Critical Criteria

The fact that only 2% of all projects make it all the way through the pipeline 1 suggests that a safer assumption is not that a project will succeed but that it will fail. This mind set then opens one to the task of determining which criteria the project is less likely to meet than other projects, the critical criteria, and to design a project plan that ensures the delivery of the critical criteria. Critical criteria are likely to be different for each project, and likely to fall into categories, such as known problems with the particular target class (e.g. druggable chemotypes for cysteine proteases, protein-protein interaction for cytokines, or high binding site homologies within the class lending to lack of selectivity for kinases), disease pathway regulatory mechanisms, therapy area (e.g. lack of validated animal models, lack of surrogate endpoints).

While some experiments need to follow others, experiments that address critical criteria and are not bound to the completion of other experiments should be performed early in the project. For example, if a critical criterion is demonstration of efficacy in an animal model and a both a relevant animal disease model and tool agent (literature small molecule or monoclonal antibody) are available, then the animal efficacy experiment should be run in the Target Discovery or Lead Discovery stage, and not wait for the Lead Optimization stage when animal efficacy experiments tend to be run. If no tool agent exists or it will take some time and expense to develop a relevant animal disease model, then the team will need to go through Lead Discovery and the early stages of Lead Optimization before it can run the animal efficacy experiment.

The committee which oversees the initiation of projects should insist on a carefully considered project plan to address such critical criteria. The committee which reviews program progress at milestones would focus attention on critical criteria. From a portfolio perspective, the company should seek a balance of projects with different sets of critical criteria. The use of critical criteria promotes the concept of “doing the last experiment first”, 2 the last experiment being the one that kills the project. A critical criteria system is not a crystal ball, but it is a system that can help to promote early failure rather than a more expensive late failure. And the critical criteria system should promote early attrition which is less expensive than late stage attrition.

  1. “Rediscovering the Sweet Spot in Drug Discovery”, David Brown, Giulio Superti-Furga, Drug Disc. Today, 8, (23) 106-1077 (Dec. 2003).
  2. A nonpharmaceutical example may be found at http://www.snacks.com/good_fun_fritolay/2009/03/doing-the-last-experiment-first.html