010;362:e6. Garber AM, Sox HC. The role of costs in comparative effectiveness research. Health Aff (Millwood) 2010;29:1805–11. Peterson ED. Innovation and comparative-effectiveness research in cardiac surgery. N Engl J Med 2009;361:1897–9. Baker SS, O'Laughlin MP, Jollis JG, Harrison JK, Sanders SP, Li JS. Cost implications of closure of atrial septal defect. Catheter Cardiovasc Interv 2002;55:83–7. Formigari R, Di Donato RM, Mazzera E, et al. Minimally invasive or interventional repair of atrial septal defects in children: experience in 171 cases and comparison with conventional strategies. J Am Coll Cardiol 2001;37:1707–12. Thomson JD, Aburawi EH, Watterson KG, Van Doorn C, Gibbs JL. Surgical and transcatheter (Amplatzer) closure of atrial septal defects: a prospective comparison of results and cost. Heart 2002;87:466–9. Hughes ML, Maskell G, Goh TH, Wilkinson JL. Prospective comparison of costs and short term health outcomes of surgical versus device closure of atrial septal defect in children. Heart 2002;88:67–70. Du ZD, Hijazi ZM, Kleinman CS, Silverman NH, Larntz K. Comparison between transcatheter and surgical closure of secundum atrial septal defect in children and adults: results of a multicenter nonrandomized trial. J Am Coll Cardiol 2002;39:1836–44. Kim JJ, Hijazi ZM. Clinical outcomes and costs of Amplatzer transcatheter closure as compared with surgical closure of ostium secundum atrial septal defects. Med Sci Monit 2002;8:CR787–91. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis 1987;40:373–83.
[16] Teng J, Mayo NE, Latimer E, et al. Costs and caregiver consequences of early supported discharge for stroke patients. Stroke 2003;34:528–36. [17] Marshall DA, Levy AR, Vidaillet H, et al. Cost-effectiveness of rhythm versus rate control in atrial fibrillation. Ann Intern Med 2004;141:653–61. [18] Finlayson G, Reimer J, Dahl M, Stargardter M, McGowan K. The direct cost of hospitalizations in Manitoba, 2005/06. Manitoba Centre for Health Policy. Available at: http://mchp-appserv.cpe.umanitoba.ca; 2009. [19] Caro JJ. Pharmacoeconomic analyses using discrete event simulation. Pharmacoeconomics 2005;23:323–32. [20] Jahn B, Pfeiffer KP, Theurl E, Tarride JE, Goeree R. Capacity constraints and costeffectiveness: a discrete event simulation for drug-eluting stents. Med Decis Making 2010;30:16–28. [21] Deniz HB, Ward A, Jaime Caro J, Alvarez P, Sadri H. Cost–benefit analysis of primary prevention of sudden cardiac death with an implantable cardioverter defibrillator versus amiodarone in Canada. Curr Med Res Opin 2009;25: 617–26. [22] Najafzadeh M, Marra CA, Galanis E, Patrick DM. Cost effectiveness of herpes zoster vaccine in Canada. Pharmacoeconomics 2009;27:991–1004. [23] Briggs AH. Handling uncertainty in cost-effectiveness models. Pharmacoeconomics 2000;17:479–500. [24] Kotowycz MA, Therrien J, Ionescu-Ittu R, et al. Long-term outcomes after surgical versus transcatheter closure of atrial septal defects in adults. JACC Cardiovasc Interv 2013;6:497–503. [25] King TD, Thompson SL, Steiner C, Mills NL. Secundum atrial septal defect. Nonoperative closure during cardiac catheterization. JAMA 1976;235:2506–9. [26] Cohen DJ, Van Hout B, Serruys PW, et al. Quality of life after PCI with drugeluting stents or coronary-artery bypass surgery. N Engl J Med 2011; 364:1016–26. [27] Neyt M, Cleemput I, Thiry N, De Laet C. Calculating an intervention's (cost-) effectiveness for the real-world target population: the potential of combining strengths of both RCTs and observational data. Health Policy 2012;106: 207–10. [28] Prieto L, Sacristan JA. Problems and solutions in calculating quality-adjusted life years (QALYs). Health Qual Life Outcomes 2003;1:80. [29] Vida VL, Barnoya J, O'Connell M, Leon-Wyss J, Larrazabal LA, Castaneda AR. Surgical versus percutaneous occlusion of ostium secundum atrial septal defects: results and cost-effective considerations in a low-income country. J Am Coll Cardiol 2006; 47:326–31.
