The Cost-Effectiveness of Belimumab and Voclosporin for Patients with Lupus Nephritis in the United States

Abstract

Background and objectives 

Despite existing therapies, people with lupus nephritis progress to kidney failure and have reduced life expectancy. Belimumab and voclosporin are two new disease-modifying therapies recently approved for the treatment of lupus nephritis.

Design, setting, participants, & measurements 

A de novo economic model was developed to estimate the cost-effectiveness of these therapies, including the following health states: “complete response,” “partial response,” and “active disease” defined by eGFR and proteinuria changes, kidney failure, and death. Short-term data and mean cohort characteristics were sourced from pivotal clinical trials of belimumab (the Belimumab International Study in Lupus Nephritis) and voclosporin (the Aurinia Urinary Protection Reduction Active–Lupus with Voclosporin trial and Aurinia Renal Response in Active Lupus With Voclosporin). Risk of mortality and kidney failure were on the basis of survival modeling using published Kaplan–Meier data. Each drug was compared with the standard of care as represented by the comparator arm in its respective pivotal trial(s) using US health care sector perspective, with a societal perspective also explored.

Results 

In the health care perspective probabilistic analysis, the incremental cost-effectiveness ratio for belimumab compared with its control arm was estimated to be approximately $95,000 per quality-adjusted life year. The corresponding incremental ratio for voclosporin compared with its control arm was approximately $150,000 per quality-adjusted life year. Compared with their respective standard care arms, the probabilities of belimumab and voclosporin being cost effective at a threshold of $150,000 per quality-adjusted life year were 69% and 49%, respectively. Cost-effectiveness was dependent on assumptions made regarding survival in response states, costs and utilities in active disease, and the utilities in response states. In the analysis from a societal perspective, the incremental ratio for belimumab was estimated to be approximately $66,000 per quality-adjusted life year, and the incremental ratio for voclosporin was estimated to be approximately $133,000 per quality-adjusted life year.

Conclusions 

Compared with their respective standard care arms, belimumab but not voclosporin met willingness-to-pay thresholds of $100,000 per quality-adjusted life year. Despite potential clinical superiority in the informing trials, there remains high uncertainty around the cost-effectiveness of voclosporin.

Visual Abstract

Introduction

Lupus nephritis, characterized by inflammation in the kidney, proteinuria, and progressive kidney damage, is caused by SLE (¹,²). Lupus nephritis is more common in women and in non-White populations (³,⁴), with most patients diagnosed between age 20 and 40 (⁵,⁶); 10%–30% of patients with lupus nephritis progress to kidney failure within 15 years of the diagnosis, requiring dialysis or kidney transplantation (^78–9). Lupus nephritis has a substantial socioeconomic effect on the US population due to the young age of patients at diagnosis and the high effect of disease on patients’ well-being.

Existing guidelines recommend treatments such as high-dose corticosteroids combined with either mycophenolate mofetil or cyclophosphamide (¹⁰,¹¹), which have a relatively low response rate among patients with lupus nephritis (¹²). Recently, the Food and Drug Administration (FDA) approved two drugs for lupus nephritis: belimumab, an intravenous B-lymphocyte inhibitor, and voclosporin, an oral calcineurin inhibitor.

In the Belimumab International Study in Lupus Nephritis (BLISS-LN) trial, belimumab increased the complete kidney response at 2 years compared with standard therapy alone, with benefits seen after the first year appearing stable at year 2. In the Aurinia Renal Response in Active Lupus With Voclosporin (AURORA) trial, voclosporin also increased the complete kidney response and the partial kidney response at 1 year compared with standard therapy alone (Table 1). Although the preliminary clinical evidence supports the clinical benefits of both treatments (¹³, ¹⁴), their cost-effectiveness has not been established.

Table 1. - Trial characteristics, outcomes, and associated assumptions for belimumab and voclosporin

Trial Characteristic	Belimumab	Voclosporin
Informing studies	BLISS-LN	AURA-LV, AURORA
Study duration	104 wk	52 wk
Inclusion/exclusion criteria	Patients 18+ with autoantibody-positive SLE that fulfills the 1982 ACR criteria, UPCR≥1 mg/mg, and biopsy-proven LN of class 2, 4, or 5 showing active lesions or active and chronic lesions in biopsy	SLE according to the ACR criteria, kidney biopsy within 6 mo of study entry confirming diagnosis of LN class 3, 4, or 5 (alone or in combination with class 3 or 4), proteinuria of ≥1.5 or ≥2 mg/mg for patients with class 5
Standard care therapies	MMF/corticosteroids or cyclophosphamide/corticosteroids	MMF/corticosteroids
Intervention therapy	Above plus belimumab 10 mg/kg every 2 wk ×3 doses, then every 4 wk	Above plus voclosporin 23.7 mg twice daily
Complete response end point definition	UPCR of <0.5 mg/mg, eGFR no worse than 10% below preflare value or ≥90 ml/min per 1.73 m2 with no use of rescue therapy	UPCR of ≤0.5 mg/mg, eGFR≥60 ml/min per 1.73 m2, or no confirmed decrease from baseline in eGFR of >20% with the presence of sustained, low-dose steroids and no administration of rescue medication
Partial response end point definition	eGFR≥10% below baseline value or >60 ml/min per 1.73 m2 and ≥50% decrease in UPCR with one of the following: UPCR<1.0 mg/mg if baseline ratio was ≤3.0 mg/mg or UPCR of <3.0 mg/mg if baseline ratio was >3.0 mg/mg, no treatment failure, and not complete kidney response	≥50% decrease in UPCR from baseline with the presence of sustained, low-dose steroids and no administration of rescue medication
Recruitment eGFR, ml/min per 1.73 m2, mean (SD)
Belimumab or voclosporin	100.0 (37.7)	92.1 (30.6)
Standard care	101.0 (42.7)	90.4 (92.0)
Recruitment PCR, mg/mg, mean (SD)
Belimumab or voclosporin	3.2 (2.7)	4.14 (2.71)
Standard care	101.0 (42.7)	3.87 (2.36)
Complete response rate
Belimumab or voclosporin	30.0	43.2
Standard care	19.7	23.0
Partial response rate
Belimumab or voclosporin	17.5	26.6 a
Standard care	17.0	28.7
Intervention-specific model assumptions	Between years 2 and 3 in the short-term model, patients remain in same response states	Between years 1 and 3 in the short-term model, patients remain in same response states

BLISS-LN, Belimumab International Study in Lupus Nephritis; AURA-LV, Aurinia Urinary Protection Reduction Active–Lupus with Voclosporin; AURORA, Aurinia Renal Response in Active Lupus With Voclosporin; ACR, American College of Rheumatology; UPCR, urine protein-creatinine ratio; UACR, urine albumin-creatinine ratio; LN, lupus nephritis; MMF, mycophenolate mofitil.

^aProportions not in complete response are presented.

Although cost-effectiveness models comparing alternative drug regimens for lupus nephritis have been developed for a range of health care settings over the last decade (^1516–17), there are no cost-effectiveness studies of belimumab or voclosporin in lupus nephritis. To assist decision/policy makers and payers in understanding the cost-effectiveness and health gains associated with these therapies, this study aims to estimate the cost-effectiveness of these drugs in people with lupus nephritis from US health care sector and societal perspectives.

Materials and Methods

A de novo economic model was developed in Microsoft Excel to estimate the cost-effectiveness of belimumab and voclosporin for a cohort of patients with lupus nephritis, with each drug compared with the standard of care as represented by the comparator arm in its own pivotal trial(s), from the US health care sector perspective. Costs, life years, and quality-adjusted life years (QALYs) were estimated using monthly time cycles and discounted at 3% per annum. This model was developed as part of the Institute for Clinical and Economic Review evaluation of belimumab and voclosporin for lupus nephritis, and input was sought from the manufacturers, patient groups, health economists, and clinical experts throughout the model development and analysis phase (Supplemental Material).

Model Overview

The model estimates the progression of lupus nephritis through the patients’ lifetime on the basis of response to treatment outcomes. It consists of two parts: (1) a short-term interpolation model concordant with data from the trials and (2) a long-term (lifetime) model on the basis of extrapolation using survival modeling.

The base case short-term model comprises 3 years for both belimumab and voclosporin. All patients start the short-term model in the active disease state and may transition to complete response, partial response, kidney failure, or death (Figure 1). The proportions of patients in the different health states during the trial follow-up are informed by key trials, the Aurinia Urinary Protection Reduction Active–Lupus with Voclosporin (AURA-LV) and AURORA for voclosporin and BLISS-LN for belimumab, with the longer-term transition to kidney failure and death from two cohort studies (⁹,¹⁸).

Figure 1.: Short-term model structure.

Key Assumptions

The key assumptions made during the modeling phase are described below, and a comprehensive list of assumptions and accompanying rationales is available in Supplemental Material. Both belimumab and voclosporin would be added on to standard therapy. Because of differences in the definitions in trials outcomes (¹³,¹⁹), belimumab and voclosporin were not compared directly with each other but with their respective control arms.

Data from the trials and studies were used directly in the short-term model, and the response rates achieved at the end of clinical trial follow-up (2 years for belimumab and 1 year for voclosporin) were assumed to be sustained until the end of the 3-year short-term model. The long-term model was informed by estimates of long-term kidney failure–free survival conditional on whether the patients achieved response in the short-term model or not.

Treatment duration was assumed to be for 3 years, except for those discontinuing due to adverse events or lack of response, which was assumed to occur at 18 months.

Model Inputs

Short-Term Model.

Table 1 presents the proportions of patients reaching response, kidney failure, or death at the end of the follow-up in the published clinical trials. The data on the complete response for voclosporin were on the basis of a random effects meta-analysis of the AURA-LV and AURORA trials (each with 1 year of follow-up) using the Mantel–Haenszel method, with the resulting risk ratio applied to the average control arm effect across the two trials (Supplemental Material). As the clinical data are only reported at specific follow-up points, the proportions of patients in interim time cycles in the short-term model were estimated by applying linear interpolation to the efficacy data in Table 1.

Treatment Duration.

On the basis of expert clinical input, we assumed that belimumab and voclosporin will be used in patients in complete and partial response states for 3 years before discontinuation (except for those stopping treatment at 18 months due to adverse events on the basis of the adverse event rates reported in their respective trials: 13% of patients taking belimumab and 11% of patients taking voclosporin). For patients remaining in the active disease state, it is assumed that both drugs would be used for 18 months before treatment was discontinued due to nonresponse.

Long-Term Extrapolation.

The long-term model used partitioned survival modeling to estimate kidney failure–free survival for the different health states (active disease and partial and complete response) on the basis of data from Davidson et al. (¹⁸), with the proportion of kidney failure events and deaths estimated on the basis of data from Chen et al. (⁹). The choice of sources in the long-term model was on the basis of combined criteria of recent data, quality of reporting, and representativeness of the US lupus nephritis population. Davidson et al. (¹⁸) included 53% Black patients and reported similar kidney failure–free survival across the partial and complete response states. The base case model applied the same risk of kidney failure to partial and complete response states but different costs and utilities to patients while they were in these states. A scenario analysis was conducted using superior kidney failure–free survival in complete response state (calculated hazard ratio 0.94) relative to partial response state on the basis of differences in outcomes at 5 years using modified Aspreva Lupus Management Study criteria definitions from the same publication.

The survival curves used in the base case analysis for long-term extrapolation are presented in Figure 2, with further detail provided in Supplemental Material. The face validity of the survival curves was confirmed by clinical experts, and the values obtained from them were as follows. In the complete and partial response health states, the mean kidney failure–free survival is 19.4 years, and the mean overall survival is 28.1 years; in the active disease health state, the mean kidney failure–free survival is 13 years, and the mean overall survival is 23.7 years.

Figure 2.: Survival curves used in the long-term extrapolation model.

Treatment, Health State, and Nonmedical Costs.

Costs and their sources are summarized in Table 2 and discussed in detail in Supplemental Material. The treatment costs for belimumab were estimated using the average sales price for intravenous (base case) and subcutaneous preparations (scenario analysis) using the standard discount in the Federal Supply Schedule (6%). For intravenous belimumab, the mean dose was estimated using the dose of 10 mg/kg body weight and the distribution of the body weights of the lupus nephritis population retrieved from the literature (¹⁴), and it accounted for vial wastage and the induction dosing regimen. In the scenario analysis for subcutaneous belimumab, the regimen of one injection (200 mg) per week was considered.

Table 2. - Key model inputs

Parameter	Input	Source
Health-related quality of life utilities in model states
Utility in complete response health state	0.80	Mohara et al. (16), Bexelius et al. (27)
Utility in partial response health state	0.71
Utility in active disease health state	0.62
Utility in kidney failure health state	0.55
Steroid-related utility increase
Increment in utilities for low-dose steroids	Utility value +0.025	Cooper et al. (32)
Increment in utilities for treatments with no steroids	Utility value +0.090
Drug costs
Belimumab cost in first month, iv form	$9811 a	ASP, WAC, FSS (33 34–35–)
Monthly cost of belimumab (months 2–36), iv form	$3560 a	ASP, WAC, FSS (33 34–35–)
Belimumab cost, subcutaneous forms	$3246	ASP, WAC, FSS (33 34–35–)
Monthly cost of voclosporin	$7686	Data from Aurinia and assumed discount of 23% (20)
Health care costs by model states
Annual cost in complete response health state	$7871	Bartels-Peculis et al. (21), Li et al. (23), Barber et al. (36), Hanly et al. (37), Medicare (data provided by the Lupus Research Alliance)
Annual cost in partial response health state	$8185
Annual cost in active disease health state	$42,510
Annual cost in kidney failure health state	$120,920
Steroid-related cost reduction compared with high-dose steroid use
Annual cost reduction with low-dose steroids	$84.5	Red book (33)
Annual cost reduction with no steroids	$126.8	Red book (33)
Nonmedical costs by health states
Annual cost in complete response health state	$5140	Cloutier et al. (24), Garris et al. (25), US Bureau of Labor Statistics (26)
Annual cost in partial response health state	$5140
Annual cost in active disease health state	$14,777
Annual cost in kidney failure health state	$24,157

iv, intravenous; ASP, average sales price; WAC, wholesale acquisition cost; FSS, Federal Supply Schedule.

^aOn the basis of FSS as of November 7, 2020.

Voclosporin costs were assessed considering the average daily dose of 39.1 mg (mean dose weighted to the duration of patients in the AURORA trial) and the price per “wallet” (containing 60 capsules of 7.9 mg each) of $3950 reported by the manufacturer. The mean discount of 23% was then applied to calculate the estimated net price for voclosporin (²⁰).

The mean all-cause health care costs per patient with lupus nephritis per year were reported as $45,469 in 2018 by Bartels-Peculis et al. (²¹) on the basis of data on 1039 patients with lupus nephritis (median age, 47 years; 83% women). The costs for active disease, partial response, and complete response were then estimated using cost ratios between the different health states and proportions of populations in each state, estimated from published literature (^2122–23). The costs in the kidney failure state were calculated as costs of people with lupus nephritis eligible for Medicare in 2016 on the basis of kidney failure alone or in combination with disability.

In the modified societal perspective analysis, indirect costs were also considered, which included costs of unemployment, absenteeism (temporary productivity loss), and caregiving (^2425–26).

Quality of Life.

The health-related quality of life utility values for the health states were obtained from published literature, incorporating feedback from clinical experts and patients, as quality of life was not reported in the informing trials. The model assumed that utility values in the complete response state were equal to utility values of the population of patients with SLE who have very low disease activity on the basis of data from a cohort of Swedish patients with SLE (²⁷). We estimated the utility values for patients in the partial response, active disease, and kidney failure states by applying EuroQoL-5 dimension utility decrements compared with the complete response state on the basis of a cost-utility analysis from Thailand (¹⁶). In the model, all utilities were capped at the general population utility for that age group (see Supplemental Material for details) to ensure they did not exceed the utilities of the general population.

For patients who have therapy with low-dose steroids (≤5 mg/d) or no steroids (≤2.5 mg/d), we applied a positive increment in utilities and a reduction in costs compared with patients on high-dose steroids to the proportion of patients in the active disease, complete response, and partial response states reported in corresponding steroid use categories in the AURORA and BLISS-LN trials (Supplemental Material).

Base Case and Scenario Analyses

Base case results were estimated from the probabilistic analysis, which was performed by jointly varying all model parameters using 1000 simulation runs. Because of the lack of data, the distributions used for costs and utilities in the probabilistic analysis were mean values ±10%.

Deterministic one-way sensitivity analyses were performed using plausible ranges on the basis of published data and expert opinion to identify the key drivers of model outcomes. For drug costs, 20% variability was applied. We conducted deterministic scenario analyses using alternative utility estimates, health state costs, survival estimates, and treatment durations, and also from a societal perspective.

Results

Cost-Effectiveness of Belimumab

The base case results for belimumab from the health care sector perspective are presented in Table 3. Belimumab treatment resulted in higher treatment costs and life years as well as higher QALYs gained, resulting in an incremental cost-effectiveness ratio (ICER) of approximately $95,000 per QALY and $113,000 per life year gained. Belimumab has 69% probability to be cost effective at a willingness-to-pay threshold of $150,000 per QALY. This probability increased to 79% at a threshold of $200,000 per QALY and decreased to 52% at a threshold of $100,000 per QALY. The sensitivity of ICER for belimumab to the variation in individual model costs and utility inputs is shown in Figure 3A.

Table 3. - Results for belimumab compared with standard care

Treatment	Total Cost, $	Life Yr	Quality- Adjusted Life Yr	Incremental Cost-Effectiveness Ratio (Costs per Life Yr), $	Incremental Cost-Effectiveness Ratio (Costs per Quality- Adjusted Life Yr), $
Base case (probabilistic) analysis
Belimumab	934,663	17.92	11.70	112,461	95,269
Standard care	886,305	17.49	11.19
Societal perspective
Belimumab	1,126,351	17.86	11.67	83,933	66,103
Standard care	1,094,193	17.48	11.18
Increased duration of active disease prior to progressing to kidney failure in the long-term model
3 yr
Belimumab	935,194	17.86	11.35	116,951	108,245
Standard care	890,385	17.48	10.93
5 yr
Belimumab	941,026	17.86	10.99	120,412	138,501
Standard care	894,891	17.48	10.66
Scenario analysis with pricing for the subcutaneuous drug form
Belimumab	920,434	17.86	11.67	88,979	70,077
Standard care	886,343	17.48	11.18
Scenario analysis with lower kidney failure–free survival in partial response state
Belimumab	1,068,154	16.76	10.16	168,134	133,250
Standard care	1,009,049	16.40	9.72

All scenarios performed using deterministic analysis.

Figure 3.: One-way sensitivity analysis demonstrating how changes to model inputs influence the incremental cost-effectiveness ratio for belimumab and voclosporin. (A) Belimumab. (B) Voclosporin. AD, active disease; CR, complete response; KF, kidney failure; PR, partial response; QALY, quality-adjusted life year. *Lower input corresponds to higher incremental cost-effectiveness ratio and vice versa.

A number of scenario analyses were performed to identify the effect of alternative inputs and assumptions on the cost-effectiveness results (Table 3). The societal perspective (including the costs of unemployment, absenteeism, and caregiving) decreased ICER for belimumab to around $66,000 per QALY gained. The results from the scenario analysis where the duration of the active disease state among those patients who eventually progressed to kidney failure was increased from 1.21 years in the base case scenario to 3 and 5 years were ICERs above the threshold of $100,000 per QALY but below the threshold of $150,000 per QALY. Worse kidney failure–free survival in partial response compared with complete response increased ICER to $133,250, and the lower cost associated with the subcutaneous form of belimumab reduced ICER to $70,077.

Cost-Effectiveness of Voclosporin

The base case results for voclosporin, using the health care sector perspective, demonstrated that voclosporin treatment results in higher costs and higher QALY gained compared with the standard care, with an ICER of $150,000 per QALY gained and $172,000 per life year gained (Table 4). Voclosporin has 49% probability to be cost effective at a willingness-to-pay threshold of $150,000 per QALY. This probability was 79% at a threshold of $200,000 per QALY and 11% if the threshold is $100,000 per QALY. The sensitivity of ICER for voclosporin to variation in individual model costs and utility inputs is shown in Figure 3B.

Table 4. - Results for voclosporin compared with standard care

Treatment	Total Cost, $	Life Yr	Quality- Adjusted Life Yr	Incremental Cost-Effectiveness Ratio (Costs per Life Yr), $	Incremental Cost-Effectiveness Ratio (Costs per Quality- Adjusted Life Yr), $
Base case (probabilistic) analysis
Voclosporin	928,107	18.42	12.64	171,927	150,334
Standard care	783,688	17.58	11.68
Scenario analyses
Societal perspective
Voclosporin	1,095,833	18.41	12.64	154,055	131,962
Standard care	968,460	17.58	11.67
Drop in kidney function within long-term time frame
Voclosporin	1,004,382	17.47	11.65	281,114	237,389
Standard care	840,567	16.88	10.96
Treatment discontinuation in nonresponse patients in 12 mo
Voclosporin	917,523	18.41	12.64	160,990	137,903
Standard care	784,416	17.58	11.68
Treatment discontinuation in patients with adverse events at midpoint in the AURORA trial (6 mo)
Voclosporin	921,463	18.41	12.64	165,755	141,984
Standard care	784,416	17.58	11.68
Scenario on the basis of efficacy data from AURORA trial only
Voclosporin	928,684	18.41	12.61	174,442	152,899
Standard care	784,454	17.58	11.66
Increased duration of active disease prior to progressing to kidney failure in the long-term model
3 yr
Voclosporin	936,174	18.41	12.17	176,661	173,127
Standard care	790,111	17.58	11.33
5 yr
Voclosporin	944,744	18.41	11.65	179,349	209,393
Standard care	796,459	17.58	10.94
Scenario analysis with lower kidney failure–free survival in partial response state
Voclosporin	1,162,891	16.73	10.36	184,444	154,934
Standard care	989,407	15.79	9.24

All scenarios performed using deterministic analysis. AURORA, Aurinia Renal Response in Active Lupus With Voclosporin.

A number of scenario analyses consider alternative modeling inputs and address uncertainty related to selected assumptions because of the limited and short-term data on voclosporin (Table 4). If a societal perspective is considered, voclosporin treatment may be considered cost effective with the threshold of $150,000 per QALY but not with the lower threshold of $100,000 per QALY. A scenario analysis assuming a drop in kidney function in the long term for patients on voclosporin (i.e., assuming a 3-year reduction in overall survival and a 5-year reduction in kidney failure–free survival in the complete and partial response states) increased ICER for voclosporin treatment to $237,000 per QALY. If patients who remain in the active disease state discontinue the treatment after 12 months of the therapy, ICER will decrease to $138,000 per QALY. Similarly, ICER remains above the threshold of $100,000 per QALY if the mean discontinuation time for patients experiencing adverse events was set at 6 months (i.e., the midpoint in the AURORA trial). The scenario on the basis of using the response rates from the AURORA trial (instead of the meta-analysis) and worse kidney failure–free survival in partial response compared with complete response did not change conclusions on cost-effectiveness for voclosporin treatment.

Discussion

With the demonstration of the clinical effectiveness of belimumab and voclosporin against standard of care (mycophenolate or cyclophosphamide) in their respective trials, this study evaluates the results from a cost-effectiveness model of belimumab and voclosporin for patients with lupus nephritis in the US setting. The incremental cost-effectiveness results were approximately $95,000 and $150,000 per QALY for belimumab and voclosporin, respectively, each compared with the standard of care in their respective trials. Our analyses suggest that there is less uncertainty in the cost-effectiveness of belimumab, whereas the cost-effectiveness of voclosporin is uncertain.

Our work builds on existing cost-utility analyses performed in a range of settings comparing therapies currently considered standard of care and utilized in the control arms of the trials of the two treatments evaluated in this study (^1516–17). Common across these analyses are the disease states of active disease, partial response, and complete response, although other model structures that reflect GFR, proteinuria, and their ability to predict outcomes are possible. From a US perspective, the model by Nee et al. (¹⁷) seems the most appropriate comparison, reporting the lifetime QALYs and costs for mycophenolate compared with azathioprine (inflated to 2019 values and without considering the reduction in mycophenolate cost) as 14.2–15.10 QALYs at a cost of $669,000–$677,000. The predictions of the costs for standard care between our model and the Nee et al. (¹⁷) model are comparable, whereas the difference in QALYs can be explained by the different approach to utility measurement applied by Nee et al. (¹⁷).

Stakeholder engagement prioritized the disutility associated with steroid therapy and the recognition that lupus nephritis differentially affects non-White racial groups in terms of incidence, outcomes, and access to therapies. It is likely that disease progression and treatment response will differ by race (²⁸,²⁹). In addition to under-representation of these groups in the relevant trials, making conclusions for these specific groups unreliable, the use of lower health-related quality of life associated with these patient groups (which is partially explained by lower socioeconomic status) could reduce access to these therapies (³⁰).

Our study is fortunate to benefit from the wider initiative to conduct longer trials in lupus nephritis, incorporating clinical effectiveness data directly from the trials and aligning this with real-world data to inform lifetime outcomes. Because data on the clinical effectiveness of the medications and disease progression by age, race, and other relevant factors were not available or uncertain, the cohort health economic model reports outcomes for an average lupus nephritis population similar to those in the informing clinical trials. Additional strengths include widespread stakeholder engagement through the model development process, which directed the study to additional information sources; the use of contemporary data; and utility estimation techniques, which yield plausible values. Despite our efforts, these analyses have important limitations, which include the lack of longer-term follow-up from the informing trials, broad health state definitions and assumed relationships between these and kidney failure–free and overall survival, and failure to capture any beneficial effects of these therapies on SLE beyond the kidney (e.g., fatigue, joint and skin involvement, cardiovascular disease, etc.). These uncertainties are reflected and explored in sensitivity and scenario analyses that evaluated different assumptions. On the basis of consultations with clinical experts, the model assumed continuation of treatments for up to 3 years, whereas the literature suggests that standard care in stable patients should continue out to 5–6 years. The real-world use of these newer therapies is yet to be observed (³¹), and more recent FDA labeling may result in patients discontinuing newer therapies because of the lack of clinical efficacy earlier than is currently considered in the model.

Further research in this area should include prospective data collection capturing longer-term outcomes, including the efficacy of the subcutaneous preparation of belimumab, rates of real-world use and discontinuation of these therapies, and methodologic consensus on how the adverse effects of steroid therapy in chronic disease should be reflected in economic evaluations. Robust utility data were lacking for the lupus nephritis population in the United States and may reveal greater value for these treatments than demonstrated in our analysis. Although these agents represent welcome additions to the therapeutic landscape for lupus nephritis, as the lupus community gains experience with these agents for this indication, additional data will be generated to inform more refined cost-effectiveness analyses. Meanwhile, given the uncertainty around its cost-effectiveness, decision makers and commissioners should consider price negotiations for voclosporin before considering reimbursement schemes for this product.

Disclosures

F. Agboola, S. Herron-Smith, and S.D. Pearson report employment with the Institute for Clinical and Economic Review. R.H. Chapman reports employment with the Innovation and Value Initiative and was employed by the Institute for Clinical and Economic Review at the time of this work. J. Fotheringham reports employment with the Sheffield Teaching Hospitals National Health Service Foundation Trust and the School of Health and Related Research, The University of Sheffield; consultancy agreements with Novartis; research funding from Novartis and Vifor Pharma; and honoraria from Fresenius, Novartis, and Vifor Pharma. O. Mandrik reports employment with The University of Sheffield and consultancy agreements with Imperial College London. S. Ren and M.D. Stevenson report employment with The University of Sheffield. P. Thokala reports employment with The University of Sheffield; consultancy fees from Abbvie, Daiichi Sankyo, GSK, Novartis, Novo Nordisk, Pfizer, and Roche; and research grants funded by Vifor Pharma but not related to belimumab, voclosporin, or lupus nephritis. J.A. Tice reports employment with the University of California, San Francisco and reports a contract between the Institute for Clinical and Economic Review and the author's university to perform technology assessments.

Funding

The study was funded by the Institute for Clinical and Economic Review (Boston, MA).