No Time To Die
There is no ‘right’ percentage for ‘failure’ in phase III. Much as the Measuring Consultancies would like it to be so, lower attrition in phase III can be bad: a 0% attrition in phase III could be achieved by simply running easy studies, or even (as this chart suggests) running lifecycle studies on 20 year old drugs. Also, a 100% failure rate would rightly be regarded as bad by most observers. So, the ‘raw’ percentages are a simple take, but still revealing for the heterogeneity in the top 30…
I’ve written before on pIII ‘failure’: However, let’s acknowledge that a drug that ‘fails’ a study could equally ask, ‘did the study fail a good drug?’ – that is, could a different question asked of the same drug have yielded a different answer?
‘Other’ and ‘unspecified’ categories include a lot of debris, washed up on expensive shores. But, when only 18 of 299 ‘failures’ were for adverse events (which, let’s remember, is what phase III is ‘for’ - to pick up signals that couldn’t be seen in smaller studies), this does show that companies are sweeping decision risk under the Development carpet until it can’t be hidden any more.
Phase III is no time for good drugs to die. As someone posted on my Twitter yesterday, the percentages hide real people - patients and people working on those studies - even when you set aside the waste of resource, those people deserve better.