Technology in Clinical Trials #3 – The Economic Potential of Generative AI in the Clinical Research Industry

It’s this time of the month again! Here comes my third newsletter and the very first one to combine all the main focus areas – artificial intelligence (AI) as well as digital solutions, innovative approaches to patient recruitment & engagement, and telemedicine.

In this edition of the newsletter, we discuss the use of generative AI (GenAI) in the clinical industry as well as how AI and digital solutions are being increasingly implemented in clinical trials.

?? The Untapped Potential of AI in Drug Discovery

[Comprehensive Report] In The economic potential of generative AI: The next productivity frontier, Michael Chui, Roger Roberts, Lareina Yee, et al from McKinsey explore the impact of GenAI on overall productivity and its potential to change the anatomy of work by automating repetitive tasks. They also discuss use cases of GenAI in the drug discovery process and briefly touch upon the use of GenAI in regulatory writing.

The Economic Prospects of GenAI

The report highlights that existing GenAI tools and other technologies have the capability to automate approximately 60 to 70 percent of the time employees spend on work activities. One of the driving factors behind this acceleration is the improved capabilities of generative AI, particularly in comprehending natural language. The report suggests that the impact of GenAI will be more prominent in "knowledge work," which pertains to occupations of higher wages and with greater educational requirements. Consequently, GenAI is expected to have a greater influence on these types of jobs compared to other work categories.

GenAI has the potential to facilitate growth in labor productivity, with estimated annual increments ranging from 0.1 to 0.6 percent until 2040. The actual increase in productivity will depend on the rate at which technology is adopted and how workers allocate their time to other activities. However, the adoption of GenAI will necessitate employees acquiring new skills, and some may need to change occupations to adapt to the evolving landscape.

In summary, the overall economic benefits of GenAI, considering the usage scenarios explored and potential productivity enhancements across knowledge workers' activities, are estimated to range between $6.1 trillion and $7.9 trillion annually (see below image).

The Impact of GenAI on the Pharmaceutical Industry

The use of GenAI in the identification and prioritization of indications (specific diseases or conditions for which a study intervention is being tested) reduces the time and effort spent on exploring potential indications, enabling researchers to efficiently select the most promising options for further investigation in clinical trials. This can significantly accelerate the drug development timeline, potentially bringing new treatments to patients faster.

Pharmaceutical companies that have adopted GenAI for indication prioritization have reported high success rates in clinical trials for the top recommended indications. These positive outcomes allow drugs to progress smoothly into Phase 3 trials, reducing the risk of failure and increasing the chances of successful market entry.

Furthermore, GenAI can contribute to more personalized medicine by accurately matching appropriate patient groups with specific indications. By considering clinical events, medical histories, and similarities with evidence-backed indications, researchers can better identify target populations for clinical trials and optimize treatment plans for improved patient outcomes.

Overall, the integration of generative AI into drug discovery and development brings benefits such as increased efficiency, accelerated timelines, improved success rates in clinical trials, and the potential for more personalized and effective treatments.

In terms of regulatory writing, GenAI tools offer the potential to significantly streamline content generation by leveraging existing documents and data sets. Furthermore, GenAI can automate the production of model documentation, identify any missing documentation, and scan regulatory updates to generate alerts for relevant changes.

Now that we have discussed the potential economic impact of AI implementation and its use in the pharmaceutical field, it is worth looking at specific use cases of how AI solutions are being tested in clinical trials.

[Magazine Article] In Three Key Imperatives Needed To Advance AI's Use In Clinical Trials, Ariel Katz dives into the topic of how AI holds huge potential for transforming clinical trials. Artificial intelligence, with its “predictive algorithms, data analysis and natural language processing (NLP) capabilities” is poised to assist in the process of clinical site identification, as well as speed up the recruitment of diverse groups of trial participants. However, before fully implementing the potential of AI in clinical trials, there are three hurdles to overcome: 1) improvement of the underlying data sources to optimize AI functionality; 2) training AI tools with pharma-specific terms; and 3) reducing hallucinations (inaccurate or misleading outcomes produced by AI models as a result of insufficient or biases in training data, or incorrect assumptions made by the AI model).

[Research Article] Liu et al published A prediction model with measured sentiment scores for the risk of in-hospital mortality in acute pancreatitis: a retrospective cohort study which describes how the group developed a predictive model for in-hospital mortality in patients with acute pancreatitis (AP) using sentiment analysis scores from nursing notes. Sentiment analysis (aka, opinion mining) is an NLP technique used to determine whether data is positive, negative, or neutral.

The study analyzed data from 631?AP patients (including the 88 patients which passed away while the study was ongoing) and found that sentiment polarity (positive or negative sentiment in nursing notes) mean was associated with a reduced risk of in?hospital mortality. To establish the model prediction and use, the score from this model was more precise as compared to the conventional Sequential Organ Failure Assessment (SOFA) and Simplified Acute Physiology Score (SAPS) II scores.

Besides changing the way clinical trials are conducted, AI solutions are being increasingly implemented in the process of patient recruitment.

???? Using AI to Improve Patient Recruitment and Engagement

[Blog Post] As described by Daniel Atieh & Olena Domanska in How AI Advances Patient Recruitment in Clinical Trials, patient recruitment is a crucial aspect of clinical trials, but researchers often encounter challenges in this process. Finding suitable participants who meet specific criteria can be time-consuming, costly, and slow down the progress of a clinical trial. However, Atieh and Domanska emphasize the transformation of patient recruitment in clinical trials through advancements in AI.

Artificial intelligence brings several benefits to patient recruitment, primarily by streamlining the identification and selection of potential participants. Through the utilization of algorithms and data analytics, AI can rapidly identify individuals who meet the eligibility criteria for a clinical trial by scanning extensive patient databases or electronic health records. This approach saves time and effort when compared to traditional manual screening methods. As noted, there was a steady decline in the number of enrollment-related terminations of clinical trials from 2010 to 2021 (see below image).

Furthermore, AI can enhance patient engagement and enrollment. Intelligent algorithms enable personalized outreach and targeted advertising, allowing for a broader reach and an increased likelihood of finding suitable candidates. Predictive analytics can identify individuals who are more inclined to participate in a clinical trial, allowing researchers to focus their efforts more efficiently.

Additionally, AI can support the informed consent process by providing patients with relevant study information, addressing concerns, and answering common questions. This improves participant understanding and enhances their willingness to join a trial, ensuring that patients are well-informed and engaged throughout the recruitment process.

Overall, the integration of AI in patient recruitment offers researchers a faster and more efficient approach to identifying eligible participants, increasing the likelihood of achieving enrollment targets and expediting the trial timeline. Atieh and Domanska conclude that by harnessing advanced technologies, clinical trials can benefit from improved patient recruitment strategies, resulting in accelerated medical advancements and enhanced patient outcomes.

?? Continued Implementation of Telemedicine in Clinical Trials

[Press Release] As reported by Newswise in Study explores the future of at-home cancer treatment, Keck Medicine of the University of Southern California is conducting a clinical trial to assess the feasibility of administering immunotherapy for non-small cell lung cancer (NSCLC) at patients' homes. Currently, immunotherapy is only administered within a medical setting via intravenous infusion. The trial will investigate whether a subcutaneous formulation of the immunotherapy drug atezolizumab can be safely and effectively administered by a nurse at the patients' homes. The study will also involve telemedicine appointments and remote monitoring through wearable trackers.

If successful, this trial could lay the foundation for future at-home cancer care. Moving immunotherapy treatment to patients' homes may provide benefits such as saving time and energy, expanding access to treatment, and reducing discomfort associated with traditional intravenous delivery. The trial aims to enroll 37 eligible NSCLC patients who will receive treatments administered by a visiting nurse every three weeks for one to two years. Remote monitoring and telehealth visits will be used to track patients' vital signs and overall health. The study will assess the feasibility, compliance, and satisfaction of patients with this (remote) approach.

The lead investigator hopes that if at-home immunotherapy proves safe and feasible for NSCLC, it could pave the way for future home treatments for other types of cancers, thus increasing access to care and addressing healthcare disparities.

?? Improving Clinical Trial Efficiency, Patient Centricity, and Data Quality with Digital Solutions

[Corporate Publication] In Streamlining clinical trials for patients and clinicians through digital solutions, Christina Duran, Magnus J?rnten-Karlsson & Liz Nevin discuss the use of digital solutions to streamline clinical trials. Their publication highlights the potential of digital technologies to improve efficiency, enhance patient experience, and accelerate the development of new treatments.

Traditional clinical trial processes can be slow and costly, often involving paper-based documentation and in-person visits. However, the adoption of digital solutions can revolutionize various aspects of clinical trials. One key aspect is the use of electronic informed consent (eConsent), which allows clinical trial participants to review and sign consent forms digitally. This eliminates the need for physical paperwork and enables remote participation, improving convenience for patients and reducing administrative burden.

Digital solutions also enable remote monitoring of patients' health data through wearable devices, sensors, and mobile apps. This real-time data collection enhances the accuracy of patient monitoring and increases the likelihood of capturing core study data.

Moreover, Duran, J?rnten-Karlsson & Nevin discuss the potential of telemedicine and virtual clinical trials, where study visits and consultations can be conducted remotely. This approach reduces geographical barriers and allows for increased patient participation, particularly for individuals who may face challenges with physical travel.

Digital solutions also assist with data management and analysis by utilizing technologies like machine learning and other types of AI. These tools can automate data processing, identify patterns, and provide real-time insights, enabling faster decision-making for researchers.

Overall, the authors stress that embracing digital solutions in clinical trials can lead to improved efficiency, patient-centricity, and data quality. By leveraging technology, clinical researchers can streamline processes, engage patients more effectively, and accelerate the development of innovative treatments for the benefit of patients and healthcare systems.

? Disclosure statement ?

Lastly, I would like to point out that I am NOT associated with and am NOT sponsored by any of the groups or organizations I refer to in this edition of the newsletter.