Improving Patient Recruitment Rates in Oncology Trials

The percentage of oncology patients participating in clinical trials is remarkably low: 8.1 percent according to a recent meta-analysis. While this is more encouraging than the commonly cited value of 3 percent, it is nonetheless disappointingly low and explains the daunting recruitment challenges that sponsors face in most oncology trials.

Low participation rates are largely attributed to access issues (the availability of trials in the right geography for a given cancer type or stage) and eligibility restrictions. However, a number of other factors often compound the situation, including a heavy patient burden (in terms of clinic visits, tissue sampling requirements, etc.), patient misconceptions concerning trials, lack of awareness on the part of patients and physicians, and a failure to reach patients with study information at the right time in their treatment journey.

Cancer patients frequently rely on their physicians to inform them of trials for which they might be eligible. Meanwhile, many investigators appear to wait for the ideal study participants to walk through their door at just the right time in their treatment progression. As a result of these two factors, which ultimately leave a key aspect of patient recruitment to chance, it is little wonder that most trials experience delays due to slow enrollment.

Fortunately, enrollment rates and study timelines can be vastly improved by understanding the patient pathway and working to meet patients where they are as they travel along it.

The site identification process for all clinical trials should take into consideration the type of treatment center where patients who fit the target profile will present for diagnosis and treatment. It is, however, especially important to do so in oncology trials due to the complexities in treatment across tumor types and disease stages. Academic centers tend to treat patients with rare or advanced cancers, while community physicians tend to treat patients who are in first-line treatment for more common tumor types. Indeed, in the U.S., 85 percent of patients receive treatment at community facilities versus 15 percent in larger academic centers.

The process of identifying the best sites for a clinical trial is data-driven. Scores of databases are available to provide insight into the top prescribers by indication, and sophisticated algorithms can be applied to prioritize investigators based on any number of criteria including their experience, capabilities, facilities and performance history. The results can then be overlaid onto a map at a granular geographic level indicating where patients with a given diagnosis are physically located. To inform site selection, the results can be displayed as a “heat map” indicating the places of greatest overlap between qualified sites and high numbers of patients. (See Figure 1.)

Further, it is possible to analyze patterns in the treatment pathway for a given tumor type to reveal which medical specialties patients are visiting and to model the local and global referral process. (See Figures 2,3.)

In this example, we identified top- and second-tier key decision makers (KDMs) among oncologists in New York state to promote a specific chemotherapy/immunotherapy drug. (Traditional approaches to KDM-mapping rely on signals available to everyone else, resulting in the same lists.) We mapped referrals from one physician to another as an indicator of recognized domain expertise and used referrals as a measure of influence. Using this approach, we were able to reveal and influence the right patient population by understanding the relationships between the numerous healthcare systems and patient facilities. The output showed the shared percentage of patient referral influence for the disease state spanning across the hospital setting, the community-based care establishments and digital engagement landscape.

As seen in Figure 3, we then scaled geo-information up and down to the national or state level and cross-referenced it with patient volume and referral frequency. This included granular filters all the way down to the physician specialty and ZIP code level, resulting in a ranking of oncologists for each geographic area by a number of salient traits, cross-validated against an external source. The decile-ranked patient populations increased awareness and informed the true disease state prevalence mapped to the diagnosis to help better position the right audience, message and channel, geographically ranked for the trial.

Other mapping analyses that can be utilized include the following:

• Multidisciplinary treatment team composition – Detailing an oncology patients’ actual team of healthcare providers (HCPs) to optimize awareness campaigns
• Drivers of and barriers to referral behavior – Deploying targeted research to identify drivers important to and potential barriers for the investigators and referring healthcare providers
• Referrer channels and content – ping the right channels and content based on historical prevalence and preference

However, simply knowing where to target patients based on where they get treatment is not sufficient to forge an adequate site selection and recruitment strategy. It is often necessary to make a basic decision: should sites that are already seeing patients be activated? Or, is it preferable— even necessary—to activate other sites that may not yet be seeing patients, but that may be more capable of handling the complexity and rigors of the given protocol and that have all of the auxiliary services and resources? In other words, is it preferable to “go where the patients are” because of their progression through the treatment journey, or to “drive patients to where you want them to be,” in effect disrupting the patient journey?

The decision to work with sites where patients are not naturally presenting requires a different recruitment strategy—one that makes strong use of referral networks and supplies referring sites with the information they need both for themselves and for their patients. In this way, the site identification strategy and the patient recruitment and engagement strategy are intertwined and informed by the patient journey.

Finding patients at the right point in the treatment pathway is not the only challenge; as patients experience cancer, they traverse both a treatment pathway and an emotional pathway. The treatment journey describes the course of the disease from diagnosis, through treatment to remission, cure or mortality. The emotional pathway, on the other hand, follows patients’ emotional states, attitudes and opinions through each stage of their treatment journey. It reflects how they feel, where they turn for support, and where they seek information. Figure 5 illustrates the typical emotional journey for oncology patients at key points along the treatment pathway.

Although the practice of using hard data to make decisions is continuous throughout the recruitment process, developing an understanding of the emotional pathway and the patient perspective layers in a more personal sensibility. This is especially critical to patient recruitment in oncology because of the range of experiences that people have with the disease and the uniqueness of each patient’s ability to cope both physically and mentally—nuances that should be appreciated when presenting clinical trials as a viable option for patients.

Insights into the emotional pathway for a given type of cancer are developed by a variety of means: social listening, interviews with key opinion leaders (KOLs), talking to primary care physicians and specialists, and discussions with patient advocacy groups. Social listening, in particular, can be very valuable in that it provides an opportunity to be privy to unfiltered, authentic conversations of both patients and caregivers. One can identify, for example, where patients go to discuss their condition and seek counsel from others, how they describe their emotional state, what challenges they’re facing, what they’re concerned about, how they’re reacting to treatment regimens, and so forth.

Such careful listening can lead to the discovery of robust audience insights that can drive the recruitment strategy. For example, if a protocol is seeking breast cancer patients who have failed first-line therapy and whose tumor has metastasized, recruitment efforts should recognize that those patients have progressed to another phase in their emotional journey. Those women who are younger may be concerned about spending quality time with their children and extended family, rather than continuing further treatment. Having ridden the waves of hope in the belief that they could be cured, they are now faced with the prospect that a clinical trial might be one of their few remaining options. Sites will be having conversations with these patients knowing that their emotional strength will be very low and that their extended family may have a strong voice in any decisions regarding participation in a clinical trial. In this instance, targeting media ads toward family members, as well as having information about study participation for families to read and understand, will help them in making that decision.