The mantra for big pharma is to fail early and fail fast in preclinical studies, then push compounds out as quickly as possible into first-in-human testing, and gather as much information as possible to determine if it is likely to succeed before putting it into expensive Phase III trials. It sounds good, but instead of lowering the cost of drug development, this has lead to an increase in costs; more technology usage in preclinical testing combined with more compounds entering Phase I and II trials mean that money spent in these areas has grown significantly. Phase III trials have become much more expensive partly due to their increasing trial size and complexity. At the same time, productivity as measured by the number of new drugs approved by the FDA has decreased.
What does it mean to fail early and fail fast?
Part 1: Use predictive medicine early in preclinical studies.
Toxicity and PK testing will be integrated with compound screening in a feedback loop to weed out compounds that may have toxicity issues or poor PK early in the process. In the future, it is hoped much of this work can be accomplished in-silico rather than actually having to run assay, speeding up this process. I believe computer modeling of these issues will be very far off considering current assays and live animal models are problematic enough in their predictive abilities.
Part 2: Develop biomarkers early in a compound’s development
This allows clinicians to detect a drug’s effect early in its treatment cycle. Biomarkers are fantastic and will be an essential part of future drug development. However, biomarker development is an expensive and slow process. Also, simply because something is modulated during drug treatment, this does not necessary predict a drug’s efficacy. Biomarkers, like the drugs themselves, need to be validated as a part of the trial process. Otherwise, pushing compounds through due to positive biomarker results simply clogs the clinical pipeline.
Part 3: Use real patients in Phase I studies
The idea is to give clinicians a head start on understanding a drug’s activity. This is simply wrong-headed. It is true, in oncology trials, actual patients participate in Phase I trials, providing safety, and sometimes, a hint of activity data. But the use of patients is due to the inherent toxicity of cancer drugs, which precludes the use of healthy subjects. In any case, oncology trials are no more successful than clinical trials for other indications- if anything- they have a higher failure rate. In studies of drugs that are to have a high safety profile, testing for safety in a sick population could confound the results of the study.
So far, compounds put into the clinic are certainly failing early and failing fast. But they are not failing cheap. Even with the leap in spending on preclinical studies, too many poorly designed compounds are making it into the clinic. It appears to me that with this new list of ADME, PK, and toxicity studies that need to be performed, companies are simply “checking the boxes”, then sending the compound on its way to an IND, showing the investment world their ability to quickly grow a pipeline.
According to Bain Consulting, between 1995 and 2000, one in eight drugs entering the clinic received FDA approval. Between 2000 and 2002, the rate was one in thirteen. Such a high failure rate is now an expected part of drug development, but a 92% clinical failure rate is clearly not feasible in the long run. Large Pharma companies from Pfizer to GSK are scrambling to build new business models. What really needs to change is their mindset. Stop focusing on clinical failure, instead, the goal should be on success.
Disclosure: No position