Spring 2007
The Acquisition and Management of Clinical Research Data in the 21st Century: A New Practical Strategy of Intelligent Data Acquisition and Management (iDAM)
In reality, current eDC is predominantly eDT (electronic Data Transmission). Most clinical trial data are first recorded on paper. Furthermore, while eDC can provide benefits over paper based data transfer, as many have found to their cost, it also brings many practical problems and shortcomings. Why? eDC today fails to address the real world and it fails to address the whole clinical trial data capture and management ecosystem from the clinic, to the patient, to the sponsor.
Background
In order to meet the competitive pressures of today’s pharmaceutical industry, Pharmaceutical and Biotechnology companies are seeking to collect quality data from clinical trials more quickly and more cheaply. In addition to shortening the clinical trial life cycle, the accelerated and accurate availability of data facilitates more complicated adaptive study designs that allow early decisions on the clinical development of individual compounds.
The pharmaceutical industry has been working on a solution to these problems for many years. In the 1980s the concept of computerised remote data entry (RDE) at site was thought to offer a solution. Although RDE systems were used selectively they failed to make a significant impact and the majority of sponsors reverted to paper when the technology proved to be inadequate.
The modern variant of RDE is electronic Data Capture (eDC) mainly using web-based applications at Investigator sites.
Generally, these systems were developed as part of the internet boom of the 1990s and many US $100’s millions have been spent developing the software and marketing it to users. Although use of eDC is growing, only a fraction of clinical trials collect data using eDC methodologies. Despite some benefits of eDC such as fewer nuisance queries, consumer (Investigator, Pharma and Biotech) resistance remains and the benefits of faster data capture and lower cost have proved elusive. Indeed, some of the shortcomings of the web based approaches have been recognised by the PhRMA EDC Task Group that provides industry guidance on eDC, who announced in March 2005 that they have redefined today’s eDC technology as “electronic data transmission” (eDT). Investigator sites do not use today’s systems to capture data, rather to transmit it from often “CRF like” paper records, where the data are originally captured.
In the setting of recent dramatic progress in the technological world more generally, it is surprising that the building blocks of clinical data capture and data management are essentially unchanged. Although web-based eDC systems have potential, their often very restricted data management capabilities means legacy clinical data management systems with origins dating back to the 1980s still remain in the background, tethering progress to the past. To make matters worse, the number and type of data sources have increased dramatically in the last 10 years, and so too have individual software systems to manage them (e.g. electronic hospital records (EHR), IVRS, biomarkers, medical devices, patient eDiaries). Acquiring, storing, transmitting, merging, validating and checking all the associated data and meta-data from these sources often reduces speed, requires more resource and causes quality issues.
A paradigm shift for the 21st Century
To achieve significant progress industry professionals should recognise that the increasing number of distinct systems involved and the wide range of types of data stored (information, status, financial, clinical data, etc.) overcomplicate and obscure the true objective of clinical data acquisition and management which is to collect, check, store, review and analyse clinical trial data from multiple sources, accurately, efficiently and quickly. Involved in this consideration is the practical acceptance that integration of EHR systems will be unrealistic on a meaningful scale for decades.
The second major shift in thinking involves considering not best in class of current technologies being employed, but instead to look outwards and forward to the ongoing 21st century information technology revolution. The world of wireless data communication and portable, powerful, lightweight, disposable, low cost hardware is now readily available, it is robust and simple to use (e.g. cell/mobile phone).
The third shift in thinking is to place the Investigator sites and their normal clinic workflow at the heart the solution. The location and nature of the doctors and clinics used in clinical trials has never been so varied. The solution has to work in their world, for them. Wired broadband with secure, reliable connectivity and dedicated workstations are not the norm around the world and hinder rather than help a normal busy clinic staff with the treatment and management of their patients.
Once accepted, these views allow much greater clarity of thought. A “fit for purpose” solution becomes achievable.
The new paradigm has to be based around a single data acquisition and “cleaning” database that unifies the whole clinical data ecosystem, from Pharma companies or their CROs, to patients in their homes, to clinics, to central laboratories etc. The use of multiple systems, technologies, different databases, data transfers, integration etc. should be avoided where the clinical data are concerned. Instead, the problem can be simplified by designing a single system that obviates or automates whole process steps and that is capable of being simply configured to work flexibly in any data acquisition capacity (eDC via cell phone or web, paper CRF data entry, direct data capture from medical devices, direct downloads of lab data etc).
This goal can be achieved by carefully choosing modern software technologies that better meet these requirements and by design principles able to build multiple user interfaces and expose only functionality to meet the needs of that particular user. Each interface must avoid users having to search for functionality they need or rummage through what they don’t need. For example, a data management user interface that helps a data manager do their job quickly, an Investigator user interface that computer naïve study nurses feel confident to use in front of a patient. The concept behind the technology is exemplified and explained further in figure 1.
The clinical data ecosystem model shows that clinical data comes from many sources and it can require merging/re-formatting/mapping between systems before the data can be used. The unification of data acquisition into a single database is a key element in breaking away from the status quo where numerous systems have legions of people working to integrate, and validate data to create a final single database. The concept of a single database accessed by a range of user interfaces is exemplified in figure 2.
When a significant change in technology and process is proposed it will only be accepted and used if the users are involved at every step of development. In this area the highest priority should be given to the needs of the Investigational site and patient as the main users/customers of the solution. It is inevitable that one data capture methodology will not be ideal for all sites, in all countries, in all indications, in all trial types so any solution must seamlessly cope with a variety of data capture methodologies – eDC by web or cell phone, PDA diaries, medical devices, and even the traditional paper CRF.
Of course Sponsor/CRO users must be considered. However if the full benefits of technology are to be realised, Pharma company/CRO management must accept that all current processes and job roles typical in clinical development may be challenged and changed. The new technological system is only a component of the overall solution, and it is the use of modern workflow thinking in system design that drives simpler processes, greater efficiency and easier compliance.
The concepts described above take thinking beyond the current view of eDC or electronic data transmission, and in order to separate these ideas from existing models (and acronyms) the new paradigm has been called intelligent Data Acquisition and Management (iDAM).
The design and realisation of an iDAM system
iDAM is no longer a concept. An iDAM system has been developed, with extensive input from Investigators and sponsor users in a series of focus groups. In 2002 a prototype system was tested at a global workshop of Investigators, site staff and study administrators. The information generated became the guiding vision for the system’s appearance and functionality, with reassessments after each study by the users. Today, it is a validated, regulatory compliant live system, called ThirdPhase, with the many anticipated benefits already achieved and proven.
The design principles for the iDAM solution have been simple and clear. These include:
Build one system with one database and one set of validation checks, so that the “curious” effects of mapping between eDC and data management systems are eliminated.
Include full enterprise data management functionality, including auto-medical coding; laboratory data loading and data review even on a portable stand alone laptop.
Hide the complexity of the technology from the user to make the user forget that they are using a computer.
Focus on those parts of the clinical trial process that are on the critical path (e.g. database configuration, trial set-up, locking, exporting and query process).
Simplify the whole data acquisition process so that the system can be installed at site without having the need to impose on the existing infrastructure of hospitals, clinics, laboratories or patient homes.
Develop the system to comply fully with 21CFR11 (audit trails, independent testing and validity) and also with CDISC and other emerging standards in mind.
Always select the “best of breed” for technology selection, technology architecture, methodology and approach and make it “fit for purpose”.
Support the technology already present and required in the real world. For example, include the ability to support current and future paper CRF usage while providing eDC wirelessly and over the web.
Allow the iDAM system to also be a patient e-diary, or handle paper diaries.
When developing system architecture: be smart and parsimonious – make the system as complex as required, but no more complex.
Develop automatic, multiple communication modalities (internet, cell GSM or GPRS, 3G etc.).
While the system will still be able to handle entry of data from paper CRFs it should also allow investigators who prefer to use their own computer hardware to submit data, as well as function on hardware provided to site – a clinical trial appliance if you will. The appliance should only need to have power and then be turned on and off by investigators, with no need to worry about the web or communication. A parallel for this approach is a comparison of the mobile or cell phone with the written letter (Paper CRFs) or land telephone line (web based eDC).
In addition to the basic software and hardware design a number of additional features were added:
Capability to capture data directly from medical devices (eg EKG).
Bar code readers – user id’s are provided by a bar code placed on a credit card size user card (no id name to forget, that leads to investigator frustration and the majority of helpdesk calls).
Finger print biometric to provide authorised access to the system (no password to forget with the advantages described in the bullet above).
Capability to capture data directly from medical devices.

Finally, the ability of the system to handle scale, enterprise deployment, stream data into “data warehouses” and similar important requirements were included. The solution had to be practical, as well as “intelligent”.

Is an iDAM solution possible and does it deliver the predicted benefits?
The authors have participated in the development of an iDAM system called ThirdPhase, able to unite the collection of data from all of the data sources described in Figure 1, within a single database. Its novel data structure and use of XML allows databases to be built more efficiently. Also, there is only one database with one set of edit checks for CRF data no matter what the source of the data – paper CRF or eCRF. Indeed the routine production of study databases is within days, in contrast to the authors’ previous experience of this activity routinely taking weeks. The use of a single database has also made the hybrid use of paper CRFs and eDC in the same study straightforward
The use of up-to-date IT architecture has facilitated the use of the clinical trials appliance concept. Studies can be completed in which all the study data are communicated wirelessly. Further, changes to the CRF in response to protocol amendment have also been communicated wirelessly to appliances at site without the investigator’s participation. This ability has allowed sites to take part in eDC studies without the use of their own clinic IT infrastructure ( i.e. no need for internet or broadband access at site). The appliance has also allowed investigators to enter data while sitting with patients, bringing eDC and e-source possibilities closer, rather than eDT.
The literature has many, broadly consistent, descriptions of the benefits of eDC in terms of reduced query numbers. In our hands a similar level of reduction in query numbers has been seen. In post-study assessment sessions sites have praised the speed and ease of use, of ThirdPhase, the ability to load laboratory data centrally and so allow them to see this data in real time. Many have also appreciated the replacement of passwords with biometric fingerprints, not just for system access but for digital signatures. The in built data management system functionality such as auto encoding of AEs and treatment terms, its fully portable nature and ease of access to data reports and data export has proved to be useful to Investigator sites, data safety monitoring boards and Pharma alike.
The involvement of Investigators in the design of the system has been most helpful; for example their request for the use of finger print biometrics to replace passwords has reduced the number of helpdesk calls. Indeed, studies have been completed without a single helpdesk call. The benefit of the involvement of data management users in system design has also helped; the data managers working with ThirdPhase on paper studies have reported taking less than half the time to complete most tasks in comparison to existing data management systems.
A formal quantification of the operational and financial benefits of the iDAM approach is not yet available, nor is the purpose of this article. However, we can report a successful design and implementation of leading technologies from other industries to produce a sophisticated, fully functional data management system that makes the iDAM concept real, and to the best of our knowledge and our collaboration partners, quite unique and like no other currently available. The data acquisition and management of over eighty studies has now been completed using the system, with very encouraging results.
Conclusion
The new iDAM strategy offers a new approach to the acquisition and management of clinical trial data that escapes the hindrance of legacy systems. The strategy to build a “fit for purpose” solution has been taken forward and developed into a fully functional and validated system that has completed over eighty studies. Initial evidence suggests that the system will deliver the benefits anticipated, further development of the system is underway to make it completely fulfil the complete promise of the iDAM strategy.
David Connelly, PhD, CEO
Andrew Griffiths, PhD, COO
Cmed Group Ltd, Horsham, UK