Machine Learning Identification of Modifiable Predictors of Patient Outcomes After Transcatheter Aortic Valve Replacement.

Background: Transcatheter aortic valve replacement (TAVR) is an important treatment option for patients with severe symptomatic aortic stenosis. It is important to identify predictors of excellent outcomes (good clinical outcomes, more time spent at home) after TAVR that are potentially amenable to improvement. Objectives: The purpose of the study was to use machine learning to identify potentially modifiable predictors of clinically relevant patient-centered outcomes after TAVR. Methods: We used data from 8,332 TAVR cases (January 2016-December 2021) from 21 hospitals to train random forest models with 57 patient characteristics (demographics, comorbidities, surgical risk score, lab values, health status scores) and care process parameters to predict the end point, a composite of parameters that designated an excellent outcome and included no major complications (in-hospital or at 30 days), post-TAVR length of stay of 1 day or less, discharge to home, no readmission, and alive at 30 days. We used recursive feature elimination with cross-validation and Shapley Additive Explanation feature importance to identify parameters with the highest predictive values. Results: The final random forest model retained 29 predictors (15 patient characteristics and 14 care process components); the area under the curve, sensitivity, and specificity were 0.77, 0.67, and 0.73, respectively. Four potentially modifiable predictors with relatively high Shapley Additive Explanation values were identified: type of anesthesia, direct movement to stepdown unit post-TAVR, time between catheterization and TAVR, and preprocedural length of stay. Conclusions: This study identified four potentially modifiable predictors of excellent outcome after TAVR, suggesting that machine learning combined with hospital-level data can inform modifiable components of care, which could support better delivery of care for patients undergoing TAVR.

Limited Access to Aortic Valve Procedures in Socioeconomically Disadvantaged Areas.

BACKGROUND: To explore how differences in local socioeconomic deprivation impact access to aortic valve procedures and the treatment of aortic valve disease, in comparison to other open and minimally invasive surgical procedures. METHODS AND RESULTS: Procedure volume data were obtained from the Healthcare Cost and Utilization Project from 18 states from 2016 to 2019 and merged with area deprivation index data, an index of zip code-level socioeconomic distress. We estimate the relationship between local deprivation ranking and differences in volumes of aortic valve replacement, which include transcatheter aortic valve replacement (TAVR) and surgical aortic valve replacement (SAVR), versus coronary artery bypass graft surgery and laparoscopic colectomy (LC). All regressions control for state and year fixed effects and an array of zip code-level characteristics. TAVR procedures have increased over time across all zip codes. The rate of increase is negatively correlated with deprivation ranking, regardless of the higher share of hospitalizations per population in high deprivation areas. Distributional analysis further supports these findings, showing that lower area deprivation index areas account for a disproportionately large share of SAVR, TAVR, and LC procedures in our sample relative to their share of all hospitalizations in our sample. By comparison, the cumulative distribution of coronary artery bypass graft procedures was nearly identical to that of total hospitalizations, suggesting that this procedure is equitably distributed. Regressions show high area deprivation index areas have lower prevalence of SAVR (ß=-15.1%, [95% CI, -26.8 to -3.5]), TAVR (ß=-9.1%, [95% CI, -18.0 to -0.2]), and LC (ß=-19.9%, [95% CI, -35.4 to -4.4]), with no statistical difference in the prevalence of coronary artery bypass graft (ß=-2.5%, [95% CI, -12.7 to 7.6]), a widespread and commonly performed procedure. In the population aged ≥80 years, results show high area deprivation index areas have a lower prevalence of TAVR (ß=-11.9%, [95% CI, -18.7 to -5.2]) but not SAVR (ß=-0.8%, [95% CI, 8.1 to 6.3]), LC (ß=-3.5%, [95% CI, -13.4 to -6.4]), or coronary artery bypass graft (ß=5.2%, [95% CI, -1.1 to 1.1]). CONCLUSIONS: People living in high deprivation areas have less access to life-saving technologies, such as SAVR, and even moreso to device-intensive minimally invasive procedures such as TAVR and LC.

Long-term risk of reintervention after transcatheter aortic valve replacement.

BACKGROUND: Transcatheter aortic valve replacement (TAVR) has surpassed surgical aortic valve replacement (SAVR) as the predominant mode of valve replacement for the treatment of severe aortic stenosis (AS). However, the long-term need for valvular reintervention after TAVR remains unknown. METHODS: Using data from the Medicare Fee for Service 100% dataset, all patients receiving TAVR between July 2011 and December 2020 were identified. Patients were categorized as receiving a valve reintervention (either surgical or transcatheter) or not using the appropriate International Classification of Diseases 10th Revision Procedure Coding System (ICD-10-PCS). A competing risk regression model was used to estimate the cumulative probability of valve reintervention. RESULTS: Of 230,644 TAVR patients were identified, of whom 1,880 received a reintervention. Patients receiving a reintervention were younger and more likely to be male. At 10 years, the crude rate of reintervention was 0.59% within a surviving cohort of 341 patients. After adjusting for the competing risk of death and other covariates, the adjusted cumulative incidence of reintervention at 10 years after TAVR was 1.63%. When the rate of reinterventions was compared between early (2011-2016) and later (2017-2020) time periods, the risk-adjusted rate of reintervention at 4 years had decreased over time (0.85% vs 0.51%). CONCLUSION: The 10-year risk of valve reintervention after TAVR is low and appears to be decreasing over time. Further research is necessary to determine the driving factors contributing to valve reintervention in the current era.

Cost-utility and cost-benefit analysis of TAVR availability in the US severe symptomatic aortic stenosis patient population.

AIMS: We evaluated the availability of transcatheter aortic valve replacement (TAVR) to determine its value across all severe symptomatic aortic stenosis (SSAS) patients, especially those untreated because of concerns regarding invasive surgical AVR (SAVR) and its impact on active aging. METHODS: We performed payer perspective cost-utility analysis (CUA) and societal perspective cost-benefit analysis (CBA). The CBA's benefit measure is active time: salaried labor, unpaid work, and active leisure. The study population is a cohort of US elderly SSAS patients. We compared a "TAVR available" scenario in which SSAS patients distribute themselves across TAVR, SAVR, and medical management (MM); and a "TAVR not available" scenario with only SAVR and MM. We structured each scenario with a decision-tree model of SSAS patient treatment allocation. We measured the association between health and active time in the US Health and Retirement Study and used this association to impute active time to SSAS patients given their health. RESULTS: The incremental cost-effectiveness ratio (ICER) and rate of return (RoR) of TAVR availability were $8,533 and 395%, respectively. CUA net monetary benefits (NMB) were $212,199 per patient and $43.4 billion population-wide. CBA NMB were $50,530 per patient and $10.3 billion population-wide. LIMITATIONS: Among study limitations were scarcity of evidence regarding key parameters and the lack of long-term survival, health utility, and treatment cost data. Our analysis did not account for TAVR durability, retreatments, and valve-in-valve treatments. CONCLUSION: Across risk-, age-, and treatment-eligibility groups, TAVR is the economically optimal treatment choice. It represents strong value-for-money per patient and population-wide. The vast majority of TAVR value involves raising treatment uptake among the untreated.

Aortic stenosis (AS) is a common and lethal heart disease. Surgical treatment has long been available, but its invasiveness limits uptake. More recently, transcatheter aortic valve replacement (TAVR) has emerged as a treatment alternative. Its minimal invasiveness has significantly increased treatment rates, but economic evaluations omit this benefit, risking undervaluation. We evaluated TAVR in elderly US severe symptomatic AS patients, using payer perspective cost-utility analysis (CUA) and societal perspective cost-benefit analysis (CBA). Both CUA and CBA incorporated TAVR's impact on treatment rates. Given patient preferences for treatment options promoting active aging, our CBA used the value of active time as a benefit measure. We found that CUA/CBA net monetary benefits are $212,199/$50,530 per patient. Across risk-, age-, and treatment-eligibility groups, TAVR is the economically optimal treatment choice over surgery and medical management. It represents strong value-for-money per patient and population-wide. Increased treatment uptake accounts for the vast share of TAVR's value.

The unintended consequences of Medicare's wage index adjustment on device-intensive hospital procedures.

OBJECTIVES: To study the association between Medicare's wage index adjustment and the differential use of labor-intensive surgical procedures and medical device-intensive minimally invasive clinical procedures across the United States. STUDY DESIGN: We combine a conceptual model and an empirical investigation of its predictions, applied to aortic valve replacement, to study the relationship between variation in Medicare wage index payment adjustment across hospital referral regions (HRRs) and the utilization of transcatheter aortic valve replacement (TAVR) and surgical aortic valve replacement (SAVR) in these areas. METHODS: Using detailed individual Medicare claims data for 2013-2018 and a novel geographical crosswalk to nest information on Medicare's wage index and utilization of TAVR and SAVR, we estimate a mixed effects Poisson regression model across HRRs to test our hypotheses. RESULTS: We find regional variation in Medicare wage index adjustment levels to be correlated with differential TAVR and SAVR utilization and growth over time. In particular, in HRRs where the wage index is half the national mean there is a 35% decline in the rate of TAVR use and in HRRs where the wage index is 50% higher than the national mean there is a 52% increase in the rate of TAVR use. CONCLUSIONS: Consistent with our framework and hypothesis, our results highlight the importance of adjusting Medicare hospital inpatient payments for device-intensive procedures. Absent such adjustment, access to appropriate interventions may be reduced in areas with low wage index, and lower reimbursement, when driven by wage index adjustment, may influence the treatment approach selected.

Contemporary Costs Associated With Transcatheter Versus Surgical Aortic Valve Replacement in Medicare Beneficiaries.

BACKGROUND: In patients with severe aortic stenosis, treatment with transcatheter aortic valve replacement (TAVR) has been shown to be cost-effective in the high-risk surgical population and cost-saving in the intermediate-risk population when compared with surgical aortic valve replacement (SAVR) in early pivotal clinical trials. Whether TAVR is associated with comparable or lower costs when compared with SAVR in contemporary clinical practice is unknown. METHODS: Using data from the Medicare Dataset Standard Analytic Files 5% Fee for Service database, patients receiving either TAVR or SAVR between 2016 and 2018 were identified. Patients were categorized as low, intermediate, or high mortality risk based on 2 validated indices-the Hospital Frailty Risk Score and the logEuroScore. Health care costs out to 1 year were compared between TAVR and SAVR among the low, intermediate, and high-risk groups, after adjustment for patient demographics. RESULTS: Nine thousand seven hundred forty-six patients were identified (4834 TAVR; 3760 SAVR) and included in the analysis. Patients receiving TAVR were older and more likely to be female. Index hospitalization costs were significantly lower with TAVR compared with SAVR across all risk strata (logEuroScore: low: $61 845 versus $68 986; intermediate: $64 658 versus $76 965; high: $65 594 versus $91 005; P<0.001 for all). Follow-up costs through 1 year were generally lower with TAVR and this difference was more pronounced in the low risk groups (logEuroScore: $9763 versus $14 073; Hospital Frailty Risk Score: $10 116 versus $12 880). Accordingly, cumulative 1-year costs were substantially lower with TAVR compared with SAVR. CONCLUSIONS: At 1 year, TAVR is associated with lower health care costs across all risk strata when compared with SAVR in contemporary practice. If long-term data continue to demonstrate similar clinical outcomes and valve durability with TAVR and SAVR, these findings suggest that TAVR may be the preferred treatment strategy for patients with aortic stenosis from an economic standpoint.

Temporal Changes in Mortality After Transcatheter and Surgical Aortic Valve Replacement: Retrospective Analysis of US Medicare Patients (2012-2019).

Background The treatment of aortic stenosis is evolving rapidly. Pace of change in the care of patients undergoing transcatheter aortic valve replacement (TAVR) and surgical aortic valve replacement (SAVR) differs. We sought to determine differences in temporal changes in 30-day mortality, 30-day readmission, and length of stay after TAVR and SAVR. Methods and Results We conducted a retrospective cohort study of patients treated in the United States between 2012 and 2019 using data from the Medicare Data Set Analytic File 100% Fee for Service database. We included consecutive patients enrolled in Medicare Parts A and B and aged ≥65 years who had SAVR or transfemoral TAVR. We defined 3 study cohorts, including all SAVR, isolated SAVR (without concomitant procedures), and elective isolated SAVR and TAVR. The primary end point was 30-day mortality; secondary end points were 30-day readmission and length of stay. Statistical models controlled for patient demographics, frailty measured by the Hospital Frailty Risk Score, and comorbidities measured by the Elixhauser Comorbidity Index (ECI). Cox proportional hazard models were developed with TAVR versus SAVR as the main covariates with a 2-way interaction term with index year. We repeated these analyses restricted to full aortic valve replacement hospitals offering both SAVR and TAVR. The main study cohort included 245 269 patients with SAVR and 188 580 patients with TAVR, with mean±SD ages 74.3±6.0 years and 80.7±6.9 years, respectively, and 36.5% and 46.2% female patients, respectively. Patients with TAVR had higher ECI scores (6.4±3.6 versus 4.4±3) and were more frail (55.4% versus 33.5%). Total aortic valve replacement volumes increased 61% during the 7-year span; TAVR volumes surpassed SAVR in 2017. The magnitude of mortality benefit associated with TAVR increased until 2016 in the main cohort (2012: hazard ratio [HR], 0.76 [95% CI, 0.67-0.86]; 2016: HR, 0.39 [95% CI, 0.36-0.43]); although TAVR continued to have lower mortality rates from 2017 to 2019, the magnitude of benefit over SAVR was attenuated. A similar pattern was seen with readmission, with a lower risk of readmission from 2012 to 2016 for patients with TAVR (2012: HR, 0.68 [95% CI, 0.63-0.73]; 2016: HR, 0.43 [95% CI, 0.41-0.45]) followed by a lesser difference from 2017 to 2019. Year over year, TAVR was associated with increasingly shorter lengths of stay compared with SAVR (2012: HR, 1.91 [95% CI, 1.84-1.98]; 2019: HR, 5.34 [95% CI, 5.22-5.45]). These results were consistent in full aortic valve replacement hospitals. Conclusions The rate of improvement in TAVR outpaced SAVR until 2016, with the recent presence of U-shaped phenomena suggesting a narrowing gap between outcomes. Future longitudinal research is needed to determine the long-term implications of lowering risk profiles across treatment options to guide case selection and clinical care.

The need for future coronary access following surgical or transcatheter aortic valve replacement.

The aim of the study was to estimate the percentage of Medicare patients needing coronary access for percutaneous coronary intervention (PCI) or coronary angiography following aortic valve replacement (AVR). Indications for TAVR have expanded to include younger and low-risk patients, raising the question of coronary access for future procedures. Medicare patients <80 years old with an AVR between 2011 and 2018 were included. Time-to-event analyses were conducted using Cox hazard models to estimate risk of coronary access up to 7 years after AVR. Model adjustments included age, sex, race, region, comorbidity, concomitant CABG, and smoking. A total of 13,469 Medicare patients (mean age 70.6) met inclusion criteria. Models estimated that 2.5% of patients at 1-year post-index and 17% at over 7 years would need coronary access. For patients who had SAVR (with or without CABG), estimates for coronary access were similar and over 15% after 6.5 years. For TAVR patients, with a previous PCI, 28% at 4.5 years required coronary access, which was higher than TAVR patients without a previous PCI. SAVR patients with and without CAD at baseline were similar; however, TAVR patients with CAD had a 22% rate of coronary access versus 7% for those without at 3 years. Approximately half of patients who needed coronary access returned to the same hospital as their initial AVR. Coronary access is required in a substantial portion of AVR patients especially those with PCI or a history of CAD undergoing TAVR. The need for coronary access may increase as transcatheter AVR becomes accessible to younger patients with a longer life expectancy.

Economic value of transcatheter valve replacement for inoperable aortic stenosis.

OBJECTIVES: To project the social value of transcatheter aortic valve replacement (TAVR) for inoperable patients with severe, symptomatic aortic stenosis (SSAS). STUDY DESIGN: This study used an economic model with parameters obtained from the literature and from US Census Bureau population projections. METHODS: Our model estimated the economic value that will accrue to inoperable patients with SSAS and to device manufacturers as a result of TAVR utilization. We estimated individual patient value as the monetized gain in quality-adjusted life-years as estimated in the cost-effectiveness literature, net of device costs and cost offsets. We estimated manufacturer value by applying an assumed profit margin to revenue from device sales. We created population-level estimates by combining these individual-level estimates with age-stratified Census Bureau population projections and estimates of the incidence of AS. We assessed model uncertainty through the use of probabilistic sensitivity analyses. RESULTS: Between 2018 and 2028, approximately 465,000 inoperable Americans with SSAS will be treated with TAVR. These procedures will yield a cumulative social benefit of up to $48 billion, with roughly 80% of that benefit accruing to patients and 20% accruing to device manufacturers. CONCLUSIONS: Policy makers and payers should take this social value into account when considering decisions related to the care of inoperable patients with SSAS.

Cost-Effectiveness of Transcatheter Aortic Valve Implantation Using a Balloon-Expandable Valve in Japan: Experience From the Japanese Pilot Health Technology Assessment.

BACKGROUND: Transcatheter aortic valve implantation (TAVI) is an innovative and effective treatment in high-surgical-risk (HR) and inoperable patients with symptomatic severe aortic stenosis. OBJECTIVES: This cost-effectiveness analysis of transfemoral TAVI (TF-TAVI) compared with surgical aortic valve replacement (SAVR) conforms with the methodological guidelines for cost-effectiveness evaluation by the Ministry of Health, Labor, and Welfare in Japan. METHODS: The cost-effectiveness of TF-TAVI using SAPIEN XT was evaluated using a lifetime Markov simulation from the national payer perspective. Comparators were SAVR for HR patients and standard of care (SOC) for inoperable patients. A systematic literature review for clinical evidence of TF-TAVI and comparators was conducted. The evidence for TF-TAVI was derived from the SOURCE XT registry and Japanese post marketing surveillance. Because there was no literature directly or indirectly comparing TF-TAVI using SAPIEN XT with comparators, the comparator data were selected from relevant published studies, considering the similarity of study eligibility criteria and patient backgrounds (eg, age and surgical risk scores). Sensitivity analyses were used to validate the robustness of results. RESULTS: The incremental cost-effectiveness ratio of TF-TAVI versus SAVR for HR patients was ¥1.3 million/quality-adjusted life-years (QALYs). The incremental cost-effectiveness ratio of TF-TAVI versus SOC for inoperable patients was ¥3.5 million/QALY. CONCLUSIONS: TF-TAVI was cost-effective when compared with SAVR for HR patients and when compared with SOC for inoperable patients, using a threshold of ¥5 million/QALY.

Incremental cost and length of stay associated with postprocedure delirium in transcatheter and surgical aortic valve replacement patients in the United States.

OBJECTIVES: To explore the impact of post-procedure delirium on resource utilization following transcatheter and surgical aortic valve replacement (TAVR and SAVR, respectively). BACKGROUND: Postprocedure delirium is associated with worse long-term survival after TAVR and SAVR. However, its effect on resource utilization has been understudied. METHODS: Using the 2015 Medicare Provider Analysis and Review File (MedPAR), we retrospectively analyzed elderly (≥80 years) Medicare beneficiaries receiving either SAVR or endovascular TAVR in the United States. Multivariate regression models estimating hospitalization cost and length of stay (LoS) were adjusted for patient demographics, comorbidities, and nondelirium complications. RESULTS: A total of 21,088 discharges were available for analysis (12,114 TAVR and 8,974 SAVR). TAVR patients were older (87 ± 3.8 vs. 84 ± 2.7 years; P < 0.001) with a higher comorbidity burden (Charlson index 3.0 ± 1.8 vs. 2.1 ± 1.7; P < 0.0001). Despite this, fewer TAVR patients (1.6%) experienced postoperative delirium during the index hospitalization compared to surgical patients (3.6%; P < 0.0001). Delirium was associated with a 4.16 [3.51-4.81] day longer hospital LoS and $15,592 ($12,849-$18,334) higher incremental hospitalization cost. When stratified by treatment approach, the adjusted incremental cost of delirium was +$13,862 ($9,431-$18,292) with TAVR and +$16,656 ($13,177-$20,136) with SAVR with an additional hospital LoS of +3.39 (2.34-4.43) days and +4.63 (3.81-5.45) days for TAVR and SAVR, respectively. CONCLUSIONS: Postprocedure delirium is associated with significantly increased hospitalization costs and LoS following AVR. TAVR was associated with a lower postoperative delirium rate compared to SAVR. Post-TAVR delirium may be associated with less resource consumption than post-SAVR delirium.