In Silico IDE: Phase 3 Staging
The most expensive bug (of many) in the pharmaceutical industry is the failed Phase III trial. Roughly 50% of drugs that show brilliant efficacy in Phase II end up failing in Phase III, wiping out billions of dollars in R“ & ”D and cratering market caps overnight.
Why does this happen?
Because although Phase II data can look incredible, it is in effect a subgroup which lends itself to overfitting. Currently, sponsors take this overfitted Phase II inference and extrapolate it directly into a global Phase III which is massive, noisy environment with thousands of unoptimised edge cases.