Recruitment in Acute Stroke Trials: Challenges and Potential Solutions

Abstract

Randomized clinical trials of acute stroke have led to major advances in acute stroke therapy over the past decade. Despite these successes, recruitment in acute trials is often difficult. We outline challenges in recruitment for acute stroke trials and present potential solutions which can increase the speed and decrease the cost of identifying new treatments for acute stroke. One of the largest opportunities to increase the speed of enrollment and make trials more generalizable is expansion of inclusion criteria whose impact on expected recruitment can be assessed by epidemiologic and registry databases. Another barrier to recruitment besides the number of eligible patients is availability of study investigators limited to business hours which may be helped by financial support for after-hours call. The wider use of telemedicine has accelerated quicker stroke treatment at many hospitals and has the potential to accelerate research enrollment but requires training of clinical investigators who are often inexperienced with this approach. Other potential solutions to enhance recruitment include rapid pre-hospital notification of clinical investigators of potential patients, use of mobile stroke units, advances in the process of emergency informed consent, storage of study medication in the emergency department, simplification of study treatments and data collection, education of physicians to improve equipoise and enthusiasm for randomization of patients within a trial, and clear recruitment plans, and even potentially co-enrollment, when there are competing trials at sites. Without successful recruitment, scientific advances and clinical benefit for acute stroke patients will lag.

Introduction and Historical Perspective:

Acute stroke progresses rapidly, whether it is due to acute loss of blood flow with lack of oxygen to brain tissue (ischemic stroke) or due to an expanding mass of blood in the brain from a ruptured artery causing both mechanical damage and toxicity to surrounding brain tissue (intracranial hemorrhage). Thus, the opportunity to improve outcome is dependent upon treatments that must be delivered within minutes to hours, with the greatest impact within the first few hours after onset of the stroke for most patients. While brain imaging has allowed for a longer time window in some patients with ischemic stroke, minimizing the physiologic time to deliver an effective treatment remains the most important determinant of outcome.

Clinical trials, which build upon the observations of basic science and epidemiologic research, are often the major rate limiting step for advances in acute stroke therapy. However, the environment for recruitment in acute stroke trials has changed since the seminal trials in the late 1980s and 1990s. The Pilot NINDS rt-PA Stroke Trial laid the groundwork and methodology for hyperacute enrollment of stroke patients within 90 minutes of onset.¹ The two subsequent randomized NINDS rt-PA trials demonstrated the effectiveness of rt-PA for acute ischemic stroke.² Remarkably, six-hundred and twenty-four patients were enrolled at eight regional centers (35 affiliated hospitals) over three years within three hours of onset; approximately half of patients were enrolled within 90 minutes. Exception from informed consent for hyperacute enrollment in trials was not available at the time of the trial because it was first created by the Food and Drug Administration in 1996.³ Several factors contributed to the success of this landmark trial: a desire to emulate recent successes in cardiovascular medicine by a core of highly motivated investigators, a new every-minute-counts approach to treatment at these centers, equipoise from the lack of any scientifically proven treatment for acute ischemic stroke, willingness of participating investigators and physicians to enroll 24 hours a day/7 days a week, highly qualified study coordinators (e.g., experienced clinical nurses), financial support for these clinical nurse coordinators to support enrollments 24 hours a day, and willingness of the investigator teams to take call and go to the hospital quickly when a potential patient arrived.

Today, the environment for enrollment of patients in acute stroke trials is substantially different. Acute stroke treatment in the U.S. is usually initiated by physicians in emergency departments who contact on-call stroke teams. Many times, the stroke team members are not initially in the hospital and much of the assessment is done virtually and remotely by looking at brain and vascular images by various software programs and using telemedicine to interface with patients and family. Research study coordinators may not have previous neuroscience experience. Recruitment of patients outside of weekday business hours is less common, even at major academic and research centers. This latter situation is driven by the economics of paying individuals for evening-hours/weekend call and a lack of research coordinators/investigators willing to take call during evening hours and weekends.

As a specific illustration of the differences between the NINDS tPA trial era and the current environment, all of the original regional site trial coordinators in the NINDS tPA Stroke Trials were clinically-trained nurses, often with clinical neuroscience experience. In the ongoing NINDS-funded Multi-Arm Optimization of Stroke Therapy (MOST) trial,⁴ 21 (31%) of the 67 site coordinators are clinically-trained nurses. While there are other difference between the MOST trial and NINDS tPA Stroke trials that impact recruitment (the COVID epidemic, greater complexity of drug administration in MOST, 24-hour call at all sites in tPA Stroke Trials, etc.), MOST has recruited 400+ subjects at 60+ sites over three years as compared to 624 subjects in the NINDS tPA Stroke Trials at 8 regional centers (35 sites) over 3 years. Both trials have a 3-hour time-window for initiation of lytic therapy.

The National Academies of Science, Engineering and Medicine recently published “Envisioning a Transformed Clinical Trials Enterprise for 2030” that emphasized patient-centeredness, diversity, use of technology, simplification of data collection, innovative trials designs, etc., as the future of clinical trials.⁵ However, the practical challenges of recruitment in hyperacute stroke trials, and clinical trials in emergency conditions more broadly, are not specifically addressed. National attention and discussion are needed by key stakeholders, including clinical trialists, the NIH, industry, policymakers, and regulatory and ethical specialists to address these practical challenges. Here, we outline challenges in acute stroke trials and present potential solutions which can increase the speed and decrease the cost of identifying new treatments for acute stroke.

Factors Impacting Recruitment in Acute Stroke Trials (Table 1)

Table 1:

Recruitment Factors for Acute Stroke Trials

Recruitment factor	Short term modifiable (Yes/No)	Potential solutions to improve recruitment	Positive features of solution	Potential limitations to solution
Study Population and Treatment Time Window
Frequency of stroke or stroke subtype in general population	No	Not applicable
Population included in trial (inclusion and exclusion criteria)	Yes	Whenever possible, limit exclusions unless strong biologic or design reason. Availability of resources to recruit all eligible patients (e.g. Spanish translations).	Less exclusions mean more generalizability, more diversity, and more eligible participants.	Participants may be less likely to benefit or more likely to have adverse events.
Treatment window from stroke onset to enrollment	Yes	Always match time window to known pathophysiology, even if it makes recruitment harder.	More available patients.	Less likely for treatment to alter outcome if treatment window exceeds pathophysiologic time window.
Trial Infrastructure and Availability of Investigators
Number of sites	Yes	Detailed feasibility assessments; monitor closely to add sites earlier rather than later.	More available participants.	Costs and time to get sites enrolling.
Inclusion of non-US sites	Yes	Use of foreign-based CRO to handle all start-up, contracting, regulatory, and payments.	More available participants and recent better recruitment rates than in US.	Costs and time to start-up site in other countries when study funded by NIH and under NIH requirements; integrating new data privacy requirements in Europe may delay contracting.
Site of enrollment – mobile stroke units	Yes	CMS reimbursement for MSU.	Quicker access to patient and LARs.	Major expense in equipment and personnel.
Availability of study team outside of business hours, Monday-Friday	Yes	Planned financial reimbursement for call or after-hour enrollments for trial coordinators, prioritize sites with 7-day week enrollment, training and use of in-hospital trial coordinators for multiple types of acute studies (stroke, brain trauma, seizures, etc.). Neuroscience training of trial coordinators.	Increased access to eligible patients.	Costs of call coverage. Availability of willing investigators.
Financial remuneration for screening and enrollment	Yes	Make sure that screening and ongoing study costs are accounted for in budget including after-hour call coverage; provide additional upfront/ongoing support to those sites which consistently demonstrated successful recruitment.	Greater likelihood sites can support personnel and recruit.	Limited financial resources economic feasibility of trial. More remunerative studies may take precedence; sites may opt out; site coordinators given too many trials may hurt recruitment and data quality in all trials.
Trial Design/Competing Trials
Complexity of study treatment	Yes	Simplify study procedures and data collection for the site, regardless of the underlying design complexity.	Less likely to have errors and more likely to be embraced by site.	Device or drugs may not be simple to deliver.
Blinding of treatment allocation	Yes	When blinding isn’t feasible for recruitment needs, consider PROBE design.	Potentially easier and more rapid enrollment.	Potential treatment biases.
Competing trials	Yes	Require clear enrollment plan at sites, document expected recruitment rate, co-enrollment in studies when feasible.	More patients recruited when populations don’t completely overlap.	Recruitment in all competing trials decreased unless co-enrollment feasible; investigator confusion between trials; patient fatigue if approached for multiple trials.
Investigator-Patient Interactions – Process of Consent
Communication between investigator/patient/family including exception from informed consent (EFIC)	Yes	Extensive training, mock enrollments, sufficient financial resources for EFIC. Centralization of EFIC workload as much as possible.	Greater likelihood of enrollment.	Setting up EFIC is a long and intensive process prior to start of enrollment, much harder during COVID.
Telemedicine assessment of stroke patients as part of practice	Yes	Training of investigators, mock enrollments, electronic consent.	Quicker access to LAR and consent, less errors with electronic consent.	Inexperience with research enrollment with telemedicine, harder to establish rapport and trust for consent.
Unwillingness of patients to participate in research	Yes	Training in research discussions for underrepresented minority populations, more diverse research investigators, increased diversity of populations of recruitment sites, public education about research participation and its societal benefits in general.	Greater diversity of patient population.	Requires investment in long-term strategies to train staff and community, and to develop staff recruitment pipelines.
Equipoise and Trial Enthusiasm
Equipoise and enthusiasm for trial	Yes	Education of participating physicians; choose only sites with equipoise.	Greater likelihood of enrollment.	Challenge to current practice and belief; potential economic loss of income for physicians if a procedure is not selected for randomization.

Major components that impact successful recruitment in an acute stroke trial include unmodifiable factors such as the subtype of stroke included in the study (e.g. ischemic stroke) and its annual incidence. Modifiable factors include the subgroup of stroke patients to be included in the study (Table 1), the untreated pathophysiology of the stroke subtype that is strongly related to the time from stroke onset to initiation of study intervention, the number and geographic location of recruiting sites, the availability of study team members during the time window for enrollment, financial support for the study team enrolling patients, the complexity of the study intervention(s) and study design, competing trials, the communication between potential participant and investigators regarding participation in the study, the use of telemedicine as part of clinical stroke practice, the regulatory environment, and the equipoise and enthusiasm of the treating clinical investigators.

Factors associated with premature termination of planned enrollment in acute stroke trials include less inclusiveness of older patients, greater complexity of trial design, industry sponsorship, pre-trial planned interim assessments for futility, and geographic location of trials (North America with higher rates of early termination compared to trials in other world regions).⁶ Futility is a planned statistical approach to end a trial that is very unlikely to be positive if recruitment of the entire planned sample size is accomplished. Futility designs do impact planned recruitment of participants but in a positive way since these patients can be potentially enrolled in other acute trials and not enrolled in a discontinued trial where benefit is highly unlikely. Not surprisingly, Industry sponsored trials were more likely to use futility assessments than non-industry sponsored acute stroke trials because decisions regarding futility of a trial can decrease the costs of new drug development.

Potential solutions:

Greater Inclusiveness of Enrolled Population While Respecting the Timing of Pathophysiology

One of the largest opportunities to increase enrollment and make a trial more generalizable is expansion of inclusion criteria without diluting anticipated treatment effects. In the planning for feasibility of the impactful Endovascular Therapy Following Imaging Evaluation for Ischemic Stroke 3 (DEFUSE 3) trial,⁷ investigators expanded the originally planned inclusion and exclusion criteria (age 90 versus 80, NIHSS ≥ 6 versus NIHSS 8–25, mRS 0–2 versus 0–1, and time from onset 6–16 hours versus 6–12 hours) because of the potential to increase recruitment and generalizability. Overall, 57% (104/182) of enrolled patients qualified only via the broadened study entry criteria. In subsequent implementation of the expanded inclusion criteria, the trial randomized 0.47 patients per site per month, nearly twice the expected enrollment rate.⁸

The primary focus for any acute stroke trial is the untreated pathophysiology of a given stroke subtype since the goal is to change the natural progression of the acute stroke. Applying a treatment during a time window after the pathophysiology is largely completed will have little or no impact on patient outcome and could even increase risks, even while it may increase patient recruitment. A great example of this is the typical time course of spontaneous intracerebral hemorrhage (ICH), where evidence points strongly to most bleeding occurring within the first few hours after onset. Yet, most trials to slow bleeding included patients substantially after this time window.^9–11 Thus, expanding the time window for an acute trial without a clear biologic rationale should be avoided, even if it enables enrollment of more eligible participants.

Enhancement of Trial Infrastructure and Availability of Investigators

The number of sites in trials for acute stroke trials is often underestimated because of over-optimistic projected rates of recruitment from sites concerning available eligible patients. Thus, detailed assessments of feasibility using objective data are critical to determine needed number of sites. These assessments can include use of epidemiologic databases that can provide known frequency of type of strokes in the general population and the effect of planned inclusion and exclusion criteria on eligibility rates as noted for the DEFUSE III Trial.¹² In addition, registries such as the Get with Guidelines Stroke Registry can provide an excellent source of preliminary feasibility data.¹² Even with excellent feasibility assessments, the enrollment projection is often an underestimate.

One of the reasons for the underestimate is that recruitment is not just based upon the number of eligible patients but also the availability of study investigators. In the ongoing MOST Trial,⁴ sites were surveyed regarding the time when the trial was open for enrollment. Those sites that were open for enrollment every day recruited participants at twice the rate of those sites who were only open for enrollment during business hours, Monday through Friday.¹³ Thus, projection of recruitment rates must account for when sites are able to enroll. One potential solution is to provide financial support for weekend and after-hour calls, not only for those sites who currently do have after-hours call but to support efforts at sites that currently don’t to encourage adding this capacity. A trial may not need more sites but simply more sites recruiting outside of business hours. The more acute stroke trials that cover these costs for after-hour coverage, the more individual sites can sustain after-hour enrollment in acute trials overall.

The widespread use of telemedicine has accelerated quicker stroke treatment at many hospitals since physicians don’t have to drive to the hospital after hours, but most physicians and coordinators have much less experience with consenting and enrolling patients via this mechanism alone. Developing a trusting relationship with patients and their families is fundamental for discussion of clinical issues and potential participation in acute stroke trials and doing this virtually, rather than in person, can be more challenging. However, the same virtual technology can provide a great mechanism for obtaining electronic and virtual consent. Electronic signed consent prompted by email (e.g., using a standard REDCap platform) permits an alternative to the cumbersome need to obtain an in-person signature in ink. Some sites have become quite proficient with this approach and have been leaders in recruitment in acute stroke trials (Personal written communication, August 2, 2022, Dr. Chris Streib, second leading enroller in MOST Trial and leading enroller in the Trial of Thrombolysis in Imaging-eligible, Late-window Patients to Assess the Efficacy and Safety of Tenecteplase (TIMELESS Trial). The REDCap platform enables site-specific storage of the electronic consent and HIPAA documents that meet the requirements for each step of the consent process. Other electronic signature platforms such as DocuSign can be used to document written consent, but the critical part is to document and safely store the electronic consent properly. Training in these techniques, mock enrollments, and team approaches to enrollment can help facilitate implementation of this method at other institutions. Virtual enrollment may make enrollments after hours and on weekends more doable with less time spent travelling to and from the hospital for the on-call coordinator and investigator. Having more than one research staff member involved in the enrollment through this process increases likelihood of successful enrollment during short time windows. Furthermore, having one staff member (typically the research coordinator) on-site often remains necessary.

Earlier notification of the clinical team and research investigators also can accelerate enrollment. Solutions include very early notification of the research/clinical team by the emergency medical system (dispatch noting that potential stroke patient is on way) and imaging applications such as Viz.ai and RAPID that notify clinical/research teams of specific imaging findings (e.g., large artery occlusion, perfusion mismatch, presence of ICH) even before emergency medicine physicians have alerted the team.¹⁴ Finally, mobile stroke units can rapidly access stroke patients in the field where they can image patients, obtain consent, and initiate a study treatment more quickly than in the emergency department.^15,16

Innovative Approaches to Patient Interactions in Emergency Settings

One of the most time-consuming elements for enrollment in acute stroke trial participants is discussion of the trial by the study investigators with the patient and, more frequently, the patient’s family or legal authorized representative (LAR).¹⁷ Identification of the key decision maker within a very short time window is often challenging and subsequent detailed discussion of the trial even more so. It is a very different situation as compared to stroke prevention and stroke recovery trials where conversations often happen over several visits. Methods to minimize time include a shortened informed consent form containing only critical information with accompanying information sheet with much more detailed information that can be read by patient and family later and a two-step consent process (much shorter form followed by a more detailed form later).¹⁸ In this latter approach used in the Tranexamic Acid for Hyperacute Primary Intracerebral Hemorrhage 2 (TICH-2) trial, when it was deemed not feasible to seek full written informed consent due to a short therapeutic time window, patients provided initial consent by signing a brief one-page information sheet (288 words) that explained the condition (ICH), treatment (tranexamic acid or placebo), blinding, and that there would be an additional computed tomography (CT) scan after 24 hours.¹⁹ This process was followed by full written informed consent (a 4-page information sheet of 2474 words) at the earliest subsequent opportunity after enrollment. These approaches can also utilize electronic virtual consent, which as noted previously, is an easier way to obtain urgent inform consent when the LAR is not able to be present at the hospital.²⁰

When the expected window for benefit for a given treatment is within 1–2 hours after onset, requiring informed consent on all participants prior to randomization and treatment is likely to fatally undermine recruitment. When informed consent would delay a time-sensitive standard-of-care treatment such as alteplase, then the trial becomes infeasible. Exception from informed consent (EFIC) or other methods of emergency consent can help patient recruitment dramatically and may even be the only ethical approach.^3,10 Under exception from informed consent, patients who cannot provide consent and for whom a LAR cannot be identified for consent initially, can be enrolled in the trial. However, efforts to obtain consent continue after enrollment until written/electronic informed consent is obtained although sometimes this is not feasible despite best efforts. The challenge of EFIC in the US is that, prior to IRB approval to enroll participants, investigators must first document community consultation in the region in which the site is located, as well as public disclosure of the trial by various means. This requires financial resources and substantial investigator time that often delays start of IRB approval by 6 months and frequently longer during the coronavirus disease (COVID) era. Other countries in Europe and in Canada which utilize emergency consent procedures do not have the regulatory requirement of community consultation and public disclosure. EFIC has been particularly challenging during COVID because gathering community members together for discussion of the trial has been challenging.²¹ In the just starting global trial of recombinant Factor VIIa for ICH within two hours of onset (FASTEST), 15 of the first 25 participants have been enrolled using EFIC or similar emergency consent procedures in other countries (personal communication – Joseph Broderick, PI FASTEST Trial). Finding creative solutions to substantially decrease the regulatory burden of community consultation and public disclosure at sites participating in EFIC studies in the U.S., compared to other countries, should be an ongoing regulatory, legislative, and research focus going forward.

Training on how to approach and discuss trial research with potential underrepresented minority participants should be part of every acute stroke trial and done by investigators experienced in this area. Training of a more diverse research work-force is also an important issue that extends beyond a given acute stroke trial and remains a larger goal for our stroke research community. Finally, inclusion of patient representatives into stroke trial design and conduct ensures that the voice of the patient is heard clearly as we consider how best to discuss trial enrollment in acute settings.

Simplification of Trial Design and Approaches to Address Competing Trials

Complexity of a trial is a barrier to enrollment within a critical time window. Currently, trial designs are becoming more complex, by testing multiple therapies within a given trial to increase trial efficiency, rather than two or three separate trials. Some of the advantages include reducing sample size with the same control group, accounting for potential treatment interactions, adding new promising treatments during the course of the trial, and creating other statistical efficiencies to get answers more quickly.²² Adaptive trial designs and platform trials have been very successful in cancer and COVID, particularly when using easy data collection (registry based or simple case report forms) and integrating research within clinical care. Thus, the overarching principle remains that study activities should be as simple as possible from the perspective of the potential participant and the site investigators and staff, in order to minimize time to treatment, minimize errors in administration of study treatments, obtain informed consent, and generate uptake by sites.

Double-blind design, whenever possible, is the gold standard for randomized trials, but in the setting of acute stroke trials often requires involvement of a research pharmacist to prepare study medication (particularly with a blinded study). Research pharmacists are often only available during business hours, Monday through Friday, and study medication may require storage in the research pharmacy, particularly if preparation of the study medication is more complex. Potential solutions to decrease time to enrollment include storage of study medication in the emergency department and training of pharmacists or other qualified personnel in the emergency department to prepare study medications after business hours, as well as step-forward randomization.²³ When double-blinding is infeasible, blinded assessment of outcomes is a reasonable accommodation to minimize potential bias.

Another factor affecting successful recruitment in acute stroke trials is competing trials that address the same or an overlapping population of eligible patients. When evaluating a site for recruitment in a trial, each site should have a recruitment plan that indicates how trial selection is prioritized at their site. Trials with a small amount of overlap in populations of eligible patients may actually help enrollment in both since the research team may be more active and attentive to enrollment and additive financial support can help support the research coordinators. However, for those trials with a large overlap in eligible patients, clear expectations and a defined enrollment plan should be set prior to inclusion of the site in a given trial.

There are multiple potential approaches to address competing trials with the same eligible patient base. One approach is a fixed allocation grid that specifies pre-set priorities (source of funding, time of day, time from onset to enrollment) and/or stratified by eligibility criteria. The grid states which trial is “up” first. It is expected that if a patient isn’t eligible for that trial, then the next trial that is “up” would be offered. The trial that wasn’t selected then moves to the top of the list. In general, acute stroke trials will be first up for consideration given the very small time-window between onset and consideration of enrollment. Co-enrollment in trials/studies can also be done if a second research study will not impact the primary outcome in the first trial (for example, genetic study in setting of patients enrolled in an acute stroke trial). If a patient can be enrolled in a prevention or recovery trial after the primary endpoint for an acute trial which is generally at three months, patients could be enrolled in both trials. Statistical methods are also available for co-enrollment in two trials but are more complex. Certain principles are key whatever framework is used: limitation of clinical bias, communication and agreement of the recruitment plan by the entire investigator team, and respect for patients and families to limit the burden of expectations and requirements.

Assessing and Enhancing Equipoise and Enthusiasm for Trials

Finally, two of the most important considerations for recruitment are clinical equipoise and the enthusiasm of investigators to randomize patients rather than treat them clinically outside of a research trial. Clinical equipoise is where either of two actions is reasonable based upon current understanding and knowledge. Clinical equipoise is generally defined by lack of certainty among the expert medical community about the preferred standard treatment,²⁴ although decisions to offer enrollment in a clinical trial are made by an individual investigator. The two states may overlap but aren’t the same. For example, there may be clinical equipoise among medical experts but an investigator may lack enthusiasm for a trial based upon personal clinical experiences, complexity of the trial design, payments for study enrollment, lack of personnel, impact upon their practices, or their own understanding of the literature, etc. However, education of investigators with new knowledge and perspective can improve their understanding of community equipoise and thereby affect their enthusiasm. Trial PIs should educate and then survey all site investigators at a potential site to determine enthusiasm for a trial and willingness to randomize. Level of enthusiasm for a trial should be an important factor in site selection. Screening logs can help determine the number of patients that are excluded, not only because of eligibility, but by the availability and willingness of the investigators to randomize a given patient into a trial. Unfortunately, reimbursement for screening activities in acute trials is often lacking and the resulting quality of screening logs are often poor.

Summary:

Recruitment into acute stroke trials remains one of the challenges for advancement of acute stroke therapy. The scientific and theoretical considerations of clinical trials are often easier and more interesting than their practical implementation. Yet without successful recruitment, scientific advances and clinical benefit for patients will lag.

Non-standard Abbreviations and Acronyms:

ICH

Intracerebral hemorrhage

EFIC

exception from informed consent

COVID

coronavirus disease

mRS

Modified Rankin Scale

rt-PA

Recombinant tissue plasminogen activator

PROBE

Prospective, Randomized, Open, Blinded End-Point