Patient Enrollment: Risks to Clinical Trial Viability

MARK B. ABELSON, MD • DONNA WELCH, RN, BSN • JAMES MCLAUGHLIN

Risk,” is one of the most important words in the clinical research lexicon. Patient risk is carefully managed and monitored by Institutional Review Boards (IRBs) and the Food and Drug Administration (FDA). But what about risk to clinical trial viability?

Patient enrollment is one of the biggest stumbling blocks in the path of a clinical trial. According to the statistics, most trials do stumble; 80% are delayed at least 1 month due to enrollment and 72% of trials are delayed more than 1 month.¹ These delays can filter through the entire drug development pipeline causing a cascade of missed deadlines.

So how should the enrollment risk be managed? In the clinic, powerful decision-making tools such as the risk:benefit ratio, treatment benchmarks, and monitoring procedures are used to guide patient care. The same approaches can be adapted to prevent and mitigate clinical trial enrollment problems.

THE PATIENT POPULATION

The broad boundaries of every patient population are set by the ethical guidelines laid out in the Declaration of Helsinki. The patient population is further restricted by enrollment criteria in a clinical trial protocol.

Every inclusion and exclusion criterion affects enrollment; that may sound obvious. But how much does a certain inclusion and exclusion requirement affect enrollment? The answer to that question is far from obvious.

The algebra used to make site enrollment estimates is relatively straightforward. Known variables, such as prevalence of disease and local population size, combined with estimates, such as the participation rate, can be analyzed with simple arithmetic. The following equation demonstrates 1 method:²

It would not be difficult to estimate the enrollment rate in a trial with minimal enrollment restrictions. Consider the following hypothetical case study:

■ Case study #1. Dr. Jane Sponsor, a retinal physician in Boston, wants to estimate the potential monthly enrollment for a neovascular age-related macular degeneration (AMD) study at her single-site practice. The protocol for the study is very thin; there are no exclusion criteria, and there is just 1 inclusion criterion (diagnosis of wet AMD). Assuming 5% participation, using the above equation, she has the potential to enroll:

or, approximately 360 patients per month in the city of Boston.

This example is not particularly relevant in the current ophthalmic clinical trial landscape, where protocols weigh over a pound, and inclusion and exclusion criteria are pages long. The recent MARINA ranibizumab trial, for example, included patients over 50 years old, who had best corrected visual acuity (BCVA) of 20/40 to 20/320, chorodial neovascularization (CNV) involving the foveal center associated with AMD, a minimally classic or occult lesion as defined by flourescein angiography or fundus photography, and many other characteristics.³

The complexity of modern protocols makes it virtually impossible to estimate trial enrollment from demographic data. The Baltimore Eye Study found 26.5% of AMD patients were in their 50s,⁴ and the Beaver Dam Eye study found 56.7% of AMD patients were female,⁵ but neither of these epidemiological studies reported the prevalence of wet AMD among patients with BCVA between 20/40 and 20/320 and minimally classic or occult lesions less than 12 optic-disc areas in size.

History is often the best indicator of inclusion and exclusion criteria practicality. Have other published studies used similar patient populations? How many sites were involved? How long was the enrollment period? Feasible enrollment criteria previously used in large, randomized, double-masked clinical trials are a good starting place for new study designs. A sponsor of a new neovascular AMD trial of an anti-VEGF agent, for instance, should carefully review the recent ranibizumab studies.

The risk: benefit ratio codified in the Declaration of Helsinki guides patient care. The declaration states, “every biomedical research project ... [should] careful[ly] assess ... predictable risks in comparison with foreseeable benefits to the subject.”⁶ The same rigorous, careful type of analysis can be used to guide clinical trial design. Each inclusion and exclusion criterion stated in a trial protocol can be considered in terms of “risks,” and “benefits.” Restrictive enrollment requirements may “benefit” the trial by making endpoints easier to reach, but may “risk” enrollment. As with patient care, the clinical trial risk: benefit ratio should always favor the “benefit,” side of the equation (Table).

The goal of a phase 3 trial, an FDA-approved indication, should also be considered in terms of “risks” and “benefits.” Small patient populations based on restrictive inclusion and exclusion criteria can lead to restrictive indications. If, for example, a drug is tested on patients with severe AMD in a successful phase 3 trial, the drug will be indicated for severe AMD. A severe AMD indication would limit adoption of the drug for milder forms of the disease. A study design with lenient inclusion and exclusion criteria can ease trial recruitment, and can also help the drug into enter the marketplace once it’s approved.

ENROLLMENT PLANS AND CONTINGENCIES

Consider another hypothetical case study:

■ Case study #2 — declining enrollment. Dr. Jane Sponsor is conducting a 2 year trial of a novel anti-VEGF agent for neovascular age-related macular degeneration (AMD). She plans to enroll 90 patients in 10 months at her practice in Boston.

During the first month of enrollment, Dr. Sponsor enrolls 12 patients in the study. This enrollment rate, 12/month, is higher than the average enrollment rate of 9/month that Dr. Sponsor will need to complete the enrollment on schedule. Dr. Sponsor is pleased with the initial enrollment, and makes no changes to the trial enrollment strategy, which is based on referrals from area ophthalmologists.

In the second month of enrollment, Dr. Sponsor enrolls 9 patients, which is the same as her targeted enrollment rate of 9/month, and Dr. Sponsor makes no corrections to the trial enrollment strategy. The third month of enrollment, Dr. Sponsor enrolls 6 patients. Although the average enrollment over the past 3 months is 9/month, the targeted enrollment rate, Dr. Sponsor is concerned about downward trend in the enrollment rate.

This phenomenon of declining rolling enrollment, known as “trial drift,” can be managed with techniques similar to those for treating a chronic, degenerative disease. In the treatment of neovascular AMD, careful monitoring and therapeutic intervention are familiar standards of care. Fluid leakage observed on fluorescein angiography has long been used as an indication for retreatment, and the PrONTO ranibizumab study recently demonstrated of value of OCT retinal imagining for determining retreatment. The combination of monitoring and therapeutic intervention can effectively be applied to trial drift as well.

Like so much in medicine, the exact etiology of trial drift is unknown. Many explanations for progressively declining enrollment have been proposed,⁷ including:

• declining interest in the study
• insufficient protocol training
• staff turnover
• competition with other studies at the site
• limited patient population.

Clinical trials for retina disease are particularly susceptible to “trial drift,” because their enrollment periods are typically long. In wet AMD trials, for instance, enrollment commonly lasts a year. During trial design, it’s important to anticipate “trial drift,” and come up with specific contingency plans to address the problem. Let’s consider our second case study again:

■ Case study #2 — treating trial drift. In the protocol of the clinical trial, Dr. Sponsor specified certain steps that would be taken if the monthly enrollment rate ever fell below 7. Since month 3’s enrollment, 6/month, was below the 7/month benchmark, Dr. Sponsor called a meeting at her Boston practice. She reintroduced her staff to the study design and reemphasized the importance of identifying patients for study screening. Dr. Sponsor also made telephone calls to the area ophthalmologists who referred patients to her practice, and printed more posters advertising the study. In the fourth month of enrollment, 9 patients were enrolled in the study.

Every study design should have appropriate benchmarks and interventions for managing trial drift. Effective plans usually include multiple benchmarks, and interventions which are proportional to the magnitude of trial drift. Studies with long enrollment periods may experience trial drift multiple times after repeated intervention. The following interventions are often useful:

• protocol retraining for site personnel
• increased advertising
• retrospective chart review to identify patients who may be eligible for the trial
• strategies for improving communication to quickly address enrollment problems
• increasing the size of the referral network
• adding additional sites (if applicable).

“The characteristically low participation rate in clinical trials has been blamed on negative media portrayals and lack of public awareness.”

NATIONWIDE TRENDS IN CLINICAL TRIAL ENROLLMENT

Enrollment criteria and recruitment strategies can be written into clinical trial protocols. One factor is out of the sponsor’s control, however; the patient participation rate.

Public perception of clinical research is a persistent problem in the United States. Only 6–12% of potential patients sign informed consent forms and enroll in clinical trials, 10–11% of Americans have participated in clinical research, and 25% drop out of before the study is completed.¹ Americans, in short, are wary of clinical trials. The public suspicion of clinical trials is conspicuously different than the public perception of medicine in general. Seventy-eight percent of recipients in 1 poll said physicians were their most trusted source of information.¹

The characteristically low participation rate in clinical trials has been blamed on negative media portrayals and lack of public awareness.⁸ Important efforts have been made by institutions such as The Center for Information and Study on Clinical Research Participation to promote understanding of the clinical research. Overall, the public perception of clinical trials appears to be improving. According to 1 poll, the clinical trial participation rate increased from 8 to 11% from between 2001 and 2004.¹ Some of the strategies used to ease clinical trial enrollment and avoid trial drift may also improve the public perception of clinical research. An emphasis on professionalism can improve the enrollment rate while simultaneously casting clinical research in a positive light.

CONCLUSION

Clinical trial design is a complicated process full of intricate considerations. Physicians who spend time in the clinic, making crucial treatment decisions, are uniquely qualified to solve trial design problems. Clinical decision-making methods are effective in the exam room, and during study design. Inclusion and exclusion criteria can be analyzed using a risk-benefit analysis. The rolling enrollment problem of trial drift can be “treated” like a chronic disease, with careful monitoring and therapeutic intervention.

Clinical trials rarely finish on time, and, according to Centerwatch, the delays are getting worse. In 1997, 18% of clinical studies were completed on schedule, but by 2003, only 6% met their timelines. Enrollment problems cause or contribute to most of these delays; 72% of trials exceed their recruitment timeline by more than a month.¹ Intuitive recruitment plans and practical enrollment criteria can keep trials on a schedule for success. RP

REFERENCES