An efficient way of demonstrating proof of concept of a new chemical entity is to develop a combined first-in-human ( SAD+MAD)+patient cohort Ph1/1b).

It depends somewhat on the nature of the proof, that is to say which question requires an answer. For example an antibiotic can be shown to work in a Petri dish in the lab before any trials are needed. Usually the question is of the nature: "Does it stop or reverse a pathological process?", in which case a trial in patients is necessary. Then the first study in patients is usually designed to answer it, so it is at Phase 2a.

Unless the compound is toxic, 2a is preceded by Phase 1 in volunteers when its pharmacokinetics and tolerability are checked because that finds the optimal dose to use in 2a. Sometimes volunteers can act as proof of principle subjects and I invented models to do exactly that. For example, using an increasing laser to generate heat on the skin, the volunteer stopping the laser as soon as a pain threshold occurs. It showed that a liquid formulation of paracetamol has a faster onset of pain relief than a tablet. The difference was small, so the precision of the model was essential for detecting the difference and so providing proof of principle. I invented a gastric monitor to study prokinetics in volunteers because again, the precision of the model, made a POP study much easier to design and conduct.

Phase 2a is usually followed by 2b to explore the best dose over a period of time, and possibly secondary indications. Phase 3 is basically designed to assess how well it works in real practice and the incidence and severity of side effects.

I hope that helps the reason why you asked the question.