This cohort study evaluates the association between Centers for Medicare & Medicaid Services transcatheter aortic valve replacement (TAVR) reimbursement rates with embolic protection device (EPD) use.
Key Points
Question
Is there an association between Centers for Medicare & Medicaid Services (CMS) reimbursement and use of an embolic protection device (EPD) during a transcatheter aortic valve replacement (TAVR) procedure?
Findings
In this cohort study using the Society for Thoracic Surgeons/American College of Cardiology Transcatheter Valve Therapy registry, 7.1% of TAVR procedures were completed with an EPD. While higher CMS reimbursement up to $50 000 per TAVR was initially associated with greater likelihood of EPD use, this association was no longer apparent after adjusting for site characteristics; however, higher TAVR volume was associated with increased EPD use.
Meaning
TAVR case volume, rather than CMS TAVR reimbursement, appears to be the predominant factor associated with EPD use.
Abstract
Importance
In the setting of uncertain efficacy and additional, unreimbursed cost, use of an embolic protection device (EPD) during transcatheter aortic valve replacement (TAVR) has had variable uptake. The Centers for Medicare & Medicaid Services (CMS) instituted a new technology add-on payment to cover EPD use in October 2018.
Objective
To evaluate the association between CMS TAVR reimbursement rates and EPD use.
Design, Setting, and Participants
This cohort study used the Society for Thoracic Surgeons/American College of Cardiology Transcatheter Valve Therapy registry to identify patients who underwent TAVR between January 2018 and September 2019. Analysis took place between July 2020 and February 2022.
Main Outcomes and Measures
The association between EPD use and CMS reimbursement was assessed using multivariable logistic regression models adjusted for patient characteristics (model 1) and patient/hospital (annualized TAVR volume and teaching status) characteristics (model 2).
Results
Among 511 institutions, CMS reimbursement for TAVR ranged from $28 062 to $111 280 with a median (IQR) of $45 884 ($40 331-$53 627). Among 84 353 patients (median [IQR] age, 81.0 [75.0-86.0] years; 46 247 male individuals [54.8%]; 3958 [4.7%] of Hispanic or Latino ethnicity; 78 170 White individuals [92.7%]) treated at the sites, 6012 (7.1%) underwent TAVR with EPD. Patient characteristics associated with EPD use included prior stroke (adjusted odds ratio [aOR], 1.13 [95% CI, 1.00-1.27]; P = .048), female sex (aOR, 0.85 [95% CI, 0.78-0.93]; P < .001), hemodialysis (aOR, 0.52 [95% CI, 0.40-0.68]; P < .001), and shock (aOR, 0.62 [95% CI, 0.41-0.94]; P = .03). Higher CMS reimbursement up to $50 000 per TAVR was associated with greater likelihood of EPD use in model 1 (per $1000; aOR, 1.08 [95% CI, 1.01-1.16]; P = .02). However, this association was no longer apparent after adjusting for site characteristics (model 2; aOR, 1.03 [95% CI, 0.96-1.11]; P = .38). Higher TAVR volume was associated with increased EPD use (per 25 TAVRs; aOR, 1.15 [95% CI, 1.09-1.21]; P < .001). There was no significant change in the odds of EPD uptake before vs after institution of the CMS new technology add-on payment across tertiles of CMS TAVR reimbursement (Wald χ2 = 3.59; P = .17).
Conclusions and Relevance
EPD use during TAVR remains infrequent and is associated with multiple patient and site characteristics. While CMS reimbursement varies significantly across institutions, TAVR case volume, rather than CMS TAVR reimbursement or the CMS new technology add-on payment, appears to be the predominant factor associated with EPD use. Ongoing work is needed to understand the economic drivers that contribute to the association between procedural volume and EPD use.
Introduction
The rate of 30-day stroke after transcatheter aortic valve replacement (TAVR) has not changed over the first 5 years of commercial TAVR in the US, and our ability to predict post-TAVR stroke is modest.1,2 The Sentinel device (Boston Scientific), a transcatheter cerebral embolic protection device (EPD), was developed to capture embolic debris during the TAVR procedure and thus reduce the risk of neurological complications. The pivotal SENTINEL (Cerebral Protection in Transcatheter Aortic Valve Replacement) study showed that the Sentinel device captured debris in 99% of patients but did not reduce cerebrovascular events.3 A subsequent meta-analysis demonstrated a trend toward lower risk of death or stroke at 30 days.3,4,5,6,7,8 In this setting, the Sentinel EPD received US Food and Drug Administration (FDA) approval in June 2017. Given its uncertain efficacy and substantial cost (approximately $3000),6,9 it is not surprising that there has been variable uptake of this device in TAVR procedures. However, beyond effectiveness, one of the factors that may influence use is TAVR reimbursement. Indeed, to help defray the device cost, the Centers for Medicare & Medicaid Services (CMS) approved a new technology add-on payment (NTAP) of $1400 in October 2018.10
Given the substantial interhospital variation in TAVR reimbursement across the US, our study aimed to examine (1) variation in TAVR reimbursement across hospitals, (2) variation in EPD use across hospitals, (3) the association between hospital reimbursement for TAVR procedures and use of an EPD during TAVR, and (4) whether there was an association between the CMS NTAP and the uptake of EPD. We hypothesized that institutions with lower reimbursement would be less likely to use an EPD during TAVR and that institution of the CMS NTAP would be associated with increased EPD use.
Methods
Study Population and Definitions
We used data collected by the Society for Thoracic Surgeons/American College of Cardiology Transcatheter Valve Therapy (STS/ACC TVT) registry between January 1, 2018, and September 30, 2019, to identify patients who underwent an isolated TAVR procedure. The STS/ACC TVT registry serves as the national database for medical device tracking of all commercially available transcatheter aortic valves implanted in the US. The American College of Cardiology Foundation institutional review board reviewed and approved the protocol and in accordance with 45 CFR 46.116(d) of federal regulations. The Advarra institutional review board waived the requirement for obtaining consent for the STS/ACC TVT registry. This study was approved by the Duke University institutional review board. Patient consent was waived because no new data were collected.
Inclusion criteria were chosen based on the FDA approval of the Sentinel device, the most recent data available in the STS/ACC TVT registry, and the NTAP payment period.11 Data collected by the STS/ACC TVT registry include patient demographic (including race and ethnicity) and clinical information, echocardiographic and laboratory data, procedural details, and patient outcomes. Site information was also collected.
Site-annualized TAVR volume was defined using all records (n = 111 384) before exclusions. TAVR diagnosis-related group (DRG) reimbursement was obtained from publicly reported data on the CMS website using Medicare severity DRG code definitions 266 and 267 (endovascular cardiac valve replacement with and without major complication comorbidity, respectively).12 For sites with missing TAVR reimbursement data (n = 157), we used the internal CMS-linked 2018 data set to calculate mean TAVR reimbursement for that site. To verify this approach, we evaluated the correlation of publicly reported TAVR reimbursement with the internal database for sites with both available (n = 390) and showed appropriate correlation (eFigure 1 in the Supplement).
Statistical Analysis
We first described TAVR reimbursement across all sites including median (Q1, Q3) and the minimum and maximum reimbursement amounts. Reimbursement was then stratified by DRG code (266 vs 267), by geographic region (Midwest, Northeast, South, West), and by teaching hospital status.
Next, we examined patient characteristics associated with EPD use. Patient demographic and clinical characteristics were described for the entire population and by EPD use. Categorical variables were presented using percentages and continuous variables using medians (25th, 75th percentiles). The Pearson χ2 test was used to compare differences in categorical variables and the Wilcoxon test for continuous variables. Sequential logistic regression models were used to evaluate the association between EPD use and prespecified variables.2 The initial model (model 1) incorporated patient characteristics (age, sex, prior aortic or nonaortic valve procedure, history of stroke, transient ischemic attack, carotid stenosis, peripheral arterial disease, current tobacco misuse, hypertension, diabetes, active dialysis requirement, porcelain aorta, atrial fibrillation/flutter, procedure acuity, body surface area, glomerular filtration rate), procedure date (per 30 days), and site TAVR reimbursement. The association between TAVR reimbursement and EPD use was nonlinear and a spline function with knot at $50 000 was used based on visual inspection of the inflection point. Model 2 included the aforementioned characteristics and added site-annualized TAVR volume, teaching status, and proportion of Medicare patients to understand the additive value of site characteristics in identifying factors associated with EPD use. The generalized estimating equation method was used to account for clustering within sites. An odds ratio (OR) with 95% CIs was calculated for each of the prespecified variables in the 2 models.
Next, we evaluated the association of the CMS NTAP availability and EPD use. Mean EPD use was graphed across the study period. Device use at the site level was calculated. A logistic regression model was used to estimate the odds of EPD use by CMS NTAP availability (beginning October 2018). Because EPD use was steadily increasing (positive slope) before implementation of the CMS NTAP, we compared the difference in the odds of uptake before vs after NTAP. To do this, we compared the difference between slopes of EPD use over time before and after CMS NTAP availability with the hypothesis that the already positive slope before NTAP would increase (ie, become steeper) after NTAP availability.
To further understand whether the impact of the CMS NTAP varied by baseline reimbursement, we divided the data set into tertiles of CMS TAVR reimbursement. Device use at the site level was calculated by tertile of TAVR reimbursement and the odds of EPD use before and after CMS NTAP availability were estimated using logistic regression models for each tertile. We then compared the change in EPD use within each tertile by examining the change in device use over time (slope) before vs after NTAP availability. This change in slope was subsequently compared across the 3 tertiles using a Wald χ2 test to determine whether there was a difference based on tertile of reimbursement. We hypothesized that any increase in EPD use following CMS NTAP availability would be most pronounced among sites with lower reimbursement. Variables used for adjustment were consistent across the CMS NTAP analyses and included those listed above for model 1 as well as site characteristics (annualized TAVR volume and teaching status).
All variables had less than 1% missing data. Simple imputation was used to impute categorical variables to the mode and continuous variables to the median value. A 2-sided P value less than .05 was considered significant. All analyses were carried out using SAS statistical software version 9.4 (SAS Institute). Analysis took place between July 2020 and February 2022.
Results
Patient and Site Cohorts
Using the STS/ACC TVT registry, we identified 111 384 patients from 647 sites. Among these, 27 031 patients were excluded (eFigure 2 in the Supplement), leaving a final population of 84 353 patients from 551 sites. The largest proportion of sites were located in the South (204 sites [36.5%]) with the fewest in the Northeast (92 sites [16.7%]). The majority (342 sites [62.1%]) of procedures were performed in an urban setting and at private/community hospitals (447 sites [81.1%]). The volume of TAVR procedures varied with a median (IQR) annualized TAVR volume of 83 (55-134) TAVRs. Patients included in the final cohort had a median (IQR) age of 81 (75-86) years, 46 247 (54.8%) were male, 3958 (4.7%) were Hispanic or Latino, and 78 170 (92.7%) were White (Table 1).
Table 1. Patient Characteristics by EPD Use.
| Characteristic | No. (%) | P value | ||
|---|---|---|---|---|
| Overall (N = 84 353) | EPD use (n = 6012) | No EPD use (n = 78 341) | ||
| Age, median (IQR), y | 81.0 (75.0-86.0) | 80.0 (74.0-86.0) | 81.0 (75.0-86.0) | <.001 |
| Male | 46 247 (54.8) | 3570 (59.4) | 42 677 (54.5) | <.001 |
| Female | 38 100 (45.2) | 2442 (40.6) | 35 658 (45.5) | |
| Hispanic or Latino | 3958 (4.7) | 199 (3.3) | 3759 (4.8) | <.001 |
| White | 78 170 (92.7) | 5616 (93.4) | 72 554 (92.6) | .02 |
| Prior aortic valve procedure | 7416 (8.8) | 558 (9.3) | 6858 (8.8) | .16 |
| Prior stroke | 9069 (10.8) | 699 (11.6) | 8370 (10.7) | .02 |
| Transient ischemic attack | 6512 (7.7) | 493 (8.2) | 6019 (7.7) | .14 |
| Carotid stenosis | 13 352 (15.8) | 863 (14.4) | 12 489 (15.9) | <.001 |
| Peripheral arterial disease | 18 990 (22.5) | 1252 (20.8) | 17 738 (22.6) | .001 |
| Current tobacco misuse | 4721 (5.6) | 330 (5.5) | 4391 (5.6) | .71 |
| Hypertension | 76 334 (90.5) | 5326 (88.6) | 71 008 (90.6) | <.001 |
| Diabetes | 32 205 (38.2) | 2166 (36.0) | 30 039 (38.3) | <.001 |
| Currently undergoing dialysis | 3138 (3.7) | 126 (2.1) | 3012 (3.8) | <.001 |
| Porcelain aorta | 1920 (2.3) | 192 (3.2) | 1728 (2.2) | <.001 |
| Atrial fibrillation/flutter | 31 065 (36.8) | 2203 (36.6) | 28 862 (36.8) | .77 |
| Acuity | ||||
| Emergency/salvage | 355 (0.4) | 13 (0.2) | 342 (0.4) | <.001 |
| Shock | 1765 (2.1) | 83 (1.4) | 1682 (2.1) | |
| Urgent | 5785 (6.9) | 463 (7.7) | 5322 (6.8) | |
| Elective | 76 448 (90.6) | 5453 (90.7) | 70 995 (90.6) | |
| Body surface area, median (IQR) | 1.9 (1.7-2.1) | 1.9 (1.7-2.1) | 1.9 (1.7-2.1) | <.001 |
| GFR, median (IQR) | 59.8 (45.3-75.5) | 61.7 (48.2-76.5) | 59.7 (45.0-75.4) | <.001 |
Abbreviations: EPD, embolic protection device; GFR, glomerular filtration rate.
CMS TAVR Reimbursement
CMS reimbursement ranged from $28 062 to $111 280 with a median (IQR) reimbursement of $45 884 ($40 331-$53 627). eFigure 3 in the Supplement shows a map of TAVR reimbursement throughout the United States. Reimbursement varied both by DRG code and by site region and teaching status (Table 2). Sites in the Northeast received the highest reimbursement for TAVR (median [IQR], $53 292 [$45 254-$60 547]) while those in the South received the lowest (median [IQR], $41 816 [$38 539-$46 812]). Teaching institutions also received higher reimbursement than nonteaching institutions (median [IQR], $47 912 [$41 941-$55 351] vs $43 095 [$38 863-$49 861]).
Table 2. TAVR Site Reimbursement.
| Characteristic | Median (IQR), $ | ||
|---|---|---|---|
| All TAVRs | TAVR DRG code 266 | TAVR DRG code 267 | |
| Overall | 45 884 (40 331-53 627) | 52 490 (47 075-61 815) | 40 086 (36 089-47 011) |
| Region | |||
| Midwest | 44 481 (40 073-50 441) | 52 275 (47 016-58 153) | 39 529 (36 255-44 749) |
| Northeast | 53 292 (45 254-60 547) | 62 591 (52 509-71 261) | 48 458 (40 061-55 568) |
| South | 41 816 (38 539-46 812) | 47 765 (44 078-52 616) | 36 400 (34 361-39 943) |
| West | 51 853 (46 063-56 764) | 59 016 (52 157-67 844) | 44 414 (40 277-50 684) |
| Teaching status | |||
| Teaching hospital | 47 912 (41 941-55 351) | 55 582 (49 285-65 584) | 42 371 (37 240-49 642) |
| Nonteaching hospital | 43 095 (38 863-49 861) | 48 971 (44 387-55 911) | 37 284 (34 373-43 185) |
Abbreviations: DRG, diagnosis-related group; TAVR, transcatheter aortic valve replacement.
Patient and Site Characteristics Associated With EPD Use
During the study period, 6012 patients (7.1%) had EPD use during their TAVR procedure. Multiple patient variables were significantly associated with EPD use (Table 3, model 1). Female individuals (adjusted OR [aOR], 0.85 [95% CI, 0.78-0.93]; P < .001), patients undergoing hemodialysis (aOR, 0.52 [95% CI, 0.40-0.68]; P < .001), and those with shock (aOR, 0.62 [95% CI, 0.41-0.94]; P = .03) were less likely while patients with prior stroke (aOR, 1.13 [95% CI, 1.00-1.27]; P = .048) were slightly more likely to have EPD use. After adjusting for patient factors, there was an increased likelihood of EPD use with increasing TAVR reimbursement (per $1000 of increasing reimbursement; aOR, 1.08 [95% CI, 1.01-1.16]; P = .02), up to a TAVR reimbursement level of $50 000 (eFigure 4 in the Supplement). Above this reimbursement threshold, there was no significant association between EPD use and increasing reimbursement (aOR, 1.01 [95% CI, 0.98-1.05]; P = .41).
Table 3. Patient and Site Characteristics Associated With Embolic Protection Device Use.
| Parameter | Model 1 | Model 2 | ||
|---|---|---|---|---|
| Adjusted odds ratio (95% CI) | P value | Adjusted odds ratio (95% CI) | P value | |
| Age | 1.00 (0.99-1.00) | .22 | 1.00 (0.99-1.00) | .22 |
| Body surface area | 1.23 (1.01-1.49) | .04 | 1.37 (1.17-1.61) | <.001 |
| GFR | 1.00 (1.00-1.00) | .69 | 1.00 (1.00-1.00) | .84 |
| Sex (female vs male) | 0.85 (0.78-0.93) | <.001 | 0.87 (0.81-0.94) | <.001 |
| Prior aortic valve procedure | 1.08 (0.86-1.35) | .51 | 0.99 (0.79-1.24) | .93 |
| Prior nonaortic valve procedure | 0.95 (0.78-1.14) | .56 | 0.92 (0.76-1.10) | .36 |
| Prior stroke | 1.13 (1.00-1.27) | .05 | 1.16 (1.03-1.30) | .01 |
| Transient ischemic attack | 1.10 (0.97-1.24) | .12 | 1.14 (1.01-1.27) | .03 |
| Carotid stenosis | 0.94 (0.80-1.11) | .50 | 1.01 (0.86-1.18) | .92 |
| Peripheral arterial disease | 0.96 (0.74-1.24) | .73 | 0.94 (0.73-1.20) | .62 |
| Current smoker | 0.96 (0.83-1.11) | .58 | 1.02 (0.88-1.18) | .78 |
| Hypertension | 0.86 (0.74-1.00) | .05 | 0.86 (0.75-0.99) | .04 |
| Diabetes | 0.92 (0.86-0.99) | .03 | 0.92 (0.85-0.99) | .02 |
| Current dialysis | 0.52 (0.40-0.68) | <.001 | 0.51 (0.38-0.67) | <.001 |
| Porcelain aorta | 1.56 (0.85-2.89) | .16 | 1.72 (0.91-3.24) | .10 |
| Atrial fibrillation | 0.99 (0.90-1.08) | .84 | 1.00 (0.91-1.09) | .98 |
| Urgent vs elective | 1.05 (0.77-1.44) | .74 | 0.88 (0.69-1.13) | .32 |
| Shock | 0.62 (0.41-0.94) | .03 | 0.60 (0.40-0.90) | .01 |
| Emergency/salvage | 0.46 (0.21-0.98) | .04 | 0.42 (0.18-0.97) | .04 |
| Procedure datea | 1.05 (1.03-1.07) | <.001 | 1.06 (1.04-1.08) | <.001 |
| TAVR reimbursementb | ||||
| <50 000 | 1.08 (1.01-1.16) | .02 | 1.03 (0.96-1.11) | .38 |
| ≥50 000 | 1.01 (0.98-1.05) | .41 | 1.02 (0.99-1.06) | .26 |
| Teaching institution | NA | NA | 1.07 (0.55-2.06) | .84 |
| Medicare procedures (≥90% vs <90%) | NA | NA | 1.59 (0.88-2.87) | <.001 |
| Annualized TAVR volumec | NA | NA | 1.15 (1.09-1.21) | <.001 |
Abbreviations: GFR, glomerular filtration rate; NA, not applicable; TAVR, transcatheter aortic valve replacement.
Per 30 days.
Per $1000 increase in reimbursement.
Per 25 increase in TAVRs.
To understand whether the association between TAVR reimbursement and EPD use was independent of both patient and nonreimbursement site factors, we added site characteristics to the logistic regression model (Table 3, model 2). When teaching institution status, proportion of Medicare patients, and annualized TAVR volume were also included, TAVR reimbursement was no longer significant (aOR, 1.03 per $1000 [95% CI, 0.96-1.11]; P = .38) but increasing annualized TAVR volume was significantly associated with EPD use (per 25 TAVRs; aOR, 1.15 [95% CI, 1.09-1.21]; P < .001). In both models, procedure date (per 30 days) was associated with EPD use, indicating an increase in EPD use over time (model 1: aOR, 1.05 [95% CI, 1.03-1.07]; P < .001; model 2: aOR, 1.06 [95% CI, 1.04-1.08]; P < .001).
CMS NTAP and EPD Use
Over half of the patients in the study (50 504 [59.9%]) were treated with TAVR after the CMS supplemental NTAP became available. Prior to NTAP availability, 1622 patients (4.8%) underwent TAVR with EPD use compared with 4390 patients (8.7%) after NATP availability (eFigure 5 in the Supplement). There was no significant difference in the rate of change in EPD use over time (slope) before vs after CMS NTAP approval (Wald χ2 = 3.59; P = .17; eFigure 5 in the Supplement).
To understand whether the association of the CMS NTAP with EPD use varied by baseline CMS reimbursement for TAVR, we performed an analysis stratified by TAVR reimbursement tertile. For tertile 1 (lowest reimbursement) and tertile 3 (highest reimbursement) only, there were increasing odds of EPD use with later procedure dates (per 30 days in the preapproval time segment) leading up to the CMS NTAP availability in October 2018 (tertile 1: aOR, 1.27; P < .001; tertile 2: aOR, 1.08; P = .05; and tertile 3: aOR, 1.14; P < .001; Table 4). However, there was an association between procedure date (per 30 days in the postapproval time segment) and EPD use after CMS NTAP availability for tertile 1 only (OR, 1.04 [95% CI, 1.00-1.09]; P = .04; Table 4). There was a significant difference in slopes before vs after CMS NTAP availability for tertile 1 (Wald χ2 = 8.3; P = .004) and tertile 3 (Wald χ2 = 7.92; P = .005). Contrary to our initial hypothesis that NTAP availability would be associated with increased EPD use, the slope for both tertiles 1 and 3 became less positive after CMS NTAP availability, suggesting a slowing of EPD uptake over time (Figure). There was no significant difference in slopes before vs after CMS NTAP availability for tertile 2 (Wald χ2 = 1.21; P = .27). When the change in slopes before vs after NTAP availability was compared across the tertiles, there was also no significant difference (Wald χ2 = 3.59; P = .17), suggesting that changes in EPD uptake before vs after CMS NTAP availability did not differ based on level of reimbursement.
Table 4. Odds of EPD Use by Tertile of CMS TAVR Reimbursement Procedure Date.
| Reimbursement tertile ($) | CMS approval | Odds ratio (95% CI)a | P value for OR | P value for change in slopeb |
|---|---|---|---|---|
| Tertile 1 (28 063-42 143) | Preapproval | 1.27 (1.12-1.45) | <.001 | .004 |
| Postapproval | 1.04 (1.00-1.09) | .04 | ||
| Tertile 2 (42 232-50 677) | Preapproval | 1.08 (1.00-1.16) | .05 | .27 |
| Postapproval | 1.03 (0.98-1.07) | .24 | ||
| Tertile 3 (50 738-111 280) | Preapproval | 1.14 (1.06-1.23) | <.001 | .005 |
| Postapproval | 1.02 (0.99-1.04) | .21 |
Abbreviations: CMS, Centers for Medicare & Medicaid Services; EPD, embolic protection device; OR, odds ratio; TAVR, transcatheter aortic valve replacement.
Odds of EPD use: per 30 days using time in the preapproval or postapproval segment.
Change in slope of EDP use before vs after CMS new technology add-on payment approval.
Figure. Embolic Protection Device (EPD) Use by Month Across the Study and Stratified by Tertile.
Raw rates of EPD use and fitted values of EPD use after adjustment for prespecified variables are reported. The reimbursement rates for tertile 1 are $28 063 to $42 143; tertile 2, $42 232 to $50 677; and tertile 3, $50 738 to $111 280.
Discussion
In this analysis of the association between site reimbursement, NTAP availability, and use of EPDs at the time of TAVR, we found that identified several key findings. Despite an increase in utilization during the study period, EPD use remained infrequent. Next, inherent to CMS reimbursement, there are significant differences across sites in reimbursement for TAVR procedures. Third, rather than TAVR reimbursement, greater TAVR volume appears to be the predominant site factor associated with EPD use. Finally, there was no evidence that the CMS NATP was associated with increased EPD use.
The decision to use an EPD during a TAVR procedure likely depends on a combination of factors including perceived device efficacy and safety, patient considerations, proceduralist experience, site teaching institution status, and cost considerations. The pivotal study that led to FDA approval of the Sentinel device demonstrated that the device was safe but not necessarily effective.3 Since that time, there has been mixed evidence surrounding the efficacy of EPD use for patients undergoing TAVR.9,13 In this setting, the ongoing Stroke PROTECTion With Sentinel During Transcatheter Aortic Valve Replacement (PROTECT TAVR) trial was designed to determine the efficacy of EPD with clinical outcomes.14 Despite the lack of clear evidence supporting the efficacy, we observed an increase in EPD use over the study period. However, EPD use remains infrequent with approximately 7% of patients undergoing TAVR with EPD during the study. We found that history of a stroke was associated with a slightly increased likelihood of EPD use but that being female, dialysis requirement, shock, and emergency/salvage procedure were associated with a decreased likelihood of use. Other studies have similarly found differences in patient characteristics among those who undergo TAVR with and without EPD use.15
Beyond patient characteristics, we hypothesized that cost considerations may be associated with EPD use. Under the current Medicare DRG system, hospitals receive a fixed payment for TAVR that is hospital specific and calculated from a base payment rate that incorporates several factors, including geographic location, surrounding wages, and teaching status of the hospital. The base payment is then adjusted for case mix and patient illness severity.6 Thus, hospital reimbursement rates for TAVR can vary significantly. In alignment with this, we found significant reimbursement differences between sites, ranging from $28 062 to $111 280; with those in the Northeast and teaching institutions receiving higher reimbursement. As a result of this variability, the contribution margin for TAVR (the difference between reimbursement and the variable cost of the hospitalization) can vary up to 20-fold across hospitals.
As CMS represents the primary payer for approximately 90% of TAVR hospitalizations,16 these variations in reimbursement are particularly relevant for TAVR. When we modeled the association between patient characteristics, TAVR reimbursement, and EPD use, we initially noted that for sites with reimbursement levels less than $50 000, each increase of $1000 in reimbursement was associated with a modest increase in the likelihood of EPD use. However, when we also incorporated site characteristics into the model, this association between TAVR reimbursement and EPD use was attenuated. While the association with reimbursement was no longer significant, increasing annualized TAVR volume was associated with a higher likelihood of EPD use.
Prior work highlights a close association between TAVR volume and site contribution margin. Specifically, institutions performing at least 50 TAVRs per year generally achieved a net-positive contribution margin per patient.17 Although the exact mechanism behind the association between TAVR volume and EPD use cannot be fully elucidated in this observational study, beyond the economic considerations, higher volume sites may have an overall greater comfort level with TAVR procedures and may have a longer duration of TAVR experience and hence may be more likely to add additional complexity and steps to cases.
Given the cost considerations associated with EPD use, we also hypothesized that the introduction of the CMS NTAP in October 2018 would be associated with an increase in EPD use. In the entire study cohort, we observed steady uptake in use, without a change in the mean rate of EPD use after CMS NTAP availability. When examined by tertile of reimbursement, contrary to our hypothesis, we found that the rate of increase of EPD use actually slowed after introduction of the NTAP at the lowest and highest tertiles of site reimbursement. Taken together, these findings suggest that the CMS NTAP did not significantly drive EPD use, regardless of the level of CMS TAVR reimbursement and suggest that clinical evidence, rather than reimbursement, may be the strongest driver for clinical decision-making for EPD use.
While there are few previous studies evaluating the efficacy of NTAP, existing analyses have shown that the program has had mixed success at improving uptake of new technologies. For example, an analysis of the NTAP for cardiac resynchronization therapy defibrillator showed only a 2.1% increase in cardiac resynchronization therapy defibrillator use associated with implementation of an NTAP payment of $16 262.50.18 Theoretical reasons why the NTAP may not have driven increased adoption of EPD include insufficient payment supplementation, lack of perceived device benefit/safety, or other confounding hospital or physician incentives. Importantly, NTAP has been used previously in the field of structural heart disease to support payment for transcatheter edge to edge repair for degenerative mitral regurgitation in fiscal years 2014-2016. Little is known about whether the use of NTAP in this scenario, where clinical trial data showed clear benefit for use, increased uptake of transcatheter edge to edge repair in appropriate patients. Further research is needed to understand the association of NTAP on use of medical devices, especially in structural heart disease, as CMS continues to explore the best process for Medicare to cover new technologies.
In the setting of conflicting clinical efficacy data8,9 as well as an ongoing trial evaluating clinical outcomes, it is not surprising that uptake of EPD use in TAVR has equilibrated at a low level. Further, as we have demonstrated, EPD use represents a complex clinical decision-making scenario with likely multiple contributions on the patient, proceduralist, and site levels. From a reimbursement perspective, although we hypothesized that the current CMS reimbursement model may lead to unintended consequences, we found site TAVR case volume to be the more significant driver of EPD use.
Limitations
We acknowledge several limitations. First, given the observational nature of our study, the rationale for clinical decision-making regarding the use of EPD cannot be elucidated. Additionally, variables beyond those available in the STS/ACC TVT registry, including anatomic considerations, were not available. Next, the cost to the site of obtaining EPD and TAVR devices was not available for this analysis. It is frequently the case in negotiated contracts that a higher volume of product purchased leads to per unit discounts on medical devices. In this way, it is plausible that at higher volume institutions, the price of the EPD device has less association with EPD use since the per-unit cost of the device is less. This hidden confounder may have attenuated the association between CMS site reimbursement and EPD use and could not be accounted for in our adjusted analyses. Lastly, our findings regarding the association of NATP with EDP use may not be generalizable to other medical devices. Specifically, the low overall rate of EPD use coupled with the modest differences in reimbursement may limit our findings to the application of this topic more broadly.
Conclusions
In this analysis of nationwide TAVR procedures in the STS/ACC TVT registry, we found that EPD use remains infrequent and that TAVR volume rather than CMS TAVR reimbursement was a significant predictor of EPD use. Implementation of the CMS NTAP was not significantly associated with EPD use, regardless of CMS TAVR reimbursement. Further work is needed to understand the economic drivers that contribute to the association between procedural volume and EPD use and to better understand the optimal design of payment structures for new technologies.
eFigure 1. Publicly Available versus TVT Internal TAVR Reimbursement Data
eFigure 2. Study Cohort Diagram
eFigure 3. TAVR Reimbursement in the US
eFigure 4. Association of EPD Use with TAVR Reimbursement
eFigure 5. Percent use of embolic protection device over time
References
- 1.Huded C, Tuzcu EM, Krishnaswamy A, et al. TCT-70 incidence, trends, and risk factors for early stroke after transcatheter aortic valve replacement in the first 5 years of US clinical practice: report of the STS/ACC TVT registry. J Am Coll Cardiol. 2018;72:B31. doi: 10.1016/j.jacc.2018.08.1161 [DOI] [Google Scholar]
- 2.Thourani VH, O’Brien SM, Kelly JJ, et al. Development and application of a risk prediction model for in-hospital stroke after transcatheter aortic valve replacement: a report from the Society of Thoracic Surgeons/American College of Cardiology Transcatheter Valve Therapy Registry. Ann Thorac Surg. 2019;107(4):1097-1103. doi: 10.1016/j.athoracsur.2018.11.013 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Kapadia SR, Kodali S, Makkar R, et al. ; SENTINEL Trial Investigators . Protection against cerebral embolism during transcatheter aortic valve replacement. J Am Coll Cardiol. 2017;69(4):367-377. doi: 10.1016/j.jacc.2016.10.023 [DOI] [PubMed] [Google Scholar]
- 4.Van Mieghem NM, van Gils L, Ahmad H, et al. Filter-based cerebral embolic protection with transcatheter aortic valve implantation: the randomised MISTRAL-C trial. EuroIntervention. 2016;12(4):499-507. doi: 10.4244/EIJV12I4A84 [DOI] [PubMed] [Google Scholar]
- 5.Haussig S, Mangner N, Dwyer MG, et al. Effect of a cerebral protection device on brain lesions following transcatheter aortic valve implantation in patients with severe aortic stenosis: the CLEAN-TAVI Randomized Clinical Trial. JAMA. 2016;316(6):592-601. doi: 10.1001/jama.2016.10302 [DOI] [PubMed] [Google Scholar]
- 6.Wendt D, Kleinbongard P, Knipp S, et al. Intraaortic Protection From Embolization in Patients Undergoing Transaortic Transcatheter Aortic Valve Implantation. Ann Thorac Surg. 2015;100(2):686-691. doi: 10.1016/j.athoracsur.2015.03.119 [DOI] [PubMed] [Google Scholar]
- 7.Lansky AJ, Schofer J, Tchetche D, et al. A prospective randomized evaluation of the TriGuard™ HDH embolic DEFLECTion device during transcatheter aortic valve implantation: results from the DEFLECT III trial. Eur Heart J. 2015;36(31):2070-2078. doi: 10.1093/eurheartj/ehv191 [DOI] [PubMed] [Google Scholar]
- 8.Giustino G, Sorrentino S, Mehran R, Faggioni M, Dangas G. Cerebral embolic protection during TAVR: a clinical event meta-analysis. J Am Coll Cardiol. 2017;69(4):465-466. doi: 10.1016/j.jacc.2016.12.002 [DOI] [PubMed] [Google Scholar]
- 9.Butala NM, Makkar R, Secemsky EA, et al. Cerebral embolic protection and outcomes of transcatheter aortic valve replacement: results from the transcatheter valve therapy registry. Circulation. 2021;143(23):2229-2240. doi: 10.1161/CIRCULATIONAHA.120.052874 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Medicare program; hospital inpatient prospective payment systems for acute care hospitals and the long-term care hospital prospective payment system and policy changes and fiscal year 2019 rates; quality reporting requirements for specific providers; Medicare and Medicaid Electronic Health Record (EHR) Incentive Programs (Promoting Interoperability Programs) requirements for eligible hospitals, critical access hospitals, and eligible professionals; Medicare cost reporting requirements; and physician certification and recertification of claims. Fed Regist. 2018. To be codified at 83 CFR §41144. [PubMed]
- 11.Centers for Medicare and Medicaid Services . New medical services and new technologies. Accessed July 15, 2022. https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Payment/AcuteInpatientPPS/newtech
- 12.Centers for Medicare & Medicaid Services . Medicare provider utilization and payment data: inpatient. https://www.cms.gov/Research-Statistics-Data-and-Systems/Statistics-Trends-and-Reports/Medicare-Provider-Charge-Data/Inpatient
- 13.Kapadia SR, Krishnaswamy A. Cerebral embolic protection in transcatheter aortic valve replacement: connecting intuition and proof. JACC Cardiovasc Interv. 2021;14(2):169-171. doi: 10.1016/j.jcin.2020.11.025 [DOI] [PubMed] [Google Scholar]
- 14.PROTECTED TAVR: Stroke PROTECTion With SEntinel During Transcatheter Aortic Valve Replacement (PROTECTED TAVR). ClinicalTrials.gov. Identifier: NCT04149535. http://clinicaltrials.gov/ct2/show/NCT04149535
- 15.Khan MZ, Zahid S, Khan MU, et al. Use and outcomes of cerebral embolic protection for transcatheter aortic valve replacement: a US nationwide study. Catheter Cardiovasc Interv. 2021;98(5):959-968. doi: 10.1002/ccd.29842 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Patel N, Doshi R, Kalra R, Bajaj NS, Arora G, Arora P. Costs of transcatheter aortic valve replacement: implications of proposed Medicare cuts. JACC Cardiovasc Interv. 2018;11(6):610-612. doi: 10.1016/j.jcin.2018.01.234 [DOI] [PubMed] [Google Scholar]
- 17.McCarthy FH, Savino DC, Brown CR, et al. Cost and contribution margin of transcatheter versus surgical aortic valve replacement. J Thorac Cardiovasc Surg. 2017;154(6):1872-1880. doi: 10.1016/j.jtcvs.2017.06.020 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Bockstedt LA. The effect of Medicare's new technology add-on payment policy. Accessed July 15, 2022. https://conservancy.umn.edu/handle/11299/93750
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Citations
- PROTECTED TAVR: Stroke PROTECTion With SEntinel During Transcatheter Aortic Valve Replacement (PROTECTED TAVR). ClinicalTrials.gov. Identifier: NCT04149535. http://clinicaltrials.gov/ct2/show/NCT04149535
Supplementary Materials
eFigure 1. Publicly Available versus TVT Internal TAVR Reimbursement Data
eFigure 2. Study Cohort Diagram
eFigure 3. TAVR Reimbursement in the US
eFigure 4. Association of EPD Use with TAVR Reimbursement
eFigure 5. Percent use of embolic protection device over time
