Pipeline Optimisation: No Pain, No Gain

One might describe the typical pharmaceutical pipeline as a cash sinkhole. A primordial factor is the high attrition rate of candidates through the development programme. But before the plug is officially pulled at the end of an unsuccessful phase, weak interim data can portend future failure, or the product may have, in fact, already failed in the eyes of regulators (and increasingly, pricing and access authorities). Whilst the first factor can be largely attributed to molecular notions that require a great deal of further in-human study to rectify, be that the poor translatability of current animal models or the heterogeneity of the patient population, the latter issue is very much culturally unique to the organisation. Hence, logic would dictate that this can be addressed much more swiftly – yet this is often not the case.

No One Wants to Hear They Have an Ugly Baby

One would think that, at the first hint of failure, companies would opt to terminate development programmes and refocus on other pipeline assets. However, it is easy to forget that pipeline projects are orchestrated by people. Clinical trials are rife with emotion, brought about by engaging with patients with desperate need for additional treatment options, as well as by devoting substantial amounts of time and effort to their planning.  Moving beyond professional investments, these individuals have personal livelihoods at stake, which they may fairly believe will be threatened by a programme’s discontinuation.

Thus, when data indicate that the asset will not meet primary endpoints or is likely to fail during the next phase, there may be an impulse to drive the trial to completion regardless, thereby incurring hefty additional costs in the process. This can be due to several reasons, not least the natural inclination to avoid failure, avoid loss of investment, and elude shareholder discontent. Furthermore, a pipeline is a balancing act between commercial and clinical teams, who both are often guided by different motives and incentives. Hence when dissonance arises between the two on how to best proceed, further inefficiencies may arise.

Pipeline Optimisation Is a Necessary Evil

Terminating a trial or aborting a project should not necessarily be regarded as a failure, particularly given that there is almost always a compelling clinical and / or commercial reason for doing so. Making a well-supported decision to terminate a programme is likely to yield alternate, and often unforeseen, future benefits. The costs of development incurred to date must not be merely disregarded as sunk costs, but as a lesson: An idea that was once worth pursuing is no longer an adequate sink for expenditure. Any future additional costs of proceeding with a futile candidate can now be spent elsewhere, possibly leading to superior return on investment. And whilst it is true that external stakeholder confidence in the company’s pipeline might falter following the announcement of a decision, often there will come soft rains.

For instance, ACTELION ($ATLN) represents a prime example of how fateful objective, unemotional decision-making can be for a company. As the company was founded by a group of former executives and friends from ROCHE ($ROG), two assets drove the company’s immediate and long-term fortunes: TRACLEER™ (bosentan) and VELETRI™ (tezosentan, at the time).  Observers of $ATLN will quickly point out that TRACLEER™ set the company on a course to rise from an already remarkable IPO worth CHF 246m in 2000 to an eventual, impressive acquisition by JOHNSON & JOHNSON ($JNJ) for CHF 30b in January 2017.  However, success did not come easy despite the success of TRACLEER™.  VELETRI™, the company’s acute heart failure candidate, exhibited excessive dose-related vasodilatory adverse events in its 2001 PHASE III trial, which led to a rather dramatic reduction of 62% in its market cap. Far be this from the end, ATLN regrouped by shelving VELETRI™ (temporarily, as the story goes), and channelling funds instead in two main directions: to ensure the TRACLEER™ launch was a success, as well as a series of acquisitions, partnerships, and launches illustrated in FIGURE 1 below.  These steps inform one of the greatest portfolio building and slimming stories in the industry over the past two decades, particularly when taking into account the company’s youth and size.  Allowing the company to recover financially and structurally, the effectual and emotional decision to halt VELETRI™’s development set a course for the company to eventually be acquired for a price 121 times its IPO earlier this year.

FIGURE 1. ATLN Investment Reprioritisation

Valuations come and go, and whilst market confidence is important, the market is fickle, and internal objectivity, decision-making, and ultimate confidence in one’s pipeline should prevail. This internal confidence must stem from knowing that the pipeline has been empirically evaluated and optimised for long-term success – even if this means that, at present, your baby is ugly.

Solving the Pipeline Problem

Pipeline optimisation must in some way quantify the level of opportunity whilst evaluating the risks and returns of each particular scenario.  Generally, development programmes should be initially inspected using a measure such as NPV (net present value).  If they yield a favourable rate of return, a more qualitative approach should be undertaken to accounts for the trade-offs within the selected scenarios.

Cash Flows and Clinical Trials

Cash flows are defined by time, size, and probability. The time value of money for therapeutics is no different to what it would be in any other industry, but pharmaceutical pipeline valuations are special in that only the cash flows of the present clinical trial phase are certain; thus future cash flows ought to account for the probability of completing the current phase and any subsequent phases.

During the R&D process, cash flows are essentially the costs of individual phases, often assumed to be evenly distributed across them. When approaching commercialisation, the company must also invest in sales force, product launch, and possibly facilities. After a launch, cash flows are mainly linked to sales and COGS. All cash flows are assumed to be negative except those relating to sales (or possibly licensing) revenue. As previously mentioned, future cash flows must account for the chance of the preceding phases succeeding. For instance, if in PHASE I, the cash flows of PHASE III should be multiplied by the probability of success in PHASE II and the probability of success in PHASE I.

The next item to discuss is the discount rate, which provides the time value of money and risk components of the NPV. The time value of money is accounted for by using the rate offered by state-issued, non-default securities such as T-bills, and a risk premium. The risk will encompass technical uncertainties in the pharmaceutical development process. The discount rate used will very much depend on the purpose of the valuation. A particular pipeline asset may warrant a higher discount rate due to its nature – for example, CNS therapies have an unusually high attrition rate of over 90% due to less translatable pre-clinical models.  Hence, a higher discount rate is appropriate, but portfolio managers may opt to use the traditional company discount rate uniformly so as to obtain an even valuation.

A positive NPV suggests that the development programme should be continued because the internal rate of return is larger than the discount rate. A negative NPV, on the other hand, suggests that a project should be discontinued because the investment does not yield the adequate return.

Positive Isn’t Always Possible

Maximising NPV may not always be the most adequate approach to pipeline optimisation. A good example of this is illustrated by the research carried out by Ding and Eliashberg at Wharton on a real company portfolio, the identity of which is less important than the conclusions it yields. We can see from the NPV-maximising approach in FIGURE 1 below that the additional expenditures incurred by increasing the number of development programmes may not be feasible or appropriate.

FIGURE 2. COMPANY X Portfolio Optimisation

This approach recommends a 767% increase in the total number of pipeline development programmes. Statistically, this increase maximises the number of therapies likely to be approved according to their respective probabilities of success, but the efforts required to do so are rather disjointed from reality.  The number of deals necessary to achieve this NPV-maximising approach is unrealistic both from financing and sourcing standpoints.

A more reasonable approach, illustrated in FIGURE 3, yields the results of a model that maximises NPV within a set budget. These results are considerably less overwhelming and even suggest terminating some development programmes in cardiovascular disease, blood disease, and immune disorders, while divesting expenditures to oncology and infectious disease instead. This model also suggests that diversification across numerous therapeutic areas may not always be in the best interests of a company – more does not always equate to better. Of course, this model will have evaluated the commercial value of the drugs, the risk of pursuing alternative approaches, and the competitive dynamics of the therapeutic areas.

FIGURE 3. COMPANY X Portfolio Optimisation – on a Budget

Ripping the Band-Aid

Killing a development programme is emotionally taxing and financially consequential.  However, objectivity allows for these decisions to be made for the greater good of the company. Using increasingly sophisticated forecasting methods and thorough landscape evaluations can allow company leadership to identify an optimal approach to discovery, development, and deal-making.

Authors