Predictive performance of front-loaded experimentation strategies in pharmaceutical discovery: a Bayesian perspective

Experimentation is a significant innovation process activity and its design isfundamental to the learning and knowledge build-up process. Front-loadedexperimentation is known as a strategy seeking to improve innovation processperformance; by exploiting early information to spot and solve problems as upstream aspossible, costly overruns in subsequent product development are avoided. Although thevalue of search through front-loaded experimentation in complex and novelenvironments is recognized, the phenomenon has not been studied in the highly relevantpharmaceutical R&D context, where typically lots of drug candidates get killed verylate in the innovation process when potential problems are insufficiently anticipated upfront.In pharmaceutical research the initial problem is to discover a “drug-like”complex biological or chemical system that has the potential to affect a biological targeton a disease pathway. My case study evidence found that the discovery process ismanaged through a front-loaded experimentation strategy. The research team graduallybuilds a mental model of the drug’s action in which the solution of critical designproblems can be initiated at various moments in the innovation process.The purpose of this research was to evaluate the predictive performance of frontloadedexperimentation strategies in the discovery process. Because predictiveperformance necessitates conditional probability thinking, a Bayesian methodology isproposed and a rationale is given to develop research propositions using Monte Carlosimulation. An adaptive system paradigm, then, is the basis for designing the simulation model used for top-down theory development.My simulation results indicate that front-loaded strategies in a pharmaceuticaldiscovery context outperform other strategies on positive predictive performance. Frontloadedstrategies therefore increase the odds for compounds succeeding subsequentdevelopment testing, provided they were found positive in discovery. Also, increasingthe number of parallel concept explorations in discovery influences significantly thenegative predictive performance of experimentation strategies, reducing the probabilityof missed opportunities in development. These results are shown to be robust forvarying degrees of predictability of the discovery process.The counterintuitive business implication of my research findings is that the keyto further reduce spend and overruns in pharmaceutical development is to be found indiscovery, where efforts to better understand drug candidates lead to higher success rates later in the innovation process.