Skip to main content

Table 2 Genetic testing for neovascular AMD. Cost-utility analysis model parameters and assumptions

From: A Value-Based Medicine cost-utility analysis of genetic testing for neovascular macular degeneration

Phenotypic Features

 Identification of 5-year progression to advanced AMD (age-related macular degeneration) using the AREDS (Age-Related Eye Disease Study) Simplified Scale for AMD [7] Identified 67.8 % of progressors to atrophic AMD, but only 36.5 % of those progressing to neovascular AMD

Genotypic Features (from Yu et al. [4])

SNPs in the following genes

Gene [4]

Hazard ratio [4] for neovascular AMD

p-value

Notes

CFH—3 SNP

1.37

2.2 × 10−4

Higher risk

ARMS2/HTRA1—2 SNPs

1.27

5.3 × 10−3

Higher risk

C2

0.60

0.14

Uncertain status

C3

1.25

0.02

Higher risk

CFB

0.57*

0.02

Lower risk

LIPC

0.57*

0.04

Lower risk

CFI

0.87

0.10

Uncertain status

TIMP3

0.79

0.30

Uncertain status

CETP

1.17

0.06

Uncertain status

ABCA1

0.97

0.79

Uncertain status

COLSA1

1.2

0.21

Uncertain status

APOE—2 SNPs

0.97

0.85

Uncertain status

Genetic profiles

 High-risk = homozygous on all genetic loci for the alleles that increase the risk of advanced AMD

 Medium-risk = heterozygous on genetic loci for the alleles that increase/decrease the risk of advanced AMD

 Low-risk = homozygous on all genetic loci for alleles that decrease the risk of advanced AMD

 As per Yu et al. [4], all genes listed above are assumed to be tested for. This conservative assumption likely biases against the financial value gain and cost-effectiveness by increasing what will likely be decreased genetic costs in the future

Progression to neovascular AMD

 Incremental, 10-year progression rate to neovascular AMD for Category 3 AMD patients with: a past smoking history, BMI of 25–29, normal fellow eye and greater than a high school education [4]

Genetic profile

10-year progression

Percent of progressors detected

High-risk

26%

90%

Medium-risk

3%

9%

Low-risk

<1%

1%

 Among people with Category 3 AMD, 20 % will progress to neovascular AMD over 10 years [4]. (SNP = single nucleotide polymorphism, * = decreased chance of progression to neovascular AMD, BMI = basal metabolic index)

Neovascular AMD therapy

 Clinical features, MARINA study [18, 20, 23]

  All participants had minimally classic or occult, subfoveal choroidal neovascularization

  Baseline vision in the affected eye: 20/40–20/320

  Choroidal neovascular lesions <12 disc areas at baseline

  Mean baseline vision: 20/80-1 in the ranibizumab treatment and sham therapy cohorts

  Mean baseline age: 75 years

  Participants randomized 1:1:1 to: (1) 0.5 mg intravitreal ranibizumab dose (n = 240), (2) 0.3 mg intravitreal ranibizumab dose (n = 238) or (3) sham injection cohort (n = 238)

  The 0.5 mg ranibizumab dose was more effective than the 0.3 mg dose, and was thus the dose approved by the Food & Drug Administration and used in the current analysis [18]

  Treatment cohort (0.5 mg ranibizumab) mean vision: MARINA Study for years 1 and 2, then LOCF (last observation carried forward) of clinical trial data for years 3–12 [23]

  Eligible, MARINA sham cohort patients were treated with ranibizumab following the end of the randomized portion of the trial after 24 months. Thus, sham treatment, control cohort data utilized mean vision in the MARINA Study [18] for years 1 and 2, and a Lineweaver–Burke plot meta-analysis control cohort from six randomized, neovascular AMD clinical trials for years 3–12 of sham therapy [25]

  Adverse event disutility QALYs, a total of 0.045 QALY, were used to calculate adverse event QALYs subtracted from total patient value gain

  The average participant received 22 × 0.05 cc intravitreal injections, given approximately monthly, over 2 years

  Mean life expectancy: 12 years for the control and ranibizumab study cohorts [49]

Value-Based Medicine©, base case, cost-utility analysis parameters for the AMD genetic screening, cost-utility analysis [18, 20, 22, 23]

 The patient value and financial value gains are those associated with genetic screening for neovascular AMD making possible the incremental earlier detection and earlier ranibizumab therapy for neovascular AMD

 Model timeline: 12 years = mean life expectancy for average neovascular AMD patient [18]

 Baseline vision in the early-treatment, ranibizumab therapy cohort was 20/40–20/80. Final vision outcome was 20/40–1 [20]

 20/40-1 vision in each eye equates with a utility of 0.789 [30, 39]

 Baseline vision in the late-treatment, ranibizumab therapy cohort was 20/160-20/320. Final vision outcome was 20/160+2.[20]

 20/160+2 vision in each eye equates with a utility of 0.0.658 [29, 30]

 Loss of vision in a first eye results in a utility loss of 0.0398 [35]

 The incremental cost-utility analysis per patient utilizes the incremental patient and incremental financial value gains associated with early-treatment ranibizumab therapy (baseline treatment vision 20/40–20/80) for neovascular AMD versus late-treatment ranibizumab therapy (baseline treatment vision of 20/160–20/320) [20]

 The costs of genetic screening were compared with the patient value gains and cost savings conferred by early-treatment ranibizumab therapy (versus late-treatment ranibizumab therapy) made possible by genetic screening

 The direct ophthalmic medical treatment costs were the same in the early-treatment, ranibizumab therapy and late-treatment, ranibizumab therapy cohorts, and therefore were not considered incremental costs

 Only Category 3 AMD eyes in patients who were 65 years of age were tested with genetic screening

 Category 3 AMD cases (drusen >125 μm) annually in the United States in a 65-year-old cohort = 944,400 [1–3]

 22.5% of baseline Category 3 cases have a high-risk genetic profile to develop neovascular AMD [4]

 The phenotypic appearance of Category 3 AMD determines the use of AREDS supplement therapy to decrease the incidence of progression to neovascular AMD [5]. Since the use of AREDS supplements in Categories 1 and 2 AMD has not yet been shown to reduce progression to neovascular AMD [5]. Genetic screening was not presumed to be of benefit to detect whether to use AREDS supplements at an earlier stage than Category 3 AMD

 Genetic testing of Category 4 AMD patients for the development of more severe atrophic changes was not presumed to be of benefit

 Genetic testing was not presumed to be of benefit if one eye was already affected by neovascular AMD or advanced atrophic AMD

 Patients underwent genetic screening at age 65, since only 5.6 % of neovascular AMD develops in patients under the age of 65 years [18]

 Baseline time: First presentation for neovascular AMD occurs at a mean age of 75 years, as per a combination of multiple clinical trials dealing with therapy for neovascular AMD [18–20, 23, 24, 50, 51]

 The outcomes included the QALY (quality-adjusted life-year) gain, percent patient value (quality-of-life) gain, and the CUR (cost-utility ratio), or dollars expended per QALY gained ($/QALY) [22]

 All eyes with neovascular AMD were presumed treated with ranibizumab, including cases presenting with bilateral disease

 Time tradeoff utilities were derived from a database of over 1100 ophthalmic patients with respective levels of vision loss [22–24, 30–38]

 Cost perspectives: societal and 3rd party insurer

 The societal cost perspective included those saved by the better mean vision outcome associated with early-treatment ranibizumab therapy versus late-treatment ranibizumab therapy. They include: (1) direct ophthalmic medical costs = AMD genetic testing costs + incremental annual ophthalmic examination and annual optical coherence tomography costs, (2) direct non-ophthalmic medical costs saved = decreased depression costs, decreased trauma costs, decreased Skilled Nursing Facility costs, decreased nursing costs and other, as yet unidentified, medical costs [40]. (3) direct non-medical costs (caregiver) saved [41] and (4) indirect medical (employment) costs saved [3, 42]

 The 3rd party insurer cost perspective includes: (1) direct ophthalmic medical costs expended and (2) direct non-ophthalmic medical costs saved

 Cost basis: 2012, average, national, Medicare Fee Schedul

 Net present value (NPV) analysis discounted patient value outcomes and costs at a 3 % annual rate. All costs were converted to 2012 US real dollars [22]

 A combined-eye model, a weighted average of first-eye and second-eye models, was utilized to calculate QALY gain per early-treatment case over late-treatment case [23, 24]

 Taking into account the annual conversion rate, the QALY gain and financial value gain accrual rate of 1st eyes with NVAMD was 85.3 % that of 2nd eyes

 Second eyes in the per patient, early treatment benefit from ranibizumab were assumed to have the same visual outcome as treated first eyes

 For the overall Category 3 cohort undergoing genetic testing, the early-treatment ranibizumab QALY gain and costs per person were multiplied by: (Percent of first eyes with presenting vision ≤20/160, or 78.0 %) × (Percent of second eyes with presenting vision ≤20/160 = 62.2 %) × (Sensitivity of genetic testing with phenotypic features for detecting 10-year conversion to NVAMD = 90 %). Thus, overall Category 3 cohort results average (78.0 % × 62.2 % × 90 % =) 43.7 % × (per patient value and financial value outcomes)

 Conversion rates of second eyes to neovaoscular AMD with Markov modeling demonstrated the patient value (QALY gain)

 It was assumed that the majority of patients converted to neovascular in one eye first and thus treated with ranibizumab initially in this first eye, though a small number of cases might have present with bilateral neovascular AMD. This assumption is conservative and biases against the analysis by decreasing the overall patient and financial gains

  1. AMD age-related macular degeneration, NVAMD neovascular age related macular degeneration, QALY quality-adjusted life-year, CUR cost-utility ratio, MARINA Minimally classic/occult trial of the Anti-VEGF antibody Ranibizumab In the treatment of Neovascular Age-Related Macular Degeneration, SNP single nucleotide polymorphism