Celine Halioua
Celine Halioua

D.Phil in progress (est. Trinity 2020)

<DRAFT> Cost-Effectiveness and Healthcare System Adoption Challenges of Ocular Gene Therapy in the United States and United Kingdom

May 2019 | Written speeding on the Eurostar in France

Tl;dr My Ph.D research uses ocular gene therapy as a case study of the potential health economic benefits and drawbacks of one-time, expensive, potentially curative therapeutics in single- (e.g., United Kingdom) versus multi-payer (e.g., United States) healthcare systems.

This is important because gene therapy has the potential to better treat a number of severe diseases, however, its extremely high costs are a challenge for health providers and may restrict access to these therapeutics.

More broadly, gene therapy is an interesting case study to explore the various economic incentives of the United States versus United Kingdom healthcare system.


The first gene therapy for an inherited retinal dystrophy received market approval in late 2018 in the United States; multiple other gene therapies are in the clinical pipeline. Thus far, gene therapy has commanded prices in the range of $500,000 to over $1,000,000 for the one-time dose. To be utilized by healthcare systems, gene therapy will at minimum need to show clinical benefit in line with its increased costs. Before longitudinal patient studies are available, model-based estimations will be necessary to project the full clinical benefit of gene therapy. To investigate the potential cost-effectiveness of gene therapy for retinal dystrophies, I am building mathematical models of ocular diseases amenable to gene therapy.

The primary questions of my thesis are:

(i) does the benefit of gene therapy potentially justify its high price tag (is gene therapy cost-effective?)

hypothesizing that gene therapy is cost-effective...

(ii) will healthcare systems afford to pay for gene therapy (what is the budgetary impact of providing gene therapy to all eligible patients?)

then, hypothesizing the budgetary impact is significant and therefore it is potentially untenable to provide gene therapy to all patients who need it...

(iii) can we design novel financing strategies to facilitate the healthcare system adoption of gene therapies?

Background: Gene Therapy

Gene Therapy

Gene therapy is a treatment strategy where a gene product is delivered to a cell to alleviate a pathological condition or otherwise modify its expression phenotype for a therapeutic purpose. This is especially adept for monogenic recessive diseases, where the entire disease phenotype is accounted for by a single mutation. With few to no traditional interventions for these diseases displaying long- term efficacy, many patients with rare monogenic diseases are presently dependent on gene therapy- based interventions succeeding in the clinic and on the market.


Ocular Gene Therapy

The eye is in many ways an ideal tissue for gene therapy. It is generally accepted that the eye is immune-privileged, and therefore less likely to have adverse reactions to gene therapy vectors, which are often viral-derived. It is also simpler to ensure sufficient tissue penetrance of the gene therapy in the eye because the vector can be applied directly to the target tissue. There are multiple genetic ocular diseases which constitute a significant unmet medical need.

Current Market Status of Gene Therapies

Five gene therapies have received market approval in the United States (US) or European Union:

Commercial name | Scientific Name | Approval | Indication | Product | Public Price

  1. Strimvelis | Autologous CD34+ 2016 enriched cell fraction that (EMA) contains CD34+ cells transduced with retroviral vector that encodes for the human ADA cDNA sequence | 2016 (EMA) | Adenosine deaminase deficiency | Ex vivo modification of autologous bone marrow CD34+ cells to express ADA (Adenosine deaminase) | £594,000 ($825,862) for one-time series of injections

  2. Glybera | alipogene tiparvovec | 2012 (EMA) | Lipoprotein lipase deficiency | LPL (lipoprotein lipase) in AAV1 vector | Approx. $1.2 million for one-time dose

  3. Luxturna | voretigene neparvovec-rzyl | 2017 (FDA) 2018 (EMA) | RPE65 mediated retinitis pigmentosa | RPE65 (Retinal pigment epithelium- specific 65 kDa protein) in AAV2 vector | $850,000 for one-time treatment of both eyes

  4. Kymriah | tisagenlecleucel | 2017 (FDA) | B-cell acute lymphoblastic leukemia | Ex vivo modification of autologous T-cells to chimeric antigen receptor to target CD19 receptor | $475,000

  5. Yescarta | axicabtagene ciloleucel | 2017 (FDA) | Diffuse large B-cell lymphoma | Ex vivo modification of autologous T-cells to chimeric antigen receptor to target CD19 receptor | $373,000

Economic Challenges of Gene Therapies

Various factors including one-time use, small patient populations, complex manufacturing and materials, lack of competition, and long periods of research and development have led to gene therapy developers to assign very large price tags to their new therapies. The prices will likely lower as the manufacturing and research and development processes are streamlined, but it is reasonable to assume that gene therapy will always have a higher price than traditional medicines. Presumably, if gene therapy is to be much more expensive than other medicines, it should be similarly be much more beneficial than these old medicines, too!

Preventative Gene Therapy

Genetic diseases are either spontaneous (a mutation which randomly occurs in the development) While most genetic diseases are also pediatric, some genetic mutations only have significant deleterious effects later in life. For these indications, gene therapy can be considered a “preventative medicine”, which allows interesting comparisons to traditional preventatives such as vaccines. The retinopathy choroideremia is an example of an adult-onset genetic form of progressive blindness.

The use of gene therapy as a preventative medicine increases the complexity surrounding reimbursement and healthcare provider adoption: in addition to nearly unprecedentedly high price points, the therapy may not provide return in investment in terms of alleviated healthcare costs to the payer for decades after treatment. There would be time for other variables to play out, such as patient lifetime, unrelated disease, or in multi-payer jurisdictions such as the United States, the patient simply switching to another healthcare provider.

We believe this looming challenge has not yet been sufficiently quantified or addressed, likely in part since the current gene therapies on market are indicated for pediatric populations and oncologic indications, where immediate treatment benefit is inherent for any efficacious therapy. The pre-market approval consideration of the economic impact of novel therapeutics may become more important in the age of gene therapies and other such expensive therapeutics.

There are many more gene therapies in the pipeline

Compounding the above, there are many ocular gene therapies in the pipeline. It is import In fact, here’s a summary of all of the active ocular gene therapies in clinical development. Here is a non-summarized list. Ample credit to the fantastic Jasleen Jolly who created a database of ocular gene therapy clinical trials from which this list was based.


Choroideremia Gene Therapy

Assuming the price of gene therapy remains high, the first question to answer is whether the benefit of gene therapy justifies this high price tag using traditional therapy evaluation methods. I have built a Markov model describing the choroideremia-associated visual degeneration over the lifetime of the average patient. Using this model, I have estimated the cost-effectiveness of choroideremia gene therapy (adeno-associated virus encapsulating the REP1 gene product, AAV.REP1) across multiple hypothetical therapy price points, patient subgroups, and WTP per QALY thresholds.

For simplicity, “cost-effective” as used in this section will be defined as the gene therapy’s ICER falling within a WTP of up to $150,000 per QALY, with the understanding that there is no strict threshold valuation of QALYs in the United States and that this is only a proxy, albeit a commonly-used one.

The Biology of Choroideremia

Choroideremia is an X-linked monogenic retinal dystrophy. Similar to retinitis pigmentosa for which the approved therapeutic Luxturna is indicated, patients experience narrowing of visual field until blindness, but significant loss of vision and blindness occurs in the middle-to-late ages in comparison to RPE65-retinitis pigmentosa’s very early onset. AAV.REP1 is entering Phase III trials. Due to its mechanistic similarity to Luxturna which allows educated estimations where data are missing or sparse, the availability of preliminary clinical efficacy data, and the late onset of the disease, we propose AAV.REP1 as an adept case study to broadly investigate the economics of preventative gene therapy.

A simplified version of the model I built of choroideremia-associated visual degeneration. Health state is defined by the combined effect of remaining visual acuity and retinal area (visual field).


A Brief Introduction to Economic Evaluation Methodology


Markov Modeling

Markov models are a type of computational model which emulates a sequence through a series of probability-driven events. Markov modeling is commonly used in health economic evaluations of therapeutics. Here, the events are usually sequential health states and the probability of progressing to a health state is directly impacted by the therapeutic in question.

Quality-adjusted Life-years (QALYs)

Quality-adjusted life years (QALYs) are a commonly-utilized strategy for quantifying the detriment of a disease on a person’s health, quality of life, and survival. A value of 1 represents a full year lived in perfect health, 0 represents death, and negative values are sometimes used to described health states which are worse than death.


Willingness to pay per QALY (WTP per QALY)

In healthcare systems such as the United Kingdom’s National Health Service, therapeutic reimbursement decisions are determined in part by the society’s threshold willingness-to-pay per QALY. This is the maximum that a healthcare system is willing to reimburse per incremental QALY a therapy is expected to deliver over the current standard of care; it can also be viewed as that society’s valuation of one ‘quality’ year of life.

The US healthcare system does not adhere to a strict threshold, and indeed it is illegal for the government to use such thresholds in healthcare decision making. However, the use and valuation of the QALY persists in US health policy literature and provides a standardized benchmark for comparing interventions.

The UK uses a willingness-to-pay per QALY threshold is £20,000 to £30,000. In the US, $50,000 per QALY is commonly used; up to $300,000 per QALY has been quoted.


There is a time preference to values: people prefer benefit from their investment sooner rather than later. When conducting cost-effectiveness analyses, you can accommodate for this by discounting the costs and health benefits of the therapies being investigated. The equation for the discounted value is

PV = CV/(1+r)^t

where PV is the present value, CV is the non-discounted value, r is the discount rate in decimals, and t is the time over which the value is being discounted. Discounting de-values far away values more strongly than closer values.


In economic analyses, health benefit and costs are usually discounted by standard accepted rates. This is because, all other things equal, a healthcare system prefers that their investment (the cost of the therapy) delivers returns (costs savings or health benefits) sooner rather than later. 3.0% per year is the standard discount rate used by NICE on both costs and benefits, with 1.5% per year used for some subsets of long-term interventions.

Incremental Cost-Effectiveness (ICER)

The incremental cost-effectiveness ratio is determined as follows:

𝐼𝐶𝐸𝑅 = 𝜆 = (∆𝐶a,b)/(∆𝐸a,b)

The difference in costs (ΔCa,b)between the two therapeutics 𝑎 and 𝑏 is divided by their difference in efficacy (ΔEa,b) as measured in QALYs.

The ICER may be directly compared to a healthcare system’s demonstrated threshold. A higher ICER indicates that gene therapy has a higher additional cost over the comparator treatment per additional QALY it is expected to generate, and hence is less cost-effective than if it produced a lower ICER. When one of the alternatives is both cheaper and expected to lead to more QALYs than the other, then it is said to ‘dominate’ it and no ratio is calculated.