OR WAIT null SECS
© 2023 MJH Life Sciences™ and Applied Clinical Trials Online. All rights reserved.
Applied Clinical Trials
Defining clinical trial feasibility and establishing a formula for patient recruitment success.
More than 80 percent of patients say they are willing to participate in clinical research studies, but only around 10 percent actually do so.1 Meanwhile, from an investigative site perspective, it is realistic to assume that at least one-third of principal investigators and their staff for a given study will fail to randomize a single patient.2 These are just two factors that make more than 76 percent of all Phase II and III studies enroll more than 90 days late; how can we develop standard procedures to address these shortfalls proactively?2 More specifically, how can sponsors accurately measure the feasibility of a protocol and plan accordingly to achieve timely and cost-effective enrollment and the quality data to advance their pipelines?
Study "feasibility" is a widely used term in the clinical trial industry among researchers, clinical operations executives, investigative sites, and patient recruitment specialists. However, it is defined differently from sponsor to sponsor—and even from individual to individual—within the same organization.
"Recruitment feasibility is not as ambiguous as the broader term of clinical trial feasibility," said Mike Neidl, Vice President of Clinical Research at Genzyme Corporation. "Seems to me, five years ago feasibility meant sending out surveys. So formal, up-front feasibility is a relatively new concept, and this is an ideal time to address it."
True clinical trial feasibility includes strategic, scientific, operational, and patient recruitment considerations—an exploration that is worthy of a series of articles. However, if we focus on establishing a platform for study feasibility through an acceptable definition and general parameters regarding the "enrollability" of a clinical trial—particularly a large-scale, multicenter, global trial—this will benefit sponsor companies and the clinical trial industry as a whole through supporting proactive, early, and efficient recruitment feasibility assessments.
This article will focus on exploring the definitions and parameters of recruitment feasibility, touching on some key considerations for all trials, and propose a flexible formula to help sponsors be more proactive with regard to their specific protocol scenarios. The goal is to establish components of recruitment feasibility and lay the groundwork for the industry-at-large to collaborate and develop a standard set of best practices—and for each sponsor to modify the standards for their own benefit based on the unique demands of their individual research pipelines.
Perhaps the best place to start is by establishing where sponsors are frequently misled and misinformed—both internally and through strategic partnerships. Whether recruitment feasibility is considered on its own or bundled in with operational feasibility, the current "standard" often consists of the following steps:
Unfortunately, from a recruitment point of view, these steps are very broad and can be defined differently by each sponsor. Likewise, each sponsor will vary in the number and type of processes they will then use within each step. Perhaps most importantly, there is no consistent application of a recruitment perspective, or "lens," to the steps. As such, these are all contributing factors that hinder up to 70 percent of trials from meeting their original recruitment targets. Ultimately sponsors pay for inaccurate forecasts either in additional time and/or money, not to mention the costliness of miscalculating the power of a study.3
"Before, the feasibility result was mainly the number of patients and the number of sites that a country can commit to," said Jurek Bojanowski, International Clinical Trial Manager at sanofi-aventis. "But this is not real feasibility. Corporate management must do some validation to check not just how many patients and sites a country can provide, but also when and how fast this enrollment can be performed."
While there are some general standard operating procedures (SOPs) available to research sites for guiding a feasibility assessment,4 sponsor companies have only just begun to define their own SOPs.
"We consider strategic feasibility to be part of a global clinical development plan for a compound or protocol," Neidl said. "It's a high-level process taking into consideration the market strategy, disease incidence and prevalence, regulatory approach, and country selection, and maybe some preliminary idea for an investigator strategy, such as specialty selection and mix of generalists and specialists."
Considering the infinite number of potential recruitment conundrums that varying protocols will create, it may seem an insurmountable task to gain a full understanding of how all of these considerations will factor into a successful study. By breaking feasibility into a few key phases, we can begin to prioritize and operationalize our efforts—a process that can be applied to any therapeutic indication or the scope of the research in question.
"We divide feasibility into two phases, the first being protocol design, and the second operational planning," said Martin Lee, MD, Executive Director of Site and Patient Recruitment for Pharmaceutical Product Development (PPD), a leading global CRO. "It's in the latter process that we consider the regulatory time lines, drug supply, and other factors that ultimately determine whether recruitment will happen in a timely fashion."
Figure 1. Example of factors assessed in a site survey.
Indeed, within the broader scope of operational feasibility for a clinical trial, recruitment feasibility is the process for determining the time frame and parameters necessary for successful recruitment.
For the purposes of exploring new methods for improving clinical trial enrollment planning and execution, we propose the following definition of recruitment feasibility.
Recruitment feasibility is the process by which a clinical study sponsor can forecast and manage the probable randomization rate (number of patients per site per month) for a specific protocol and determine realistic parameters for site enrollment months (number of sites multiplied by number of open-to-enrollment months).
With this in mind, study planners must consider each of the following factors through a recruitment lens.
Protocol design. How would each group of study community members respond to this protocol—regulators, investigators, coordinators, project managers, monitors and patients? In what ways might the protocol design be off-putting to one or more of these groups, and can that be changed, or must it be mitigated? Can we afford to prioritize one group over another?
Country. While there may be many business reasons to include a specific country—such as market importance, regulatory barriers affecting speed to market, existing relationships with investigators, cost of conducting—none of these exempts a country from a recruitment feasibility assessment. From a recruitment perspective, the goal is not merely to determine which countries should be included, but what will it take within each country under consideration to succeed? What level of recruitment support may be required to offset protocol design challenges?
Sites. What types of investigators are most likely to be high enrollers for this study? Will that vary by country? Do we have a platform for contracting specific sites in each selected country? What are the likely enrollment rates for sites in each country? Site surveys play an important role in assessing recruitment, however most current models co-mingle recruitment feasibility with site selection—a near guarantee for collecting misleading information.
Patient. While it may seem obvious to include patient perspectives in the recruitment feasibility process, it is a rare sponsor that consistently asks patients for their input. More often than not, investigators, key opinion leaders, or country managers are used as surrogates for patients.
Brendan O'Neill, Manager in the Patient Recruitment Specialist Group at Merck and Co., agreed with this sequence, "We approach recruitment feasibility in that order—country, site, and patient. It's great to consider the patient level, but if the protocol won't be approved in a country, you're wasting everyone's time."
Indeed, from a patient recruitment perspective, these four key considerations are the root of most enrollment barriers and certainly a thorough analysis of a given protocol will provide the necessary insight regarding ideal study countries, sites, and ways and means to reach the patients most likely to value study participation. Assuming that we know our protocol and our patients well, let's focus on some key considerations for country and site assessment and selection and how qualitative data collection strategies can be applied universally for all studies, and ultimately applied to a feasibility formula to generate quantitative insights for successful study planning.
By focusing on business drivers, study planners often prevent a careful examination of recruitment barriers when deciding which countries to select for a global trial. Our recommendation here is to develop a method for a comprehensive assessment of countries that includes business drivers but also many other recruitment-related criteria.
"We work with the clinical teams as they're designing the protocols to add our recruitment knowledge before it gets out to the field," O'Neill said. "We then work with our international affiliates to make sure, first, that the procedures align with the standard of care so we can assure approval in the country, and second, that the patient populations are available. We take it a step further and work with some of our experienced investigators to determine the likelihood for successful execution of the protocol."
Objective. Assess recruitment suitability of 25 countries for a Phase III study seeking to enroll 800+ treatment-naïve stage III non-small cell lung cancer (NSCLC) patients.
We began by conducting research in the 25 potential countries to understand the environment for a study with an active comparator investigating concurrent radiotherapy followed by consolidation therapy. Variables identified as important to recruitment included incidence, healthcare system, attitudes toward the comparator, and the number of competing clinical trials. Once the key variables were in place, we weighted them according to their significance to impact recruitment (Table 1). Using the results of the market research, we rated each of the countries and ranked them from most to least likely to deliver recruitment success.
Table 1. Each factor is weighted for importance in affecting recruitment and a consistent scoring system designed.
Results. The research revealed that many physicians in the countries targeted for participation rarely used the active comparator for first-line treatment and in the majority of countries, only some sites had the infrastructure necessary to conduct concurrent radiotherapy.
For this study, we recommended that the sponsor reconsider the choice of active comparator to align with current standards-of-care in the majority of countries and to focus country and site selection on physicians who continue to practice consolidation therapy.
Perhaps the most flawed element in the current standard for selecting investigators is the expectation that sites can provide accurate estimates without enough time, information, or compensation for their efforts. Lee referred to an analysis conducted by PPD on site estimates, and he concluded, "We've looked at whether site surveys predict enrollment rates. We don't think they do, if you simply ask them how many patients they can enroll in 12 months. Our data shows that there's almost no correlation between site predictions and their actual enrollment performance."
Many voices in the industry advocate for a more robust assessment of site enrollment capabilities, including provisions for more detailed protocol criteria and payment for time spent. "I do think sites should be compensated for the time they spend on feasibility estimates," said Neidl, "but it's also important to have some sort of quality assessment on the value of the data returned."
Even assuming such support, the site selection process is still in need of a broader, more consistent process that can be applied across trials and companies.
To start, it is important to throw the widest net when considering sites for running a trial. Because every protocol is unique, even choosing sites with prior experience may not ensure enrollment success. Thus, expanding the pool of sites to include some the sponsor or CRO has not worked with can be strategically important.
In addition, as in country selection, the list of criteria for site assessment must be expanded for truly effective selection. These factors should be weighted by their potential impact on patient recruitment performance for the particular protocol (Figure 2). The results of the analysis will reveal each site's overall recruitment aptitude.
Figure 2. The benefits of defining recruitment feasibility and developing standard procedures are varied, but they must accommodate the needs and goals that are unique to each study and sponsor.
Objective. Identify 80 sites for a Phase III study of a treatment for immune thrombocytopenic purpura (ITP).
Barriers to success included the low incidence of the condition, the high number of competing ITP studies, and a protocol design that included a placebo arm. At minimum, the ideal investigator would be a hematologist specializing in ITP whose treatment philosophy aligned with the novel scientific approach of the investigational drug.
The challenge lay in developing a site selection process that would uncover the true treatment philosophy and standard of care of each potential site. Of concern: would sites contracted for previous studies reveal their preferences? Would past performance be repeated resulting in the need for 80 or more sites to meet the study goals?
Results. By targeting sites unfamiliar to the sponsor, the online survey produced more than 90 potential sites that qualified as "superior." An additional level of analysis determined that, if well supported, only the top 40 would be needed to cost effectively enroll the study, saving time and expense.
Many sponsors use trend analysis as the basis for enrollment projection modeling and the assumptive means for expediting patient recruitment.5 Unfortunately, trend analysis is contingent upon historical data: since no two trials are alike, past performance for a given prostate cancer or asthma trial will never serve as an accurate measure for future success—even with the next phase of research for the same compound.
Even for an ongoing study, if a sponsor relies on trend analysis derived from most electronic data capture programs, interactive response systems, or clinical trial management software, by the time enough enrollment data has been analyzed to reveal that a study is off-track, it's often too late to recover the original patient recruitment plan. The divergence from the original enrollment time line is simply a realization that the study is behind: the quality of this intelligence alone is not sufficient to provide insights as to how the recruitment effort should be adjusted to help the sponsor meet the original study goals. As the divergence gets worse, time-consuming tactical readjustment decisions must be weighed. This can mean costly extension of study time lines, adding additional sites, or funding contingency outreach expenditures—often with blind faith that the selected approach will jump-start enrollment.
"Miscalculations of mere trend analysis can add many months to the ultimate enrollment of a given clinical trial," said Sarah Schoen, Clinical Trial Coordinator for the Eastern Cooperative Oncology Group. "Missed protocol data locks can blunt the market viability of a given compound or franchise and have equally broad implications for a sponsor's overall competitive footing," Schoen said. "It's a scenario where poor upfront planning can wreak havoc for sponsors—especially for midsized or smaller companies who are relying heavily on just a few studies to help them secure market standing or potential corporate partnerships."
Countries and sites are two prime examples of universal factors that influence patient recruitment planning for any study. When we include protocol specific considerations and the ideal patient population as well, we can begin to develop measurements that will yield enrollment rates to inform study time frames. While each sequence of knowledge gathering will vary from one study to the next, every study's recruitment or enrollment feasibility can be quantified through the use of a formula called the primary enrollment principle6 : e = r *t *s. Where: e = the number of randomized patients needed; r = the randomization rate (the number of patients per site per month); t = the enrollment time period in months; s = the number of sites actively randomizing patients.
In order to forecast and manage the probable randomization rate, and determine realistic parameters for site enrollment months, we use the formula as follows:
The difference between r1 and r2 determines the level of support or intervention needed to make the study feasible from a recruitment perspective.
In essence, the primary enrollment principal is designed as a common denominator for calculating enrollment success regardless of the breadth and scope of a given study, and regardless of the indication or nature of the investigation drugs or procedures. It is the foundation for establishing standard recruitment feasibility practices.
We can put the principal to work. You need to recruit 1,800 trial participants for a type 2 diabetes study and you have been allotted 10 months and 50 sites to get this accomplished. Therefore, you know the following: e = 1,800 patients; t = 10 months; and s = 50 sites.
Using the primary enrollment principal, we now solve for r: e=(r t s); 1,800 = (r x 10 x 50); 1,800 = r x 500; r = 1,800/500; r = 3.6 patients/site/month.
To meet this study's enrollment goals, you would need to enroll four patients per site/per month. Can sites in every country accomplish this? Will each country be recruiting for all 10 months? How realistic is this randomization rate for each country and for all countries in the aggregate? Does averaging the differences from country to country actually have meaning—and would you bet $12 million dollars on it?
If 25 sites are in the United States and five sites are in Australia, Argentina, Belgium, Mexico, and Spain, how would this impact the enrollment projection model?
There are certain operational considerations that impact recruitment time frames. These considerations affect t (the enrollment time period in months), and s (the number of sites actively randomizing patients), which when multiplied by each other produce the site enrollment months (the cumulative number of all open-to-enrollment months for the study). The considerations affecting t and s include can be seen in Table 2.
Table 2. Breakdown of elements that could affect a trial.
Conducting preliminary market research will help to determine the structure and content of the site survey instruments that are typically developed by many sponsors and CROs to determine r. Through an analysis of the protocol, we can pinpoint the essential questions to focus on in the survey and how to weight their importance. For example, access to a large patient panel may be critical in a metastatic breast cancer study, in which other physicians are unlikely to refer, while for a diabetes trial, the more important factor might be patterns of treatment and levels of medication dosing.
Most importantly, by using market research methodology, we can ensure that the structure and design of the survey questions anticipate investigator bias and prevent the impact of investigators' intent (often unconscious) to confuse, mislead, please, impress, or safeguard reputation. How? By building fail-safes into the survey design. For example:
Market research-style surveys thus produce the blend of quantitative and qualitative data that can be objectively analyzed to provide overall ranking and scoring of site compatibility with the protocol from a recruitment perspective.
While there are common elements that make up the foundation for good feasibility analysis, there is no one model that is more worthy than the other. Organizations must adopt the model and the SOPs that best suit them (Figure 2).
Once sponsors establish their own working models for optimal feasibility analyses, perhaps then we can elevate the conversation toward developing industry-level standards that have broad applications across therapeutic areas. Bojanowski sums it up well, "Companies all have their own SOPs and methodology. We need to share our experiences, similar to what was done with marketing 20 years ago when IMS served as a conduit to collect and share sales figures for pharmaceutical companies, thus providing a clear picture of what is on the market and how the market is structured. This is missing in research. The barrier is that there is not enough information flow."
The components, processes, and tools outlined in this article are suggestions toward developing industry-wide language for assessing and improving enrollment feasibility predictions and management strategies. They are offered in the spirit of collaboration and an opening of a dialogue to which all interested parties are invited. The need for best practices in recruitment feasibility is clear, and could contribute to increased achievement of enrollment success across many sponsor companies. The primary enrollment principle formula is one means for establishing a common feasibility language that can address protocol-specific concerns for all sponsors.
Such SOPs will ensure common internal expectations and performance between individuals and across departments and the organization as a whole. Building and following a process will promote veracity, accuracy, trust, and confidence among all parties, whether sponsor personnel or site staff. Such a process also ensures that suppliers are able to comply with company requirements. Overall, consistent application of the definition and processes developed will produce reliable and measurable success results.
Matthew Kibby is Global Operations Leader for BBK Worldwide, 320 Needham Street, Suite 150, Newton, MA, e-mail: firstname.lastname@example.org.
1. BBK Healthcare, Inc./Harris Interactive, "The Will & Why Survey." Reinventing Patient Recruitment: Revolutionary Ideas for Clinical Trials Success, (Gower Publishing, Surrey, UK, 2006).
2. Thomson Medstat, report, "Using Data and Metrics for Clinical Trials," (2004), interest.healthcare.thomsonreuters.com/content/DownloadLibrary-Research.
3. R. E. Carter, "Application of Stochastic Processes to Participant Recruitment in Clinical Trials," Controlled Clinical Trials, 25 (5) 429-436 (2004).
4. Individual medical research institutions; Standard Operating Procedure for Assessing Protocol Feasibility, store.centerwatch.com/pdfs/samples/sop03_ss201.pdf.
5. J. Pellegrino, R. Smith, "Predictive Modeling in Clinical Trial Enrollment," white paper, Acurian Patient Recruitment Solutions, (2009),http://www.ngpsummit.com/media/whitepapers/Acurian_NGPUS.pdf.
6. Joan F. Bachenheimer and Bonnie A. Brescia, Reinventing Patient Recruitment: Revolutionary Ideas for Clinical Trial Success, (Gower Publishing, Surrey, UK, 2007).