Budget impact analysis of transcatheter aortic valve replacement in low, intermediate, and high-risk patients with severe aortic stenosis in Saudi Arabia

Abstract

Aims

A budget impact analysis (BIA) comparing transcatheter aortic valve replacement (TAVR) with SAPIEN 3 and surgical aortic valve replacement (SAVR) for severe, symptomatic aortic stenosis among patients of low, intermediate, and high surgical risk from the perspective of the public and private sectors in Saudi Arabia.

Materials and methods

A Markov model was developed with six states to calculate the budget impact from time of either TAVR or SAVR intervention up to 5 years. We compared the budget effects of new permanent pacemaker implantation (PPI), new onset atrial fibrillation (AF), major/disabling stroke (MDS), and surgical site infections (SSI). One-way sensitivity analyses (OWSA) were performed on cost and probability inputs.

Results

Analysis of the base case parameters suggests TAVR vs. SAVR is budget saving among intermediate- and high-risk patients at 5 years. TAVR vs. SAVR for low surgical risk reaches budget neutrality at 5 years. TAVR is associated with higher costs for PPI and budget savings for MDS, AF, and SSI. TAVR also results in savings for non-device costs due to fewer human resource uses and shorter procedure durations. Similarly, TAVR is associated with cost savings due to shorter hospital intensive care unit (ICU) and non-ICU stays. The OWSA consistently revealed that SAVR non-device theater costs were the leading cost driver across all surgical risk levels.

Limitations

This is the first budget impact analysis of its kind in Saudi Arabia and future research is needed on costing TAVR and SAVR procedures, the economic impact of SSI, and corroborating estimates for the public and private sectors.

Conclusions

Payers, providers, and policymakers increasingly turn to results of BIA to inform technologies affordability decisions. TAVR with SAPIEN 3 appears to generate savings vs. SAVR from a budget impact perspective across various surgical risk levels in Saudi Arabia.

Keywords:

• Budget impact analysis
• Saudi Arabia
• severe aortic stenosis
• transcatheter aortic valve replacement
• surgical aortic valve replacement

JEL codes:

• D81
• D8
• D
• E37
• E3
• E

Introduction

Aortic stenosis (AS) is an end stage disease characterized by a narrowing of the aortic valve commonly due to calcification but also associated with infection, rheumatic fever, or a congenital heart defect¹. As a result, a patient’s blood flow from the left ventricle to the aorta is reduced, leading to an increased likelihood of mortality and reductions in quality-of-life, whilst increasing healthcare costs for the patient and healthcare system^2–6.

The prevalence of AS in patients >60 years is 2–7% in advanced economies and increases with age up to 13.3% among patients >75 years^7–9. In the Kingdom of Saudi Arabia (KSA), increasing life expectancy is anticipated to increase the number of AS cases¹⁰, and the total cost to the healthcare system. Although some patients are medically managed due to multiple comorbidities and higher surgical risk⁵, most undergo either surgical aortic valve replacement (SAVR) or transcatheter aortic valve replacement (TAVR). Patients that undergo SAVR experience longer recovery times and longer hospital lengths of stay (LOS) compared with TAVR procedures that are less invasive and associated with improved patient outcomes^11–15.

Since April 2018, patients with severe AS of all surgical risk levels in KSA have access to the balloon-expandable SAPIEN 3ⁱ System (hereafter S3) for TAVR procedures. S3 is the third generation of the SAPIEN valve and builds on numerous trials and evidence across various surgical risk groups^16–28. Overall, TAVR procedures with S3 vs. SAVR are associated with improved survival, reduced complications, shorter hospital LOS, and reduced post-procedure wound infections. However, the acquisition cost of S3 is higher compared to valves used in SAVR leading to the core question on the budget impact assessing potential peri- and post-procedure cost offsets due to reduced healthcare consumptions in KSA.

Payers, providers, and policy-makers increasingly turn to results of budget impact analyses (BIA) to inform decisions on the adoption of new healthcare technologies^29–31. In KSA, BIA is not yet required for reimbursement of medical devices. This will change when BIA becomes an essential part of dossiers submissions requirements with the formalized introduction of a health technology assessment (HTA) system in KSA^32,33. There are no published results of a budget impact analyses among severe AS patients in KSA or the Middle East. Payers, providers, and policymakers lack decision-analytic tools to determine budget offsets associated with innovative technologies such as S3 and the impact on healthcare resource consumption. To this end, determining the budget impact of S3 is a challenge in KSA due to the lack of epidemiological data, costing of SAVR and TAVR procedures in both the public and private sectors, and patient registries with clinical outcomes for TAVR and SAVR.

This study therefore employs the best available evidence and validated assumptions to analyze the budget impact associated with S3 among severe AS patients in low, intermediate, and high surgical risk patients from the perspective of the public and private sectors in KSA.

Methods

The BIA estimates the financial consequences of adopting S3 for TAVR vs. SAVR procedures over a 5-year time horizon from the perspective of both public and private sector payers in KSA. Guidelines by the International Society for Pharmacoeconomics and Outcomes Research (ISPOR) for BIA were used to guide the development of the model and presentation of the results²⁹.

Model structure

A Markov model compares TAVR with S3 to SAVR among severe symptomatic aortic stenosis patients across low-, intermediate-, and high-risk surgical levels. The model calculates the expected cost that a patient accumulates as they progress through mutually exclusive states defined by a patient’s journey from the time of either a TAVR or SAVR intervention. These states include an in-hospital, post-operative 30-day period covering all outcome-related costs, a state in the outpatient setting for patients at risk of long-term complications, three maintenance states for outpatients with either a new permanent pacemaker implantation (PPI), new onset atrial fibrillation (AF), or major/disabling stroke (MDS), and a final absorptive state for aggregating mortality in the model (see Figure 1). Monthly cycles and transition probabilities determine how patients move between states over a duration of 5 years. The model includes a half-cycle correction to account for over counting in Markov processes.

Figure 1. State transition diagram.

Patient population and clinical outcomes

All patients start in the state “in-hospital, post-operative (30-day period)” and enter the model at different ages based on low, intermediate, or high-risk surgical levels (Table 1)^18,22,23,26. In this initial state, patients at risk of death transitioned based on published mortality rates, and up to 1 year thereafter in any subsequent state, based on period-specific transition probabilities^{11,18,22,23,26}. TAVR-related mortality is consistently lower across all surgical risk levels compared with SAVR. Thereafter, and for the remaining duration of the model, patient mortality reverted to the general population life tables for the KSA³⁴.

Table 1. Model parameters – patient population and clinical outcomes (base case parameters).

Variables	Period	TAVR	SAVR	Sources
Start age (mean, years)
Low-risk	–	73.3	73.6	^26
Intermediate-risk	–	81.8	81.6	^18^,^22^,^23
High-risk	–	83.1	84.5	^17^,^18
Mortality − 30 day
Low-risk	30-day	0.004	0.011	^26
Intermediate-risk	30-day	0.011	0.04	^22
High-risk	30-day	0.022	0.065	^11^,^18
Mortality − 1 year
Low-risk	1 Year	0.01	0.025	^26
Intermediate-risk	1 Year	0.074	0.13	^22
High-risk	1 Year	0.127	0.268	^11^,^17
Atrial fibrillation (new onset)
High-risk	30-day	0.073	0.160	^11^,^16^,^19
Intermediate-risk	30-day	0.050	0.283	^22
Low-risk	30-day	0.050	0.395	^26
High-risk	1 Year	0.073	0.171	^11^,^16^,^19
Intermediate-risk	1 Year	0.059	0.292	^22
Low-risk	1 Year	0.070	0.409	^26
Major/disabling stroke
High-risk	30-day	0.009	0.021	^11^,^17^,^18
Intermediate-risk	30-day	0.01	0.044	^22
Low-risk	30-day	0.006	0.024	^26
High-risk	1 Year	0.030	0.024	^11^,^17^,^18
Intermediate-risk	1 Year	0.023	0.059	^22
Low-risk	1 Year	0.002	0.009	^26
New PPI
High-risk	30-day	0.133	0.036	^17^,^18
Intermediate-risk	30-day	0.102	0.073	^22
Low-risk	30-day	0.066	0.041	^26
High-risk	1 Year	0.145	0.05	^17^,^18
Intermediate-risk	1 Year	0.124	0.094	^22
Low-risk	1 Year	0.075	0.055	^26
SSI
All risk levels	30-day	0.01	0.08	^42

Patients that undergo TAVR or SAVR procedures are at risk of complications. The co-authors prioritized a list of complications based on clinical relevancy and higher healthcare resource consumption; and included PPI, MDS, AF, and surgical site infections (SSI). Under clinical relevance we considered complications with statistically meaningful differences in event rates and high absolute event rates relative to other complications. Paravalvular leak, acute kidney injury, endocarditis, major vascular complications, life-threatening bleeding, and history of AF prior to either TAVR or SAVR were all excluded due to low, insignificant event rates when compared to the major complications, or low healthcare resource impact. Treatment goals for TAVR include minimizing complications and reducing downstream, recurring expenses. Event rates for patients at risk of in-hospital or outpatient complications were derived from published clinical evidence (Table 1).

TAVR procedures are associated with lower event rates for AF^{11,16,19,22,26} across all surgical risk levels during the in-hospital, post-operative 30-day period, and the state for patients at risk of long-term complications in the outpatient setting. Similarly, TAVR is associated with lower event rates for MDS across all surgical risk levels except among high-risk patients at 1 year where the event rate for SAVR is lower compared to TAVR^{11,17,18,22,26}. Although numerically lower, this is most likely not statistically significant due to data sourced from different clinical studies with different patient samples. PPI event rates are higher across all risk levels for TAVR with SAPIEN 3 vs. SAVR (30-day period and 1 year) – except in low-risk patients with numerical but not statistically significant differences reported between TAVR with SAPIEN 3 and SAVR. We assumed that published 1 year event rates for any complication are relevant for the duration of the model.

SSI among patients undergoing TAVR or SAVR is an important determinant of mortality, hospital length of stay, and hospitalization costs (intensive care unit (ICU) and non-ICU)^35–45. Wound disorders are significantly lower in TAVR vs. SAVR (1% vs. 8%, p = 0.035)⁴², and deep surgical wound infection and mediastinitis after median sternotomy are associated with mortality as high as 50%³⁵. Overall, patients undergoing TAVR develop fewer infectious complications and consume fewer healthcare resources. In the absence of published evidence, we assumed that incidence rates for SSI were uniformly applicable across all surgical risk levels (Table 2).

Table 2. Model parameters – costs (in SAR) and hospital length-of-stay (base case parameters).

State	Frequency	TAVR	SAVR	Sources
In-hospital, post-procedure (30 day)
Procedure/Theater
Device cost	Per procedure/event	120,995.36	18,182.34	Edwards Lifesciences
Non-device cost	Per procedure/event	28,757.53	72,237.70	^46–49
Post-procedure
ICU	Per procedure/event	5,400.00	5,400.00	Expert Opinion
Non-ICU	Per procedure/event	1,640.93	1,640.93	^59
In-hospital complications
New PPI	Per procedure/event	5,432.16	5,432.16	Expert Opinion
New onset AF	Per procedure/event	2,673.68	2,673.68	Intermarket Cost-Ratio Analysis
Major/disabling stroke	Per procedure/event	11,411.65	11,411.65	Intermarket Cost-Ratio Analysis
SSI (ratio)	Per procedure/event	1.10	2.00	^38^,^39
Outpatient – long-term complications
New PPI	Per procedure/event	60,090.00	60,090.00	Expert Opinion
Major/disabling stroke	Per procedure/event	103,884.00	103,884.00	^61
New onset AF	Per procedure/event	16,626.38	16,626.38	^62
Outpatient
New PPI	Per month	1,192.51	1,192.51	Intermarket Cost-Ratio Analysis
New Onset AF	Per month	1,291.89	1,291.89	Intermarket Cost-Ratio Analysis
Major/disabling stroke	Per month	16,211.22	16,211.22	Intermarket Cost-Ratio Analysis
Hospital LOS (in days)	Setting	TAVR	SAVR	Sources
High-risk	ICU/Non-ICU	4.4 / 6.6	14.4 / 21.6	^50
Intermediate-risk	ICU/Non-ICU	1.6 / 2.4	4.6 / 6.2	^22
Low-risk	ICU/Non-ICU	1.2 / 1.8	2.8 / 4.2	^26

Costs and resource consumption

The model considers only direct costs in 2021 Saudi Riyal (SAR) from the perspective of both the public and private sectors. Costs for the in-hospital, post-operative 30-day period were disaggregated into Procedure/Theater covering Device and Non-Device Costs; Post-Procedure for the total ICU and non-ICU stays, and In-Hospital Complications for PPI, AF, MDS, and SSI. Cost for long-term complications in the outpatient setting were disaggregated per complication. In the first two states (Figure 1 – #1 and #2), costs were calculated per procedure or per event. In-hospital complications (under the first state, Figure 1 – #1) considered only the additional cost, while in the outpatient setting, long-term complication costs included hospitalization. The remaining maintenance states aggregated monthly costs to manage patients with PPI, AF, or MDS.

Procedure/theater and post-procedure costs

Device costs included the private sector acquisition of S3 for TAVR and assumed SAVR procedures used Inspiris Resilia (Edwards Lifesciences) (Table 2). Private sector acquisition costs result in conservative results as prices are lower when acquired through the centralized procurement agency National Unified Procurement Company (NUPCO) for the public sector. An additional 10.2% cost was added to the device cost, based on expert opinion, to cover consumables, such as catheter balloon, introducer sheath, device vascular closure, etc.

The model included non-device costs to estimate human resources for TAVR and SAVR during the initial state such as surgeon/cardiologist, anesthesiologist, and theater assists. Total costs were based on the mean of published procedure durations^46–48, number of human resources per procedure, and a resource cost per minute⁴⁹. Shorter TAVR procedure durations are associated with lower total non-device costs compared to SAVR (Table 2). We made no assumptions on capital costs that could be included under non-device costs but would vary, considerably, between the private and public sectors. We assumed that non-device costs remained fixed for each surgical risk level.

The model considers that SSI are initially identified in-hospital and a cost per event accrues to each patient. The average hospital cost of postoperative sternal infection was double that of noninfected patients³⁸. In a different study, costs were 2.8-times higher for patients with sternal would complications compared with uncomplicated postoperative cases³⁹. We assumed SSI in SAVR doubled the post-procedure total costs per admission (only ICU and non-ICU stay) and, in the absence of available evidence, we assumed a 10% increase for TAVR. The final estimate for additional SSI costs were different per surgical risk level given differences in hospital LOS.

Hospital length of stay

Hospital LOS is an important multiplier for total hospitalization costs and accounts for significant differences between TAVR and SAVR^22,26,50. When a data source did not report ICU and non-ICU stay separately, the total LOS was split between 40% for ICU and the remaining 60% for non-ICU stay based on published data⁵¹. TAVR is associated with substantial reductions in LOS consistent across all surgical risk levels (Table 2).

Complication costs

The model accounts for state costs covering in-hospital and long-term complication-related costs. Given the lack of structured costing data on TAVR vs. SAVR in KSA, three independent approaches to data collection included intermarket cost-ratios, published evidence, and institutional inputs from co-authors. An intermarket cost-ratio analysis assumes that costs in KSA are proportional to other countries⁵². In this first approach, costs reported in published studies^49,53–58 were extracted and expressed as a ratio of in-country device costs. Unlike prior studies that indexed the KSA vs. the United Kingdom (UK) physician visit costs⁵², this study estimated event costs in KSA as a proportion of device costs in KSA.

An intermarket cost-ratio analysis may bias estimates due to differences in healthcare delivery systems (public vs. private) and level of economic development. This approach was strengthened with searches on PubMed and Google Scholar for cost inputs related to PPI, MDS, AF, and SSI in KSA. Inputs from all available studies^52,59–62 were extracted, adjusted to the present value, and compared with calculated estimates. These were complemented with institutional costs provided by co-authors and included expert opinion on reimbursement tariffs using Disease-Related Groups (DRG) for the private and public sectors. AR-DRG codes F03A/B, F04A/B/C, and F19A/B/Z used in KSA describe open vs. percutaneous approaches in cardiac valve replacement interventions. Although the AR-DRG is a well-developed medical coding framework, TAVR currently groups to the same DRG “as cardiac valve replacement interventions performed using an open approach”⁶³ (p.27). Addressing weaknesses in this approach is a work-in-progress.

Final cost estimates for all states were subjected to internal consistency and face validity checks among co-authors. Published evidence was preferred as a source, followed by expert opinion, and then Intermarket Cost-Ratio Analysis. No discounting was applied to cost estimates as recommended by ISPOR Guideline for BIA²⁹. All costs and prices are based on the Saudi Arabian Riyal (SAR), equal to 3.75 SAR per United States Dollar (USD).

Sensitivity analysis

A challenge in KSA is uncertainty around cost estimates. One-way sensitivity analyses were performed on cost estimates with broad 50% bands covering values from the lower to the upper extremes. All one-way sensitivity analyses on cost inputs were arranged in a tornado diagram to facilitate analysis. Probabilities in the model were analyzed in a separate tornado diagram but with inputs varied by ±20% given that these were derived from published evidence.

Results

Overall results of the BIA by surgical risk level and disaggregated by period (30-day and annual) are presented in Table 3.

Table 3. Budget impact results (in SAR) disaggregated by period and cost categories.

Overall budget impact results (in SAR), by period				In-hospital, post-procedure (30-day period) disaggregated costs				Procedure/theater costs disaggregated by device, non-device, and LOS
Period	TAVR	SAVR	Cumulative difference	Cost category	TAVR	SAVR	Difference	Procedure/theater costs	TAVR	SAVR	Difference
Low surgical risk
30 day	159,848	117,451	42,398	Procedure	159,187	112,432	46,755	Device	120,995	18,182	102,813
1 yr	167,866	134,328	33,538	New PPI	357	220	137	Non-device	28,758	72,238	(43,480)
2 yr	177,621	153,076	24,545	MDS	68	271	(203)	LOS ICU	6,480	15,120	(8,640)
3 yr	187,980	171,465	16,515	New onset AF	133	1,044	(911)	LOS non-ICU	2,954	6,892	(3,938)
4 yr	198,680	189,189	9,490	SSI	103	3,483	(3,380)
5 yr	210,230	207,418	2,812
Intermediate surgical risk
30 day	163,261	132,401	30,860	Procedure	162,331	125,434	36,897	Device	120,995	18,182	102,813
1 yr	176,632	154,433	22,199	New PPI	548	381	167	Non-device	28,758	72,238	(43,480)
2 yr	193,010	178,812	14,198	MDS	113	482	(369)	LOS ICU	8,640	24,840	(16,200)
3 yr	210,129	202,676	7,453	New onset AF	132	726	(594)	LOS non-ICU	3,938	10,174	(6,236)
4 yr	227,179	225,128	2,051	SSI	137	5,378	(5,241)
5 yr	244,733	247,287	(2,554)
High surgical risk
30 day	185,713	221,366	(35,653)	Procedure	184,343	203,624	(19,281)	Device	120,995	18,182	102,813
1 yr	200,292	233,000	(32,707)	New PPI	707	183	524	Non-device	28,758	72,238	(43,480)
2 yr	217,308	245,601	(28,293)	MDS	100	224	(124)	LOS ICU	23,760	77,760	(54,000)
3 yr	234,661	258,466	(23,806)	New onset AF	191	400	(209)	LOS non-ICU	10,830	35,444	(24,614)
4 yr	251,522	270,905	(19,384)	SSI	372	16,935	(16,563)
5 yr	268,469	283,319	(14,849)

Abbreviations: SAR, Saudi Arabian Riyal; ICU, Intensive Care Unit; New PPI, new permanent pacemaker insertion; New onset AF, new onset atrial fibrillation; SSI, surgical site infection.

Analysis of the base case parameters suggests TAVR vs. SAVR in KSA is overall budget saving among high- and intermediate-risk patients at 5 years. However, TAVR vs. SAVR is not budget-saving for low surgical risk but approaches budget neutrality after the fifth year. Results are determined by initial in-hospital, post-procedure (30-day period) costs offset by downstream costs. While procedure costs are higher among low- and intermediate-risk groups, these are offset by cost reductions in all cost categories except New PPI. Cost savings in high-risk patients are eroded as the model progresses due to significantly lower TAVR mortality. During the initial phase, procedure costs are the main contributor, costing more in TAVR vs. SAVR among low- and intermediate-risk patients, but costing less in high-risk patients. TAVR vs. SAVR is associated with higher expected costs for PPI and budget savings for MDS, AF, and SSI (Table 3).

Device costs impact total procedure/theater costs the most and remain fixed across all surgical risk levels (Table 3). However, TAVR results in savings for non-device costs due to lower human resource use and shorter procedure durations. Similarly, TAVR is associated with substantial cost savings due to shorter hospital ICU and Non-ICU stays, with the greatest savings among high-risk patients.

Sensitivity analysis

Results of the one-way sensitivity analyses (Figure 2) of costs and probabilities among low-risk patients suggests that SAVR non-device theater costs, monthly maintenance cost to manage SAVR patients with AF, and TAVR non-device procedure costs are the three most impactful variables in the model.

Figure 2. Tornado diagram – TAVR vs. SAVR (costs and probabilities), low-risk.

Among intermediate-risk patients, SAVR non-device theater costs, SAVR monthly maintenance cost of stroke, and SAVR monthly maintenance cost of AF are the most impactful variables (Figure 3).

Figure 3. Tornado diagram – TAVR. vs. SAVR (costs and probabilities), intermediate-risk.

Finally, among high-risk patients, SAVR non-device theater costs, TAVR monthly maintenance costs for stroke, and TAVR non-device procedure costs are the top three cost drivers (Figure 4).

Figure 4. Tornado diagram – TAVR. vs. SAVR (costs and probabilities), high-risk.

Discussion

To the authors’ knowledge, this is the first BIA in KSA to calculate the financial consequences of using TAVR vs. SAVR in low-, intermediate-, and high-surgical risk patients. The results are specific to SAPIEN 3 and were modeled on clinical evidence and costing inputs specific to this technology. Results are therefore not generalizable to other technologies that may be used in TAVR procedures. The analysis demonstrates that TAVR is cost saving due to reductions in MDS, AF, SSI, and LOS-related costs. TAVR vs. SAVR in intermediate-risk and high-risk patients is budget saving after 5 years. TAVR in low-risk patients may become budget saving within 5 years if non-device procedure costs are reduced to same-day procedures^64,65.

The one-way sensitivity analyses consistently revealed that SAVR non-device theater costs were the single, leading cost driver in the model across all surgical risk levels. These costs relate to human resource use and procedure durations. The cost of complications varied in impact across risk levels, with monthly maintenance costs of stroke patients as a leading contributor in this category.

In the literature, there are no other budget-impact results in KSA or the Middle East to compare these results. However, comparing costs only reported in prior studies^49,53–58 suggests results in this analysis are comparable. SAVR LOS and the cost per day of hospitalization were also leading cost drivers in a tornado diagram for intermediate risk patients in Australia⁵⁴. The Health Information and Quality Authority (HIQA) in Ireland reported TAVR vs. SAVR was budget neutral in a mix of 67% low-risk and 33% intermediate-risk patients⁵⁶. Their analysis included an up-front capital cost for additional catheterization laboratory capacity which was offset by reduced procedure costs in subsequent years. The Norwegian Institute of Public Health (NIPH) reported a cost-neutral impact of TAVR vs. SAVR in low- and intermediate-risk patients⁵⁷. The French HTA agency, Haute Autorité de Santé (HAS), reported budget savings of €67 million among low-risk patients (TAVR with SAPIEN 3 vs. SAVR) and positive recommendation for the reimbursement of the technology⁶⁶. A key difference compared to other studies is higher non-device procedure/theater costs in Canada vs. KSA (CAD34395 vs. USD8511, respectively). Differences in costing hospitalizations and whether capital costs are included may determine overall budget impact results across countries.

Unlike prior studies comparing TAVR vs. SAVR, this analysis included the budget-impact of SSI. Sternotomy wound infections (deep or superficial) in SAVR are costly events with high mortality rates. Although the costs and probabilities related to SSI were not among the most impactful variables in the model, they nonetheless contributed higher cost savings for TAVR than the other complications: PPI, MDS, and AF. The cost of managing SSI may be an important goal in aortic valve replacement procedures, both in KSA and other countries.

Limitations

Assumptions were used to facilitate this analysis, so results should be considered carefully before adoption by payers, providers, and policymakers in KSA. First, collecting costs in KSA is challenging given the absence of structured costing data for TAVR vs. SAVR. The costing approach in this analysis included an intermarket cost-ratio analysis, published evidence, and expert opinion. Despite efforts to strengthen the costing methodology, there may still be bias included in the costing estimates. In the future, a health system-wide adoption of DRGs or a micro-costing analysis may help refine results. Similarly, additional analysis may be required to determine the contribution of capital costs under non-device procedure/theater costs on the overall budget impact of TAVR vs. SAVR.

We assumed that non-device costs remained fixed for each surgical risk level. In reality, and if published evidence were available, non-device costs may be lower among low-risk patients due to shorter procedure durations, and higher in high-risk patients. The result of this assumption is conservative budget impact results, meaning that TAVR may become budget saving in low-risk patients within 5 years. We also assumed that SSI in SAVR doubled the post-procedure total costs per admission. This was based on dated studies (2000 and 2001) and since then wound infection practices and technologies may have improved. To this end, the accuracy of this analysis may improve with future research on SSI in TAVR and SAVR patients in KSA and other countries.

The results reported herein are from the perspective of both the public and private sectors in KSA. Each sector is different and repeating this analysis per sector may yield different results. Moreover, given the absence of epidemiological data in KSA for AS, a separate analysis that quantifies the prevalence and incidence across surgical risk groups, age, geographical region, public and private sector, and changes over time, may add to our future understanding of TAVR vs. SAVR budget impact. However, the sensitivity analysis included broad 50% bands (low and high values) and may already include cost estimates relevant in separate analyses. The results reported herein point to cost drivers that may lead to critical differences between the public and private sectors. Future research may corroborate the relevance of these estimates for either sector.

Conclusion

The results of this analysis suggest that the higher, initial cost of S3 is offset by cost savings in MDS, AF, SSI, and LOS. The leading cost driver is SAVR non-device theater costs due to longer procedure durations and human resource use. TAVR is budget saving among intermediate- and high-risk patients and approaches budget neutrality among low-risk patients. Payers, providers, and policymakers increasingly turn to results of BIA to inform technologies affordability decisions. TAVR with SAPIEN 3 appears to generate savings vs. SAVR from a budget impact perspective across various surgical risk levels in Saudi Arabia.

Transparency

Declaration of funding

This work was supported by a grant from Edwards Lifesciences.

Declaration of financial/other interests

JLC is a consultant for Edwards Lifesciences. HAA, FA, and SAS are supported by their employing institutions. PC is an employee at Edwards Lifesciences.

Peer reviewers on this manuscript have no relevant financial or other relationships to disclose.

Acknowledgements

The authors would like to acknowledge the initial conceptualization of the project by Sarrah Barakat, former employee at Edwards Lifesciences.

Notes

i SAPIEN 3 is a registered trademark of Edwards Lifesciences Corporation, Irvine, CA, USA.