This study aimed to determine the financial implications of isolated transcatheter aortic valve replacement (TAVR) versus surgical aortic valve replacement (SAVR) in the Brazilian Public Health System (SUS) using time-driven activity-based costing (TDABC). We performed a retrospective observational cohort study at a single center from 2019 to 2022, including 180 patients (TAVR: n = 65; SAVR: n = 115). TAVR patients were older and had higher surgical risk. The total cost of TAVR was significantly higher than SAVR (INT$ 40,776.9 vs INT$ 12,393.7, p <0.001), primarily driven by the prosthetic valve cost. However, TAVR was associated with a shorter postprocedural length of stay (5 days vs 7 days, p <0.001) and lower absolute 1-year mortality rates (3.1% vs 6.1%), although this difference did not reach statistical significance (p = 0.60). Most TAVR cases were accessed through litigation due to the lack of formal incorporation in the public system during the study period, which contributed to higher baseline frailty. In conclusion, while TAVR remains more expensive due to device costs, it offers clinical advantages in high-risk populations within a developing healthcare system.
The management of severe symptomatic aortic stenosis has been revolutionized with the advent of transcatheter aortic valve replacement (TAVR), providing an alternative to the traditional surgical aortic valve replacement (SAVR). ,, Initially restricted to nonsurgical candidates and high-risk patients, TAVR was proven to be noninferior if not superior to SAVR in patients with intermediate and low risk. ,,,,,,,, This innovation brings to the forefront not only the clinical outcomes but also the economic ramifications within the healthcare system.
Despite TAVR’s broadening application, which now includes patients across a spectrum of surgical risks and age, its financial implications remain largely unquantified. Many countries have evaluated the cost-effectiveness of TAVR compared to SAVR in high, intermediate, and low-risk patients. ,,,,,,, Although the cost-effectiveness is almost unanimously of a favorable recommendation in turned-down surgical candidates and high-risk patients, the results in intermediate and low-risk patients are still controversial.
In Brazil, the prevalence of calcific aortic stenosis was 23.7 for every 1,000 people, and 4.5% of the patients were 75 years or older in a population base study that looked into 2019 data. Although TAVR has been a well-established procedure for several years, its use in the Brazilian Public Health System (SUS, acronym in Portuguese), is very limited due to its high cost. Given the constrained resources of the Brazilian Public Health System, it is imperative to establish a comprehensive analysis of the direct medical and hospital costs related to both TAVR and SAVR.
The Brazilian healthcare landscape is divided between the Unified Health System (SUS), which provides universal public coverage, and a private supplementary system. During the period analyzed (2019–2022), TAVR was not formally incorporated into the SUS list of procedures. Access was primarily achieved through litigation (judicial mandates), where the public system was legally required to fund the intervention for specific patients. This unique scenario often led to extended waiting periods for legal clearance, during which patients remained hospitalized, often reaching the procedure in a state of advanced clinical deterioration and physical deconditioning.
This study employs a combination of macro and micro-costing methods to provide a broad economic assessment of TAVR and SAVR within the Brazilian Public Health System (SUS). , It focuses on the direct costs incurred from cardiac prostheses, delivery systems, and hospital services, as well as professional fees and the financial impact of postprocedural complications. The aim is to compare the direct costs and 1-year clinical outcomes of TAVR and SAVR within the SUS, focusing on the elderly population with severe aortic stenosis. The analysis will detail the procedural cost differentials and their outcomes, accounting for material and labor expenses, and hospitalization rates in the intensive care unit (ICU) and ward. Through this approach, the study will highlight the critical components of procedural expenses, thereby offering a nuanced view of the economic dimensions of TAVR and SAVR.
Methods
Study design and setting
This study was a retrospective cohort analysis conducted at the Hospital de Clínicas de Porto Alegre (HCPA), a tertiary university hospital located in the South of Brazil. It involved a comparison of all consecutive patients who underwent TAVR or SAVR between January 2019 and December 2022. The study included patients who had undergone either isolated TAVR or isolated SAVR during the specified period and was approved by the Institutional Research and Ethics Committee (CAEE 67080722.1.0000.5327). Patients were selected for TAVR if they were considered at high or prohibitive surgical risk by the institutional Heart Team, following the clinical requirements often necessitated by the judicial process in Brazil.
It excluded patients with combined surgical interventions or active endocarditis. According to the General Data Protection Law (Lei Geral de Proteção de Dados, LGPD) in Brazil, patients who had not been given authorization to have access to individual data were contacted, and only those who agreed to participate and signed an informed consent form were included. Based on institutional norms, patients who died were included regardless of prior agreement.
The study hypothesis was that the reduction in LOS and postoperative resources with TAVR would partially offset the high initial acquisition cost of the prosthesis.
Data collection
Data collection was retrospective, using the electronic health records (EHR) of HCPA system. Variables not only related to the surgical procedure but also encompassing a 1-year follow-up period were gathered. Research personnel conducted data acquisition, which included micro-costing variables and relevant clinical information from the preoperative and postoperative periods. Preoperative information collected were: age, gender, Society of Thoracic Surgeons predicted risk of mortality (STS-PROM) score, New York Heart Association (NYHA) class, presence of syncope, angina, atherosclerotic coronary disease, hypertension, diabetes, dyslipidemia, prior myocardial infarction (MI), stroke, and previous TAVR/SAVR. Postoperative information included length of stay (ward and ICU), incidence of stroke, myocardial infarction (MI), permanent pacemaker implantation (PPM), need for rehospitalization, death, acute renal injury, and endocarditis.
Hospital costs
To evaluate hospital direct costs, the steps from the time-driven activity-based costing (TDABC) method were followed, including: mapping the flow of care within main activities to which the patient is submitted; identification of all resources and departments used by the patient; estimation of the overall expenses of every resource identified in the flow of care; estimation of the hourly capacity of every resource or department, and unit cost rate (UCR) calculation; analysis of the time spent on the patient for every resource, and structuring of the time and costs equations; cost per patient calculation; and statistical analyses. All costs were initially obtained in Brazilian Reais (R$) and then converted into international dollars (INT$). The conversion rate was based on the most recent data available from the World Bank’s Purchasing Power Parity database, with 1 INT$ equivalent to R$2.44 and 1.00 American Dollar (2023).
Mapping the flow of care for the main activities to which the patient is submitted
The mapping of flow of care was performed by a multidisciplinary team, including physicians from the Cardiology and Cardiovascular Surgery services of the site participating in the project, as well as researchers from the health technologies evaluation area.
Identification of all hospital resources and departments used by the patient
Based on the review of clinical information in EHR, site services where patients were treated were identified, as well as the equipment and workers involved in healthcare provision.
Estimation of the total cost for every resource identified in the hospital flow of care
For each resource identified, annual mean costs were calculated. For hospital infrastructure, areas where the patient received care were surveyed by the hospital’s financial department, considering fixed costs of depreciation, energy, supporting materials, taxes, and system licenses. For labor costs, data from 2024 was used, including charges by mean professional category from the institution.
Estimation of every resource or department’s hourly capacity and unit cost rate calculation
The operation room and catheterization laboratory capacities, where patients with aortic stenosis were treated, were estimated considering the availability of rooms, medical and nursing personnel, utilization of the installed capacity, material, and devices used. From the cost per resource and capacity data, it was possible to calculate the cost-capacity rate (CCR) for each resource.
Analysis of time spent per patient for every resource and structuring of the time and cost equations
The mean time every professional spends involved with consultations and procedures performed on patients was estimated based on reports from healthcare professionals.
To evaluate the individual cost per patient, time spent for each resource was multiplied by the time CCR and, subsequently, by the costs of medications and exams. For medications, individual consumption was extracted from EHR, as well as all exams and additional tests performed. Costs documented at the site were used as the basis for financial information. All collected data were aggregated into a comprehensive electronic database specifically constructed for this micro-costing study of TAVR and SAVR.
Follow-up costs estimation
To estimate health-related costs within 1-year postprocedure, clinical records detailing SAVR and TAVR, ward and ICU hospitalizations length-of-stay (LOS), reasons for readmissions or type of complications, and subsequent hospitalization times were analyzed.
Costs were converted to International Dollars (INT$) using the World Bank Purchasing Power Parity (PPP) conversion factors for Brazil.
Data Analysis
Continuous variables were described using means and standard deviations or median and interquartile ranges, depending on the data distribution. Categorical variables were reported as relative and absolute frequencies. All variables were compared between the 2 surgical procedures, and outcomes were also graphically displayed.
Variables between the procedure groups were compared using generalized linear models based on the distribution of each variable or using the Kruskal-Wallis test. Clinical outcomes were compared between groups using the same class of models, according to the distribution of each outcome, or with the Kruskal-Wallis test. The distribution of the data and the distribution of the residuals from the models were assessed graphically. All analyses were performed using R software, version 4.3.2.
Results
From January 2019 to December 2022, a total of 180 patients were included, 65 underwent TAVR and 115 SAVR ( Figure 1 ). The median age and STS-PROM score were higher in the TAVR group (83 [76–86] vs 65 [57–72] years old, p <0.001; and 3.2% [2.4–5] vs 1.4% [0.9–2.5], p <0.001), respectively, highlighting that TAVR patients had higher surgical risk. The STS-PROM score dispersion is shown in Figure 2 . Prior stroke was identified in 29.2% of patients in the TAVR group and 6.1% in the SAVR group. Previous SAVR was present in 10.8% (ViV TAVR) and 9.6% (SAVR reoperation) of the groups. The baseline characteristics of both groups are summarized in Table 1 .
Study flowchart.
STS-PROM score according to aortic procedure performed (TAVR or SAVR).
Table 1
Clinical and demographic characteristics
| Variable | Total | TAVR | SAVR | p-value |
|---|---|---|---|---|
| Age. median (IQR) | 71 (61–80) | 83 (76–86) | 65 (57–72) | <0.001 |
| Age. mean (SD) | 69.8 (13.8) | 80.8 (7.8) | 63.6 (12.5) | <0.001 |
| Gender. n./n.total (%) | 0.048 | |||
| Female | 70/180 (38.9) | 32/65 (49.2)* | 38/115 (33) | |
| Male | 110/180 (61.1) | 33/65 (50.8) | 77/115 (67) | |
| Pre—Procedure | ||||
| STS-PROM score. median (IQR) | 2.1 (1.1–3.3) | 3.2 (2.4–5) | 1.4 (0.9–2.5) | <0.001 |
| STS-PROM Score. mean (SD) | 2.8 (2.7) | 4.1 (2.9) | 2.2 (2.4) | <0.001 |
| Functional classification (NYHA), n./n.total (%) | <0.001 | |||
| 1 | 21/94 (22.3) | 0/26 (0) | 21/68 (30.9) | |
| 2 | 18/94 (21.6) | 3/26 (11.5) | 15/68 (22.1) | |
| 3 | 44/94 (42.3) | 21/26 (80.8)* | 23/68 (33.8) | |
| 4 | 11/94 (11.7) | 2/26 (7.7) | 9/68 (13.2) | |
| Syncope, n./n.total (%) | 23/179 (12.8) | 8/64 (12.5) | 15/115 (13) | 0.999 |
| Angina, n./n.total (%) | 23/179 (24.7) | 6/63 (9.5) | 38/115 (33) | 0.001 |
| Coronary artery disease, n./n.total (%) | 55/180 (30.6) | 34/65 (52.3) | 21/115 (18.3) | <0.001 |
| High blood pressure, n./n.total (%) | 134/180 (74.4) | 54/65 (83.1) | 80/115 (69.6) | 0.069 |
| Diabetes, n./n.total (%) | 54/180 (30) | 21/65 (32.3) | 33/115 (28.7) | 0.734 |
| Dyslipidemia, n./n.total (%) | 79/179 (44.1) | 37/64 (57.8) | 42/115 (36.5) | 0.001 |
| Prior myocardial infarction, n./n.total (%) | 9/180 (5) | 7/65 (10.8) | 2/115 (1.7) | 0.021 |
| Prior stroke, n./n.total (%) | 26/180 (14.4) | 19/65 (29.2) | 7/115 (6.1) | <0.001 |
| Previous procedure SAVR, n./n.total (%) | 18/180 (12.7) | 7/65 (10.8) | 11/115 (9.6) | 0.999 |
IQR = interquartile range, SD = standard deviation.
The median [Interquartile range] ward length of stay (LOS) postprocedure for TAVR and SAVR was 4 (2–5) and 4 (3–5) days, respectively. The ICU median LOS was 2 (1–2) days for TAVR and 3 (3–4) days for SAVR (p <0.001), and the total LOS for the index hospitalization was 5 (3–8) for TAVR and 7 (6–8.5) for SAVR (p <0.001), highlighting that despite higher STS-PROM scores, TAVR patients had lower ICU and hospitalization length of stay ( Figure 3 ). The length of postprocedure hospitalization is detailed in Table 2 .
