Combination therapies are becoming increasingly used in drug development for
a range of therapeutic areas such as oncology and infectious diseases, providing
potential benefits such as minimising drug resistance and toxicity. Typically, a
pharmaceutical company will have multiple treatments in different stages of
development in their portfolio and the problem of portfolio decision-making
will include decisions such as which studies to initiate and how to prioritise
studies.

This problem is more complex for portfolios of combinations since sets of combination studies may be related, for example if they have at least one treatment in common and are used in the same indication. However, in this setting, value can be gained by sharing information between related combination studies in terms of improving the treatment effect estimates and improving the portfolio-level decisions. We discuss the challenges of portfolio decision-making for a portfolio of combinations and present methodology to assist with this.

One of the key estimates that is used in decision-making regarding a clinical
study is the probability of study success. We present a framework that allows
the study success probabilities of a set of related combination therapies to be
updated based on the outcome of a single combination study. This allows us
to incorporate both direct and indirect data on a combination therapy in the
decision-making process for future studies.

Existing methods for portfolio decision-making do not account for the differences between single agent and combination drug development. We extend the existing methodology to consider the relationship between combinations and the effect that observing certain outcomes may have on the portfolio decisions we make. This is achieved by updating the study success probabilities throughout the decision-making process whenever a relevant outcome is observed.