Top Tips for CDISC Implementation

Peter Van Reusel, Business Unit Director of CRO Services at Business & Decision Life Sciences, and Albert Chau, Director of Datacision, joins Pharma IQ to discuss CDISC standardisation.

Pharma IQ: To start off with, with regards to standardisation, Peter, do you think that there are any viable alternatives to CDISC?

P Van Reusel: Viable alternatives to CDISC - a very good question. When I think a couple of years back, actually quite a few years back, to the beginning of data management, I’m talking about the 1980s and 1990s, we can see large companies and medium companies who have already seen the light and they’ve already seen the advantages of data standardisation. Those companies over the years have tried to standardise their data internally in order that they can reuse a lot of their data standards, their templates, their standardised programmes. They can reuse that over and over again.  Now, since 2003/2004 it’s the first time that we’ve seen industry data standards, data standards specifically for pharmaceutical companies and biotechs. There is CDISC and the CDISC data standards. We are talking about SDTM, ADaM, to define XML. I believe there’s no alternative. I think it’s the first time that we’ve seen that in the industry and I think it’s here to stay. 

A Chau: I share a very, very similar view to Peter, really. There are companies that have been working on those different standards; however, all those different companies have their own ideas how the standards will work. So, when it comes to exchanging data between companies or even regulators, everyone has to relearn each other’s standards, which is not a very efficient way of working. Having CDISC, it means that every company would be working towards the same standard and it would facilitate the exchange of the data in the whole industry.

Pharma IQ: Are there any tips that you can give someone who is about to implement CDISC across their organisation? We’ll start with your views on this, Albert.

A Chau: I suppose the planning is actually very, very important to start and not to underestimate the amount of resource that needs to be in there just to get it kick-started. Obviously they need to make sure that the management is buying into that and fully support the whole CDISC implementation. Also, when it comes to actually implementing it and actually developing the standards, you need to make sure you have the right staff on board, to train them and make sure they all understand and communicate with each other to understand the general philosophy of the CDISC. That’s about it, yes.

P Van Reusel:I totally agree with the topics of Albert and so getting your management buy-in perfect in trying to plan for that. Do not underestimate implementing CDISC across the organisation. One thing I would like to add, and Albert said staff and training, is you do need experience in this. So, you need to get experienced staff. Unfortunately, if you look around in our industry, there aren’t that many experts on CDISC, on SDTM or ADaM. More and more of our industry people or data managers are getting experience in that but usually there is some experience with one data model or another data model. If you’re talking about implementing a data standard across the organisation, I would say get help. There are a number of expert companies out there that can advise you with, for example, the CDISC road map and I think that would be a good place to start for a start-up company or if you really need to consider, okay, where do I start? What’s the beginning?

Pharma IQ: Are there any good examples of standardisation that you can talk us through, Peter?

P Van Reusel: Good examples of standardisation that I’ve seen? Well, yes, of course. We’re going to talk about that a little bit later also. When you look around we can see that the CDISC data models, and I’m thinking about SDTM and ADaM right now, they’re quickly becoming the de facto standard for your data back and forth between the sponsors and the CROs out there. Now, in the past we’ve seen quite a few problems because the SDTM standard, for example, is very open for interpretation. So, even though the CROs were using the industry standard, we could still see that, for example, a study that was being sent to one CRO or another CRO would give back completely different results, even though they were using the same standard.

Now, what are more and more companies doing right now? They’re building their internal libraries and they’re using SDTM 312 as a background. So, they’re using that as their internal standard but they’re going to build on that. They’re going to make it more clear. They’re going to make it less open for interpretation. They’re going to give more exact instructions to their partners, not only talking about CROs but also talking about the EDC vendors. So, more and more sponsors are giving more exact instructions on how to use these standards in their environment and they’re talking about building an EDC library or an ECRF library, building an internal data standards library. We’ve seen some good examples of sponsors out there doing that and being successful in that. I would say that’s a good example of standardisation.

A Chau: Yes, I would totally agree with Peter, first of all. Certainly in some of my experience we have seen studies where we start implementing the CDISC standard, which helps to speed up the set-up of subsequent studies and also in terms of both service trial and also when it comes to the reporting stage where we could use the standard, the data centre structure, and then to do all the analysis associated with it. Also, it has actually improved the efficiency in terms of reviewing the data along the way as well and also improved the data quality. So, that has actually helped tremendously.

Pharma IQ: Albert, I know that it’s perhaps not your key area of expertise but from where you stand, do you think that offshoring your data management can be advantageous and what problems, if any, can people expect to encounter with this?

A Chau: I think one of the motivations of offshoring is the cost. As you mentioned, I haven’t really had firsthand experience of this; however, from what I observe, it’s mainly driven by cost so far. However, there are problems that I can see that people can encounter. First of all, the communication becomes very important because people working offshore, they will be working on different sites. Sometimes they are just really seeing isolated groups, so communication is absolutely key. And also to make sure that everyone’s talking to each other and also understanding the actual business process because a lot of the time the offshoring offices may not have the first-hand experience of the business processes like what we have in Western Europe or in America. So, that becomes very, very important, for them to understand why certain things have to be done in a certain way rather than just telling them you just have to do this and just force-feed them with information and just say, go with that. 
[inlinead]
P Van Reusel:Very good question, again, offshoring data management, and we could think about more types of services to offshore. I think there are a number of problems that people can expect to encounter. What we are seeing in real life usually, when we are talking about offshoring, is that we’re often offshoring to a different culture and also to a different time zone and that usually affects people. So, working with a different time zone, that’s something to get used to. Now, I believe in our industry people know what a different time zone is – people from Europe working with people from the US – but if you’re also working with, for example, India or China, that’s another thing to deal with. So, that’s certainly one. From what I’ve heard from offshoring, there’s also that cultural shock sometimes. People tend to think a little bit differently, specifically towards roles and responsibilities. I think when you’re working with an offshore company you definitely need to exact your roles and responsibilities and say, okay, this is what we expect from you. From a distance there is that cost attractiveness. So, typically these services can be very low in cost. However, we also need to think about the second cost that belongs to that. There is a certain overview task that the sponsor will still need to do in order to guide its offshoring company through the process and that can definitely increase the cost. So, that’s really something to consider. We shouldn’t buy into one and forget about the other one.

Pharma IQ: Working with CROs can have its problems when it comes to data standards. Have you got any best practice tips, Peter, that you can give data managers or sponsors to ensure an effective working relationship?

P Van Reusel: I’ve already touched on that a little bit earlier. Working with CROs and data standards from a sponsor’s perspective can have a number of problems because some of the standards out there, they’re very good industry data standards but still open for interpretation. I think as a sponsor you need to add on that. You need to specifically tell your CROs, that’s how I want my CRO partner to implement these data standards for me. So, that means that a sponsor library of CRS, a sponsor library of their metadata, for example their SDTM data models and maybe their ADaM value level metadata,all of that internal standardisation will help the CROs out there because there is less room for interpretation. I think that wriggle room that we all know in SDTM and ADaM and CDISC models, we need to try to remove that as much as possible and then you’re going to see more consistency in your deliverables back from the CROs. 

A Chau: Yes. Again, I agree with Peter, that the internal library, the metadata standards, they will all help. The other thing that I have seen more and more with the sponsors as well, they start to have data stewards. Even if they outsource the data management part to the CROs, internally they would still have a data steward, a data expert, to make sure that everything is consistent and also if the CRO has any questions on a particular data issue where they’re not sure how a data standard should look, this is where the data steward can come and really ensure that they have the answer to that for the sponsor company. So, I think that would really help. And also just to make sure that the sponsor company and CRO work together towards a standard and communicate as much as they can.

Pharma IQ: Looking at the landscape in five years’ time, what do you think will have changed, Albert?

A Chau: I think, first of all, there’s more and more requirements for the data standard, and certainly, FDA is asking more and more to have the data to be submitted in a certain way for the SDTM, for sure, and also going to be for the analysis data in the near future as well. So, there needs to be moving towards more adoption of the CDISC standards and with all the work that the industry has been doing, there’s going to be a lot more clarity in the next couple of years. So, I think that will make the adoption pace go faster. Also, there will be more expertise around. So, that will really help because I think in the last couple of years we’ve seen that the adoption is starting to take up, however there’s a lack of understanding or not enough people understanding it. So, the pace is a bit more difficult but I think in the future it will get easier. Also, I have seen that there’s a few tools that are coming out and the technology side will really start to help on getting the standards in place as well.

Pharma IQ: Some useful projections there. And, Peter, how do you see the landscape changing in five years’ time?

P Van Reusel: Actually, I totally agree. So, in our industry we are going to see more experience for people. Since 2003/2004 people are becoming more and more exposed to CDISC data standards and they’re getting experience and that’s going to help us. So, the adoption is indeed going to increase. Also something fascinating that we’ve seen over the years, since 2005 I would say, we can see companies out there at different rates and at different tempos implementing SDETM and right now we are noticing, in 2011 already, the same thing about ADaM.

So, companies are also implementing that data standard so that in five years from now SDTM and ADaM are common and companies have implemented. As a result legacy data conversions, which has been around for quite a while, they’re going to be disappear slowly. So, that’s something that we are going to see. Another thing that we have seen in the development of data standards, I would say, is that SDTM and ADaM specifically, they started with the common ground for data standards. That means basically data around safety, a little bit about efficacy. More and more we see that CDISC is extremely heavily involved with the creation of data standards around therapeutic areas and disease-specific data standards, the real data standards that companies are using out there. So, we’re going to see also a quicker development of the data standards for us all to use.

Pharma IQ: And leading on from that, I guess, are there any breakthrough technologies or strategies that you expect to see in the next few years, Peter?

P Van Reusel: Breakthrough technologies, I think we’re going to see that the tools out there, that the tools that everybody is using – data collection tools but also metadata tools – that they are already more and more CDISC aware or data standards aware. I think all the tools that are common out there are going to be data standards aware. Strategies, I can see all companies out there buying into the CDISC standardisation. I can see all companies in the next five years implementing these CDISC standards.

Pharma IQ: What technologies or strategies do you expect in the next few years?

A Chau: Starting with the technology, I can see the different strands of CDISC that will be starting to come together a bit more. So, for example, like up to now different people will be implementing the CDISC SDTM and the terminology maybe slightly separately and also slightly separate from the CDASH standard as well. However, I can see that with the technology, that’s going to streamline that part a bit more in the future. In terms of strategy, I think with the companies around, some of them will start rethinking about the whole clinical trial process. So, again, making use of different standards and try to streamline the whole process from start to end.

Pharma IQ:Why have some companies been reticent to take on CDISC standardisation in your opinion?

A Chau: I think, first of all, maybe there’s a lack of understanding in some companies exactly what CDISC is about and also there is going to be an investment in the resources that you will need to put in place to make sure you implement the CDISC standard in there and when you do that you will have learning curves there. It’s quite a steep learning curve for a lot of people. When that comes about, that means that there’ll be a disruption to the business and no one likes that, really. Also, I have seen some people saying, why bother changing when up to now they have been submitting drug submissions and they’ve got the drug approved, why do we need to change it? So, there needs to be a change of mindset. There are multiple reasons why people are reluctant to take on the CDISC standards up to now.

Pharma IQ: Peter, do you have any additional views on why companies have been so reluctant about CDISC?

P Van Reusel: Yes, indeed, if you’re looking into the practice, I totally agree with Albert. Companies are seeing the investment and maybe they’re not measuring the value of an industry standard correctly or just in time. We can see that more and more are changing. I’m talking about the investment because this takes time. Companies out there have been operating longer than CDISC. So, CDISC came on board really as an industry standard in 2004/2005 and they produced standards that you could really use, really implement. But companies have been around much longer.

That means that companies, sponsors out there and CROs, had their internal data standards which they have been developing for the last ten years or 15 years and they really like their own internal data standards. So, there is also reluctance from people to start using another data standard. It’s also an investment because there are legacy data centres but also legacy tools. If you look at these big companies specifically, you can see that the tools are built on their internal data standard and if the data changes, that also means that the tools and the processes change around that, so legacy data standards and legacy tools and I would also say, as Albert said, legacy people. Why would we change? They’re usually reluctant because it’s going to interrupt their business processes.

Pharma IQ: In many ways, standardisation is a mind-set. What would you say to those higher up in pharma companies to persuade them of the benefits of standardisation of clinical data? For this one we’ll start with you, Peter, and your views on this.

P Van Reusel: Indeed, standardisation is a mind-set. What can I say? I would say, okay, come up with an alternative. What is the alternative? There is really no other way than implementing data standards. Data standards are good, we all know that, and try to make a business case, let’s use data standards and look at the results, let’s not use data standards and look at the results, and I think that the choice will be quickly made. There is no other alternative.

A Chau: I think sometimes it’s quite difficult to try to sell saying it is an investment but it’s an investment that will get you the returns. It may not be an immediate return, however, there’s a lot longer gains. For example, by having standardised data, one, hopefully it will improve the data quality and, secondly, it will improve the efficiency. Also, by doing that in future with all the regulatory requirements, especially with the FDA requesting such data, it will speed up their review process as well. So, hopefully that will be speeding up the time to market for the drugs.

Pharma IQ: Now, your peer, Philippe Verplancke says that standardisation will benefit from generational shift. Do you think that this is true and why, Albert?

A Chau: I think partly it’s true that it could benefit from generational shift. I think partly because of what Peter and I have discussed before, that there’s always the legacy with the people thinking, why fix a problem when it has been working fine up to now. I think it’s more a change of mind-set. There are certainly people who have worked with the previous, well, the non-standards before and can see the problem and they are already adopting the CDISC side quite heavily. For example, myself, actually I’m totally bought into that. There are people who are already in a different mind-set; they just need to be the champions in a company or an industry to do that. However, certainly with the younger people coming in, they will have touched on the CDISC already. So, they will be totally familiar with that. So, I think it could potentially benefit from generational shift but that is not the only reason.

Pharma IQ: Sure. And, Peter, do you think that generational shift will make a difference?

P Van Reusel: I do and I think for the same reason that Albert said. Now, when Philippe Verplancke made that remark, I’m not really sure what he meant but I do know how Philippe thinks about metadata. So, I’m going to try to use that assumption and see how it fits. I believe until now our current or our past generations thinking about data and metadata in a tabulary level. In a tabulary level, that means that when they are thinking about a systolic blood pressure or a diastolic blood pressure, they’re thinking about a variable in a table, so a tabulary representation of data and also metadata.

I think more and more professionals are getting involved with these new data standards. They see that metadata and data works in multiple dimensions. For example, a systolic blood pressure does not stand alone. There is also a position that belongs to that, there’s also a method that belongs to that and other attributes. So, they’re thinking more and more about data as a central point of data but then with attributes around that. It’s actually the way of thinking of what’s currently on development, CSHARE from CDISC. It’s a way how to make our metadata interpretable, in fact. Give more meaning to our data through metadata. I think that was something he was thinking about when he made that statement.

Pharma IQ: Just to finish off, you’ll both be speaking at the forthcoming Clinical Data Standardisation and Management conference. Could I ask what will be the key learning points of your presentation? Peter, if you could start please.

P Van Reusel:Thanks a lot. My presentation is going to be about implementing data standards. So, it’s about a concept. What do you need to implement a data standard in your environment? In working with CROs, working with internal departments, what are the key elements, what are the processes and what do you need to have in place? I’m inviting all people to come listen to that. I am sure we are going to get some interesting questions and thoughts.

Pharma IQ:What are you looking to gain from being present at the event?

P Van Reusel: For me it’s really a networking event. It’s really talking with people about data standards. What are you doing? What are you thinking about? How can we help? How can we interact with each other?

Pharma IQ: Albert, what will you be focusing on in your presentation? Can you give us a quick taster of what you’ll be covering?

A Chau:Yes. My presentation is very practical and I’ll be basing it on the implementation of the CDISC standards in the company that I worked for before. So, over a period of four years how we have implemented from absolutely nothing to some of the CDISC standards and the benefits we’ve seen. Also how we have worked with CROs and external partners to make sure that the data standards are there. So, that’s what I’m going to talk about.

Pharma IQ: We look forward to hearing more about the subject. And why do you think that an event of this type is important right now for the industry?

A Chau: It’s important because a lot of people are starting to implement CDISC. There are still a lot of questions out there and I think to have a platform like this, that people can exchange ideas, exchange their experience, I think is actually very, very helpful.