I joined our Clinical Research Unit in Vietnam as an IT assistant in 2002. At that stage the role was to provide IT services and support to computer users. Besides helping scientists with their computer issues I was asked to create databases for clinical trial data entry by using either Foxpro or Access. It seemed an ordinary task within IT support because at that stage I had no idea about clinical data management, clinical trials or good clinical practice and all the specific requirements of clinical trial data bases.
Everything changed when I was appointed as the Regional Data Manager for the South East Asia Clinical Infectious Diseases Clinical Research Network (SEAICRN). As the Data Manager my role was to lead the data management team in Vietnam to handle data management activities and provide support to multiple sites within Vietnam. This was not only a challenge but also an opportunity for me to get involved professionally in clinical data management because we had to conduct the trials complying with such international standards as ICH Good Clinical Practices and FDA’s Title 21 part 11 which is about electronic records and data capture. I attended a 3-day workshop of Research Ethics and Good Clinical Practice that enabled me to better understand this interesting subject of clinical research. It was also my first introduction to the guidelines that ensure clinical research trials are conducted ethically and gather high quality data. I also had an opportunity to visit FHI (Family Health International) for 'Lead Data management Training' and meet international recognised data management experts. Here I learnt about the types of systems and practices used in actual trials in many varied settings across the world. It was the first time I used and had exposure to a real data management platform, in this case it was the commercial product called ClinTrialTM . This was an impressive platform that was designed specifically for clinical trials so meeting the guidelines whilst also being functionally good. As that point I realised the benefit a system like that could bring to our clinical trial programmes in Vietnam.
Following this I became head of our newly established data management department and I was given the mission of building a professional team to provide data management services according to international best practices. The new data management centre had been working on ClinTrial as a remote data entry system for SEAICRN studies and provided data management services for scientists for other studies. Alongside this MS Access was still our core application for developing databases for trials. Unfortunately this is a general data management package and it is very difficult to programme in such as way that it would comply with Good Clinical Practice and regulatory requirements. Activities such as double data entry, concurrent access, audit trails and quality checks are important yet not practically feasible and technically very difficult using non-specialist software. We had reached the point where a centralised data management platform was required for my department, but programmes such as ClinTrial were too expensive for an academic group such as ours. I had tried to determine whether any similar systems were available in Vietnam that we may use but the result was very pessimistic. Finally, I took a difficult and challenging decision and set out to design and develop our own centralised data management system. As an IT expert in database management and software development I was confident that I could lead my team to a successful project. However the reality was that it was a challenge.
I began by analysising the requirement specifications for the system, then selected the database server and programming platforms and finally drew up and implementation plan for the enterprise. Requirement specifications analysis was the most critical part of the work as it would determine the final outcome of the project. I studied the requirements of ICH GCP guidelines and FDA Title 21 Part 11 – Electronic Records, Signature Records. I also reviewed and assessed databases I had worked on before to develop a framework of what functions the system should provide and how it might comply with international best practices. This process led to these key requirements needed for the project:
- A centralized databases system to manage all databases for clinical trials.
- A secured web-based application that can be accessed from anywhere.
- With bilingual interfaces.
- To support electronic data capture from sites or fields and double data entry at data management offices.
- To have tools to manage data entry processes such as case report forms tracking and reconciling, data entry discrepancies resolution, data entry queries, etc.
- To have tools for QA-QC monitor such as real-time data quality check to validate the data during data entry or the data validation for post entry to validate the data consistency after it has been entered into the database.
- To be able to trigger all data changes and generate read-only audit trails.
- To have the capacity to be specifically programmed for some trials for special feature like a calendar, email, etc.
- To have dataset extraction tools for PI and biostatisticians.
The next decision was to select database server and programming platforms. This was not a hard decision because we had an installed Microsoft SQL Server for other applications in our information system and we did not want to have other server platform as that would be complicated. We then selected Microsoft Studio.Net as our programming language, IIS (Internet Information System) for web server and Microsoft ISA Server for the Firewall. And then, the most time-consuming and difficult work started - the actual designing of the system and database infrastructure and programming application.
It took us nearly two years to produce the first mock-up and during this period we encountered many difficulties and challenges. Our team had just four people. I designed the system infrastructure and led the project, we had one graphics and web designer who helped to design web layouts and make up the interfaces and we had two core programmers. All this was being conducted alongside routine ongoing IT and data management work. This project was technically very challenging as we needed to produce a database application that would be adaptable for different trials. Each trial has its own design of data collection forms or CRFs and those forms vary from research areas such as TB, Dengue or respiratory. For each trial a data capture form is designed and then the database structure for that trial is developed from the data collection forms with the necessary added features such as audit trails, data validation, double data entry and batch tracker. Batch tracker will work for all generated databases.
In our region many languages are spoken and therefore bilingual support brings another issue. We needed to think about language at the outset as written text is everywhere within a database. These are just some examples of the issues we needed to resolve.
At last it was built and we were ready to use the system live with the first trial. It was a dengue study. We were pleased that the investigator allowed us to pilot the application with her study. The outcome was very good and so both the investigator and I were happy to work on the new platform for the whole study. We began with the final version in October 2009 with this dengue trial. We are now delighted to report that is this one of ten studies currently having their data management run with this new system.
But what is the system called? We have still not decided and I am calling it CliRes until a better name is found. Does anyone have any suggestions? We write this guidance to firstly set out the challenges involved in clinical trial data management and secondly to encourage others to review their data management systems as it is an important part of running a high quality trial. This has been a good solution for us and we would be pleased to share our experience and resources with others if you would like to get in touch through this programme. There are now also open source solutions to clinical trial data management such as open clinica. We do not have experience of this but others have written about it on this website.
This is a very insightful piece Hien, and thank you for sharing your experience. It is interesting that the same challenges exist for other research groups and having headed a data management unit, I do identify with this. Against so many odds e.g. limited resources you have come up with a custom solution that works for you and this is very encouraging for other data managers seeking data management solutions.
Thank you very much for sharing experience which I enjoyed because it reminds me of me. I am also computer training who took the head of a department of clinical data management so I understand how difficult it is to integrate into the field of public health. I have also wanted to get involved in the design of a centralized system for managing data from our research center, but I realized I could make OpenClinica with more time and spend less money.
We congratulate Hien for writing this article. It is very helpful for several reasons. Firsly is provides a first hand example of how data management is a difficult problem. This is a common discussion topic on this website and so this account is very instructive for others. The reviewers also felt that this article gives encouragement to those working as data managers as this is a good career within clinical trials and Hien is certainly as success story himself who can be an excellent role model for others.