Molecules don’t self-determine their direction or their destination - people do. The IDEA philosophy, the logical conclusion of the longest ever study of innovation in pharma, is that the early choices both enable and disable choices that will be made later. It may seem like the early choices are the obvious ones: is it safe enough to put into humans?, for example. But there’s an entire industry of pre-decision making (“it can’t be oral? There’s no room for an injectable…”) that narrow the exploration opportunity well before we know anything about the molecule.

And, we need to know a lot about two things - the molecule and the destination - before we make any choices. Knowing too little about the destination puts those early decisions at risk of being found wrong later.

I was intrigued when I rediscovered this chart, which takes everyday language, and maps it to another axis - knowing that a drug ‘probably’ works in an indication leaves 25% doubt that it does…

As soon as someone decides to turn a molecule into a product, and writes it down (in the form of a TPP, or some other template), a whole set of consequential decisions (studies, regulatory endpoints) are cast. That illuminates the decision risk - if someone starts a phase II study, all of the ‘other’ options are lost. That may be invisible, if alternative paths to market weren’t examined, but that doesn’t mean they don’t exist.

So, work back: if the highest decision risk is just ahead of phase II, you’d imagine that it would also be the point of highest decision support. You’d imagine our approach to decision science in early phase would be widespread, but it is still rare. And that isn’t because the process the industry has always used works - almost all of the time (‘almost certainly’ on the chart above), it fails our drugs. The industry tends to rely on McK-style over-analysis of the wrong things, rather than appropriate analysis of the right things (opportunity, alternative paths).

As soon as a molecule accumulates formulation, dose, indication, patient population and more, it becomes a product. Unfortunately for most companies, those decisions are made without the necessary amount of thought, and by default, and then the analysis comes in. This has to change: there’s too much riding on badly informed decisions.