Fixing the patient recruitment “leaky funnel”: blending data with a patient and site-centric approach to address the challenge

Most people in our industry are familiar with the “Leaky Funnel” analogy that describes the model where we approach a large number of patients for inclusion in a study, but they leak out of the pipe at every juncture. The loss of patients through planning, screening and execution of the study is not only costly, but it can also significantly delay or even prevent the successful execution of an otherwise valid study.

We recognize this specific challenge as an opportunity for us to address using our substantial, differentiated assets and capabilities. Specifically, we see our unparalleled data modeling/analytics and patient insights as being key to reducing the leaks in this pipe. More than just bringing these capabilities to bear in a siloed approach, we believe that the real solution comes from blending all these capabilities for an approach that is both site-centric and patient-centric – the whole being greater than the sum of the parts.

Let’s consider how combinations of data, analytics and deep patient insights can address specific leaks in the pipe.

Identification – Optimizing Protocol Design

We focus on the identification of the participant population by seeking to optimally match the protocol requirements to the real-world patient population. We use data-driven approaches to select sites of known performance to drive enrollment and reduce the misalignment with potential participants’ expectations to increase retention.

Key to study success is designing out the “flaws” in protocols that can lead to recruitment delays, or at least allow us to be aware of them up-front and account for them in our planning. We aim to design trials that “can be done.”

Protocol designs can mandate a very specific subset of a patient population: those who do not suffer from common comorbidities have had limited treatment exposure, and other parameters set within the inclusion/exclusion criteria. While the disease may be common, the population meeting the criteria may be minimal, increasing the probability of inaccurate planning and missed recruitment targets.

The trifecta of optimizing protocol design is to understand the addressable patient population, be able to accurately forecast recruitment rates, and understand the potential participants’ appetites to enroll and remain in the study.

Through our unique access to Labcorp’s lab results, we can obtain real-world diagnostic data including 30 billion test results across greater than 5,000 assays on approximately 50 percent of the U.S. population. Using these de-identified data sets as the basis for an advanced modeling system, we can not only show the general population pool for a very specific set of inclusion/exclusion criteria, but also provide suggestions to increase the potential pool if necessary. These suggestions can be modelled in real-time with a sponsor to enable the team to increase the potential pool without jeopardizing the study objectives. Our outcomes suggest this approach is more effective when compared to historic approaches where we would ask sites to estimate addressable populations.

The Central Lab result data provides access to aggregated site performance derived from data on more than 50 percent of all global studies. These insights include patient enrollment rates, patient densities and markers of site quality by indication. This allows us to assess by region whether a protocol’s requirements are realistic in relation to the study objectives. It also gives us a unique insight into historic performance.

Lastly, the teams have access to insights about the probability that a participant will enroll – and stay enrolled. Our global Patient Intelligence database, Xcellerate^® Patient Intelligence, contains information about a combination of existing and naïve participants and their tolerance to specific study criteria. Labcorp’s ability to assess its customer base for these insights adds very broad, real-world insights addressing common and often unseen barriers to participation.

The combination of access to these unparalleled data sources, award-winning analytics, deep understanding of the patient, and our experience in trial execution are key to increasing the probability of designing protocols that can be successfully executed.

Enrollment

Leaks in the funnel related to enrollment tend to be focused around understanding the optimal patients and sites. Prospective participants need to see that the benefits of a study outweigh any inconveniences, and the optimal sites need to be selected for the specific protocol.

Patients and Participants

Understanding patients’ perspectives on the reasons why they would or would not take part in a study is an important factor in designing a study that can be successfully carried out in terms of both enrollment and retention. Although we seek to design the protocol to support optimal recruitment, there are many occasions when the protocol has been finalized before we are involved in the study and without due consideration of patients’ perspectives or other factors. This may sound like a simple fix, yet it remains a barrier to patient enrollment today. In these cases, a deep understanding of the patients’ tolerances is even more important. For example, we have found that areas such as availability of medication post trial or the number of inconvenient procedures have a profound impact on patient willingness to participate and stay in trials. Our Xcellerate Patient Intelligence database includes insights from clinical trial participants from across the globe to help plan for this critical area, including even simple data such as how far they would be willing to travel for a site visit.

Along with fostering enrollment, we are working to reach a broader number of potential patients. We are innovating on several initiatives, including working through Labcorp to grow our “direct-to-consumer” recruitment model rather than just relying on sites to perform outreach to potential patients. In this model, where patients have opted-in, we are able to pre-screen patients by looking at their longitudinal diagnostic lab data and then inviting them to participate in a study at a nearby site. A growing number of Labcorp consumers have already shown an interest in being enrolled. We continue to invest in this exciting new outreach model that further expands the opportunity for patients to consider research as a treatment option and support sites with additional recruitment opportunities.

Sites

The Central Labs generate more clinical trial data than any other organization in the world. With a line of sight to over 50 percent of all global studies at any given point in time, one of the key attributes of these aggregated data is the ability to assess the historic performance of 175,000 unique investigators by indication. In our analysis work, we have seen that experience does not automatically equate with performance. When we look across our data set, we find high-performing sites at all levels of experience. The age-old presumption of going to more experienced sites does not guarantee results.

Specifically, we can analyze aggregated performance of these sites’ recruitment rates, screen failure rates, and patient densities within a region and within an indication. We find that historic performance is a key predictor of actual performance and, put simply, having the vast number of data points that we do gives us the ability for greater accuracy in selecting the best sites for a study. Given the challenges of selecting sites without data and the number of sites that continue to recruit no patients across the industry, we find this an invaluable tool in “fixing the leaky pipe.”

Retention

There are several factors related to decreasing the leaks around retention, and they focus generally on reducing barriers to participation and inconvenience for the patient and their families. Basically, we need to execute studies in a way that keeps the patient engaged and motivated throughout the study.

With access to unique data from both active and naïve subjects from over 20 countries, Xcellerate Patient Intelligence data helps identify undue burdens on the patient. For example, recent and common themes where our more patient-centric approach avoids recruitment delays include:

▶ Understanding how often and far a patient will travel by indication and what types of procedure they’ll tolerate.

▶ Working to make trial medication available post-trial – especially so where existing treatments are unaffordable.

▶ Helping the patient understand the benefits to the broader population of the study.

We also have a number of other unique capabilities that directly drive better retention rates. For example, we can forecast kit placement with 98.3 percent accuracy. Ensuring kits are available when a patient arrives for a scheduled visit reduces unnecessary frustration – for the patient and the site.

We are also working to leverage the vast network of Labcorp patient service centers (PSCs) for specific types of patient visits, including some PSCs located in Walgreens’ stores. Visits to these PSCs tends to be much more convenient, and potentially pleasant, for the patient may reduce the time and cost of the study.

We believe our unique data sets and innovative approaches, combined with our deep understanding of the clinical trial business, has been instrumental in our ability to get to Last Patient In (LPI) 4.2 months faster than the industry average.*

*Phase I-IV, Global Oncology Site Activation to Last Patient In.

About The Author
Ben Quartley, PhD, vice president, head of feasibility, patient recruitment and engagement, has spent more than 20 years in the pharmaceutical industry and has broad experience in project management and clinical operations, having worked at a site management organization, large pharma and CROs. Early in his career, he saw the challenges of accurately planning and delivering patient recruitment on time from both the site and sponsor perspectives and has had an interest in this key interface ever since, holding multiple positions in this area. Ben is highly engaged in leveraging the power of data to support optimal trial planning.