Increasing Efficiency of Recruitment in Early Parkinson’s Disease Trials: A Case Study Examination of the STEADY-PD III Trial

INTRODUCTION

Development of effective strategies to slow progression of Parkinson’s disease (PD) is an area of major unmet need. Clinical trials of disease modifying therapies are one of the most active areas for clinical research in PD, with an average of 2 studies per year, over the past 25 years [1]. Unfortunately, no studies to date have yielded positive, conclusive results. A still nascent understanding of the pathogenesis and heterogeneity of the disease combined with a lack of biomarkers and suboptimal animal models contribute to lack of success. Other challenges include slow recruitment into complex and often lengthy clinical trials. This is further compounded by the desire to enroll individuals who are newly diagnosed with PD, often before they have initiated symptomatic treatment, as most clinical trialists believe that this stage of the disease is the best therapeutic window for slowing or halting disease progression. In light of these challenges, this paper will examine methods that enabled full recruitment in STEADY PD III, a National Institute of Neurological Disorders and Stroke (NINDS)-funded Phase 3 study of isradipine in PD.

MATERIALS AND METHODS

RESULTS

A study enrollment report (Fig. 1) shows a steeper slope of actual vs. anticipated enrollment, reflective of a recruitment period accelerated by 6 months. In addition, the prespecified goal of 10% minority recruitment was met.

Fig.1

Actual vs. Projected Enrollment.

Recruitment success in STEADY-PD III was bolstered by a multitude of subject referral sources as reflected in (Fig. 2). Referral sources were recorded at the time of screening and logged into case report forms. These data indicate the top four referral sources as: Site personnel (53.8%); Neurologists (24%); FTF (10.2%); and MJFF Communications (3.9%).

Fig.2

Referral Sources: Enrolled STEADY PD III Volunteers.

Analysis of MJFF communications that took place prior to and throughout the recruitment period provides insight into the role of mixed media in generating awareness of the trial, and directing individuals to resources for learning more about participation. Starting in March of 2014, MJFF, with study leadership, released a podcast that reported the compound Isradipine was moving to Phase III testing, and recruitment would begin later that same year. The podcast was downloaded by a total of (2,043) iTunes listeners and followed by an uptick in STEADYPD3.com site traffic beginning in May 2014 and peaking in July 2014 (Fig. 3). One of the steepest peaks occurred in January 2015 following a December 2014 webinar by MJFF that focused on therapies with the potential to slow or stop Parkinson’s progression and highlighted the STEADY PD III trial. A third peak took place November 2015 following an MJFF webinar and podcast on studies to slow/stop Parkinson’s disease on October 2015. While proprietary ownership of web analytics makes it a challenge to compare web traffic of STEADY PD III to that of other studies, a better understanding of visitor traffic to and behaviour on clinical trial websites would help to elucidate their utility as a recruitment tool and remains an important area for future research.

Fig.3

STEADYPD3.Com and MJFF Mixed Media Outreach.

Comparison of actual enrollment vs projected enrollment for those sites that filled out recruitment activity questionnaires (n = 54) revealed that (20.4%) met their goals, (24.1%) exceeded their goals, and (55.6%) fell below projected goals. Major themes from the qualitative analysis were as follows: 1) goal setting; 2) site team communication; 3) anticipated recruitment barriers; 4) use of recruitment toolkit materials; 5) community engagement; and 6) patient tracking methods. Our analysis revealed that a consistent challenge, across sites, was the ability to accurately predict recruitment goals. That only 20% of sites met their recruitment goal compared to the 80% that exceeded or under-enrolled further supports this finding. Communication was similar across sites that exceeded enrollment and under-enrolled, with both groups citing weekly check-ins between the investigator and the coordinator. All sites, independent of recruitment results, reported similarly on anticipated barriers, which included: concerns around finding individuals that met the eligibility criteria (i.e. de novo, not already taking medication and/or early PD), and issues with logistics, such as transportation, scheduling, workload levels, and finding time to contact patients. Sites that exceeded or under-enrolled most frequently reported no potential barriers, which may have contributed to or mitigated successful achievement of enrollment goals. The recruitment strategy for all three enrollment groups involved a multi-modal approach, with particular emphasis on outreach to clinician groups and placement of materials in areas of high volunteer traffic (primarily at the site or in clinics). All enrollment groups included an element of social media (list servs and websites) and participation in local events, although the latter two elements were not emphasized as much as the former. In regards to community engagement, connections with patient support groups were cited more frequently than with physician referral networks. Finally, all enrollment groups reported some form of tracking method for conducting outreach to and screening of volunteers, with the majority of sites reporting use of a paper log and or excel spreadsheet. Only one or two sites in the exceeded enrollment and under enrollment groups reported use of a secure database.

DISCUSSION

STEADY PD III SC was able to successfully recruit full target population six months ahead of schedule due to a pre-specified comprehensive recruitment plan and involvement of key stakeholders early in the planning phase of the clinical trial. These actions allowed SC to identify study and site level barriers that had the potential to negatively impact recruitment, and develop a strategy to mitigate them. Key components of this strategy consisted of: 1) a reduction in protocol complexity; 2) implementation of a comprehensive outreach and awareness campaign; and 3) rigorous selection and activation of clinical trial sites.

The fact that the protocol complexity was identified as one of the study level barriers is not surprising. Clinical trial complexity is an issue that has been on the rise over the past 20 years. Research by the Tuft’s Center for the Study of Drug Development found that in this time, the average number of inclusion criteria have increased threefold and the number of unique procedures per protocol have increased by 6.5% [4]. This growing level of complexity has negatively impacted recruitment, resulting in a 53% increase in the times between study initiation and completion, and a 16% reduction in enrollment rates [4]. The authors posit that this trend is attributable to a shift in focus to chronic diseases, where clinical endpoints are difficult to measure and the target population for recruitment is highly specific. Additional factors, such as study design lend themselves to more complex protocols, with studies of efficacy tending towards greater stringency in eligibility criteria to allow them to recruit for a more homogenous patient population and facilitate assessment of the outcome of interest [6]. STEADY PD III, based on the study objectives, was inherently more complex – seeking to demonstrate efficacy in a disease state that currently lacks biomarkers and among a patient population that can be challenging to recruit (i.e. those recently diagnosed and not yet on ST). Modifications to the study protocol addressed this barrier by loosening the stringency of the inclusion and exclusion criteria without jeopardizing scientific integrity. This in turn, increased the number of individuals who were eligible to participate in the study, lowered the burden for potential participants, and broadened the pool of individuals from which sites could recruit. The positive impact of these modifications on recruitment can be seen in a comparison of the low rate of screen failures from STEADY PD III (18.6%) to that of parallel studies of disease modification in patients with early Alzheimer’s disease, which can reach rates upwards of (85%) [7].

Gaps in patient knowledge around the availability of clinical research and the benefits of research participation were also identified as a barrier to recruitment. A 2001 survey by Harris Interactive found 85% of patients were unaware that participation in a clinical trial was an option at the time of diagnosis, and 75% of these patients would have been willing to enroll had they known it was a possibility [8]. Central to addressing this barrier was involvement of key stakeholder groups who leveraged their unique capabilities to engage with communities and raise awareness. Stakeholders such as PDF maximized peer to peer engagement via the PAIR program and local events such as the Brain and Health Fair and the Unity Walk; NPF harnessed the power of social media through webinars and press releases; MJFF leveraged technology such as Fox Trial Finder to connect volunteers to trial teams; and The Muhammad Ali Foundation focused on increased participation of historically underrepresented minority populations via community engagement, translation of materials, and outreach via the “Southwestern Parkinson’s Newsletter.” The use of localized, grassroots events, and social media activities, combined with a proactive approach to recruitment helped to engage and make aware a broader swath of the population than would have been possible for clinical trial sites alone. This approach also enabled study teams to connect with a more diverse patient population who obtain their information from a variety of media and news sources. The impact of these efforts on facilitating recruitment for the STEADY PD III trial are challenged by the data reflected in (Fig. 2) on self-reported referral sources. We posit that this is less about the efficacy of these efforts and more about the challenges stemming from memory recall bias in referral source attribution. Greater efforts such as interviewer training, better referral source definition, and alternative means of data collection should be considered for future recruitment campaigns to improve the accuracy of attribution [9].

A comparison of STEADY PD III enrollment figures to 2013 Tufts Center for the Study of Drug Development (CSDD) benchmarks demonstrated that the number of STEADY PD III sites that exceeded enrollment goals (24.1%) were nearly double that of the CSDD benchmark (13%), while the percentage of STEADY PD III sites that met their goals (20.4%) was nearly half that of the (39%) CSDD benchmark [10]. The STEADY PD III sites that under-enrolled (55.6%) were ∼19 percentage points above the CSDD average (37%) [10]. These data suggest that: 1) there were a number of high performing sites, approximately one quarter of those participating in the study, that played a large role in meeting recruitment goals; 2) study leadership accurately identified site level characteristics that were needed for successful recruitment and targeted sites that exhibited those characteristics; and 3) site activation processes, such as customized recruitment plans, helped to address the majority of site-level barriers, but some still remain. These data also indicate that sites may experience challenges accurately predicting their recruitment goals, and it may be of benefit for study leadership to consider making recruitment competitive from trial initiation to avoid any inefficiencies or additional costs [11]. Two additional areas for site level enhancements are: greater automation of tools and organizational systems to track recruitment efforts, which would allow sites to more quickly respond to high or low volumes of study participants; and increased engagement with clinician groups to foster long term referral networks.

While the STEADY PD III approach to clinical trial recruitment was a success, a few key limitations should be taken into consideration. A reduction in protocol complexity may not be scientifically feasible for all clinical research studies. If considering, the impact that the modifications will have on scientific rigor and validity should be weighed before attempting. STEADY PD III also had a relatively low level of participant burden, which could make these results somewhat less applicable to studies requiring more complex procedures. Another limitation, especially when trying to quantify the impact of partnerships on recruitment numbers is that referral sources may suffer from memory recall bias; multiple sources could have led to an individual’s awareness and participation in a study. Consistent and descriptive categorical definitions should be provided to site staff to better assist individuals in determining the source that led to their study involvement. Finally, the drug being tested for STEADY PD III had been repurposed, meaning it had been FDA approved for other indications which may have positively skewed attitudes and perceptions of safety for participants.

CONCLUSION

STEADY PD III serves as an important example of methods that can be used to increase clinical trial recruitment. Involvement of key stakeholders early on is important for learning about constituent motivations and barriers to participation. There is growing interest in improving the value and efficiency of the clinical trial process, including the identification of site qualities (geography, investigator characteristics) and the multi-modal methods (e.g., recruitment tools, social media platforms) needed to optimize recruitment success and improve clinical trial efficiency [12]. Development of a comprehensive recruitment strategy can be used to build awareness and bridge knowledge gaps; and technology, both new platforms and social media, has an ever-growing role in the recruitment landscape. Additional research is needed to better understand the impact of multi-modal outreach efforts (e.g. channels that are most effective for recruitment efforts and how they compare to one another; messaging that resonates with volunteers and why), and how these lessons might be applied to future recruitment campaigns. Gaps remain on ways to improve site infrastructure and methods to enhance recruitment, and on-going efforts should be dedicated to increase participation of minorities in clinical research.

CONFLICT OF INTEREST

The authors have no conflict of interest to report.