OR WAIT null SECS
A key function in clinical trials, patient enrollment, has fallen behind during a time where technology has played a vital role in the industry. Adaptive patient recruitment allows for clinical data to be collected and reviewed in real-time as to improve enrollment outcomes as they are taking place.
Technological advances in clinical trials are streamlining operations and data collection, yet the most essential function-patient enrollment-often remains behind the curve despite the advances technology can bring. Research suggests that half of investigative sites under-enroll, 11% of sites fail to enroll a single patient in a clinical trial, and a mere 13% exceed their enrollment target. What’s more, Phase II-IV study timelines are frequently extended to nearly twice their original length to reach the study’s enrollment goals.1 Now, with clinical data flowing into purpose-built solutions from multiple sources, it is finally possible to use this information to improve enrollment outcomes through real-time corrections during the recruitment process. Known as adaptive patient recruitment, this approach leverages technology to support recruitment aspects of studies as they are happening. The model enables changes to a recruitment plan so a study can stay on budget and on track instead of sliding into rescue mode, a costly proposition that can put the entire study at risk.
The idea behind “adaptive” is to use data in real time to protect on-time enrollment in a constantly changing recruitment landscape. This concept is rooted in the notion that as clinical trials unfold, things inevitably change. It comes with the territory, making quick access to data all the more critical.
Data on the status of patient enrollment need to be collected to inform decision makers charged with overseeing study progress. This has been standard practice for a long while, but oftentimes, stakeholders rely on information gathered after the fact, such as the number of patients enrolled and randomized per site per month. What is more relevant is real-time feedback on what’s happening in the field now instead of waiting for monthly data figures. A lot happens during the month, and it is important to have visibility into the impact on recruitment from actions such as the investigator meeting, monitoring visits, distribution of study materials, and launch of the advertising campaign. This provides an up-to-the minute, data-driven picture of what is working and what is not. Significantly, these data are available from systems that many sites already have, such as the interactive voice response system (IVRS) and clinical trial management system (CTMS).
This article describes the value in taking data from these systems and using them early on in the service of patient recruitment and enrollment. If everyone’s goal is to have on-time enrollment, adaptive recruitment is a key option enabled by technology that facilitates transparency and removes the guess work as to whether a course correction is needed.
Defining adaptive patient recruitment
Over the past decade, there has been a sharp focus on the subject of adaptive clinical trials. The basic concept is that adaptive designs acquire knowledge about a study as the clinical trial data accumulate. With statistical analysis, this knowledge can be used to make modifications that optimize study execution without affecting validity and integrity.2 In 2010, the FDA released a draft guidance on adaptive design, noting that there is great interest in this methodology as a way to make clinical trials more efficient through shorter timeframes and fewer patients.3 Similarly, the European Medicines Agency (EMA) put forth a Reflection Paper in 2007 on adaptive design, explaining that it has the potential to accelerate the clinical development process by allocating resources more efficiently without lowering scientific and regulatory standards.4
The adaptive approach is most effective when implemented early in the clinical program, and as described in the literature, it can improve decision making, yield cost savings2,5, and reduce developmental timelines.6 At a time when there are more than 183,000 trials listed on the clinicaltrials.gov website7, many seeking patients, well-founded recruitment and enrollment strategies remain as critical as ever.
Similar to adaptive design of clinical trials, when applied to patient recruitment, this effort involves extensive pre-planning, starting as early as protocol development. This new approach differs dramatically from the traditional stance of delaying plans for patient recruitment until it is time for site initiation. The goal of the adaptive method is to avoid the all-too-common study rescue, which is costly and wreaks havoc with timelines. It is also a discipline that allows the clinical trials community to work in a more collaborative and transparent environment. By removing siloes, data sharing flows more efficiently, and stakeholders can make informed decisions faster.
The data to be analyzed, such as number of patients enrolled per month per site, number of patients in the site’s database that meet the eligibility criteria, and enrollment success rate in previous studies from the same therapeutic area are housed in technologies such as electronic health records (EHRs), the IVRS, and CTMS. Identifying potential subjects for a study can be extracted from the site’s database structured from EHRs. Next, IVRS data can provide site-specific information on screening, consent, randomization, and retention rates. And once a study is underway, information about completed site visits and procedures are available from the CTMS.
These data, when combined, are powerful. They can be used to craft an algorithm that tracks current enrollment status. In addition, if one more piece of data is added, namely planned activation date of the site versus actual date, the ability to determine if sites are at plan increases. Because site initiation and the study timeline are based on the planned activation date, this factor becomes a surrogate for the real data, and offers stakeholders more power to accurately predict patient recruitment status. Moreover, for the clinical community not to know the planned activation date or not to use this information puts stakeholders at a disadvantage as to when to trigger adaptive action.
There is also the issue of the recruitment budget. To enable prompt adapting of the recruitment plan, establishing a contingency budget from the beginning is a best practice. Calculating this budget requires an understanding of the therapeutic area, the number of patients needed, the planned screen-fail rate, and more (see Table 1). Successfully securing a contingency budget upfront for recruitment purposes generally requires educating management on the value of this approach. Failure to do so often results in working frantically after the fact to secure funds. The education process starts by confirming everyone’s goal of on-time enrollment followed by explaining to stakeholders that changes to the recruitment plan are inevitable.
A key aspect of explaining the need for an upfront budget entails assuring the sponsor or contract research organization (CRO) that the contingency funds will only be tapped if needed.
Letting go of old models and embracing change
The fluidity offered by adaptive patient recruitment, a major asset, is often at odds with the classic concept of a clinical trial or any scientific experiment, which is to make as little change as possible. Specifically, protocols are intentionally written to clearly define the study population, the study windows during which patient visits are to occur, procedures to be done, and data to be collected and transmitted. Anything other than meticulous following of the protocol can result in a violation, raising concerns about patient safety, possible introduction of bias, and flawed conclusions.8 Adaptive patient recruitment challenges this paradigm for stakeholders who are often uncomfortable with changes to a scientifically based study. But the scientific method, with its rigid structure, is not working in the constantly changing environment of patient recruitment, as evidenced by the continual failure of many trials to enroll the planned number of patients on time.
There are a number of factors that contribute to a changing patient recruitment environment, rendering an adaptive approach all the more relevant. These include delays in drug delivery, and slower than expected country start-up and site initiation, as described in Table 2. The impact of any one of these factors on the patient recruitment and enrollment timeline can be significant, but when more than one are operating at the same time, the impact is multiplied and timelines spin further out of control. As a result, making use of data becomes even more essential if there is any hope of meeting enrollment targets on time.
Building an adaptive study
Launching successful adaptive patient recruitment starts with the building of a strategic plan that helps stakeholders engage in identifying the best sites for a specific study. This requires a data-driven approach that validates and measures enrollment success by asking the right questions. The questions might center on the profiles of prospective investigators vis-à-vis the eligibility criteria to determine which investigators might champion the study and why. Is it because of the science? Are there competing studies? Does the investigator have a large database of patients with the condition being studied?
Using purpose-built technology that is approachable and flexible, it is possible to encode these survey results to create an ideal site recruitment profile by ascribing a weighting to each response. For example, on the question of the size of the investigator’s patient database, a +1 could translate to “most compatible with the protocol”; a “0” could be “neutral”; and “-1” could reflect “not compatible with the protocol.” This method results in a ranking of sites, which is information to be shared with the sponsor and CRO. Once the stakeholders select the sites, the adaptive plan is built based on the sites that have been selected.
As the study unfolds, the data will invariably show a range of site performance, from top enrollers to those that are struggling. In true adaptive fashion, looking at results across individual sites helps stakeholders make decisions about the most effective way to invest resources to spur enrollment, possibly modify strategic branding and advertising, and ultimately, avoid rescue.
It may make the most sense to implement a multi-tiered plan to optimize the sites that are doing well and help further boost their enrollment rather than dedicating more resources to less successful performers. Those that are high enrollers may only need to take modest additional steps, such as extra advertising or more outreach to the medical community. Sites with moderate enrollment delays might benefit from conference calls and merit visits from the monitor to discuss the study in an effort to uncover why the site is lagging. And for those who have not screened a single subject after a significant period of time, it may make sense to remove those sites from the study, but still work to maintain the relationship the site has established with the monitor.
The sooner the better
Lengthy development timelines and slower adoption of technology are associated with higher costs and lower productivity gains.9 Research indicates that clinical development of a new therapeutic has recently been taking approximately 6.8 years, a 13% increase from the six year timeframe reported from 1999 – 2001.10 Patient recruitment plays into these extended timelines, yet with careful advance planning and an adaptive approach, it becomes possible to sidestep at least some of the pitfalls associated with slow enrollment. Too often, however, sponsors are reluctant to engage on this subject early on, assuming that the sites will work out the recruitment and enrollment challenge. Frequently, this mindset fosters an unwillingness to fund a contingency budget upfront meant to enable adaptive action when there are bumps in the road.
Yet, if a study falls into rescue mode, stakeholders are often suddenly quite willing to fund efforts to boost enrollment. At the same time, they are open to discussions about adaptive recruitment for the current trial and view it as a technique to be used in upcoming trials, hoping to avoid rescue in the future. Once stakeholders with enrollment responsibility understand the value proposition of adaptive recruitment, and see a strategic plan rooted in data, they can approach management and explain that the original feasibility assumptions were likely flawed, and a new approach that starts early in the process is strongly indicated.
The clinical trials industry has spent the better part of the last three decades professionalizing the patient recruitment and enrollment process.11 What was originally dismissed as an afterthought with a minimal budget, patient enrollment has emerged as a major topic of discussion, and an intractable challenge in need of new thinking. Adaptive patient recruitment offers a viable solution.
Like adaptive clinical trials, adaptive recruitment works best when everyone has on-time enrollment as a goal and it is implemented from the beginning. By letting the clinical trial data tell the story as it flows in from the CTMS and IVRS systems, the clinical team can evaluate data in real time and make informed changes to the strategic plan. In sharing this information across the team, it finally becomes possible to address the patient enrollment dilemma in a data-driven, meaningful way, benefiting patients and stakeholders alike.
Joan F. Bachenheimer is Founding Principal of BBK Worldwide