Business problem
Clinical trials take longer time due to poorly selected trial sites
Opportunities
Machine learning, business process design, data governance and management
Published by
ibm.com
Successful clinical studies hinge on efficiently recruiting and retaining diverse participants. Yet, clinical trial professionals across the globe grapple with notable challenges in these areas. In this chapter of the IBM series on clinical trial innovation, we spotlight key strategies for boosting recruitment speed, helping to ensure diversity, and harnessing digital advancements. Seamlessly integrating these elements is essential for leading-edge success in clinical development. Recruitment difficulties are the leading reason for trial terminations. While the overall clinical trial termination rate has decreased over time, low accrual rates within trials remain the most common termination reason. The public is often unaware that they have the option to participate in clinical trials. This knowledge gap is even more pronounced among minority populations. Of people who enroll in a clinical trial, the majority say they motivate themselves to stay engaged, as seen in Exhibit 1. Industry analysts report that dropout rates in phase 3 clinical trials can sometimes reach 20% to 30%. This underscores the need to redefine the roles of trial administrators and investigators in the process.
However, high turnover rates among clinical trial investigators also contribute to inefficiency, instability and increased costs. Our analysis of the voluntarily reported Form FDA 1572 BMIS database reveals a potential lack of sustainability in the investigator pool, both in the United States (US) and globally (Exhibit 2). The number of first-time clinical investigators has declined, especially among non-US based investigators.
Lastly, addressing the lack of demographic diversity in clinical trials is crucial. In 2022, less than 10% of trial participants for FDA approval were Black, fewer than 12% were Asian, under 13% were Hispanic, and women constituted less than 50% (Exhibit 3), not reflective of the current US population. Recognizing this gap, regulators emphasize the importance of greater diversity.
For instance, the FDA released guidance in November 2020 titled, “Enhancing the diversity of clinical trial populations.” In April 2022, they issued another draft guideline, “Diversity plans to improve enrollment of participants from underrepresented racial and ethnic populations in clinical trials: Guidance for industry,” aiming to provide recommendations for sponsors to increase enrollment of underrepresented populations
5 barriers to efficient patient recruitment and retention
There are several key factors contributing to the challenges of inadequate patient volume and sluggish recruitment speed in clinical trials:
- Complex trial protocols: Delays often stem from intricate or unrealistic trial protocols. It’s crucial to evaluate the feasibility of trials and refine protocols using evidence-based strategies.
- Barriers to patient accessibility: Numerous challenges like geographical constraints, transportation issues, scarce trial site availability and physical disabilities restrict potential participants from accessing trials.
- Patient pool expansion hurdles: Despite efforts to broaden participant inclusion, clinical trials still face hurdles in identifying and engaging new patient demographics, especially underrepresented groups.
- Ineffective outreach: Clinical trial marketing efforts sometimes miss the mark. Lack of awareness and trust among potential participants underscores the need for enhanced communication and trust-building strategies.
- Site underperformance: Many clinical trials face interruptions due to suboptimal performance at trial sites. Predicting site performance, spotting underperforming sites and formulating timely interventions are essential.
5 moves to boost recruitment speed, patient volume and diversity
1. Optimize protocols using historical and synthetic data.
Complex and stringent protocols are notorious for delaying clinical trials and eroding patient engagement. Ensuring early assumptions resonate with real-world execution is paramount. Enter the age of data-driven protocol assessment: using benchmarking tools and predictive modeling to gauge protocol intricacies and forecast eligible patient numbers, which then inform protocol adjustments.
Diving deep into historical trial data with a protocol complexity rating also reveals golden insights, especially around patient-centric elements. Key facets to spotlight in a protocol’s design include the investigational product’s nature, study design, endpoint definition, eligibility criteria, administrative burden, the presence of redundant processes, and the time that a patient would need to invest to participate. Grasping these dimensions sharpens the recruitment lens. Refining trial protocols isn’t a once-off; it’s an evolving, multidisciplinary quest, enriched by lessons from the past to shape future (more effective) trial designs.
Learning from historical protocol data and using synthetically generated scenario events to optimize inclusion and exclusion criteria can be powerful for achieving efficient trial design. By fine-tuning these criteria, protocols can help attract a targeted and more relevant patient group, speeding up recruitment.
When patients align with the inclusion criteria more accurately, their willingness to enroll increases. The FDA’s 2020 guidance emphasized expanding eligibility criteria and reducing unnecessary exclusions. Broader eligibility criteria not only streamline recruitment but also promote greater diversity, helping to ensure a more comprehensive and inclusive clinical trial.
The latest developments in large language models (LLMs) have the potential to significantly expedite protocol design processes. The current, labor-intensive manual approach can compromise the timeliness, accuracy and validity of results. LLMs demonstrate a superior understanding of the semantic relationships between entities within inclusion and exclusion criteria. They also possess query generation capabilities that can automate the process of identifying matching patients with trials, expediting the trial start-up process.
Additionally, generative adversarial networks (GANs) can be used to simulate real recruitment scenarios, further optimizing protocol design. These technological advancements promise substantial improvements in protocol design, ultimately boosting patient enrollment.
2. Embrace decentralized approaches for expanded reach and efficiency.
Decentralized clinical trials (DCTs) are gaining traction for their prowess in dismantling traditional hurdles in patient participation in clinical research. By removing geographical limitations, increasing accessibility and broadening the participant base, DCTs not only improve recruitment and retention but also foster greater diversity, welcoming participants from underserved communities.
The FDA, in its May 2023 draft guidance, backed the adoption of DCTs across drugs, biologics and medical devices, highlighting their merits such as enhanced patient convenience, diminished caregiver burden, broader access to varied demographics, amplified trial productivity, and support for research on rare or mobility-restricted patient groups.
Integral to DCTs are digital health technologies and software. The rise in the deployment of electronic patient-reported outcomes (ePROs), electronic clinical outcome assessments (eCOAs), and electronic informed consent (eConsent) from 2020 to 2021, primarily driven by contract research organizations underscores this shift.
Incorporating telehealth, real-time monitoring via devices such as activity trackers, blood pressure monitors, and other digital tools is now commonplace across many therapeutic areas. Augmented reality (AR) and virtual reality (VR) devices are increasingly playing a role and can be integrated into DCTs. The swift progression of these technologies is revolutionizing clinical trial paradigms.
Digital health technologies and software do more than just enhance accessibility and efficiency in clinical trials. They also pave the way into the realm of digital behavior data. This vast data set can provide insights into patient behaviors. In some instances, one wearable device can collect 120 million data points per day for each patient. Access to such a massive volume of daily behavior data provides a comprehensive understanding of each patient, promoting personalized engagement.
This pivot towards patient-centric care bolsters clinical trial patient recruitment and retention. Moreover, by transitioning away from the traditional site-centric model, clinical trials can tap into nationwide data, pinpointing underrepresented populations and thus encouraging greater diversity within clinical trial cohorts.
3. Partner with primary care: A goldmine for patient recruitment.
Forging alliances with community-based primary care physicians can dramatically enhance clinical trial participation. Given their longstanding patient relationships and in-depth understanding of patient history, primary care providers offer a doorway to a vast, diverse reservoir of potential trial participants. The bond of trust between patients and their primary care team cannot be understated.
A nod from a trusted doctor can greatly sway a patient’s decision to participate in a trial, significantly boosting enrollment figures. Engaging the primary care team not only enhances recruitment but also elevates the overall quality of trials.
Primary care doctors have access to vast amounts of patient health and medical data, including both structured and unstructured information, as well as medical images and videos. Machine learning and deep neural network models can effectively analyze this data to identify patterns, correlations and relationships, which is particularly useful for understanding a patient’s unique profile.
Computer vision models, such as convolutional neural network models, can assist doctors in detecting and classifying diseases in 2D and 3D medical images. Recently developed computer vision foundation models have significantly improved the accuracy of image classification tasks.
The amalgamation of artificial intelligence (AI) with primary care offers significant advantages in the realm of clinical trials. By deriving insights from diverse patient data formats, primary care doctors can achieve a more profound understanding of patient profiles. Such medical insights can be instrumental in refining trial protocols to align more closely with genuine patient experiences and help ensure continual oversight regarding patient safety. When patients engage in trials under the continual care of their physician, their likelihood of sustained involvement increases, consequently reducing attrition rates.
4. Refine marketing tactics to elevate awareness and foster trust.
Based on data from the 2020 Health Information National Trends Survey, 41.3% of the 3772 surveyed US adults reported not knowing about clinical trials. Elevating this awareness demands a targeted marketing thrust, using tools like social media promotion, engaging with key opinion leaders, and impactful campaigns to bridge the gap with prospective patients.
Studies over the past 10 years underscore the profound role of trust in determining clinical research participation, especially among underrepresented groups. A pivotal insight reveals that trust, or the lack thereof, is a primary determinant of participation. Prevailing trust-related apprehensions encompass fears of mistreatment, exploitation and unintended consequences.
These 3 tactics have proven to be effective:
- AI-powered social media advertising: Enhance the effectiveness of social media outreach for clinical trial promotions by employing AI algorithms on platforms such as Facebook, Instagram and Twitter. These algorithms can help curate highly personalized advertisements and content tailored to the desired audience. Through in-depth AI analysis of user behaviors and patterns, promotional messages can be fine-tuned to resonate with specific age groups, geographic locations and health interests, amplifying the relevancy and impact of the outreach. By harnessing these AI capabilities, clinical trial promotions on social media can precisely target the right audience, delivering the appropriate message at the optimal moment. This strategic approach not only elevates awareness but also fosters a sense of community within the target audience, heightening engagement and the likelihood of participation in the clinical trial.
- Engage with healthcare influencers and advocacy groups: Forge partnerships with trusted healthcare influencers and patient advocacy entities. Their expansive reach and credibility in patient circles make them invaluable allies. By collaborating, their endorsement can effectively expand the message reach and engagement levels.
- Targeted campaigns at recruitment locations: Execute campaigns that are precisely calibrated for individual recruitment sites and their associated communities. Such specificity helps ensure that the outreach resonates with the unique attributes of each site or community, capturing the attention of potential participants.
A sharp, tailored marketing approach elevates clinical trial visibility. Moreover, it’s crucial to address and build the trust factor, as it plays an essential role in influencing participation decisions. The strategies listed are instrumental in widening awareness and fostering trust among potential participants.
5. Streamline site performance and enrollment with AI.
Integrating AI-enabled capabilities in biopharma operations transforms clinical trial site selection, promotes scalable AI expertise and helps ensure cost-efficiency. AI algorithms consistently outperform traditional methods by analyzing intricate recruitment data, helping to ensure precise forecasting for study, indication and country-specific enrollments. By accurately predicting enrollment rates, AI has the potential to minimize financial risks, refine enrollment strategies and support budgeting to preclude potential setbacks and delays.
Moreover, gaining instantaneous insights into site performance keeps stakeholders informed about enrollment dynamics, quickly identifies potential bottlenecks and paves the way for agile decision-making and necessary adjustments. The AI automation enables real-time site performance tracking, sends prompt alerts and helps ensure streamlined reporting.
Additionally, the next best action mechanisms have the potential to provide real-time recommendations on the most impactful measures to enhance site performance. This agility helps to ensure uninterrupted trials, reduces disruptions and empowers stakeholders to adeptly navigate unforeseen challenges.
Embracing AI technologies strategically.
In the intricate landscape of clinical trials, the dual challenges of recruitment and retention persist, often becoming significant roadblocks to pharmaceutical progress. However, with the strategic embrace of AI technologies, we can collectively reshape this narrative. IBM is at the forefront of adopting AI for the pharmaceutical business, showcasing our commitment to refining this domain.
Through tailored protocol designs, decentralized trial models, enriched primary care collaborations, strategic marketing endeavors and the powerful precision of predictive engines, we can surge past these barriers.
The quest for faster, diverse and robust clinical trials is not just an aspiration, it’s an achievable reality. Clinical professionals globally have the tools and insights and now is the time to wield them with intent. For those ready to revolutionize the world of research and development, remember that innovation is not just about technology; it’s about harnessing every available resource to usher in a new era of clinical excellence.
Co-authored with Andrea Dobrindt, AI/ML Competency Leader for Public Markets at IBM Consulting. Original article is here.