Assessment of clinical and economic impact of rivaroxaban plus aspirin vs. aspirin alone as a secondary prophylaxis in patients with chronic and symptomatic peripheral arterial disease in the United States

Abstract

Aim

The objective in this study was to assess the clinical and economic implications of the inclusion of rivaroxaban as a secondary prophylaxis in patients with chronic or symptomatic peripheral artery disease (PAD) in the United States (US).

Methods

A cost-consequence model was adapted to evaluate the economic impact of rivaroxaban plus aspirin in a hypothetical 1-million-member health plan. The model inputs were taken from multiple sources: efficacy and safety of rivaroxaban + aspirin vs. aspirin alone were abstracted from COMPASS and VOYAGER randomized clinical trials; the prevalence of chronic and symptomatic PAD and incidence rates of clinical events (major adverse cardiac events [MACE], major adverse limb events [MALE], and major bleeding), were abstracted from the analysis of claims data; healthcare costs of clinical events and wholesale acquisition costs for rivaroxaban were abstracted from the literature and Red Book, respectively (2022 USD). One-way sensitivity analyses and subgroup analyses were also conducted.

Results

Over one year, with a 5% uptake of rivaroxaban, the model estimated rivaroxaban + aspirin to reduce 21 MACE/MALE events in the PAD patient population. The reduction in these clinical events offsets the increased risk of major bleeding (16 additional events), demonstrating a positive health benefit of the rivaroxaban addition. These benefits led to a $0.27 incremental cost per member per month (PMPM) to a US plan. The major driver of the incremental cost was the cost of rivaroxaban. In a subgroup of patients with the presence of any high-risk factor (heart failure, diabetes, renal insufficiency, or history of vascular disease affecting two or more vascular beds), the incremental PMPM cost was $0.13.

Conclusions

Rivaroxaban + aspirin was found to provide positive net clinical benefit on the annual number of MACE/MALE avoided, with a modest increase in the PMPM cost.

Keywords:

• Aspirin
• peripheral artery disease
• cost-consequence analysis
• real-world evidence
• rivaroxaban
• major adverse cardiac events
• major adverse limb events

JEL Classification Codes:

• I10
• I1
• I
• I12
• D61
• D

Introduction

Peripheral artery disease (PAD) is a slow but progressive circulation disorder characterized by stenosis and/or occlusion of medium and large arteries that carry blood from the heart to the distal extremities^1,2. The lower extremity vessels are more commonly impacted by PAD than upper extremity vessels, resulting in recurring fatigue, cramping sensations, or pain¹. Nearly 6.5 million Americans (∼5.8%) aged 40 years and older have PAD^3,4. The prevalence increases with age and PAD is predominant in high-risk individuals (aged >70 years or aged 50–69 years with a history of cigarette smoking or diabetes) with a prevalence of 29%¹. Given the symptoms associated with PAD, it is very common for these patients to have reduced functional capacity and increased risk for cardiovascular morbidity and mortality¹.

The annual direct medical burden associated with chronic and symptomatic PAD is also significant^5,6. In 2015, the total economic cost among patients with PAD in the US was between $224 and $414 billion^7,8. This cost is primarily attributed to the healthcare burden associated with major adverse cardiovascular events (MACE) and major adverse limb events (MALE)⁸. A study published by Berger et al. in 2019 demonstrated that 22.9% of patients with chronic PAD experienced ≥1 MACE and/or MALE over a median follow-up of 2.3 years, resulting in a rate of 12.2 events per 100 person-years⁹. Similarly, a study published by Desai et al. in 2021 demonstrated that 17.4% of the patients with symptomatic PAD had a major atherothrombotic vascular event (i.e. composite of acute limb ischemia, major non-traumatic lower-limb amputation, myocardial infarction, ischemic stroke, and cardiovascular-related death) during a median follow-up of one year¹⁰.

The management of PAD includes lifestyle modifications (e.g. smoking cessation, exercise), medical management, endovascular repair, or surgery. Medical management of PAD is multifaceted and includes pharmacotherapy to manage contributing factors (e.g. cholesterol reduction therapy, antiplatelet therapy, anticoagulants, peripheral vasodilators, and antihypertensives)¹¹. In combination with aspirin, rivaroxaban, a direct inhibitor of factor Xa, has emerged as a promising strategy for reducing cardiovascular and limb events^12,13. The efficacy and safety of rivaroxaban were assessed in two phase 3, multinational, double-blind, randomized, placebo-controlled trials, namely the COMPASS trial that included chronic PAD patients and the VOYAGER trials that included symptomatic PAD patients^12,13. The COMPASS trial reported better cardiovascular and limb outcomes, and more major bleeding events in the rivaroxaban (2.5 mg BID) + aspirin (100 mg QD) arm compared to the aspirin alone (100 mg QD) arm¹⁴. Similarly, the VOYAGER trial reported better cardiovascular and limb outcomes but similar major bleeding events in the rivaroxaban (2.5 mg BID) + aspirin (100 mg QD) arm compared to the aspirin alone (100 mg QD) arm¹².

To the best of our knowledge, there has not been a study focused on evaluating the economic benefit of rivaroxaban + aspirin in the aggregated population of chronic or symptomatic PAD in the US setting. A prior cost-consequence model¹⁵ that leveraged real-world evidence data⁹ in the US to estimate one-year economic implications of preventing MALE and MACE with the use of rivaroxaban plus aspirin as a secondary prophylaxis in patients with chronic coronary artery disease (CAD) and/or PAD showed reductions of MACE and MALE in patients with rivaroxaban plus aspirin vs. aspirin alone, indicating a net health benefit for this drug regimen¹⁵. These reductions were achieved at an incremental per-member per-month (PMPM) cost of $0.16, mainly because of rivaroxaban’s acquisition cost. Furthermore, the model also demonstrated that the contribution of rivaroxaban in reducing the event rates of MACE and/or MALE would be greater in patients with ≥2 risk factors for MACE or MALE¹⁵.

While the prior cost-consequence model¹⁵ helped shed light on the burden of chronic PAD and the costs of MACE and MALE events, it did not provide comparable information on those with symptomatic PAD (i.e. VOYAGER-like cohort). Furthermore, the prior model did not provide information on the degree to which the risk profile for MACE and MALE event changes based on four specific risk factors that were identified in a post-hoc analyses of COMPASS trial (i.e. heart failure, diabetes, renal insufficiency, and history of vascular disease affecting two or more vascular beds).

The overarching goals of this study were to develop a cost-consequence model that can assess expected clinical and economic outcomes associated with the use of rivaroxaban plus aspirin vs. aspirin alone in the aggregated cohort of COMPASS-like chronic PAD patients (referred as COMPASS PAD-like cohort from hereafter) and VOYAGER-like symptomatic PAD patients, as well as in the subgroup of patients with the presence of one of more risk factors¹⁶ (i.e. heart failure, diabetes, renal insufficiency, and history of vascular disease affecting two or more vascular beds) for cardiovascular and limb events.

Methods

A previously published cost-consequence model¹⁵ was adapted to assess the clinical and economic implications of using rivaroxaban plus aspirin as a secondary prophylaxis in patients with chronic as well as symptomatic PAD (an aggregated cohort of the COMPASS PAD-like cohort and VOYAGER-like cohort) in a US payer formulary over a one-year timeframe. The model assessed the following two scenarios¹: without rivaroxaban—a scenario with patients receiving aspirin 100 mg daily; and² with rivaroxaban—a scenario with patients receiving rivaroxaban 2.5 mg BID in addition to the 100 mg aspirin QD. The incidence of clinical events (any MACE/MALE and major bleeding) and direct costs (i.e. drug acquisition and costs of clinical event management) were estimated under both scenarios. The net clinical benefit and the net economic benefit of rivaroxaban plus aspirin compared to aspirin monotherapy were calculated as the difference in incidence of clinical events and direct medical costs under the two scenarios, respectively.

Model inputs

While a description of the methodology for the analysis used to inform this model was published previously^9,15, in summary the inputs of the model (Table 1) were obtained from various sources. The efficacy and safety data for rivaroxaban plus aspirin were extracted from COMPASS and VOYAGER trials^12–14. The real-world incidence rates of clinical events (any MACE/MALE and major bleeding) in both COMPASS PAD-like and VOYAGER-like cohorts were generated from the analyses of Optum’s de-identified Clinformatics Data Mart claims database (Optum, Eden Prairie, MN, USA), which included patients with PAD from January 1, 2010 to June 30, 2021. The description of the cohort selection criteria, definition of clinical events, and analytical approach are provided in the Supplementary Methods. The model inputs included data on real-world incidence of clinical events which was obtained from analysis of medical and pharmacy claims using two mutually exclusive COMPASS PAD-like and VOYAGER-like cohorts, published literature for healthcare costs^9,15, and the Red Book for wholesale drug acquisition costs¹⁷.

Table 1. Model inputs.

Inputs for Model	Aggregated Cohort	Subgroup analysis: patients with ≥1 high-risk factors^a
Patients Observed in the Optum Clinformatics Database, n	710,088	339,459
COMPASS	672,567 (95%)	315,068 (93%)
VOYAGER	37,521 (5%)	24,391 (7%)
Enrollees with chronic/symptomatic PAD in a Plan of One Million Members, n^b	15,651	7,482
Incidence rates per 100 person-years^c
Any MACE (include CV-related death)/MALE
COMPASS	8.9	10.5
VOYAGER	16.3	19.0
Major Bleeding
COMPASS	3.5	4.0
VOYAGER	8.2	9.0
Hazard Ratio (95% Confidence Interval) RIVA + ASA vs. ASA^d
Any MACE (include CV-related death)/MALE
COMPASS	0.69 (0.56–0.85)^eCitation14
VOYAGER	0.85 (0.76–0.96)^fCitation12
Major Bleeding
COMPASS	1.61 (1.12–2.31)^14
VOYAGER	1.43 (0.97–2.10)^12
Drug cost, WAC for 30-day supply, $^17
Rivaroxaban	387.45^g
Aspirin	1.01
Incremental Annual Healthcare Cost per Event, $
Any MACE/MALE	39,904.58^h Citation9,Citation18
Bleeding event w/o MACE/MALE	20,218.45^i Citation15,Citation18

Abbreviations. ASA, aspirin; CV, cardiovascular; MACE, major adverse cardiovascular event; MALE, major adverse limb event; PAD, peripheral artery disease; RIVA + ASA, Rivaroxaban 2.5 mg BID + aspirin; WAC, wholesale acquisition cost.

^aHigh-risk factors include heart failure, diabetes, renal insufficiency, and history of vascular disease affecting 2 or more vascular beds¹⁶.

^bTotal number of unique members in the Optum Clinformatics were 45,369,951 between January 1, 2010 to June 30, 2021.

^cDerived from the analysis of claims data from the Optum Clinformatics between January 1, 2010 to June 30, 2021; details of the analysis are provided in the supplementary material.

^dThe efficacy and safety of rivaroxaban plus aspirin versus aspirin alone in subgroups was assumed to be equivalent to the overall trial population.

^eHazard ratio reported for the key composite outcome - Cardiovascular death, stroke, myocardial infarction or major adverse limb event including a major amputation – in PAD¹⁴.

^fHazard ratio reported for primary efficacy outcome of acute limb ischemia, major amputation for vascular causes, myocardial infarction, ischemic stroke, or CV death¹².

^gReflects a 25% rebate to the published WAC price of $861 for a pack of 100 rivaroxaban tablets (2.5 mg each).

^h2016 US dollar values were inflated to 2022 values using a factor of 1.166255¹⁸.

ⁱ2019 US dollar values were inflated to 2022 values using a factor of 1.084967¹⁸.

The analytical cohort for the model analysis comprised an aggregated cohort of patients with COMPASS PAD-like and VOYAGER-like symptomatic PAD patients. The claims analysis reported a total of 672,567 (95%) patients who met all the selection criteria for the COMPASS PAD-like cohort, while 37,521 (5%) patients met all the selection criteria for the VOYAGER-like cohort (Supplemental Figure 1). The size of the analytical cohort in a hypothetical health plan of 1 million members was calculated by scaling the number of patients in the COMPASS PAD-like cohort and VOYAGER-like cohort (N = 710,088) identified in the claims database analysis (N = 45,369,951 covered lives) to a 1-million-member health plan (N = 15,651 = 710,088/45,369,951 × 1 million). The plan size was assumed to remain constant, and all patients in the analytical cohort were assumed to receive treatment for the one-year model time horizon (i.e. 100% adherence and no treatment discontinuation). The model time horizon is similar to that used in prior analyses conducted in the US¹⁹. The incidence rates of clinical events for patients receiving aspirin monotherapy under the two modelled scenarios were informed by the analysis performed using Optum Clinformatics Database and are reported in Table 1. We derived the incidence rates of clinical events for patients receiving rivaroxaban plus aspirin by using hazard ratios(HR) reported in the COMPASS and VOYAGER trials. The HRs were applied to the incidence rates of clinical events for patients receiving aspirin monotherapy (Table 1). A detailed methodology has been published previously¹⁵.

Clinical event costs were based on the reimbursement amount reported by Berger et al.⁹ (any MACE/MALE) and Hernandez et al.¹⁵ (major bleeding). These costs were adjusted to 2022 US dollar values using the medical care component of the US Consumer Price Index¹⁸. The clinical event costs used in the model represent incremental events costs (i.e. calculated as a difference in annual healthcare costs between the patients who had an event and those who did not have an event; Table 1).

The 2022 wholesale acquisition cost (WAC) for rivaroxaban and aspirin was derived from the Red Book (WAC for a 30-day supply: aspirin $1.01; rivaroxaban $516.60). A rebate of 25% was applied to the rivaroxaban acquisition cost in the base case. This results in an annual acquisition cost of $12.15 for aspirin and $4,661.55 for rivaroxaban + aspirin (calculated as a 30-day supply WAC of rivaroxaban × 25% rebate + 30-day supply WAC of aspirin, multiplied by 12). No patient cost-sharing was assumed.

Model analyses

In the base-case analysis, we assumed a breakdown of the market share of 5% and 95% of the target population for rivaroxaban plus aspirin and aspirin alone, respectively. For each scenario, the model reports the following outcomes: number of any MACE/MALE events, number of major bleeding events, total costs, and costs by category (drug acquisition costs, healthcare costs associated with the incidence of clinical events). Incremental outcomes included cost per MACE/MALE avoided and total cost PMPM.

One-way sensitivity analyses were conducted to assess the impact of varying key model parameters one at a time on the incremental cost PMPM. The following parameters were tested: rebate for rivaroxaban acquisition cost, rivaroxaban market share, HRs of rivaroxaban plus aspirin, and incidence rates and healthcare costs of clinical events. The parameter values used in the one-way sensitivity are shown in Supplemental Table 1.

In a post-hoc analysis of the high-risk group of CAD or PAD patients enrolled in the COMPASS trial, patients with the highest risk of recurrent vascular events included those with disease in ≥2 vascular beds, a history of heart failure, renal insufficiency defined as an EGFR < 60 ml/min, or diabetes¹⁶. Based on this finding, a subgroup analysis was conducted in the cohort of patients with the presence of any of these high-risk factors. The model inputs specific to subgroups are shown in Table 1.

Results

In a 1-million-member US health plan, 15,651 patients in the aggregated cohort were estimated to receive aspirin monotherapy under the without rivaroxaban scenario (Table 1). Under the with rivaroxaban scenario, 5% (n = 782) of these patients were assumed to receive rivaroxaban plus aspirin, and the remaining 95% (n = 14,869) received aspirin alone. The availability of rivaroxaban was estimated to result in a reduction of 21 MACEs/MALEs compared with aspirin alone, which translated to reduction in annual clinical event costs of $791,643 ($756,945 and $34,698 in COMPASS PAD- and VOYAGER-like cohorts, respectively). The decrease in clinical event costs provided a complete offset to the additional costs of major bleeding events (16 additional major bleeding events under the “with rivaroxaban” scenario compared to “without rivaroxaban” scenario; annual incremental cost of $325,272). The decreased clinical event costs also provided partial offset to rivaroxaban’s acquisition cost, leading to a net increase of $0.27 PMPM attributable to the addition of rivaroxaban to a US third-party payer’s formulary (Table 2).

Table 2. Base-case results for aggregated cohort.

Parameter	Patients with chronic or symptomatic PAD
	Treatment without rivaroxaban	Treatment with rivaroxaban		Net benefit (treatment with vs without rivaroxaban)^a
	Aspirin (N = 15,651)	Aspirin (N = 14,869)	Rivaroxaban + aspirin (N = 782)
Number of events
Any MACE/MALE, COMPASS, N	1319	1253	46	−20
Any MACE/MALE, VOYAGER, N	135	128	6	−1
Major bleeding, N	522	496	42	16
Cost, $
Any MACE/MALE, COMPASS	50,372,926	47,854,280	1,761,701	−756,945
Any MACE/MALE, VOYAGER	4,963,661	4,715,478	213,485	−34,698
Major bleeding	10,549,286	10,021,822	852,736	325,272
Drug acquisition	192,934	183,287	3,701,110	3,691,463
Total costs	66,078,806	62,774,866	6,529,032	3,225,092
Incremental outcomes, $
Cost per MACE/MALE avoided				150,278
Cost PMPM				0.27

Abbreviations. MACE, major adverse cardiovascular event; MALE, major adverse limb event; PAD, peripheral artery disease; PMPM, per member per month.

^aNegative value indicates benefit/savings due to rivaroxaban.

Baseline characteristics and incidence rates of the aggregated cohort and the subgroup of patients with the presence of any of the high-risk factors are presented in Supplemental Tables 2–4.

Subgroup analysis

The results of the subgroup analysis are presented in Table 3. In this subgroup of patients with the presence of any risk factors (i.e. heart failure, diabetes, renal insufficiency, history of vascular disease affecting 2 or more beds), the annual incremental cost PMPM was estimated at $0.13. Under this analysis, savings due to a reduction in any MACE/MALE incidence was greater in the “with rivaroxaban” scenario compared to the base case, leading to an increase in the offset of the acquisition cost of rivaroxaban plus aspirin.

Table 3. Subgroup analysis results.

Parameter	Patients with chronic or symptomatic PAD + presence of any 1 or more risk factors of heart failure, diabetes, renal insufficiency, history of vascular disease affecting 2 or more beds
	Treatment without rivaroxaban	Treatment with rivaroxaban		Net benefit (treatment with vs without rivaroxaban)^a
	Aspirin (N = 7482)	Aspirin (N = 7108)	Rivaroxaban + aspirin (N = 374)
Number of events
Any MACE/MALE, COMPASS, N	729	693	25	−11
Any MACE/MALE, VOYAGER, N	102	97	4	−1
Major bleeding, N	283	269	23	9
Cost, $
Any MACE/MALE, COMPASS	27,621,484	26,240,410	968,344	−412,730
Any MACE/MALE, VOYAGER	3,712,211	3,526,601	159,967	−25,644
Major bleeding	5,730,264	5,443,751	463,098	176,585
Drug acquisition	92,232	87,621	1,769,323	1,764,711
Total costs	37,156,192	35,298,383	3,360,731	1,502,922
Incremental outcomes, $
Cost per MACE/MALE avoided				124,536
Cost PMPM				0.13

Abbreviations. MACE, major adverse cardiovascular event; MALE, major adverse limb event; PAD, peripheral artery disease; PMPM, per member per month.

^aNegative value indicates benefit/savings due to rivaroxaban.

Sensitivity analyses

The market share of rivaroxaban under the “with rivaroxaban” scenario was the most impactful parameter on model results (incremental cost PMPM). Model parameters, such as rebate for rivaroxaban’s drug acquisition cost, rates of clinical events, treatment effects of rivaroxaban plus aspirin, and annual healthcare costs of clinical events, had minimal impact on the base-case incremental cost PMPM (Supplemental Table 1).

Discussion

In this analysis, a cost-consequence model was used to estimate the clinical and economic benefits of rivaroxaban plus aspirin vs. aspirin alone as a secondary prophylaxis in patients in the COMPASS PAD-like and VOYAGER-like cohorts in the US. In the base-case scenario, treating 5% of the aggregated cohort with rivaroxaban plus aspirin resulted in preventing 21 MACE/MALE events. This prevention in MACE/MALE events was accompanied by 16 major bleeding events, which leads to a reduction of $448,911 in annual clinical event costs. Balanced against this reduction is the acquisition cost of rivaroxaban itself, resulting in an incremental annual cost PMPM of $0.27 in a health plan with 1 million members. These incremental costs should be viewed in the context of low-cost generic aspirin as the comparator. The market share was the most impactful driver of the results with the incremental costs PMPM in direct proportion with changes in the value of the market share.

Approximately 48% of the aggregated cohort included patients with the presence of one or more high-risk factors of heart failure, diabetes, renal insufficiency, or a history of vascular disease affecting two or more beds. In this subgroup of patients with the presence of any risk factors, the incremental PMPM cost of rivaroxaban plus aspirin was approximately 50% lower than the base case because of the increased offset of rivaroxaban’s acquisition cost by savings generated from the greater reductions in MACE and MALE in patients at higher risk for these clinical events.

We believe our estimates of incremental costs PMPM are conservative in two key aspects. First, the model assumes 100% adherence to treatment, no treatment discontinuation, and no patient cost-sharing. The HRs of treatment effect used in this analysis were derived on the intention-to-treat basis in all randomized patients, regardless of whether they received the treatment assigned to them in the COMPASS and the VOYAGER-PAD trials^12,14. Therefore, the HRs reflect the impact of treatment discontinuations for any reason (including adherence, death, etc.) observed in the clinical trials. Hence, the assumption of 100% adherence to treatment and no treatment discontinuation results in conservative estimates of the incremental cost of rivaroxaban plus aspirin to a health plan. Second, the annual clinical event costs used in the model do not capture the additional costs associated with rehabilitation due to chronic/acute limb ischemia, limb amputations, or stroke given a 1-year time-horizon of the analysis.

The results should be interpreted in the context of limitations. Briefly, the model does not account for the impact of clinical events and treatments on quality of life as well as excludes non-cardiovascular mortality. It should be noted that currently, there is no evidence to suggest that taking rivaroxaban + aspirin or aspirin alone affects non-cardiovascular mortality. Prior cost-effectiveness analyses of rivaroxaban conducted in patients with coronary or peripheral artery disease by the Institute for Clinical and Economic Review (ICER) in the US²⁰ used the lifetable approach to inform non-CV mortality which supports the lack of evidence on treatment effect on non-cardiovascular mortality. Hence, in the context of our study, the non-cardiovascular mortality would be similar under both the “without rivaroxaban” and “with rivaroxaban” scenario and will be cancelled out in the incremental analysis. Additionally, given the short analysis horizon of one-year, we assumed time-independent rates for clinical events and excluded any potential impact of previous event history on the risk of future clinical events. Given the lack of information on over-the-counter sales, the use of aspirin could not be captured in the claims database such as Optum Clinformatics. Hence, all patients were assumed to be on aspirin in this study as patients with PAD are typically expected to be taking aspirin per the treatment guidelines²¹. The analysis is also impacted by certain limitations specific to the analysis of Optum Clinformatics Database (provided in the Supplementary Methods). However, despite these limitations, our study is unique as to our knowledge, this model is the first economic evaluation of patients from the COMPASS PAD-like and VOYAGER-like cohorts from a US third-party payer perspective.

Conclusions

Our model estimated the real-world clinical and economic impact of outcomes observed for chronic PAD and symptomatic PAD patients in the COMPASS and VOYAGER clinical trials, respectively. The analysis leveraged an observational database study to inform incidence of clinical events rather than clinical trial estimates of such events reported under controlled environment.

Our analysis indicates that PAD patients on aspirin experience significant clinical burden due to the higher incidence of MACE/MALE and costs associated with these events. The actual burden is likely to be higher given our analysis excluded potential costs associated with rehabilitation, skilled nursing facility, long-term events, and productivity loss. Rivaroxaban plus aspirin can offer an effective treatment strategy option for patients with chronic or symptomatic PAD. The net clinical benefit associated with rivaroxaban plus aspirin was more pronounced in patients with the presence of risk factors for recurrent vascular events.

Transparency

Author contributions

All authors contributed to the study design, drafting, or revising the article, have agreed on the journal to which the article will be submitted, gave final approval of the version to be published, and agree to be accountable for all aspects of the work.

Reviewer disclosures

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

Previous presentations

None

Declaration of funding

Funding for this study was provided by Janssen Scientific Affairs.

Declaration of financial/other interests

AS, FD, and SS are employees of Evidera, a consulting company, which received funding from Janssen Scientific Affairs to conduct the study. AK and VA are employees of Janssen Scientific Affairs and hold Johnson and Johnson stock options. The authors report no other conflicts of interest in this work.

Data availability statement

The source data for this study is available online with the exception of data obtained from the Optum^® Clinformatics^® Data Mart. Restrictions apply to the Optum data, which were licensed by Janssen for this study. The Optum database can be licensed for a fee (https://www.optum.com/business/life-sciences/real-world-data/claims-data.html).