Planning and Running the e-Clinical Trial

The interest in exploiting the Internet and its associated technologies to revolutionize the conduct of clinical trials continues to be discussed at length in journals and at conferences.^1-4 Although increasing numbers of clinical trials are being commissioned and performed as e-studies, the general take-up is still low when compared with traditional methods. Estimates suggest that approximately 5-10% of all clinical trial data are collected and managed currently as e-trials.⁵

Figure 1. Standardized framework for the implementation and operation of an e-clinical trial.

Key earlier, practical reasons given for not undertaking a study using e-clinical trial methods have been addressed and solved. Thus, e-trial-enabled software from various vendors is now readily available, and the regulatory issues, mainly concerned with subject confidentiality, have been addressed.^6,7 Various authors and speakers have demonstrated that e-trials can bring value to a study.^8-10 The improvements that may be seen are usually in speed (and availability) of data collection and collation, increased data accuracy (fewer clinical data queries), and benefit of the availability of the immediate communications infrastructure.

Why, therefore, is the uptake of these approaches still so relatively low? A number of factors may be inhibiting a more rapid uptake. First, e-trials require a fundamentally different approach to both the design and procedural aspects of a study when compared with traditional methods. This is often organizationally difficult to achieve quickly, and the difficulty may be exacerbated if previous commitments to other technologies over, for example, a five-year cycle have been made. E-trials are often presented as too expensive because of the higher up-front costs associated with design and implementation. And the real or imagined risk of committing "my study" to these methods also suppresses uptake. An external factor of significance is that investigational sites may not have the will or the skill to use these methods, and sponsors are unwilling to invest in the extra training costs required.

The very wide range of areas where e-approaches can add value often further complicates these issues. For example, the study design can benefit (by using sequential statistical designs), the research data collection can benefit (by using an electronic case report form [e-CRF]), the management of study sites can benefit (with documents online, direct online monitor reporting), the drug supply chain can benefit (using online requests and batch confirmation), and subject medical monitoring can be improved (by earlier adverse event and laboratory data alerts). The value of using an e-trial approach can then easily be lost in the complexity of integrating so many specific areas where value can be added. Finally, it is now well understood that the introduction of new systems without parallel process changes will not guarantee operational improvement.

This article presents a standardized model to help with reviewing in part or whole any function that could benefit from "e-ization." The key objectives of each element of the model are presented, together with a discussion of key practical factors that will lead to success in implementing and running the e-clinical trial.

A standardized framework

In order to enable any clinical trial function (whether it be a specific part, the whole trial, or the trial plus all the support functions) to benefit from e-ization, it is essential to have a common framework for developing, reviewing, and testing each required operational element or the system as a whole.

Figure 1 presents such a standardized framework, one that can be used to review and develop an e-clinical trial. The model consists of seven top-level interacting elements:

• Validated processes

• Human resources

• Qualified staff

• Design and build

• Production configuration

• Production processing

• Close and archive.

The first three elements represent the key implementation and support functions that an e-clinical trial requires, while the last four elements represent the e-trial project processes. For any project each element must be reviewed, put in place (if necessary), and its products and procedures validated.

In the figure, the links between the elements represent a requirement by one element for a product of another element. Links can be uni- or bidirectional. Orange lines show elements requiring products of the validated processes element; green lines show elements requiring products of human resources and qualified staff elements; blue lines show elements requiring products of design and build, production configuration, production processing, and close and archive elements; and purple lines show production processing elements requiring repeated use of those processes.

Table 1 Involvement of different operational groups in e-clinical trial implementation and operations

The model is of value in developing and operating e-clinical trials for the following reasons. First, it can be applied at the micro and macro levels without modification and used to support small or large projects. Second, the output from any element is a product (not a link between processes), and therefore does not require procedures to be explicitly stated. Procedures emerge from the requirement that one element in the framework requires a product from another element before it can proceed, and that the creation of the product must be controlled by procedures. Third, the model recognizes the need to provide properly project-qualified staff to the e-trial, and these staff must meet both basic qualification criteria and be available to work on the project. Thus, a box around all seven elements can represent either a department (responsible for implementing and running e-clinical trials), or a specific part of the trial, for example the collection of just the research data using an e-CRF. In each case, all seven elements apply and need to be addressed to meet the objectives of the exercise, although to different degrees in each case.

The model also accommodates and easily assimilates other views of the e-trial processes and outcomes that a clinical trial requires. For example, the requirements demanded by computer systems validation, where other standardized models are available (examples include installation qualification [IQ], operational qualification [OQ], and production qualification [PQ]), immediately become part of one or more of the frameworks, elements.¹¹ (In the case of Stokes's model, IQ and OQ become part of validated processes, and PQ becomes part of production configuration and production processing, respectively). Investment in other approaches is therefore maintained, and, it is hoped, enhanced.

Unlike process-oriented methods, this method operates fundamentally with work products, not procedures. The links between elements represent a product requirement, not a process flow. For example, Figure 1 shows that before it can proceed, the production configuration element requires products from the design and build element (such as the online data collection system, which must be built before the production configuration element can be configured for a specific investigational site). Although the model does not demand any order to be stated, in practice an order naturally emerges. Thus, the last four elements of the above list represent the central operational or project process, which starts with design and build activities, whose products are used by the production configuration and production processing elements, which in turn provide the products required by the close and archive element. These elements therefore represent the core intent of the system (in this case, running an e-study). The other elements support this core set and, depending on the specific case, may be required to provide products to these four elements before they can begin any activity. The most critical is the validated processes element, which can be linked to all elements (because documented evidence is required to show all e-trial functions are operating as intended).

TABLE 2 Key objectives and products of model elements

In contrast, the human resources and qualified staff elements, representing the staffing requirements for the e-trial, do not have to be addressed at any particular time. The model clearly shows, however, that a failure to address them prior to the activating core element tasks will lead to problems with project resourcing, both in terms of availability and qualifications.

The rest of this article uses these elements as the basis for identifying and discussing the factors for success when implementing and running an e-clinical trial. It discusses the key objectives and success factors from the perspective of a department initiative to set up and run its first e-clinical trial.

General success factors

For any e-trial, three general factors will always govern the overall success. These are that it is known how e-izing any component of the trial will benefit the overall outcome.

These are that

• it is known how e-izing any component of the trial will benefit the overall outcome.

• it is understood who is doing what, and when they need to do it.

• all of the tools (systems), training, and study configuration are completed and ready for use before being needed for the first time.

In a recent paper, Nickel presented the wide range of areas where e-izing can benefit the clinical trial.¹² The range, which seems never-ending, includes the direct collection of the clinical research data via electronic data collection or e-CRF, Internet-enabled study clinical laboratories services, departmental support systems (such as pharmacy and drug supplies, trial monitoring and reporting, adverse event reporting, investigator portals and support systems, through-to-patient/subject identification and recruitment, and so on). In each case it can be argued that an advantage will be gained through e-ization, either in time or quality, or both. This does not, however, guarantee overall success.

Two examples illustrate typical issues. First, if an automated-request, just-in-time-delivery drug supply system is triggered by a clinical research associate responding to a laboratory result on site during each visit, it is clear that the study will not gain significantly in time, even though the pharmacy may enjoy better quality. Second, if any e-ized component is mirroring (or duplicating) another system, then it may be necessary to confirm equivalence between the data in the two systems, an overhead that may be out of proportion to the value added. Thus, the key point for the project is to ensure that each e-ized component is really adding value to the clinical project, knowing where this occurs, and that it will not be compromised by a need to interact with other contributors to the project. Finally, it is a good idea to have available clearly defined and measurable methods to monitor success.

Even with the advantages of the e-clinical trial clearly stated, understood, and measured, a successful outcome will not emerge if the staffing needs are not met when required. Unlike the traditional, paper-based study, where requests for activity and transfer of information can be, and often are, managed through a project manager orchestrating each departmental contribution, the e-clinical trial demands a service-oriented, "please-complete-the-defined-task-now," approach. This changes the project manager's role from one of orchestrating an ongoing activity, to one of preparing the players to respond-and thereafter monitoring that they undertook the tasks. Understanding who is involved with the e-trial from both the core and extended project team members is therefore critical to e-trial success. Having identified each role or required resource, one must also define clearly what that person is contributing and what he or she must be able to contribute. Planning must then ensure that all team members, both core and extended, understand the stages at which they have to be available to the e-trial.

Table 1 illustrates a range of contributors to the e-clinical trial and their involvement over the course of delivering an e-trial project. Using the standardized framework developed earlier, it shows the involvement of a typical set of clinical trial operational groups in establishing and delivering a study. For each element the group with the primary responsibility for delivery is identified, together with those involved directly or indirectly in the delivery activities. What is immediately clear from the figure is the wide range of skills needed to deliver an e-clinical trial, and the varying levels of involvement of each group over the course of the study. It also illustrates how the primary responsibility for delivery of the e-trial products moves between different groups at different times and shows how the e-trial may have no "study champion." In the figure, no one is identified as having a primary project responsibility for delivering qualified human resources.

The final general success factor for all e-clinical trials is to ensure that all the tools are available when they are first needed. Unlike traditional approaches, where subject recruitment is started before many key products are available, e-clinical trials must have all tools and systems available before the first subject's first visit (FSFV). Failure to achieve this simply means delay until all e-trial components are in place. It is helpful to separate clearly what is required to be in place to support any e-trial (the infrastructure requirements) from what is required to deliver the specific project. In terms of the framework described in Figure 1, this means that all validated processes (and the systems supporting them) must be in place before any design and build, production configuration, production processing, or close and archive activities occur. It may or may not mean that human resources or qualified staff are available, but as discussed above, failure to address the e-trials staffing requirement early will affect the ultimate outcome of the e-study.

Specific success factors

The general factors for success discussed above are not specific enough to help with the day-to-day tasks needed to implement and operate an e-clinical trial. With their wide-ranging implications over a large part of the organization, these general factors belong predominately to the senior owners of the e-trial project. Presented here are the major objectives of each element, together with a discussion of the factors that are central to the successful outcomes that need to be achieved.

In the following sections, the example of a small- or medium-sized clinical operation needing to establish and run an e-clinical trial has been selected for discussion. Because of the wide range of operational and organizational methods used in our industry, the discussions may not always apply directly in a particular organization. However, the general principles should always be applicable.

Table 2 presents the major e-trial objectives for each of the elements of our model. Each objective or product would be satisfied or created using documented standard operating procedures, some of which will be e-trial specific and some that will not be. An early practical exercise that can contribute significantly to the success of an e-trial project is to question of each element, "What products and/or objectives and what supporting procedures required to run an e-trial are currently not available?"

Validated processes. This element establishes the basis for running the e-trial. The key objectives are to establish the e-trial internal infrastructure (systems, software, procedures, and so on) to confirm that they are operating as required (computer systems validation, and so on) and regulatory requirements and guidelines are met. This element should also be able to present and define clearly the scope of the e-ization and determine or develop the metrics and other review tools against which the systems will be measured.

E-clinical trials have the potential to extend both the range and complexity of the specific items that need to be addressed here. The range is extended by the requirement to assure compliance with general regulations (ICH Guideline for Good Clinical Practice [E6] and 21 CFR 312.62), while at the same time conforming to computer systems requirements (Guidance for Industry: Computer Systems Used in Clinical Trials) and electronic records regulations (21 CFR 11, Electronic Records and Signatures). In some jurisdictions it may also be necessary to adhere to data transfer or data privacy rules. The important key success point here is to know which regulations are relevant, to confirm that your e-trial systems conform, and to know how they are to be defended.

It is critical during this stage to understand the general and study-specific training requirements that an e-trial brings. Leaving the identification and planning for training (not, however, the specific scheduling; this should be managed elsewhere) to later stages only compromises the availability of the all-important human resources later on.

Finally, ensure your initiatives incorporate and define fallback positions for all activities. All e-trials rely on technologies that will fail sometime, somewhere. Ensure that where the e-application is supporting regulatory or subject safety issues (adverse events, for example), backup procedures are in place to respond nontechnically to the e-system failure, and nontechnical recovery procedures are in place to reestablish the e-system fully.

Human resources. The primary purpose of the activities in this element is to identify and commit the human resources needed to run an e-clinical trial. Although most organizations manage this along departmental lines, this element does not necessarily represent a human resources (HR) department. HR can and should be involved, however. The primary internal goal here is to train staff in the basic e-trial skills and, if possible, to identify investigator sites sympathetic to e-trial methods and to confirm and/or train the staff in the general site e-trial requirements.

A key to future success is ensuring that all participants, both within the core and extended e-trial project team, understand their roles and responsibilities, what will trigger their being on e-trial duty, and how and to whom to communicate that tasks are completed. Omitting these points leads to resources not being available, additional unplanned work for the project team, and potential development of processing backlogs. The immediacy of the information flow in e-trials can make it very hard to recover from this position, if it is allowed to develop.

Qualified staff. The qualified staff element is designed to take the "ready for an e-trial" human resources and convert them into "ready for this e-trial." The key goals here are to first identify and then train staff in the specific skills (e-trial or otherwise) required by the study. This type of training is often mixed with basic skills training. You should take care to ensure that both are addressed satisfactorily.

Attention to three areas will help considerably in minimizing subsequent issues. First, train all staff to recognize the difference between clinical and technical issues and omissions. Clinical issues can be presented as technical problems, and technical issues as clinical problems. Ensure that the project team can identify each and organize the correct response. Second, train investigator sites as close as possible to their site's FSFV. And third, ensure that sites understand the extended range of responsibilities that working with an e-trial brings, especially from the technical viewpoint (the protection of user names and passwords, for example).

Design and build. E-trials rely on well-designed, well-built, and well-configured systems and software. In this element, all the study-specific tools and procedures must emerge, with the key goals being

• the analysis of requirements and e-trial specification development such as e-CRF, supplies logistics, and so on.

• building and testing these systems to meet regulated data rules, in support of nonregulated data collection, and so on.

• establishing the communications specifications for the systems.

It is also increasingly important to have available the tools for reviewing data, creating reports, and, what is often most critical, integrating with other systems.

Key to success here is first to ensure that the study will benefit from an e-trial approach. After confirming this, you will reduce potential issues further down the line by making detailed analysis of the systems requirements and confirmation of specifications early and before system builds begin. From the application standpoint, ensuring that site e-trial tasks dovetail with normal clinical practice can avoid the emergence of many usability issues, as will investing in detailed reviews and confirmation of the study's site-connectivity requirements. Attention to developing focused, standardized user documentation, and contributing and/or reviewing training materials, can reduce many calls for support later on. Starting to plan for the inevitable second- and third-level service support calls during the production phase will ensure that urgent and important support calls do not compromise other projects.

Finally, test and validate the systems for real, by using methods that emulate "in the small" the study in the field. Establish a site away from the development environment, add users, and develop and test subject data scenarios using role-playing. The value of this type of exercise is enormous, because it highlights early where the problems lie, and gives a window for correction before a study starts. The use of two environments-one for validation and, thereafter, training, and one for production operations-should always be considered.

Production configuration. In this element the products of the design and build activities are configured for production use for the study. The main activities are to configure all the systems with study details. These tasks may range from making specific study documents available, to loading study-specific fixed data.

The procedural control of study fixed data loads is especially important here. For example, unintended use of the randomization schedule prior to study start could require considerable rework. It is helpful to develop a study-specific production configuration checklist to control this work, and then to test the checklist.

During this phase, site e-trial configuration should be scheduled. This should include technical surveys-which will identify the appropriate additional tasks required to upgrade, configure, and install site equipment as required-and training requirements. Efficient use of time can be gained if these activities are undertaken in parallel with design and build tasks, but this requires site technical specifications to be ready early.

The most critical goal for this phase is the management of user access to e-trial, an activity that can be used to advantage by making it the final and only task before the study systems go live. It is important to ensure that users are again properly advised of their e-trial responsibilities, especially when the e-trial tools are to be electronically signed.

Finally, it is helpful during this phase to establish the first-level information technology support desk (call center or helpdesk), and to use it to manage the production configuration tasks. It is also helpful to arrange for all extended project team members to be alerted at this point that the project is now in production, and to be ready to respond to project requests or tasks.

Production processes. The key outcome for this element is to run the e-trial as designed. This may range widely in scope-from the collection and collation of the clinical data, through the management of the study logistics and reporting laboratory results, to reporting progress and responding to problems.

Overall, the success of the e-trial depends heavily on completing these tasks as planned. And that is highly reliant on all previous tasks being completed satisfactorily, as they all come together at this point. Many issues at this point will be highly study-specific, and care should be taken to ensure that staff members know how to get help with technical and clinical questions quickly and easily. If it has not already been put in place, a first-level, project-specific support desk should be considered.

Staff working in this phase should establish service level agreements with all "customers" and monitor responsiveness against them. This in its turn will draw attention to issues such as resourcing and lack of qualified staff availability, as well as software and systems issues only discovered during this phase. An important consideration here is the availability of backup methods, workarounds, and recovery procedures, recognizing that once in production, the e-trial "hub" has little control over the use of the systems by the "spokes," who now have a high expectation of availability and robustness.

Access to a rapid change control is also critical for this element. If changes to e-trial applications are needed (from, for example, protocol amendments, procedure changes, or software bugs), they must be able to be addressed swiftly, accurately and, if possible, with no disruption to the study's production tools. If changes must happen, careful attention to the consequences for data preexisting in the system can save much rework if considered early.

Close and archive. The objectives for this element are to close all e-trial operations in an ordered and compliant manner. This will include removing user access, locking data operations, enabling any final data review and confirmation to take place, and recovering equipment and data, if necessary.

Along with standard end-of-study actions, this element should include a plan to release sites from immediate project e-trial obligations, and to instruct sites in long-term e-trial project obligations.

The key to success is to ensure that clear technical and procedural plans are in place for all e-trial closure and archive tasks. It is helpful to develop clear definitions of what initiates closedown activities, in answer to questions such as "What triggers removing investigator site (direct) access?" or "Who may still update data?" It is often the case that by the time this phase is reached, many of the extended study project team members will not be concerned primarily with this project. Therefore, ensure that appropriate second- and third-level support is available when needed.

Get results

The e-clinical trial offers many advantages, both general and specific, for the conduct, efficiency, and accuracy of clinical research projects. Its potential to assist in such a wide-ranging and diverse way, across many functional areas, may be contributing to some of the reluctance or inability to introduce its use more widely. This article presents a framework for reviewing, planning, implementing, and operating e-clinical trials that we have found of practical value in the planning and execution of studies using these approaches. It can be used at both the macro (organizational or project) and micro (study application) levels to identify and develop all the components needed, and is particularly valuable in recognizing resourcing and training requirements.

E-clinical trials challenge both the organizations and functional departments. Some of the main factors for a successful e-trial have been presented for each of the model's elements. In this paper it is possible to present only the most significant factors, and then in the most general of terms. Specific cases and circumstances will highlight other critical items, or may negate the relative importance of some factors presented here. What can be assured is that for any successful e-trial, each of the elements (staff members and/or departments) needs to be present somewhere in the organization, and identified and ready to participate in the e-study.

Practical experience suggests that the following points are of most value to ensuring a successful e-study.

• Have confirmed systems in place before studies begin.

• Separate e-trial infrastructure activities from e-trial project tasks.

• Ensure all project start-up and delivery activities are planned in detail and ready before FSFV.

• Ensure all the core project team members understand their roles and responsibilities.

• Ensure all necessary second- and third-level support will be available to the project when required.

• Resource the project properly, because e-trials offer little slack with which to work.

• Concentrate design and build effort on defining and analyzing e-trial study requirements and specifications.

• Don't make assumptions about investigator sites' e-trial knowledge or capabilities.

E-ization of the clinical trial may be the trigger to radically alter and, it is hoped, radically improve, clinical operations practices, which in some areas have changed little in 20 years. To support this process all team members will need to be able to identify and systematically review the components required. The use of unambiguous methods, such as those presented here, will be essential to building confidence in those taking on the e-trial.

References

1. Applied Clinical Trials European Summit. Lyon, France (April 2002).

2. Henry Stuart Conference: Regulatory Requirements for e-Clinical Trials. London, UK (February 2001).

3. J. Scibetta, "Internet-Intelligent Clinical Data Management: Using the Internet for Clinical Data Capture," Drug Information Journal, 35 (3) 737-744 (2001).

4. R. Scognamillo et al., "Clinical Trial Management and Remote Data Entry on the Internet," Drug Information Journal, 33 (4) 1061-1065 (1999).

5. Pharmaceutical R&D Statistical Sourcebook 2000 (Parexel, Rose Tree Corporate Center, 1440 N. Providence Rd., Ste. 2000, Media, PA 19063).

6. Various, "7th Annual Special Resource Issue," Applied Clinical Trials, December 2001.

7. Various, "Electronic Data Capture," Applied Clinical Trials Supplement, August 2002.

8. A. Richardson and R. Drury, "Experiences Using Internet EDC for a General Practice Study," Presented Paper: Electronic Data Capture, IIR Conference. Geneva, Switzerland (April 2001).

9. S.J. Rangel et al., "Development of an Internet-based Protocol to Facilitate Randomized Clinical Trials in Pediatric Surgery," Journal of Pediatric Surgery, 37 (7) 990-994 (2002).

10. L. Marks and E. Power, "Using Technology to Address Recruitment Issues in the Clinical Trial Process," Trends in Biotechnology, 20 (3) 105-109 (2002).

11. T. Stokes, The Survive and Thrive Guide to Computer Validation (Interpharm Press, Englewood, Colorado, 1998).

12. W.U. Nickel, "The Challenge of e-Clinical Trials in the Future: Perspective from a Sponsor," Presented Paper: 12th International Conference on Pharmaceutical Medicine. Cancun, Mexico (May 2002).

Acknowledgements

This article is adapted from a paper presented at the Applied Clinical Trials European Summit, Lyon, France, April 2002, titled, "How to Lay the Foundations for Future Success in e-Trials."

Andy Richardson, PhD, is a partner at Opttimus Consulting, PO Box 7718, Hungerford, Berkshire, RG17 0YJ, UK, +44 7748 185 176, fax +44 1488 684 082, email: andyr@opttimus.com, www.opttimus.com.