Annuity Payment

Annuity Payments for One-Off Therapies

High-cost therapies given once, such as gene therapies, ask a payer to fund in a single year a health gain that is expected to last for many years. The payer then has two linked choices, whether to pay the whole price at the point of treatment or defer part of it, and, if it defers, whether later payments should depend on the therapy continuing to work. The instalment follows from a standard formula, the schedule has to be entered correctly in both a cost-effectiveness model and a budget impact analysis, and the financing rate decides whether spreading raises or lowers the real price.

Why one-off therapies strain annual budgets

Health budgets are usually set year by year, and the costs of a chronic treatment arrive at roughly the same pace as its benefits. A one-off therapy breaks that link, because almost all of its cost falls at treatment, while the health gain and any savings from treatments no longer needed accrue over the rest of the patient's life. A therapy can therefore be cost-effective over a lifetime horizon and still be hard to fund from one year's budget, particularly when many existing patients are eligible at launch.

The upfront structure also concentrates risk on the payer. If the benefit fades after a few years, the payer has already paid for decades of effect that did not happen. Spreading payments, and making later payments depend on continued response, move part of that cost and risk into the period in which the benefit is observed.

Setting the instalment with the annuity formula

In a fixed annuity the instalment is chosen so that the stream of payments has the same present value, at an agreed financing rate, as the upfront price. When the first instalment is paid at the point of treatment, as in the set-up used by a recent simulation study of advanced therapies, the relevant form is an annuity due, with each payment at the start of a year. Rearranging the present value of an annuity due for the payment gives the following.

$$A = \frac{P \times i}{\left[1 - (1+i)^{-n}\right] \times (1+i)}$$

where $A$ is the instalment paid at the start of each year, $P$ is the upfront price that the schedule replaces, $i$ is the annual financing rate agreed between the parties, and $n$ is the number of instalments. When payments fall at the end of each year instead, the final division by $(1+i)$ is dropped; that ordinary annuity is the version used to annualise capital costs. As $i$ approaches zero the instalment approaches $P/n$, an equal share of the price.

The financing rate is set in the commercial negotiation rather than by HTA methods guidance. It reflects the cost to whoever finances the deferral, usually the manufacturer, of waiting for the money and of the risk that later payments are never made.

Fixed and outcomes-based annuities

Two kinds of scheme share the annuity label. In a fixed annuity the payer owes every instalment once the patient is treated, so spreading changes the timing of the cost but leaves the payer carrying the risk that the therapy fails. In an outcomes-based annuity each later instalment is due only if the patient still meets a pre-agreed response criterion on that date, and payments stop when the criterion is not met.

The outcomes-based form combines financing with performance-based risk sharing, and a modelling study of England's budget impact test classes this outcome-linked form as an outcomes-based managed entry agreement. The same study treats purely financial tools, such as loans to payers, as a separate category, because the manufacturer then bears none of the risk about long-term benefit; a fixed instalment plan with no outcome condition is closer to that financing category. The outcome-linked form needs an outcome that can be measured at each payment date, a data system that records it, and contract rules for patients who die, move to another payer or are lost to follow-up. The expected cost to the payer is then no longer the full price, but the sum of the instalments weighted by the probability that each one falls due, as follows.

$$E[PV] = \sum_{t=0}^{n-1} \frac{S_t \times A}{(1+r)^t}$$

where $E[PV]$ is the expected present value of the payments, $t$ is the number of years since treatment, $S_t$ is the probability that the instalment due at year $t$ is payable (with $S_0 = 1$ for the payment made at treatment), $A$ is the instalment, $r$ is the discount rate used in the evaluation and $n$ is the number of instalments.

How a payment schedule enters a cost-effectiveness model

In an economic evaluation the cost of the therapy should be the present value of what the payer is expected to pay, discounted at the evaluation's own rate, rather than the list price or the nominal sum of the instalments. The result then turns on how the financing rate in the instalment compares with the discount rate used by the HTA body.

If the financing rate equals the evaluation discount rate, the present value of a fixed schedule equals the upfront price and the incremental cost-effectiveness ratio does not change. If the financing rate is higher, the schedule costs more than paying upfront in present value terms. If it is lower, including an interest-free split, the schedule works as a hidden price reduction. For an outcomes-based schedule the response probabilities used to weight the payments should be the same ones that drive the modelled health benefit, so that a patient who stops responding stops generating both QALYs and payments.

The NICE manual (PMG36) sets a reference-case rate of 3.5% a year for costs and health effects. It allows the committee to consider 1.5% where the technology is for people who would otherwise die or have a very severely impaired life, is likely to restore them to full or near-full health, and has benefits likely to be sustained over a very long period. In those cases the manual adds that the committee must be satisfied that irrecoverable costs, including acquisition costs, have been captured in the model or mitigated through commercial arrangements. An outcomes-based instalment schedule is one way of limiting that exposure, because money not yet paid can be withheld if the benefit ends.

Worked example of a £1 million gene therapy paid over five years

The figures in this example are illustrative. A one-off therapy has an upfront price of £1,000,000. The manufacturer offers five annual instalments, the first at treatment, with a financing rate of 5%, and the evaluation discounts costs at 3.5% a year.

1. Calculate the instalment. The annuity due formula gives the following.

$$A = \frac{1000000 \times 0.05}{(1 - 1.05^{-5}) \times 1.05} = \frac{50000}{0.227298} = 219976$$

where 219,976 is the instalment in pounds paid at the start of each of the five years. The nominal total is 5 × £219,976, or £1,099,880, about 10% more than the upfront price.

2. Value the fixed schedule at the evaluation discount rate. The discount factors at 3.5% for years 0 to 4 are 1, 0.9662, 0.9335, 0.9019 and 0.8714, so the present value is as follows.

$$PV = 219976 \times (1 + 0.9662 + 0.9335 + 0.9019 + 0.8714) = 219976 \times 4.673 = 1027948$$

where $PV$ is the present value in pounds of the five fixed instalments. Because the financing rate of 5% is above the evaluation rate of 3.5%, the schedule costs about 2.8% more than paying upfront. An interest-free split of five payments of £200,000 would instead have a present value of £934,600 (200,000 × 4.673), about 6.5% below the upfront price.

3. Value an outcomes-based version. Suppose, for illustration, that the probability that the patient is alive and still meets the response criterion at each later payment date is 0.90, 0.85, 0.80 and 0.75. The expected present value is then as follows.

$$E[PV] = 219976 \times (1 + 0.90 \times 0.9662 + 0.85 \times 0.9335 + 0.80 \times 0.9019 + 0.75 \times 0.8714) = 219976 \times 4.0381 = 888285$$

where $E[PV]$ is the expected present value in pounds of the payments under the outcomes-based schedule. The expected cost is about 11% below the upfront price. That reduction reflects the payments avoided for patients whose benefit ends, and the same patients contribute fewer QALYs in the model, so it should not be read as a price discount.

4. Interpret. The same headline price of £1 million can enter the model at about £1.03 million, £0.93 million or £0.89 million, depending on the financing rate and on whether payments depend on response.

Budget impact and what spreading does and does not change

Spreading payments mainly changes the timing of spending, and over the long run it does not reduce it. A budget impact analysis shows this clearly when the number of patients treated each year is stable. Continuing the illustrative example with 20 patients treated every year, each new cohort adds its instalments to those still being paid by earlier cohorts, as follows.

The schedule lowers spending in the first four years, but from year 5 five cohorts are paying at once and annual spending is 10% higher than under upfront payment, which is the financing charge. Spreading helps most when a large group of existing patients is treated soon after launch and later inflow is small. A recent simulation study reached a similar conclusion, finding that annuity-based payments were most suitable when initial prevalence greatly exceeds annual incidence and the financial risk premium does not exceed the social discount rate.

England has a formal test of early-year spending. When the budget impact test was introduced in 2017, it applied where the net budget impact of a technology appraised by NICE was expected to exceed £20 million a year in any of its first three financial years in the NHS. Under the NHS commercial framework for new medicines, updated in January 2025, the threshold is now £40 million a year in any of the first three years of use, and exceeding it allows NHS England to enter commercial discussions with the company to manage affordability. A modelling study of the original £20 million test found that a full upfront payment could limit the number of patients treated, and concluded that annuity-based payments combined with outcomes-based payment could increase access without exceeding the threshold.

How payers and HTA bodies have approached instalments

Few instalment agreements have been described in primary sources that can be checked, and the examples below are limited to those. In May 2019 AveXis, a Novartis company, announced United States access arrangements for its spinal muscular atrophy gene therapy Zolgensma, with a wholesale acquisition cost of USD 2.125 million. It quoted an annualised cost of USD 425,000 a year for five years, which is the list price divided by five, without saying whether pay-over-time contracts carried a financing charge. A pay-over-time option of up to five years was arranged with a partner, Accredo, to ease short-term budget constraints, especially for states, small payers and self-insured employers, alongside outcomes-based agreements of up to five years in which, according to one health plan, a portion of the cost was placed at risk.

In England, NICE commissioned a mock appraisal, using a CAR T-cell therapy as the exemplar, to assess whether its methods and processes needed changing for regenerative medicines. The modelling study of the budget impact test reports that this NICE study highlighted payment at the time the health outcome occurs, such as per year a patient remains free of disease, as a potential way to address affordability, and that it explored annuity-based payments, described as leasing, as a way to reduce payer uncertainty. A later review adds that the mock appraisal indicated payers need clarity on how far the interest charged on deferred payments raises the price.

Annuity payments, outcomes-based rebates and capital annuitisation

Several nearby arrangements protect the payer in similar ways but differ in timing and in who finances the delay, so each enters the model and the budget differently. An outcomes-based rebate keeps the full upfront payment and returns money if the therapy fails, so it transfers performance risk without smoothing the budget. The Cell and Gene Therapy Access Model run by the Centers for Medicare and Medicaid Services in the United States is of this type and supports outcomes-based agreements for sickle cell disease gene therapies, with pricing discounts and outcomes-based rebates negotiated with manufacturers on behalf of state Medicaid agencies. Technology leasing, one of the options explored in the NICE mock appraisal, ties payment to a period of use, and the modelling study of the budget impact test describes it as a form of annuity-based payment.

The same formula has an older use in costing. In economic evaluation, the capital cost of equipment or buildings is often converted into an equivalent annual cost by annuitising it over the asset's useful life at the evaluation's discount rate. No instalments are paid in that case. An annuity payment for a therapy is a contractual schedule, and its rate is a financing term rather than the evaluation's time value of money assumption.

Practical barriers and common misreadings

A review of the literature on outcome-based spread payments for one-shot curative therapies found that the main barriers were agreeing financial terms within 12-month budget cycles and the possible conflict with international accounting rules. Correcting payments for outcomes was held back by the need for extra data collection, the lack of clear governance structures, and the resulting administrative cost.

Several errors recur in reports and models. One is to enter the nominal total of the instalments, or the list price when payments are conditional, instead of the expected present value. Another is to assume that spreading lowers lifetime spending, when with steady patient numbers it raises long-run annual spending by the financing charge. A third is to treat an outcomes-based schedule as resolving uncertainty about effectiveness. It moves part of the financial risk to the manufacturer but produces no comparative evidence unless the agreement also collects data that can answer the question.

Sources

• Centers for Medicare and Medicaid Services. Cell and Gene Therapy (CGT) Access Model. Web page, accessed 1 October 2026. https://www.cms.gov/priorities/innovation/innovation-models/cgt
• Drummond MF, Sculpher MJ, Claxton K, Stoddart GL, Torrance GW. Methods for the Economic Evaluation of Health Care Programmes. 4th ed. Oxford: Oxford University Press; 2015.
• Hettle R, Corbett M, Hinde S, Hodgson R, Jones-Diette J, Woolacott N, Palmer S. The assessment and appraisal of regenerative medicines and cell therapy products: an exploration of methods for review, economic evaluation and appraisal. Health Technology Assessment. 2017;21(7):1-204.
• Jørgensen J, Kefalas P. Annuity payments can increase patient access to innovative cell and gene therapies under England's net budget impact test. Journal of Market Access and Health Policy. 2017;5(1):1355203.
• Michelsen S, Nachi S, Van Dyck W, Simoens S, Huys I. Barriers and opportunities for implementation of outcome-based spread payments for high-cost, one-shot curative therapies. Frontiers in Pharmacology. 2020;11:594446.
• National Institute for Health and Care Excellence. NICE technology appraisal and highly specialised technologies guidance: the manual (PMG36). London: NICE; 2022, last updated 13 October 2026. Section 4.5.
• National Institute for Health and Care Excellence. Procedure: budget impact test and varying the timescale for mandatory funding. Technology Appraisal and Highly Specialised Technologies Programmes. London: NICE; 2017. Paragraph 2.4.
• NHS England. Budget Impact Test threshold consultation: summary of response (PRN01610iii). 29 January 2025.
• NHS England. NHS commercial framework for new medicines (PRN01610). Published 29 January 2025, updated 31 January 2025.
• Novartis. AveXis announces innovative Zolgensma gene therapy access programs for US payers and families. Media release, 24 May 2019.
• Van Dyck W, Michelsen S, Veredas D, Huys I, Luyten J, Simoens S. When do annuity-based payments help to address the affordability challenge of funding advanced therapies? Insights from a budget impact simulation. Journal of Market Access and Health Policy. 2026;14(2):23.