Connecting Data to the Patient: Paul McKenzie on J&J Pharma R&D's Lab-to-Patient Program

Once subject matter experts understand the overall goals, they can go in and explain them to their colleagues, in a language they’ll understand. They really get it, as in “So, sample preparation really is a significant series of unit operations that, if we do well, can help the tech transfer and help compare data across the continuum.”

So it’s really just a matter of spending the time. 

I think most people really want to make technology transfer successful. As an R&D scientist, there is no better feeling than seeing your product or your method executed in a plant or in a QC lab. And the more products we can transfer more seamlessly, the easier it is to manage our product over its lifecycle, and the more products we get to work on as R&D scientists. And that is really where you sell the whole concept.

PhM: To what degree have you been able to connect clinical, CAPA and adverse patient response data to sales forecasting, manufacturing and R&D? What approaches and technologies will be needed to make this connectivity between the end user, and operational siloes, possible within the industry?

PM: That connectivity poses a ripe opportunity for us. We still have pillars of say, clinical data, CAPA data, sales and forecasting data.  Where we’re trying to concentrate on right now is to start to understand the questions of interest across those pillars. 

For instance, let’s say I would like to know where the process that I made with Process Approach A was used clinically. I’d have to have my process database that has A, B, C, and D in it and then we have a clinical database of how the trials were run. How do we get a common vocabulary across those so that I can quickly query them to get that question of interest satisfied?

At this point, we’re generating these questions of interest.  Based on those questions, we can then set up a standard data structure which would allow us to extract data quickly. 

This work is at a very early stage, but we’re starting to optimize by asking the cross-pillar questions so that once we get a good cross-pillar solution the pillars, themselves, should be more ready to allow those cross-pillar questions of interest to be answered very readily. 

The amount of data that is generated in all those areas is huge. To mine it well requires standing back and asking, “What kind of mining would we want to do?”

J&J's Pharma R&D’s Informatics group is really taking this on to make sure we can take a more holistic look at this cross-pillar integration of information so that we can make better and more rapid data mining activities a reality, as well as make decisions quicker from that data mining. 

PhM: Yes. Do you see FDA encouraging this type of work or is it still too soon?

PM: Oh, I absolutely think the FDA and other agencies expect us to go move in this direction. A lot of their ability to approve drugs relies on our understanding the interaction between the process, the analytical and the clinical. The better we understand that interaction, and the better the tools are that we put in place to show that interaction very quickly, the more the entire industry would benefit.