|
Spring 2010
Issue No. 2
Xtra
Daljit Cheema, Vice President of Technology at Perceptive Informatics, discusses the essential platform components needed to build for eClinical.
Current state of the market
Clinical development requires the use and coordination of a range of applications. They include data acquisition and management (electronic data capture, medical imaging, electronic patient-reported outcomes and safety management), trial control and logistics (randomization and trial supply management), planning and administration (clinical trial management systems, document management and simulation) and data analysis and reporting.
Many organizations have procured these applications based on the relative merits of an individual application in comparison to its alternatives, thereby supporting procurement based on a “best-of-breed” selection process.
How do we improve end-user experience and productivity of trial conduct?
When looking at adoption of technology in a complex domain such as clinical research, there are several challenges that need to be overcome. Some advances have been made by standards bodies such as HL7 and CDISC that have primarily focused on data interchange and submission standards.
Whereas this cooperative standards building effort has improved some of the back office and planning aspects of the data processing workload, the emphasis on the end-user of the applications (investigators, pharmacists, trial managers, monitors, other sponsor staff) has largely lagged behind in terms of features that make their jobs simpler and more productive on a daily basis.
In reviewing the needs of users of clinical applications, there are some basic principles that would make a difference to the end-user experience and productivity. We are after all interested in efficiency through automation as well as the effectiveness that results from better control and management of information and improved decision making.
Some of these principles would include:
Minimization Principle - Reducing the number of applications that a user has to access or minimizing the need to learn different multiple application interfaces would reduce complexity.
Usability Principle - Making systems that are intuitive to learn and use and fit for purpose. Some of the other more obvious attributes around ‘fit for purpose’ include reliability, availability and scalability which are critical and will affect end-user experience when they do not meet expectations.
Value Adding Principle - Probably the most complex principle to decipher because of the various dimensions that it comprises. From a computing perspective, this represents a spectrum of potential benefit derived from basic automation, interoperability via data and process integration and to more complex interactions between systems that share and understand the meaning of data that they process (semantic interoperability).
Addressing these three principles helps us to design applications and suites of solutions that improve user experience and productivity. Perceptive’s definition of eClinical is focused on addressing this objective. At the highest level, we consider eClinical to mean business process-driven solutions optimized for the way our customers manage trials. Perceptive’s eClinical solution development is geared around not simply data integration but how to make the process of using technology, in particular multiple technologies, simpler for the site and sponsor end-user. To deliver this requires an enabling infrastructure and platform.
eClinical platform and essential components
The construction of target solution architecture needs to take into account the state of play of the vendor landscape, investment decisions and current application portfolios while building for a better future state. Some of the desirable characteristics that this platform would encompass include:
Suite of Clinical Applications - This can be thought of as the clinical multipurpose tool where distinct applications are used for specific purposes. An application built for complex image processing may have very different characteristics to other applications used for web data entry or other multichannel applications like IVR/IWR systems. A one-size–fits-all paradigm does not sit comfortably with the range of data processing needs found in the industry.
Common Interface Layer - Standardizing user interfaces has proven benefits when you look at the past successes such as the Windows® operating system that spawned third-party applications and many successful application companies providing software for Windows® operating systems. Users were presented with applications with common user interface elements which reduced some of the learning time required. The success of the web in itself is down to the relatively small number of standards that are successfully used to deliver a rich set of interactive applications. Other interface models include the use of portals to aggregate and unify access across a range of applications.
Information Access Layer - The lifeblood of any application is the ability to access the information it manages. This includes basic operational reporting through to rich active dashboards that may aggregate data from multiple reports or systems providing business intelligence. In addition, the ability to access both canned reports as well as create ad hoc reports and extract data for offline use are seen as important to empower users.
Integration Layer - The integration layer likens itself to that of a road network where major trunk roads would provide the main arterial channels for traveling from one part of the country to another. This distribution system supports a wide variety of vehicles with different payloads. In a similar fashion, the integration layer would facilitate both data and process integration between different applications and provide a reliable mechanism of tracking such interactions.
Repository Layer - This layer has the responsibility for the storage of data and provides mechanisms for access and maintaining data integrity. It needs to consider both the immediate operational needs of an application in addition to longer term retention and support restructuring of data destined for data marts and warehouses.
Security Layer -Provides the ability to control access to applications and data based on common security models such as role-based security and different levels of granularity from coarse grained (such as access to an application to view data for all system users) through to fine grained where access to an individual field of data might need to be restricted. A constant source of frustration in the digital age is the numerous passwords and user credentials that need to be remembered and maintained in order to access different applications and in some cases the same applications but different clinical trial projects. A first step in reducing some of this complexity is for systems to embrace more generic methods of managing a user identity. This would allow a single identity to be associated with a user which is then used to provide a common model of authentication across a range of applications. What a difference it would make if the user needed only one identity and one password to access a whole range of clinical applications.
Infrastructure Layer - This layer provides the means in which software operates and needs to be designed to support the operational and service level needs of each application and understand the business, technological and regulatory risks associated with electronic information processing. The support of multiple deployment models increases the complexity since multiple configurations have to be tested, certified and supported.
eClinical has multiple definitions in the marketplace today. This article suggests that from a technical perspective eClinical should be viewed as a composite computing model that has to be engineered to provide a broad range of services and highlights many of the essential components required to assemble this platform. |
