Medtech Snapshot: VP R&D Arif Iftekar - Predicting Product Failures Before Clinicals

Clinical trials are costly—how can R&D teams predict failures before they happen? In this MedTech Snapshot, Arif Padamsee explores the balance between preclinical testing, simulations, and real-world data to set new products up for success.

Transcript:

Travis: Well Arif, as of last week, the poll that we put out, which, as you know we were really addressing: how do you determine ahead of time, and predict what potential product failures you're going to experience most importantly before you get into clinical trial. 
 
Arif: Travis, you know an important topic for the industry, and for R&D in particular, there's no question that when you look at it from the outside in the most important signal uh one can get about the maturity of a new product, a new therapy, is, you know, important milestones like: first, in human, an early feasibility study, an IDE (Investigational Device Exemption) that's approved and rolling.  
 
There's no question that a clinical study is ultimate proof of safety, of performance, and of the commercializability of a new product and therapy. Having said that, it's very expensive to get there in terms of what you've put together. And, two, it's very risky if you have surprising observations in the clinical studies. It's really hard to get back to the drawing board at that point.  ​

Screenshot of the poll being referenced. Question: What is the best way to predict medical device product failures on the bench before entering clinical trials? 1) Upstream marketing dictates testing - 32% 2) Utilize machine learning - 4% 3) Outsource FMEA to experts - 28% 4) Other (comment below) - 36%

So the question that you have pushed in the poll,--very pertinent--is how do you set yourself up for success before that? Through pre-clinical testing? bench testing? animal work?, etc. And that is, I think, where there is a tension. How much do you do on the bench before you get into clinical study?, versus how much are you prepared to learn from a clinical study? You can move forward or come back to the drawing board.  
 
So, most people would say to get your testing done correctly, do it thoroughly. But, as you saw in the poll, there's some divergence on even what that looks like. Is it more DB (database) testing? is it more simulations? is it getting your inputs right?--and it's all of the above. The real answer I think is this iterative process: be prepared for surprises and find a balance between how much you can de-risk for a clinical study, but allowing yourself to learn a few things.