Making clinical development more investable with “financially adaptive trials”
Want to get smarter about biotech clinical trials, R&D strategy, and financial modeling? Join the thousands of biopharma execs and professionals who subscribe to our free email newsletter. Read a sample issue, and then sign up here.
Why are clinical trials of new drugs such a terrible investment? Let us count the ways:
Drug trials are risky, expensive, and time-consuming – making them unattractive for all but the most clinically savvy and deep-pocketed investors
Cash-strapped biotechs often design small, poorly-powered trials to conserve funds – thus increasing the risk of failure
Many senior pharma execs are forced to choose which trials to fund based on their apparent “riskiness” – thus starving many potentially impactful projects
But there is, in fact, another way: "financially adaptive trials", which transparently link trial design to financial analytics. "Financially adaptive trials" are based on three basic premises:
Trial design is intimately linked to cost, time, and risk – which together define the trial’s “investability”
There is not a single “best” trial – instead, one can use features of adaptive trial design (such as interim analyses and sample size re-estimation) to balance these metrics in different ways
Linking trial design and financial analytics can help companies and investors make transparent trade-offs between cost, time, and risk – which could make clinical development a more attractive investment
I’ve summarized below two examples that my colleagues* and I have developed to illustrate the potential applications of this approach: a biotech seeking external financing, and a large pharma optimizing an R&D program.
* * *
Example #1: Managing risk and cost for a clinical-stage biotech
Sunesis’s VALOR trial illustrates how linking trial design to financial analytics can open up new investment opportunities. (See here for more detailed background.)
Sunesis was struggling to balance cost and risk in its pivotal study of vosaroxin in acute myelogenous leukemia (AML). In the optimistic case, vosaroxin would have such a large clinical benefit that only a small sample would be needed. However, this lean trial would be under-powered to detect a smaller (but still clinically meaningful) effect. How, with limited resources, could Sunesis avoid “overspending” on the trial, while not missing a clinically important outcome?
By working closely with Royalty Pharma, Sunesis was able to structure a “financially adaptive trial” that linked the trial design and outcomes to the deal structure. Sunesis and Royalty Pharma jointly designed a smaller study that incorporated an interim analysis to guide further investment.
The biotech and the investor agreed on pre-specified rules linking the data at the interim analysis to the deal terms (see figure below). These terms gave Sunesis the ability to run the smaller trial with a commitment for extra funds if it was necessary to increase the sample size. In return, Royalty Pharma got the opportunity to make its investment dependent on the de-risking of vosaroxin with the interim data.
Importantly, Sunesis and Royalty Pharma collaborated on both the trial design and the deal structure – as opposed to the traditional investment model, in which the parties work closely on deal terms, but not (typically) on the trial design.
But the extra complexity and effort paid off for both parties. Sunesis limited its near-term R&D spend without sacrificing its probability of success, while Royalty Pharma got the chance to invest in a program that otherwise would have been too risky for its portfolio. (In fact, the interim analysis fell in the "promising zone", triggering the expansion of the sample size and the milestone payment.)
Example #2: Defining clinical development trade-offs for a large pharma
What should pharma execs do with “borderline” assets that are scientifically and clinically promising, yet too risky, low-revenue, and/or expensive? Often these programs are killed outright in favor of others, or they “die by starvation” through meager R&D investments that waste time, money, and future value. But is there a better, more flexible way to fund and develop these assets?
To explore this question, we worked with a large pharma’s R&D group to model a prototypical Ph2-ready asset in a niche cancer indication. The “base case” plan was long and relatively risky, yielding an eNPV of less than $6M for a $47.4M investment – not likely to be a winning combination to senior management!
We designed and analyzed two alternatives that could be more attractive to corporate execs. In one, we tried to maximize the program’s NPV. In the other, we sought instead to reduce the time before we could re-assess the program based on new clinical data. For both situations, we iteratively refined and analyzed the clinical plan to optimize the key metrics.
In both cases, we were able to generate better options than the “base case” development plan. In the first scenario, we boosted the eNPV almost nine-fold (to over $40M) with a trial that was larger, yet faster and overall more likely to succeed. The trial in the second scenario, optimized for speed to a first “get-out” decision point, generated actionable clinical data more than three years earlier than the original design.
Note that "average time" is calculated from the trial simulations, and thus does not agree exactly with the times in the schematic in the prior figure.
This example demonstrates how “financially adaptive trials” can bridge the gap between clinical and financial decision-makers in pharma portfolio management. Although franchise heads and corporate execs often evaluate financial analyses of proposed development plans, they are rarely presented a suite of design options for a single program, and an integrated analysis of the financial and strategic implications of optimizing specific variables.
“Financially adaptive trials” could enable management, R&D, and finance to openly discuss which variables (cost, time, risk) are most important to optimize for a particular asset, and the value (and trade-offs) of optimizing them. And, if applied systematically, this approach could allow companies to better balance these factors across the entire R&D portfolio.
* * *
At their core, “financially adaptive trials” in biopharma transparently link trial design and financial analytics. This enables R&D and financial stakeholders to iteratively and collaboratively define the “best” trial for a particular situation and set of priorities – whether they are in the same company (like a large pharma) or not (like a biotech and an investor). All of the key trial design and analytic elements are currently accepted in both R&D and finance – this approach provides a way to bring them together.
There are bound to be implementation challenges of implementing “financially adaptive trials” – but the hurdles appear surmountable. Although regulators increasingly support adaptive designs, many pharma R&D leaders and senior execs are not yet aware of this fact, and education is clearly needed. In addition, this approach requires the key stakeholders – a biotech and an investor, or management and R&D within a large pharma – to agree on the definition and interpretation of “probability of success”, which combines data-driven and subjective components.
But overall, “financially adaptive trials” increase not only the breadth of R&D investment options for small and large companies, but also the transparency by which they are analyzed and presented. This approach would significantly improve the quality of R&D and investment decision-making across the drug development industry – which could only be a good thing.
* I've had the good fortune to collaborate on this work with Nitin Patel and Zoran Antonijevic of Cytel, an adaptive trials consultancy, and Kraig Schulz and Sarah Bobulsky of Ernst & Young’s Transactional Advisory practice. For more details behind the information in this post, please review the presentation deck from the DIA webinar Nitin and I recently led, or contact us directly. You can also refer to a chapter on this topic in a book on pharma portfolio optimization.
I've also recently co-authored a paper in Nature Reviews Drug Discovery on this topic - see here. Note that "average time" reported in Fig. 4 above is distinct from "median time" reported in the NRDD paper.
Many thanks to John LaMattina and David Sable for helpful discussions during the preparation of this post.
Edit history: Amended footnote to account for book publication; made minor text edits for clarity (July 6, 2015). Added NRDD citation (July 24, 2015).