How Janssen uses big data to develop new diagnostics: Interview with Nigel Hughes

Nigel Hughes

In the lead up to IQPC's Big Data in Pharmaceuticals conference, Nigel Hughes, Global Director Marketing, Health Information Technology Strategy Leader, at Janssen Diagnostics talks about how his company is using data in R&D, the key challenges facing the use of data in the pharmaceutical industry and why the people challenges are bigger than the technical ones.

PEX Network: How are you using Big Data at Janssen Diagnostics?

Nigel Hughes: First, I’ll say let’s just say data. Big Data is used very often now as a term, but, actually, we’re still really talking about just data.

We’ve had a long history of accessing clinical data - under all the right regulations and restrictions, of course - for many years.

Data - the new superhero helping to identify all sorts of diseases?

I’ll give you a couple of examples. First, over the years we were involved in HIV resistance interpretation. If a physician was faced with someone living with HIV whose drug therapy was failing, maybe due to resistance, it was a challenge to interpret what was happening and work out the next best drug combination to combat resistance and suppress the virus.

It’s less of a challenge now in many markets, thankfully, because the drugs have become so much better.

But back in the day, the forerunner to Janssen Diagnostics was a company called Virco, which looked at data derived from clinical centers treating people who were living with HIV, clinical trials and samples. We combined all of this data to create a bio-informatic tool called vircoTYPE, which allows physicians to more simply interpret resistance to HIV and then work out the next best drug regimen and so forth.

I think that was very useful in paving the way for physicians to manage HIV in the way that they do today. It was important to be able to access so many patient records to analyze specific aspects of their resistance, their clinical outcomes, their treatment and so on. Then, by linking that data with lab-based results analyzing the viral genome and its behavior, its phenotype, we could create a bio-informatic tool which would allow interpretation in minutes versus up to a month if trying to interpret it in a lab-based setting.

That was one area where we have some heritage from doing that.

Then, more formally and more recently, we’re now working with clinicians, clinical centers and accessing clinical data around other infectious diseases, including HIV, but also hepatitis-C.

We always base this around a so-called quid pro quo, so we always want to ensure that it’s a relationship where clinicians are providing data under the right conditions. We provide value back to them in terms of interpretation, generating insight, working with them to do further research, audit or benchmarking, but meanwhile we would like to generate insights which will tell us things related to the market and help us in terms of research and development for new diagnostics and, working with our pharma colleagues, in terms of developing new drugs and new approaches to treatment as well.

So there are a number of different use cases. Accessing clinical data is absolutely vital because otherwise we really wouldn’t be able to develop new diagnostics, new drugs and new information tools.

PEX Network: What do you see as the most challenging aspects with regards data, facing pharmaceutical organizations today?

Nigel Hughes: There are many, unfortunately. I’ll try to summarize a few that I think are significant. First and foremost, because we’re talking about the pharmaceutical industry, and, unfortunately, there are clearly perceptions and issues around trust and confidence in our access to data where industries are concerned, it’s absolutely essential that when we’re accessing clinical data it is under the right conditions, as I was mentioning earlier – this includes legal, privacy and security constraints – but also that it’s open, honest and transparent.

When the industry is working with external data from the market, working with clinicians, working with patients and patient groups it’s important that’s done in a way that people can clearly understand what it is we’re doing with the data, what the intent is regarding the data access, and, importantly, what that data eventually will be used for. I think there is, as I say, a key issue around perception - that’s one critical challenge. Once you get through that, there are, obviously, issues around privacy, security and confidentiality.

Regulations are very varied, depending on the market; United States is very different to Europe. Within Europe, every member state has its own interpretation of privacy directives, for instance. This can really make it challenging in terms of setting up studies, for instance, observational, non-interventional studies, or accessing data for other research. It can vary enormously in terms of the time, energy, resources and finances required, even country by country, just in Europe alone, let alone in the United States, Asia or other parts of the world.

Infrastructure is also very different and that creates challenges. For example, we did a pilot registry - a longitudinal collection of data from a small cohort of people with chronic hepatitis-C - but in the countries we were working in, in one country they were using electronic health records and other databases; in others it was pure paper records, and you were having to access the source data through a third party, a data clerk, which was a very laborious process. There are extremes of issues of data access right through to just logistics and tactical aspects of data, which are very challenging.

I think, overall, the key here is to not underestimate how difficult it can be to work with real-life data. The rewards and the outcomes can be significant and extremely useful, but it’s not an easy process and one shouldn’t take it lightly.

PEX Network: You once said that collaboration is key to success in terms of sharing information. How do you think the pharmaceutical industry and physicians can better collaborate on data-sharing?

Nigel Hughes: It comes back to the point I think I made earlier, but one of the key aspects is this relationship is based on a quid pro quo. The balance between someone providing data, under whatever conditions, and those receiving data needs to be equitable, needs to be equal. If we receive data within Janssen or just a pharmaceutical company in general, we will generate some value from doing that: generating insight, helping us develop new products and services, and so on.

Meanwhile, for those providing data – we could pay them for that data and that might be fair, but I don’t think it creates much value for individual clinicians or clinical centers, and hospitals and so on. Instead, what we try to do is look at is how can we assist clinicians in generating insights from their own data?

Although physicians and other healthcare workers are collecting data, they may not be in a position to really maximize the insights they generate from that data. If we can provide those services to the physician, working collaboratively in a neutral area of interest around research, audit and benchmarking, we can help generate those kinds of benchmarking tools and business analytics that support clinicians in understanding how their outcomes compare to a wider context of other clinical centres.

PEX Network: The final question I have is what are the actual technical considerations that need to be in place in order to achieve some of those aims?

Nigel Hughes: The most challenging part isn’t the technical consideration, it’s the people bit. It’s developing the relationships. It’s working with clinical centers and clinicians. It’s ensuring everything’s done under the right conditions where all appropriate requirements are met. It’s contracting, it’s going to ethics committees, it’s all of these processes, which are quite challenging and take most of the time when you’re setting up access to external data, real-world data.

That’s not to say that the technology isn’t challenging. But there are so many tools available now, either off the shelf or proprietary or in-house, that are available. There’s a big move within the pharmaceutical industry to move to cloud-based tools and services.

For instance, our electronic health record system and other products that are working in that environment are all web-based and accessed through a secure browser login. You can then through that process develop further architectures. You don’t touch the primary architectures dealing with the clinical data but you can develop secondary architectures to build in business analytic tools, for instance.

That’s just one example; there are many, many others as well as areas of overlap in interest and value for both clinicians and patients and for the industry based around these kinds of technology tools.