Issue:June 2023

CLINICAL TRIALS – How Conversational Data & Listening at Scale Improve Clinical Recruitment


It’s no secret clinical trials have a recruitment problem. Re­search shows 85% of trials fail to retain enough patients. Recent news revealed COVID trials lacked diversity and female repre­sentation — fewer POC were recruited, and it’s common for women to be underrepresented in trials because of the potential risk to their fertility or impacts on future pregnancy.

Recruiting more diverse clinical trial participants shouldn’t be a secret — nor is it a problem to ignore. Without representation from all populations, researchers fail to gather the data needed to fully understand the effects of the drugs they’re developing.

One of the most effective ways to gather valuable, relevant, and useful information from a clinical trial is by listening to con­versations at scale. This approach isn’t just valuable for under­standing participants’ mindsets. It also helps researchers understand the psychosocial factors influencing and affecting a specific patient population.

Organized conversational data can answer questions about patient frustrations and fears and uncover transportation chal­lenges; caregiver, familial, and societal factors; and other social determinants of health (SDOH) impacting patients’ participation in a trial.

By listening to their participants at scale, life science compa­nies can improve their clinical trial recruitment success, seeking underrepresented populations with more intentionality.


Listening to recorded conversations between participating providers and patients has the potential to generate incredible value in understanding the mindset of patients diagnosed with specific illnesses, conditions, or diseases. During these emotional conversations, patients’ vulnerabilities, fears, and questions sur­face.

Currently, this area of research uses structured data sources like lab, clinical, and claims data. We have a good sense of im­proving comorbidities and how clinical diagnoses can lead to cer­tain health outcomes. Where we’ve not had enough innovation is in understanding the psychosocial factors of healthcare. Con­versational data offers a wealth of insight to better understand participants in their own words.

Using AI, organizations can listen at scale to surface patient emotions. Machine learning algorithms can be trained utilizing both text and sound wave analysis to detect tone and intent, which can indicate certain emotions like frustration and confusion. There are nuances in understanding the difference between these emo­tions and the implications on how your patient population re­sponds.

This active feedback is critical because listening gives in-the-moment feedback, whereas surveys and other data-driven sources are after-the-fact. They’re an aggregate of the whole — which is important — but you lose the individual voice. One of the most important things we can do is to deeply understand the lives of the patient population — and be very intentional about seeking out populations underrepresented in the data source.


The US government has recommended changes to drive compliance and reward increased representation in clinical trials as well. The Committee on Improving Representation of Women and Underrepresented Minorities in Clinical Trials and Research offered several recommendations to re­cruit with intention. They included the fol­lowing:

  • Forming a new Department of Health and Human Services task force to ex­amine equity in research and ensure proper data collection.
  • Implementing new FDA requirements for recruitment plans when submitting an application for investigating a new drug or new device exemption applica­tion.
  • Standardizing requirements for submit­ting demographic data to the Clinical­ database.
  • Updating the coverage guidelines con­sidering representation from the Cen­ters for Medicare and Medicaid Services.

Accelerating representation, recruit­ment, and retention of diverse patients re­lies on the sponsors, study sites, and investigators. Sponsors, particularly, can help clinical development teams under­stand the barriers to participation and re­tention by first defining and then identifying the targeted patient popula­tions.

By looking through the lens of emo­tions and diverse topics, you can tease out the SDOH from that data source — data that helps sponsors identify which patients to vet for clinical trial participation. Spon­sors also gain a complete picture of the demographic, geographical, and social factors causing hesitancy and confusion.


Treatment regimens researchers find effective in clinical trials can’t necessarily be applied confidently to all populations if certain groups lack adequate representa­tion during those trials. For example, the human clinical trials for the COVID vac­cines had an insufficient number of Black and Asian participants. Lower numbers led to vaccine hesitancy among these popula­tions. Other examples that lacked repre­sentation included studies for breast cancer, new asthma medication, blood thinners, and seizure drugs.

But part of the challenge in attracting historically marginalized groups to partic­ipate in clinical trials has come from lim­ited racial reporting, the underutilization of available medical resources, and hesi­tancy or mistrust in the medical system.

The US Congress recently commis­sioned a report, Improving Representation of Women and Underrepresented Minori­ties in Clinical Trials and Research, to study the severity of underrepresentation. The report discovered that while progress has been made with the inclusion of white women, racial and ethnic minority popu­lations continue to be left out of clinical re­search and trials, as have:

  • Members of the LGBTQIA+ community
  • Older adults
  • Pregnant and lactating individuals
  • People with disabilities


Many factors contribute to the chal­lenge of finding patients to participate in clinical trials. The number of drugs in the market is increasing, and saturation has contributed to the dwindling pool of po­tential subjects.

Many trials fail to reach their recruit­ment goals. Phase 1 and 2 trials need hundreds of patients, and some Phase 3 trials require thousands of patients. Yet up to 85% of clinical trials may not reach those targets within the specific time pe­riod. Other barriers include the following:

  • Financial limitations
  • Insufficient infrastructure support
  • A lack of physician awareness about current trials
  • A lack of appropriate trials for commu­nity-based settings
  • The patients’ negative beliefs or attitudes toward research

Successful patient criteria requires de­signing a trial reflecting patient needs while continuously keeping the patient’s perspective in mind. And the more trans­parent you are during the recruitment process, the better — even though that transparency is difficult to achieve.

Listening to conversational data, how­ever, can help researchers not only recruit participants but also evaluate whether the trial, once underway, addresses the symp­toms and challenges patients find mean­ingful. These conversations capture what’s top of mind for patients, including the clin­ical trial’s goal, why one trial is preferred over another, and whether the clinical trial is researching the cure for their condition — or new treatment options.

Communication is key during the re­cruitment process. Patient recruitment strategists must review protocols, verifying whether the endpoints are adjustable once the trial launches and that the trials meet patient needs. For example, benefits of participation might include receiving care from an expert in their condition — or hav­ing the opportunity to help researchers learn.


Psychological safety and cultural com­petence are essential for including diversity in clinical trials. Psychological safety en­compasses the belief that everyone is safe from judgment, humiliation, or punish­ment if they speak up, ask questions, or admit a mistake.

Cultural competency includes recog­nizing a practitioner’s biases, power im­balances, and barriers impeding effective clinical care. It involves recognizing and acknowledging disease incidence and prevalence, unique health beliefs, and treatment outcomes in diverse popula­tions.

It also introduces the idea of cultural humility. That is, a healthcare practitioner or researcher in a clinical trial recognizes the differences in cultural values and does­n’t assume one or another is the norm. Organizations and individuals embracing cultural competency intentionally minimize the impact of implicit biases and work to eliminate their own unconscious biases.

Conversational data uses ML algo­rithms designed to detect words indicating emotion and nuances of tone helping to inform training needs. With conversational data, you can listen for phrases like:

  • “I’m excited.”
  • “I’m hopeful.”
  • “I’m fearful.”
  • “I have questions, but I know I need to explore something that might be in a clinical trial rather than try something already available commercially.”

By analyzing conversational data in combination with participant demographic data, organizations can also identify con­cerns expressed by a specific community. Listening at scale provides additional con­text that enhances already gathered de­mographic data. Addressing those concerns sooner shows transparency, helps build trust, and can increase some­one’s willingness to participate in a trial.

Listening directly to the voice of the customer facilitates an even more granular analysis of understanding how tone, paired with words, may indicate specific emotions. For example, in the commercial space, the top two negative emotions con­versational data most commonly identified are frustration and confusion. These emo­tions have certain implications depending on whether your patient population is frus­trated or confused.

Instances of noted confusion include a possible messaging problem or the need for enhanced clarity on directing patients’ next steps. Frustration implications are a bit different and refer to a CX or journey problem in which a participant is unsure what to do despite trying to call and com­municate with the provider.

But applying a generic label of nega­tive sentiment isn’t good enough for figur­ing out what that negative sentiment means — or what to do about it. Focusing the research and data science work on diving more deeply beyond a sentiment score and pairing those emotions with the conversation’s key topics creates a more holistic picture that allows organizations to understand at scale.

A single data point of negative senti­ment doesn’t provide enough context for clinical trial sponsors to address potential participant concerns. But direct, voice-of-the-customer conversations allow organi­zations to pinpoint with precision what factors are driving the caller’s emotion and gain valuable insights and feedback.

You must listen to patient perceptions because learning about those experiences is the best way to deepen that understand­ing. Unstructured conversational data:

  • Identifies SDOH and other social pres­sures patients — and potential trial par­ticipants — face.
  • Pinpoints quality and access indicators, helping clinical trial designers to find the most effective ways to improve par­ticipant satisfaction.
  • Analyzes different interactions to identify challenges, mitigate risks, and refocus communications and programming.
  • Identifies training opportunities and areas of improvement — and positive recognition.

Life sciences organizations and drug companies need to continue their efforts to make care more accessible, starting with populating clinical trials in all phases with more diverse participants. You can’t get the health outcomes you want without seeking the truth of the human side of the equation.


The incredible value of listening to patient voices in a new, different, scalable way provides strategic value to those plan­ning and conducting clinical trials. Drug companies empowered to listen to their potential trial participants will be more in­formed, empathetic, and enlightened when approaching the task of populating trials. The result? Trial programs will be more prepared to recruit and serve sub­jects that more closely reflect the diversity of the patients the drugs are designed to treat.

Amy Brown is the Founder and CEO of Authenticx – the software platform analyzes and activates patients’ voices at scale to reveal transformational opportunities in healthcare. She built her career as a rising executive in the healthcare industry, during which time she advocated for underserved populations, led and mobilized teams to expand healthcare coverage to thousands of Indiana residents, and learned the nuance of corporate operations. In 2018, she decided to leverage her decades of industry experience to tackle healthcare through technology. She founded Authenticx with the mission to bring the authentic voice of the patient into the boardroom and increase positive healthcare outcomes.