Abstract:
Critical strategic decisions are made based on valuation models of pharmaceutical assets. Of the many assumptions that underpin a valuation model, probability of success is a key input, but one that often receives little thought or consideration. Several publications have analyzed historic probability of success in the pharmaceutical industry. Each study uses different datasets or different methodologies, often leading to a significant difference in conclusions. We examine the literature, provide a summary of the different datasets and urge model developers to think carefully about which assumptions to use. Depending on the profile of the drug, there can be as much as a 280% difference for the final valuation. Failing to account for the characteristics of the molecule, which we believe is best done with a thorough review of the data by seasoned drug developers, can result in a significantly skewed picture of valuation and misguided investment decisions.
About the authors
Serena Zhou worked as an analyst at Alacrita during the summer of 2018.
Rob Johnson is a co-founder of Alacrita. His areas of expertise include business development and licensing, market assessments and commercial due diligence.
Paper Download
Click here to download a PDF of the full white paper.
Introduction
The estimated cost of developing pharmaceutical drugs has increased from $1.04 billion in the 1990s through the mid-2000s to $2.56 billion in the 2000s through the mid-2010s1. One primary factor of this increase in cost is the risk associated with bringing a drug from preclinical into the clinic and ultimately to market. Much of this uncertainty comes from the likelihood, higher in some indications and lower in others, that the drug in development may be terminated due to any number of factors, including efficacy, safety, or commercialization concerns2. These drug development risks must be considered when performing pharmaceutical valuations for companies and products, as they comprise a basis of the industry standard risk-adjusted net present value (rNPV). For pharmaceutical valuations, an understanding of the historical probability of success of a therapeutic through to approval provides an understanding of whether an investment is in a company’s strategic interest.
This report provides a review of the published literature and their methodologies on the pharmaceutical probabilities of success from 1993 to 2015 and insight into any potential trends that have emerged in the past 22 years.
Executive Summary
This white paper summarizes the probabilities of success from eight publications reviewing therapeutic products from 1993 to 2015. While differences in years covered, methodology, and sample source did result in differences between the various publications, several general trends emerged.
Of the four stages of clinical development, Phase II to Phase III had the lowest probability of success because it is the first stage during which efficacy is assessed. While efficacy and safety are two of the primary reasons for clinical trial termination, commercial reasons such as rationalization of the company portfolio also play an
important role in the low probability of success values, affecting as many as a third of all trials terminated in Phases I and II from 2005 to 2010. Of the 16 major therapeutic areas covered by the publications, psychiatry and oncology had the lowest overall probabilities of success, although only five indications exceeded an overall likelihood of approval (LOA) of one in five.
Further breakdowns to determine the role of a lead indication demonstrate that while lead indications typically have improved success when considering Phase III to Approval and overall LOA, there are inconsistent effects elsewhere, particularly in Phases I to III. Oncology, one of the largest therapeutic classes and consisting of up to as much as a third of the data evaluated by some publications, has a large role in decreasing the overall probability of success values often reported without differentiating by
therapeutic area. Among oncological tumor types, hematological tumors had lower probabilities of success in Phase I, owing to greater safety risks, but solid tumors had lower probabilities of success in Phases II, III, and Registration to Approval, most likely due to the issues of tumor penetration, toxicity, and mechanistic insufficiency.
Other trends were present when considering orphan indications, biomarker usage, modalities, and drug origin. Orphan indications’ probability of success was inconsistent in terms of Phase III to Approval and overall probability of success with the various sources differing on whether orphan indications improved the likelihood of approval. In contrast, biomarkers for patient selection improved probability of success across all clinical stages but resulted in a decrease in metabolic and endocrinology indications due to the low sample sizes. When biomarker identification and evaluation trials are included, biomarker usage did not cause a large and consistent increase in the probability of success.
Among the various drug classes or modalities, NMEs had the lowest probability of success, but when oncology and non-oncology modalities were considered, oncology vaccines had by far the lowest overall LOA. Products that were licensed-in to a top 50 pharmaceutical firm had higher probabilities of success than both self-originated and licensed-out, demonstrating the importance of partnering to the advancement of clinical programs for patients.
Clear understanding of the orphan or biomarker status and careful selection of a probability of success value will enable the development of valuations more in line with the true value of a proposed drug.
Additional Articles
Alacrita's pharma valuation expertise
Our expertise in performing business and asset valuations covers a wide range of technology types, including small molecules, biologics and gene therapies across a variety of clinical indications.