Sergio Sánchez-Gambetta

Jill Guary, Patient Centricity Data Manager, Experience Optimization Team, IQVIA

Shana Hull, Director, Patient Recruitment Strategy, IQVIA

Sergio Sánchez-Gambetta, Director, Design Analytics, IQVIA

Patients and their caregivers know their unique needs and challenges better than anyone else. Though clinical trial sponsors cannot always anticipate the right path forward, actively listening to patients to understand which (and how) specific protocol elements can hinder or enhance engagement can help shape trial design. A comprehensive “patient empathy mapping” strategy can implement several key methodologies to help sponsors, their contract research organization (CRO) partners, and study teams gauge which patient-friendly elements— whether traditional, hybrid, or decentralized—can or should be considered in trial design and may be beneficial to the current target population.

Uncovering areas of perceived burden and patient preferences creates a strong foundation that can be adjusted and updated during trial design to best meet patient needs and to maintain their interest as valued partners in the process. This can be achieved through a multipronged approach, including activities described ahead, that provides the sponsor with a holistic scope of intelligence to better prioritize which trial design elements patients truly need or want.

Direct surveys: Capturing the patient voice

Hearing directly from patients and caregivers for real-time insights into participation motivators and barriers is critical to successfully preparing for patient-first trial design. Survey groups may be found in proprietary, company-specific patient communities and registries, advocacy groups, online communications organizations that reach rare-disease populations, and similar databases.

Because patient populations vary in therapeutic need, age, geography, race/ethnicity, and socio-economic status, surveying provides a holistic view of what barriers impact which subpopulations more and how to prioritize solutions accordingly.

For example, to learn from patients with cardiovascular disease or related risk factors about their experiences with and opinions of trials, IQVIA surveyed more than 600 screened and qualified patients (primarily 60-years-old and older with cardiovascular disease or risk factors) to determine potential participation deterrents. Trial length, number of required study visits, and mode of medication administration (e.g., oral, injection, etc.) were not priority concerns. Instead, financial motivators (e.g., participation compensation, lack of treatment fee, etc.) were the top-cited factors for deciding to participate. For patient enrollment strategies, this may mean prioritizing how we address financial barriers, such as including the patient stipend and related details in the informed consent form and ensuring that site coordinators can walk participants through financial concerns as part of site education during consent.

When gauging how decentralized a trial may need to be to ensure participation, the survey found that more than half of respondents recognized that virtual trial solutions like telehealth may reduce participation burden, and 82% said they’d be comfortable using a mobile device to participate. However, only 15% of respondents preferred in-home visits by a medical professional over in-person visits for routine study assessments, which require travel to and from sites. So, while we recognize the convenience of these decentralized trial components, it may not be a priority for this specific patient group. Rather, nearly half (49%) of the respondents said they would like travel assistance to and from site visits, which again emphasizes the value placed on financial support.

There are more than a billion unique non-identified patient records currently available. By leveraging an extensive breadth of real-world data (RWD) (e.g., claims data, medical records, etc.), sponsors can gauge how potential patients may view trial procedures based on their current experiences and usual disease care. These insights can show how many patients are familiar with trial procedure and how frequently they may experience it. Comparing that to how many times these procedures may be required within a trial, which could be more frequently, should provide a better understanding of the patient’s level of willingness to participate.

While drafting trial protocols, understanding patients’ willingness to participate based on what they believe is expected of them can improve recruitment and retention. By leveraging patient-centric insights gained through surveys, virtual, or face-to-face focus group discussions, RWD, and more, it is possible to apply analytics to assess patient burden based on trial design aspects that impact willingness to participate in studies or may require more aggressive mitigation.

 Using those analytics, protocol elements can be scored to quantify patient burden. Using burden-scoring for each element, sponsors can examine whether protocols are at the expected level of burden when compared to trial protocols for a similar phase, therapeutic area, disease, etc.

This analysis can be further broken down to provide insights according to race and ethnicity to identify and reach better patient subpopulations. This type of scientific assessment should more reliably predict patient satisfaction and recruitment success.

Analytics and artificial intelligence/machine learning for design

Sponsors have access to unprecedented amounts of detail on patient journeys via multiple data sources, so it is important to understand data source collection approaches and data standardization and curation processes that allow enhanced analyses. Specifically, understanding potential limitations and opportunities with individual data sources is a key first step. For example, claims data do not allow for capturing all meaningful variables and relevant endpoints, such as death dates to calculate overall survival in oncology patients.

When we understand what data sources can and cannot provide, artificial intelligence (AI)/machine learning (ML) approaches can help aggregate data from multiple sources for fine-tuned analysis, resulting in patient-centered insights. AI/ML-driven analysis can help sponsors understand the real-world patient journey, identify subpopulation needs with different or better clinical response to the same therapy, and then adjust eligibility criteria accordingly. Advanced ML methodologies can capture a large group of patients’ voices and analyze this data to create and model potential scenarios to improve clinical responses and/or identify better diagnostic techniques.

One example is the use of the Shapley values technique in ML, which practically applies the Shapley game theory by understanding how each variable or feature contributes in a final outcome.

 This methodology can be used to identify a subgroup of patients with severe asthma who are at higher risk for acute exacerbations as well as to show how patients who may respond differently to the same therapy can potentially benefit differently, too.

 By recognizing these unique patient needs, sponsors can tailor clinical development plans and specific trial designs with the best elements for the target patient population.

Supporting patients through necessary trial elements

Because some trial protocol elements are necessary regardless of how patient-friendly they are or aren’t, sponsors, CROs, and study teams must think through the brunt of the burden for that necessary element and how to best support the patient, site or other stakeholder through the process. In these instances, additional patient education may help to show why this element is critical.

Animated study explainers to view on-site or printed brochures to take home and discuss with their families are key examples of tailored communications to help the patient through what may be an engagement roadblock.

In some cases, protocol adjustments will have to be made during the trial based on ongoing collection of patient insights. For example, a study examining a weight-loss treatment for patients with obesity may require volunteers to capture daily food intake in an electronic diary. However, patients may say they are discouraged by this process. Sponsors will then need to determine whether it’s necessary to secure the data from patients.

If yes, sponsors may adjust the design of the e-diary/decentralized clinical trial technology, or enhance motivating messaging within the data capturing process to better support patients in their journey. This is where gathering patient intelligence on trial procedures and elements prior to trial startup is beneficial.

A listening-first approach to patient outreach

When patient preference insights are available early in trial design, sponsors, CRO partners, and study teams have a tremendous opportunity to “get it right” from the start, including how, when, and where to approach potential patients for participation.

Patient preferences can be reflected in study branding materials, marketing campaigns, and media channels, all the way down to the very last detail on how outreach emails and other messages are written, including the best hour of the day to send them and what patient subpopulation(s) will react to which channels of communication.

Listening to patients prior to outreach provides a chance to gather feedback about the messaging that is used to communicate with patients (e.g., advertising). For a cardiovascular study, if patient insight activities have already informed the sponsor that the number of electrocardiograms will present a challenge for trial interest and retention, tailored patient outreach could ask for feedback or preferences regarding support services that may ease that known burden, such as extended office hours to avoid time off from work or participant stipends to defray childcare costs for longer visits.

Because patient-first trial design will never be a one-size-fits-all approach, it is vital to understand the deep variations in patient populations and their preferences to best cater to their needs, starting with outreach communications. Questions to answer to account for variations, include:

Based on potential participant location, how accessible are stable internet connections? Are they likely to connect via mobile devices, laptops, or other smart devices?

How much do they know about research in general, much less the trial at hand? What are health literacy levels, which indicates the level of educational messaging needed?

Are participants receiving information via culturally preferred messages and channels?

What type of media do patients consume? Do they prefer to read, listen, or both, and how much time do they spend?

Is there a community leader and/or patient association representing patients in this town/state/country who should be engaged first?

Guiding current and future patient-first trial design

Always keeping patients’ needs top of mind and constantly gathering and evaluating insights before, during, and after trials can help inform future pipeline programs and create building blocks to improve patient-centered trials.

Whether it is decentralized trial elements, financial support, or educational information, patients are speaking up about what they need to participate in research that can impact their health and others’. They want to be informed about study progress, results, and related implications. As such, sponsors, investigators (medical practitioners), CROs, and study teams are well-positioned to ask for insights and translate findings into an experience that genuinely works for the patient, ensuring they are engaged in the process, feel valued and motivated to do more, and will become advocates of clinical research.

, patient centricity data manager, experience optimization team, 

, director, patient recruitment strategy, and 

, director, design analytics; all with IQVIA

Cameron, D.; Willoughby, C.; Messer, D.; et al. Assessing Participation Burden in Clinical Trials: Introducing the Patient Friction Coefficient? 

https://pubmed.ncbi.nlm.nih.gov/32741647/

Game Theory. Lecture Notes by Y. Narahari, Department of Computer Science and Automation, Indian Institute of Science, Bangalore, India. October 2012, 

https://gtl.csa.iisc.ac.in/gametheory/ln/web-cp5-shapley.pdf

Walls, C. Identifying and Characterizing Asthma Subgroups at High Risk of Severe and Life-Threatening Exacerbations with Machine Learning and Longitudinal Real-World Data. Pharmaceutical Management Science Association. May 1, 2023, 

https://www.pmsa.org/past-presentations/2023/1793-identifying-and-characterizing-asthma-subgroups

Articles

Approaching protocol design and operational planning with patient empathy. 

Sergio Sánchez-Gambetta

Articles

Executing Trial Design With Patients—Not For Patients