Sunday, October 16, 2011

Challenges in Comparator Studies

As a statistician, I have been involved in the design and analysis of many blinded clinical trials. Being part of the clinical team, the discussion usually evolves around the choice of comparator (e.g. matching placebo or active control). If active control is used, there are usually some further considerations, e.g.:
  • which drug / dose / schedule should be used, as this may differ in terms of clinical practice and regulatory acceptance depending on the region and countries.
  • what is the treatment difference to be detected (for superiority trials) or the non-inferiority margin (for non-inferiority trials), as this would have an impact on the sample size which in turn would affect the time and cost of the trial
  • how to handle the randomisation and blinded kits in the IVRS, to ensure that the subjects receive the intended treatment; and how to deal with subjects who have not received the intended treatment.
Rarely do statisticians get into the more practical details on how to actually produce the blinded comparator treatment. Therefore it was fascinating for me to attend a conference on comparator studies and gain an insight from clinical supplies and sourcing colleagues on the technical aspects and challenges on producing blinded treatments. Ultimately their role is to ensure the integrity of the comparator is maintained and blinding can be achieved as far as possible, bearing in mind that the cost, timing and technical challenges may become a major hurdle in the process of blinding a comparator treatment.

Some of the common methods to blind an oral drug include:
  • Over-encapsulation
  • De-inking
  • Over-coating
  • Over-printing
  • Mill-and-fill
While one method may work for one oral formulation of a drug, it may not work for another oral formulation of the same drug. For example, over-encapsulation may not be appropriate for a fast-acting release of a drug product.

For other types of treatment administrations, there may be other methods such as relabelling, masking or exchanging certain secondary components.

Often there are additional aspects that need to be considered and dealt with e.g. taste, odour, consistency, colour, drug delivery device etc.

With more and more late-phase clinical trials going global, there needs to be consideration on whether to source the comparator drug direct from the original manufacturer (which may be difficult as it's a competitor product), from a generics manufacturer (only if the comparator is off-patent already and the generic form is available) or from wholesaler. The decision on whether to source the comparator drug centrally or locally should not be taken lightly: for the former, one of the challenges is the logistics of moving the comparator from the central source to different countries; while for the latter, the comparator may have slight variation in terms of formulation and availability between countries, and can become a major headache if there is a problem with meeting the demand of the drug locally.

Bearing in mind that there are so many areas to consider, it makes sense to get clinical supplies/sourcing involved early on in the clinical trial planning process once we have an idea on what are the possible comparators, even if we don't know what the comparator is going to be in the final trial design. It takes time for these guys to source, blind and repackage the comparator product. During this process, there may be additional development work required to ensure the method of blinding has no or little impact on the stability, bioavailability and dissolution profile of the comparator drug. The time and effort required in gathering all the necessary regulatory documentations should not be underestimated either.

Saturday, June 25, 2011

Grocery Shopping & EDC

I consider myself as pretty technology-savvy, and barcode scanners really shouldn't be a problem for me. However, this morning I was cursing the grocery store that had these self-service checkouts installed.

I usually do my weekly grocery shopping at Waitrose which has handheld barcode scanners, so shoppers would scan the items and put them into shopping bags as they go round the supermarket. At the end of shopping, just scan the "finish shopping" barcode at the checkout machine and pay. Quick and simple - no need for taking all the shopping items off the trolley and repacking at the other end of the conveyor belt. When I first joined this scheme, I had a quick 2-minute training by the staff, and everything on the scanner was intuitive. When I have a problem with scanning any items, I just go to the checkout and always get this sorted quickly by the staff. It almost makes the chore of weekly food shopping fun!

This morning I went to another supermarket just to pick up a few items. This shop installed the self-service checkouts. The process has not fundamentally changed from the manned checkouts - you go round the supermarket, put the items in the trolley, then take them out of the trolley at checkout, and then put them in bags.

So what has this got to do with EDC (Electronic Data Capture), I hear you ask! With the changes in the way clinical trials and data management are performed, EDC is becoming more and more commonplace. Many people naively think that once the technology is in place, the benefits can be seen immediately without examining (and possibly changing) the process and also putting in place the right support. I am going to draw some parallels to my grocery shopping experience:

(1) Process: The idea of self-service scanning (whether using a handheld scanner or at checkout) is to reduce the long queues often found at checkouts. Most supermarkets have opted for self-service checkouts but the fundamental process of shopping has not changed. Waitrose, on the other hand, must have stood back and  looked at the whole process, and has opted for a change in the way people would shop. From what I have seen, Waitrose seems to have managed to cut the queues dramatically because customers can put their shopping into the bags as they go along, and once they have finished, they just pay and go. In the drug development world, you cannot simply introduce an EDC system based on a paper-based CRF process, and you need to stand back and look at how the technology may affect various parts of the business, not just data management. I have seen a company suggesting to me that they needed to print out the eCRF at the end of the study and do a QC against what's on screen - when I asked why, they said that it was their normal process in a paper-based study and they just carried on with EDC studies!

(2) Training & User-friendly device: I was given a very short but helpful training on the handheld scanner the very first time I used it, and by the end of the first shopping adventure using this device, I was already pretty proficient with it. The scanner is simple - only a few functions and any messages on screen are pretty self-explanatory. Can't scan the item? No worry - just take it to the customer service at the end of the shopping and they will sort it out. With the self-service checkouts,when the barcode scanner going blind and doesn't recognise the item (so you can't proceed any further until an assistant comes and helps you, which usually means a 5-minute). or when the phrase "unexpected item in the bagging area" pops up on screen (or worse when the machine announces that loudly), you can't figure out what was the unexpected item. In the EDC world, if I were an investigator, I would find it frustrating if I got bombarded with complex messages in a system that has a steep learning curve. With data getting more complicated, we have a tendency to design CRFs that are too complex - ask yourself this question: Do you really need to collect all these data (or are they "nice-to-have" or "just-in-case-we-need-it")?

(3) Workflow: There was one time when I was expecting a mixture of notes and loose change for my shopping, and I spent a good minute looking for my £5 note close to where the machine spit out the coins, before realising that the note was in an obscure place on the other side of the machine! With EDC, the navigation of the eCRF needs to be clear and intuitive, and user testing would help.

(4) Support: With the self-service checkouts, it seems that there's usually one assistant assigned to every 6 checkout machines. As many customers seem to have one problem or another with the scanning, the poor assistant is always rushing around trying to firefight. To date I have only managed to complete the whole scanning and paying process without the intervention of the assistant once! The record waiting time for an assistant to help was 5 minutes - I nearly gave up and walked out of the shop. Similarly with EDC, user support is important.

Although I have been comparing my grocery shopping with EDC here, the same lessons can be applied to other technologies, whether it's in clinical trial or in other business areas or industries.

Wednesday, June 15, 2011

IT Tips for Start-up / Small Businesses

When you move from a corporate environment with dedicated IT support to a small business start-up (with a few people working together) where you suddenly have no in-house IT support and have to be even more vigilant about spending money, you can either spend the money to replicate your previous working environment, or take a step back and look at what's on offer and see what can make life easier. Here are a few applications that I have found useful:

E-mail/Calendar: Most corporate users are already familiar with Microsoft Outlook or Lotus Notes, but the infrastructure to support these software don't come cheap. If you are already using Google Mail or Yahoo or another web-based e-mail, then you have already experienced a bit of cloud computing. With Google, you can set the e-mail up so that you can have the e-mail address with your own domain name rather than the usual or The beauty of this is that you don't have to carry your computer around to download and check your e-mails, and you can check your e-mails and calendar using most of the mobile devices as long as you have access to internet. Google even have an offline access so that you can download the e-mails and work on them if the internet access is unreliable (eg if you are on a train and about to go through a few tunnels!)

Phone Calls: Many people are already using Skype for PC-to-PC calls, but it's worth purchasing one of the Skype paid plans to allow you to call landlines or abroad cheaply. This is especially useful if you have to call a "toll-free" number abroad - with your normal phone, you still have to pay, but with Skype, I have participated in many conference calls outside the UK. Skype is also great if you want to do a teleconference with people from multiple locations. For a PC-to-PC call, you can also share the screen with the other user which is handy if you need to review something together or show the other user how to do something on screen.

File Sharing / Collaborative tools: Most people still like sending files by e-mail but if this can quickly get out of control if you are working on draft documents and the files are updated constantly by multiple people. Even if it's on a server, usually only one person can edit the document while the others have read access only. With cloud computing, multiple people can edit the same file and the updates can be seen in real time - making the collaborative effort a lot easier, and since there's only one file, you don't end up with multiple versions with the risk of the modifications spreading over different versions. Google Docs is one of such tools available if you want to do simple word processing, spreadsheet or presentation stuff. It can also generate a representation of your Microsoft Word, Excel or Powerpoint file but the layout may be messed up depending on how you upload the files. One way to really combine Google Docs and Microsoft Office and harness the power of the combination is to have Google Cloud Connect or OffiSync - imagine using Google Docs as simply your file sharing area, and you still use Microsoft Office to open and edit your documents. Obviously Google Docs or other online file-sharing applications (eg DropBox, etc) may not be suitable, or you need to assess the risk vs benefit, if you need specific server configuration to meet certain requirements or compliance.

Web Conferencing/Webinar: If you need to share slides or screen with a group of users over the internet, many corporate users immediately think of WebEx. However this may turn out to be quite costly, and there are a few applications around that can do an adequate job: one tool I have tried out is Present Online Now which allows you to set up a web conferencing room very quickly and it's even free if you have no more than 10 participants at any one time.

There are many other open source or cloud computing applications that are cost effective and can simply things in life (eg before you compile a large spreadsheet with your customer contact details, consider a customer relationship management tool like Insightly ; or if you just need a tool to convert files to PDF occasionally, consider PDF 995). Google Apps provides a wide range of tools to help you to get started in your business, from Gmail to Google Docs for file sharing to Google Sites for a very intuitive way to build websites. Keep an open mind and look around - I'd be interested to hear your experience and thoughts, especially if you are a small business or about to leave the corporate environment to set up your own business.

Monday, June 13, 2011

CDISC - Is there an alternative?

In the past few years, on several occasions, I have been asked the question on whether there's an alternative to CDISC?

The question comes from people from different backgrounds and affiliations: Large companies with their own well-established data standards, small companies with no internal standards and do not have the resource, knowledge or motivation to adopt such standards, companies who recognise the need of standards but feel that the CDISC standard is still changing constantly and prefer to wait, and companies which have been on the journey of CDISC adoption but have found that it was a rather bumpy ride. It seems that the majority who asked the question prefer to maintain the status quo - after all, we have managed to submit clinical data to FDA and obtain regulatory approvals without any data standards up to now, why should we go through the major upheaval and invest the money and time on adopting a standard that is still evolving?

The FDA Critical Path Opportunities List report has already mentioned CDISC SDTM (see item 44 in the report) as a data standard to streamline drug development and for exchange of healthcare information. In addition, the CDASH initiative in CDISC will help to promote the consensus on standards for case report forms (item 45 in the report). Development and implementation of such standards cannot be achieved overnight, but not embracing it means the risk of falling behind your competitors and greater difficulty in the path to regulatory submission and approval.

With the pressure on healthcare costs, outsourcing and offshoring seem to be the way forward to reduce costs. However, I believe that efficiency gain by the implementation of standards at various stages of the clinical trial process will drive the cost down also. For a sponsor, while you can create your own "alternative" standards and insist on all CROs to follow it, you need the resource to maintain such standards anyway and you would lose out on the leverage you can get from CROs that are familiar with the CDISC standards. There is a cost associated with the implementation of CDISC standards initially, and this investment is only worthwhile if you have a long-term vision and commitment to the implementation and adoption of standards. Of course when and how you implement CDISC will be dependent on your own circumstances, and you will have to decide the choices of CDISC standards at various stages e.g. you would probably not want to convert everything to CDISC from your legacy company-specific standard if you are just about to submit your drug to the FDA in the next few months (unless the FDA has asked you to provide the data in CDISC SDTM format).

When there are multiple standards, one always comes up trump - remember the good old days of VHS vs Betamax video tapes (for those old enough to even remember video tapes), or more recently blu-ray vs HD-DVD discs? As far as I am aware, there is not another standard that competes with CDISC currently, and even if there is, it will have a pretty high hurdle if it is to come out on top of CDISC, as it does not only have to demonstrate that it has a distinct advantage over CDISC, but it will have to compete with the large number of CDISC followers to gain acceptance.

Saturday, June 11, 2011

Welcome to the Datacision blog

Welcome to the blog at Datacision Limited!

I have written various blogs over the years (most recently my Chow With Chau food blog) and it's about time that I start one for all work-related stuff that I am passionate about. This could range from data management and statistics issues faced in clinical trials and drug development, to my views on turning data into knowledge that would play a key part in making informed decisions, to my experience in running a small business.

This is not a one-way communication - I hope that you will share your thoughts and experience with me, and I would appreciate valuable comments that you leave on this website. So, enjoy!