Automating Phase I Trials

The automation of Phase I trials takes on many shapes and sizes. While electronic data capture (EDC) remains the most familiar type in Phase I, many other areas in early phase clinical research can be automated. For example, the typically mobile staffers need to know precisely where, when, and how they must execute each step in the protocol process. Samples, especially pharmacokinetics (PK) samples, must be processed accurately. And recruiting full groups is critical to the success and profitability of a unit. Automated solutions can help in each of these areas.

To ensure the smooth implementation of automation technologies in Phase I, it's important to develop a tailored implementation plan. Only through a managed introduction of new technologies into this highly pressurized environment can the entrance hurdles be overcome.

Why automate?

While an automated clinical environment is intuitively appealing, the reality is that the shift from a paper-based world does not happen without considerable planning and resources.¹ While the best solutions will harmonize easily with current clinical procedures, the need to validate, test, and train in areas that form the heart of a 24/7 enterprise provide a significant hurdle to clear. Two primary benefits drive companies to clear these hurdles: safety and financial.

A discussion with those employed in the Phase I space will quickly lead to safety— either the need to assess the safety of a new drug or the need to keep the volunteers safe. Focusing on those solutions that can increase the quality, visibility, and speed of data collection (not just EDC but those that improve the quality of the operational process) allows drug assessments (dose escalation decisions and other ongoing safety reviews) to be made quickly and more accurately, while also enabling rapid intervention if the safety of the subject is impacted. Having this higher quality data available online makes it readily available to all global partners—sponsors, CROs, the clinical unit—and facilitates the involvement of those best placed to make assessments.

The financial benefits can be more subtle but still provide a critical driver toward automation. From a day-to-day perspective, the reduction of errors and the earlier identification of problems when they do occur both reduces or removes the cost of query resolution, which was estimated to be as high as $60 to $100 per query as long ago as 2001,² and, in more serious cases, the need to repeat procedures from a single subject's blood pressure to an entire study.

The more operational solutions reduce staff costs, as employees are moved away from the scheduling of studies, recruiting of volunteers, and general paper trail that drives your average unit. This frees the staff to return to clinical activities. Collecting data in electronic format (eSource) at bedside reduces the resources required compared to traditional paper collection.

From a practical perspective, the ability to make more efficient use of staffers time is key. Phase I trials are becoming more complex; at the same time, the drive to run a strong business operation is requiring that units take on more trials. Automation can boost productivity so that Phase I units can efficiently run more trials, which translates to more revenue per unit capacity. In addition, an automated unit with this increased level of safety and lower financial cost base starts to differentiate companies from their competitors, enhancing the firm's reputation and delivering a financial impact of its own.

Where to prioritize

To inform and drive an implementation plan, it is important to identify and document those benefits that are sought, or if easier, the challenges that the clinical unit seek to overcome. These targets can then drive the selection of solutions and, as importantly, help identify the order in which they are implemented. For the purposes of this article we will look at automated solutions that aid the information flow that drives operational processes and those that facilitate the capture and processing of clinical data. The challenges that fall under this broad definition include:

Recruiting volunteers. Trials generally require processing six to 12 volunteers as a group. The failure to recruit full subject groups can delay a trial but, more importantly for the clinical unit, have a significant adverse effect on the financial side as all activities are repeated for those volunteers required to complete the group.

Complex workflow. In Phase I, data is collected at a rapid pace and in a highly structured manner, however, each trial is different. Time is of the essence, and the highly mobile staffers are typically multitasking within the unit, as they are usually handling more than one study.

Scheduling studies. Key to accommodating a new trial is the availability of beds and staff to conduct the trial. In fact, the similarities with a hotel are marked, with the exception that you cannot move volunteers to a neighboring establishment if you overbook. Once started, trials often stop and restart, and trial designs change. This complex picture requires solutions that can model this flexibility and the ability to make real-time changes.

Sample integrity. In the maelstrom of activity that typifies a Phase I unit, it's important that clinical samples (such as blood) be processed accurately. If these samples are mixed up or improperly stored, or if collection times are missed, the volunteer's data cannot be used and his or her participation in the study can be invalidated.

Evidence of processes. While a strong set of standard operating procedures integrated with evidence of training can support claims of good practice, the ability to present an auditor with contemporaneous data—of freezer temperatures, for example—can provide the definitive picture that attracts both sponsors and regulatory bodies.

Accurate data capture. Missing or erroneous data can invalidate results and necessitate a partial or complete repetition.

Historically, these challenges have been addressed with increasingly intricate paper processes. These studies are, however, becoming more complex. The unit must accommodate the requirement that early phase trials deliver a first look at efficacy and stakeholders' desire to share data rapidly. Meanwhile, the units aim to take on larger numbers of studies. All of these factors are leading to the simple conclusion that paper processes are no longer practical.

Available solutions

Once the priorities for automation are identified, the team can select appropriate solutions to meet those challenges. Early in the days of Phase I clinic automation, organizations attempted to modify solutions originally developed for later phases for use in Phase I. Today, however, there are many new solutions available built around Phase I requirements.

The mobile behavior of staff and wide variety of clinical settings requires portable solutions. These have become possible as the variety of hardware available has increased—touch screen tablet PCs, bacteria and drop resistant laptops, and smaller devices like the PDA and mobile phones—ensuring a tailored solution can be available at the required location.

Importantly, in parallel with these advances has come the acceptance of processing operational data directly in electronic format. Many homes now have their own wireless networks, and Internet use consumes a large part of the working and recreational day. Historical problems with network speeds are being overcome as high-level broadband becomes widely available. In areas where it is not as available, network accelerator companies provide the required boost. In short, the idea of pushing operational data directly via device and onto a remote database via wireless network is easily accepted by most.

A natural extension sees the collection of clinical data directly into electronic format. To do this topic justice would require a paper in its own right. Positively, regulatory and industry bodies are starting to provide some guidelines toward such use of eSource, and issues previously seen as barriers are now seen as clearable hurdles.^3,4,5

Key areas where solutions have been developed to support the operational and process management challenges highlighted earlier include:⁶

Electronic schedule to drive operations. To fully automate a Phase I trial, it is important to codify the trial design framework. How many subjects, time of dosing, what clinical samples and data will be collected, and at what time relative to dosing? While the protocol and related documents typically convey this information, it can be captured electronically in just a few hours. This electronic schedule can then drive key operations. The schedule can be sent to the clinical staff via wireless devices, using alerts with the "who, what, when" info that's critical to running the trial. An update to the master schedule (e.g., a one-hour delay to dosing when the pharmacy encounters a delay with the drug) can be cascaded to the team, updating all subsequent activity times. As adaptive trials become more common, the use of an electronic schedule will allow the clinical staffers to accommodate planned and ad hoc trial design changes.

Information can flow in both directions, with staffers reporting that steps have been completed. A Web-based dashboard can present comparisons of scheduled and actual times, allowing immediate remedial action when needed and preventing protocol deviations.

A simple extension of this single-study framework allows the management of resources across studies. With the addition of a graphical interface, rapid assessments can be made on how to optimally schedule a new study—ensuring that sufficient beds, appropriately trained staff, and equipment are all available.

Sample tracking for data integrity. The basic technology of barcode scanning is familiar to everyone from a trip to the supermarket. It also addresses one of the most fundamental issues in Phase I research: proper identification of subjects' PK (and other) samples. Barcodes can be printed on labels attached to the volunteer, the nurse, and the sample tubes. A three-way scan at the bedside can instantly verify that the right tube is being used for the right subject at the right time, while recording who carried out the procedure.

An extension to this model allows the samples to be tracked throughout the entire sample processing. For example, when the sample goes into and comes out of a centrifuge, the barcode can be read along with the equipment ID. The transfer of the sample into other tubes can be recorded. The location of a sample can be tracked down to the specific shelf within a freezer. If samples are shipped to an external lab, the information can also be tracked by barcodes. Ultimately, to provide a snapshot of sample processing and location and to identify problems in real time, all the information can be loaded into a master database and monitored via dashboard.

Visualizing data for at-a-glance review. One of the most significant benefits of clinic automation in Phase I is the ability to review, monitor, and query the data in real-time. Given the wide range of data being collected, the benefits here extend from those commonly seen from EDC (including simple CRF representation in electronic form, with standard and ad hoc summary reports) to comprehensive dashboards to present operational data (as discussed). Taken together, these options provide data that can be reviewed in great detail or at a high level by the various global players who have an interest in a trial's progress and results.

Sophisticated tools for volunteer recruitment. Recruiting full groups, as noted earlier, is important to satisfying the requirements of the trial, which is in turn critical to the financial viability of the unit. A clinical unit's volunteer database is considered one of its prized assets.

More sophisticated systems built for this precise function can drive recruitment campaigns, improving success rates and allowing staff to be deployed to more billable tasks. The basic details tracked (contact info, volunteer history at the unit, and exclusion parameters) can be expanded into a full subject profile. Such a system makes it possible to query the database using selection criteria to identify eligible volunteers for a new trial. Those selected can then automatically be sent an email notification. As potential subjects call in, the recruiting officer can see all relevant data supported by confirmation of the date and trial type of the last trial the subject completed. When the subject is accepted for a trial, their details and history can instantly be loaded into the database. Appointments can be set up based on clinic openings, and automated reminders can be sent out to subjects.

Electronic data capture. For most, the rapid and accurate capture of clinical data is the most familiar component of the automation picture. The benefits of EDC technology for automating the clinical trial process are now widely understood.⁷ And EDC fits as well if not better in the Phase I space. Key to success, however, is the adoption of processes appropriate to this phase—applying a Phase III EDC approach to Phase I trials would hamper success, much as applying paper-based Phase III data management processes to Phase I trials is not effective.

Beyond replacing paper CRFs with eCRFs, other techniques can be used to automate data capture. For example, with ports on new generations of medical equipment, such as sphygmomanometers for measuring blood pressure, data can be captured directly into the electronic database with no staff interaction other than the push of a button. Barcode scanners can rapidly scan standardized answers from a wide variety of data collection activities, such as categorical responses associated with the severity of an adverse event.

Getting started

When considering how to automate Phase I trials, there are a number of preliminary steps that can ensure a smooth implementation. Here are some key issues to consider.

Agreement on the primary benefits and goals. While the many benefits discussed in this article can easily justify an automation project, certain issues may be particularly critical to your organization. Getting agreement across the team up front may influence precisely how you implement automation technology and how rapidly you do it.

Work with your vendor to gain in-house experiences. Before committing to a major implementation exercise, you can work with your vendor to run sample projects. Setting up mock studies can be helpful to ensure that all team members understand the implications of the new solutions and that they are right for you.

Establish an implementation team. Create a project team with representatives from each area of the operation—not just IT and data management. These representatives can serve as ambassadors for the system, and are essential to "buy-in" from the clinical teams. They can provide invaluable details to ensure the implementation matches your environment. Don't rely on vendors to provide subject matter expert (SME) support. Nominate staffers to get inside the system from the start and get them to act as local SMEs.

A mindset for success. Has your team thought through the steps that will be required to transition to an automated unit? It's important that the team be progressive in its thinking regarding eSource, for example. Many staffers still feel nervous about moving away from paper, and worry about fallback plans if a device breaks or the wireless infrastructure goes down. Staff must be open to learning new technologies and processes for this type of initiative to succeed. And even with the flexibility afforded by newer systems, adapting to new processes will still be required.

A realistic timescale to implement. Clinical units are 24/7 environments. Many clinical units adopt a phased approach, automating one area at a time rather than all at once. For example, it may make sense to start with just one trial, automating the PK sample processing lab, and then moving to the recruiting process and the clinic floor. It is important to also allow sufficient time to work in partnership with your vendor as they mentor you through the initial trials.

The right technologies. In transitioning to a software-based environment, putting the right technology infrastructure in place is a key ingredient for success. Working in conjunction with your software vendor, you can determine the best barcode readers, laptops, and handheld devices for optimal interoperability. Upfront costs for new hardware must be part of your budget. If there are concerns about the use of new technologies—for example, using a wireless network to facilitate direct capture of data at bedside—get this on the table early. Work with your regulatory department to assess concerns and conduct a risk analysis.

Develop a training plan for staff. Depending on the nature of your team, it may work best to train staffers operation-by-operation and then bring the team together to provide an overview of the entire process. It's also important to cultivate a resident expert who can help train other staffers as new professionals come on board. While people often prefer the work procedures they're accustomed to, most clinical staffers understand the importance of keeping current with technology in terms of their professional development.

Conclusion

Several factors have come together to make Phase I automation practical and viable. Wireless approaches are now well proven and when combined with improved Internet access, new hardware, and the wide range of new solutions, the technological infrastructure is now ready. Clinical units are increasingly aware of the opportunities open to them through these new solutions. With the ability to improve the quality of safety data while reducing costs, these units are focusing on overcoming the challenges inherent in new technology adoption and moving to an automated future.

Rob Nichols is Director, Commercial Development, at Phase Forward, Waltham, MA, email: [email protected].

References

1. S. Adcock, M. Willett, J. Rosenblum, "The Need for Electronic Data Capture in Phase I," Applied Clinical Trials, October 2006.

2. Electronic Data Management Forum presentation, "Cost Benefit of EDC," September 2001, http://www.eclinicalforum.com/content/Knowledge/Articles/Cost%20Benefit%20of%20EDC_Sep01.pdf.

3. P. Bleicher, "Computerized Systems Guide a Go at FDA," Applied Clinical Trials, June 2007.

4. S. Bishop, "Unraveling the eSource," Pharmaceutical Executive, November 2006, http://pharmexec.findpharma.com/pharmexec/Articles/Unraveling-the-eSource/ArticleStandard/Article/detail/382540?contextCategoryId=39722.

5. U.S. Department of Health and Human Services, Food and Drug Administration, Office of the Commissioner, "Guidance for Industry—Computerized Systems Used in Clinical Investigations," May 2007, http://www.fda.gov/downloads/Drugs/Guidance ComplianceRegulatoryInformation/Guidances/UCM070266.pdf.

6. A more detailed review of automation can be found in the author's article, "Phase I: The Power of Automation," GCPj, 16 (1) 21-24 (January 2009), http://www.phaseforward.com/news/media/documents/GCPJ%20Jan09%20pp21-24%20Phase%20Forward%20HIGHRES.PDF

7. A. Bhat, "Has EDC Finally Come of Age?" Clinical Discovery, 8, 14-15 (September-October 2007).