MechanizedIntelligence

Bruce Booth has three thought-provoking suggestions about what he’d do if he was in charge of Pharma R&D at a large company: “Invert the periphery and the core”, “Get the rest of the company out of the way of research”, and “Establish a longer-term, science based approach”. If you’re interested in the details, follow that link and take a look at the discussion on Derek Lowe’s blog, In the Pipeline.

In the meantime, I’ll throw in my two cents about a fundamental problem specific to Pharma R&D that I don’t see mentioned much. Evidence indicates that pharmaceutical research is a stochastic process with two really brutal attributes: 1) High project failure rates and 2) An extremely long lag time between actions and outcomes — seven years or more from when a project is started to when the new drug successfully hits the market. The high failure rates make it tough to figure out what works — and for even the best scientists and managers the feedback loop on their efforts is long (several years) and not at all direct. The result? It’s awfully hard to know if the strategy or the organizational scheme you used was any better or worse than any other alternative. The behavior of dynamic systems predicts that with variable processes and delayed feedback loops a system has the potential to oscillate wildly. Throw in a management team too impatient to wait for a complete outcome before making another change, and a company trying to Change Things will flail about. Which is exactly what we’re seeing across the industry — continual shifting strategies and repeated restructuring efforts that are implemented before enough time has elapsed to evaluate the outcome.  Hard to see how that will result in substantial improvement.