A Model to Explore the Potential Budget Impact of a Novel Screening Tool for the Detection of Subclinical Rejection among Kidney Transplant Patients

Disclosure: This project was funded by Immucor, Inc.

IMMUCOR20925 Crossroads CircleWaukesha, WI 53186

T262-290-8534 www.immucor.com

1. Avalere Health, LLC, Washington, DC2. Immucor, Inc. Waukesha, WI

Timothy J. Inocencio, PharmD, PhD2

Kevin Jaglinski, BA1

Hiroshi Uchida, PhD1

Kathleen E. Hughes, MBA2

Background & Objectives /

1 Acute rejection is associated with long-term effects. Previous studies have suggested that the occurrence of acute rejection and increasing frequency of acute rejection results in decreased long-term renal allograft survival among kidney transplant patients.2-4 Routine monitoring of kidney function includes monitoring for an increase in

detected after substantial damage has occurred.5

Detection of subclinical rejection (SCR) can give providers the opportunity to intervene earlier before the presence of

histological changes specific for acute rejection on screening or protocol biopsy, in the absence of clinical symptoms or signs.6 However, blood-based gene markers for kidney rejection may occur before histologic abnormalities are found through biopsy and offer an opportunity for earlier intervention and adjustment to immunosuppressive regimens.

The Kidney Solid Organ Response Test (kSORT) assay is a 17-gene set non-invasive molecular assay that measures blood-based gene markers for transplanted kidney rejection. In the recently published Acute Rejection in Renal Transplantation (AART) study, acute rejection was detected by kSORT up to 3 months before detection by biopsy7, giving providers an earlier opportunity to modify immunosuppression to prevent subsequent rejection. Previous data have shown that more frequent monitoring for SCR results in

8

The objective of this exploratory analysis is to evaluate the potential budget impact of the kSORT assay from a commercial payer perspective.

Methods /kSORT assay as a monitoring tool for patients who have undergone a renal transplant. The model projects the impact of adoption of this product over the course of two years. It considers costs associated with the kSORT assay in relation to the cost impact of improved detection, development of alternative strategies, and subsequent management of subclinical rejection (SCR). The model was built as an interactive spreadsheet using Microsoft Excel 2010.

Clinical and economic inputs and assumptions were based on information from the peer-reviewed and publicly available literature (see references) across a number of clinical scenarios. When published sources were not available, data and assumptions were based on the consensus of a multidisciplinary panel of experts. See Tables 1 and 2 for inputs around costs and probabilities.

These scenarios include the following: A. No monitoringB. Monitoring through protocol biopsy (PB) onlyC. Monitoring through kSORT assay onlyD. Monitoring through combination of PB and kSORT

The monitoring strategies and their assumed mixes before and after kSORT coverage are outlined below:

Assumptions1. This model assumes equal probabilities of disease progression for the pediatric and adult

populations.2. As a starting point, the base case scenario assumes equal diagnostic performance of PB and

the kSORT—a conservative assumption.3. Diagnosis of SCR results in a change in management by physicians to optimize drug regimens,

thereby modifying the effect of SCR on acute clinical rejection (ACR) and graft failure (GF).4. It is assumed that a diagnosis of SCR results in a 90 percent reduction in ACR.5. Diagnosis of SCR results in a 50% reduction in GF, independent of ACR.6. The model includes the costs of the tests, per their indications for use, only before ACR or

GF has occurred.7. The prevalence of acute SCR decreases through time9, and assumes that prevalence of SCR is

between approximately 15 to 30 percent.10 8. Active SCR results in an increase in the risk of ACR (probabilities are calibrated in model).10 9. The development of SCR and ACR (given SCR) are both time-dependent, and are associated

with decreased probabilities as time passes.10. The kSORT assay was assumed to be performed for incident kidney transplant patients during

months 1, 3, 6, 9, 12, 18 and 24 post-procedure, while PB was assumed to take place during months 3 and 9.

Cohort SizeThe size of the commercial plan is assumed to be 5 million, with a mix of 78% of adult and 22% pediatric (obtained through the Current Population Survey in 201311). The incidence of kidney transplant was estimated to be 70.3 per million for adults and 10.3 per million in the pediatric population (taken from the SRTR/OPTN 2012 report1, and adjusted using U.S. Census Data). For a plan of 5 million

ansplants incorporating the aforementioned incidence data was estimated to be 285 patients.

Model StructureA 2-year exploratory semi-Markov cohort model with monthly cycles was constructed to model the relationship between SCR, ACR, and GF. Four separate Markov models were created, each representing one of the four scenarios described above, while incorporating the months during which a patient receives the respective monitoring tool (i.e., kSORT, PB, or both).

Patients enter the model in the “No SCR” health state, and can either develop SCR or GF. Patients who are in the SCR health state can enter a “post-SCR” state where they no longer have active SCR, or they can experience ACR and enter a post-acute rejection state (PARS), after which patients may continue in that state or progress to GF (See Figure 1a). Patients may enter the death state at any time from any health state. For patients who have kSORT and/or PB monitoring, patients may transition from the “No SCR health state and instead may transition to a “Detected SCR” health state (See Figure 1b),

progression to these health states.

Transition ProbabilitiesDetailed data on transition probabilities, especially those related to SCR, were not available for this patient population. As such, data were calibrated to approximate actual rates of ACR and GF, obtained from the SRTR/OPTN registry data1, while ensuring that modelled rates of SCR were within the range of prevalence data obtained from the literature. Transition probabilities are presented in Table 1, while comparisons between the modelled and actual acute rejection and graft failure rates are shown in Figures 2, respectively.

Figure 1a / Markov Model Diagram — Base Model

Figure 1b / kSORT and PB Monitoring Cohorts—Monitoring Submodel

Results /In the base case scenario, kSORT is expected to produce a minimal budget impact of $0.0057 PMPM

in Table 3. Total costs for the plan during Year 1 and Year 2 are provided in Figure 3.

Table 2 / Cost Inputs*

Background Costs Source

Year 1 $2,171.39 SRTR/OPTN 2012 Report1

Year 2 $1,213.19

Marginal AR costs

Year 1 $2,148.00 Gheorghian et al.13

Year 2 $1,094.99

Marginal GF Costs

$6,644.36 Gheorghian et al.13

Drug Costs

Monthly Drug Costs for First 3 Months $2,373.80 SRTR/OPTN 2012 Report1 and Red Book AWP prices14Monthly Drug Costs After 3 Months $2,050.47

Percent Increase in Drug Costs for Diagnosed SCR 2% AssumptionPercent Decrease in Drug Costs for No SCR 2% Assumption

Protocol Biopsy Costs $3,878.00 Nankivell and Chapman10

kSORT Assay $1,500.00 Modeled Price**

Table 3 / Budget Impact Estimates for kSORT by Year

BUDGET IMPACTTotal Year 1 Year 2

Total Cost Difference $ 686,696.07 $ 340,282.41 $ 346,413.66 Per Member $ 0.1373 $ 0.0681 $ 0.0693 PMPY $ 0.0687 $ 0.0681 $ 0.0693 PMPM $ 0.0057 $ 0.0057 $ 0.0058

*The PARS health state represents state after ACR event has occurred** Patients may transition to the ‘death’ state from any health state.

Graft Failure

Detected SCR2

No SCR

PARS*

Death**

SCRNo

SCR1No

SCR2

Death**

PARS*

Graft Failure

Monitoring Submodel

Figure 3 / Costs With and Without kSORT

Test Costs ACR-Related Costs GF-Related Costs Rx Drug Costs Background Costs

$35$30$25$20$15$10

$5$0

Tota

l Cos

t (M

illio

ns)

Baseline (No kSORT)

With kSORT

Year 1

$17.0 $17.3

$30.5 $30.9

Year 2

Baseline (No kSORT)

With kSORT

Figure 2 / Comparisons of ACR and GF Rates Between SRTR/OPTN Rates and Modeled Rates

16141210

8

642

0

Perc

ent

Month0 5 10 15 20 25

SRTS/OPTN 2012 GF Rates

Modeled GF Rates: No Diagnosis

SRTS/OPTN 2012 ACR Rates

Modeled ACR Rates: No Diagnosis

MONITORING APPROACH BEFORE KSORT COVERAGE AFTER KSORT COVERAGE

Routine monitoring 60% 50%

PB only 40% 30%

kSORT only -- 10%

PB + kSORT -- 10%

Table 1 / Model Transition Probabilities

Parameter Value Parameter Value

Probability of Detecting SCR, Given SCR Probability of GFc

Protocol Biopsya 0.9 Months 0 to 3 0.01450kSORTa 0.9 Months 3 to 12 0.00120

Reduction in ACR, Given Dx of SCRa 90% Months 12 to 24 0.00001Probability of SCR Resolutiona 0.5 Increase in GF, Given SCRa 0%

Probability of SCRb Months 3 to 12 0.00624 Month 0 (Immediate) 0.2 Months 12 to 24 0.00005

Months 1 to 3 0.25 Reduction in GF, Given Dx of SCRa 0.5 Months 3 to 12 0.2 Monthly Probability of Deathc 0.002223

Months 12 to 24 0.15Probability of ACR, Given SCRc

Months 0 to 1 0.115 Months 1 to 3 0.017

Months 3 to 12 0.034 Months 12 to 24 0.01

Monthly Probability of Deathc 50%

SCR: subclinical rejection; ACR: acute clinical rejection; Dx: diagnosis; GF: graft failurea. Assumptionb. Calculated from Nankivell and Chapman10

c. Calibrated based on rates obtained from SRTS/OPTN 2012 Report1

d. Calculated from Meier-Kriesche et al.12

Sensitivity AnalysisBecause of the small number of individuals in the plan with kidney transplant, the overall budget impact

e model was most sensitive to kSORT costs (Figure 4).

Figure 4 / Tornado Diagram for One-Way Sensitivity Analysis Around Costs** and Transition Probabilities

*Background costs are fixed, and do not change according to the method used for surveillance**Costs were varied between +/- 25% of their base case value

0.002

kSORT Costs

Probability of ACR, given SCR (+/- 25%)

Probability of SCR (+/- 50%)

Probability of GF, Given ACR (+/- 25%)

Probability of SCR Resolution (0.3 to 0.7)Reduction in ACR, Given Diagnosis of SCR (0.4 to 0.95)

Marginal GF Costs

Reduction to ACR, Given SCR (0.2 TO 0.8)

Probability of Diagnosis of SCR: kSORT (0.8 to 1)

Marginal ACR Costs – Year 1Probability of Death (+/- 25%)

Drug costs – Post 3 Months

Marginal ACR Costs – Year 2Percent Increase in GF, Given SCR (0% to 10%)

Probability of Diagnosis of SCR: Protocol Biopsy (0.8 to 1)

Protocol Biopsy Costs

Drug costs – First 3 Months

Background Costs*

-0.001 -0.0005 0 0.0005 0.001 0.0015

Conclusion /This exploratory model indicates that the kSORT monitoring assay is expected to produce minimal

in AR or GF. The low budget impact is attributed to the small patient population within plans with renal transplants every year, in addition to relatively low acute rejection and graft failure rates. Although the budget impact is small, additional clinical data showing how kSORT improves patient outcomes are needed. Long-term data may be of particular import in showing true value of the technology.

References /1 Matas AJ, Smith JM, Skeans MA, et al. OPTN/SRTR 2012 Annual Data Report: kidney. Am J Transplant. 2014;14 Suppl 1:11-44.2 Wu J, Chen J, Wang Y, et al. Impact of acute rejection episodes on long-term renal allograft survival. Chin Med J (Engl).

2003;116(11):1741-1745.3 Emiroğlu R, Yagmurdur MC, Karakayali F, et al. Role of donor age and acute rejection episodes on long-term graft survival in cadaveric

kidney transplantations. Transplant Proc. 2005;37(7):2954-2956.4 Matas AJ, Gillingham KJ, Payne WD, Najarian JS. The impact of an acute rejection episode on long-term renal allograft survival (t1/2).

Transplantation. 1994;57(6):857-859.5 onic renal allograft injury in a randomized trial on steroid

avoidance in pediatric kidney transplantation. Am J Transplant. 2012;12(10):2730-2743.6 Kidney Disease: Improving Global Outcomes (KDIGO) Transplant Work Group. KDIGO clinical practice guideline for the care of kidney

transplant recipients. Am J Transplant. 2009;9 Suppl 3:S1-155.7 Roedder S, Sigdel T, Salomonis N, et al. The kSORT assay to detect renal transplant patients at high risk for acute rejection: results of the

multicenter AART study. PLoS Med. 2014;11(11):e1001759.8 J Am Soc Nephrol.

1998;9(11):2129-2134.9 Nankivell BJ, Borrows RJ, Fung CL, O’Connell PJ, Allen RD, Chapman JR. Natural history, risk factors, and impact of subclinical rejection in

kidney transplantation. Transplantation. 2004;78(2):242-249.10 Am J Transplant.

2006;6(9):2006-2012.11 United States Census Bureau. Current Population Survey. Table HI01. Health Insurance Coverage Status and Type of Coverage by Selected

Characteristics: 2013. http://www.census.gov/hhes/www/cpstables/032014/health/hi01_1.xls. Accessed April 28, 2015.12 Meier-Kriesche HU, Ojo AO, Hanson JA, et al. Increased impact of acute rejection on chronic allograft failure in recent era. Transplantation.

2000;70(7):1098-1100.13 Gheorghian A, Schnitzler MA, Axelrod DA, Kalsekar A, L’italien G, Lentine KL. The implications of acute rejection and reduced allograft

function on health care expenditures in contemporary US kidney transplantation. Transplantation. 2012;94(3):241-24914 Micromedex THA. Red Book. 2014; http://www.redbook.com/redbook/index.html. Accessed December 1, 2014.

Base Model

*Costs are reported in 2014 U.S. dollars, using the medical component of the Consumer Price Index (http://www.bls.gov/cpi/) to account for inflation**Price for the kSORT assay has not been finalized at the time of this publication; estimate represents the modeled price for purposes of this analysis