‘No Fate but what we Make’: AI Offers Great Potential for Clinical Trial Recruiting and Engagement

  Jonathan Bloom
Copywriter

Machine learning, deep learning and algorithms holds potential to reduce time and costs of studies

Artificial Intelligence (AI) is proof of Isaac Asimov’s prediction that, “Today’s science fiction is tomorrow’s science fact” … only when it comes to AI, tomorrow is today.

We already leverage the benefits of AI daily, but most people don’t recognize it. AI is in use in many industries where we would expect to find it — robotics, science, the military, surveillance, finance and even entertainment — but it is also present in the background of common, everyday services such as predictive texting on mobile phones and in apps like Uber and Google Maps.

Just as AI supports those apps and industries today, so too can AI offer significant enhancements to patient recruitment and engagement that can reduce the time and cost of clinical studies.

Saying “Hasta la Vista” to AI of “The Terminator”

Before looking at what the technology can do for clinical studies, one first must understand what AI isn’t. 
 
In spite of its already wide-spread usage, AI remains commonly misunderstood. A poll conducted by The Mirror in 2020 showed only about two in five (41%) people in the United Kingdom understand AI. The report even showed that 19 percent of people polled still believe “The Terminator” movie series is “a prime illustration of the technology”…even though the first film debuted almost 40 years ago.
 
AI pioneer Yoshua Bengio scoffs at how “The Terminator” and other movies depict the technology. Of the AI in movies, Bengio told the BBC in October 2019, “They paint a picture which is really not coherent with the current understanding of how AI systems are built today and in the foreseeable future. We are very far from super-intelligent AI systems, and there may even be fundamental obstacles to get much beyond human intelligence.”
 
The AI depicted in movies like “The Terminator,” and even in more recent ones like “The Avengers: Age of Ultron,” is vastly different from reality. Unlike what is depicted in the movies, AI is not automation that it is not poised to eliminate humankind. Rather than replace the human element, AI is a tool for human use rather than a substitute for human intelligence.

AI is Not Self-Aware

John McCarthy — considered the father of AI, with contributions to the technology dating back to the 1970’s — provided a similarly less fantastical explanation of AI. Rather than striving to transcend human intelligence, the late Stanford University computer scientist explained that AI uses computers to understand human intelligence. It combines computer science and robust datasets to understand and produce human-like problem-solving and decision-making. AI also incorporates the fields of machine learning and deep learning, which employ algorithms that analyze data or directions input by humans to formulate predictions, classifications or whatever type of outcome is sought.
 
The true value of AI, however, is that it is not limited to what is physiologically possible. AI locates, incorporates, compares and considers far larger blocks of data faster than the human mind. This empowers AI to yield better, more highly informed conclusions more efficiently than the human mind, which offers exciting potential for clinical studies.
 

One Possible Future

In a February 2023 article, Forbes contributor Joe McKendrick said, “Healthcare may lag other industries in [AI] adoption, but may be the ultimate proving ground where AI can really demonstrate its worth across many dimensions.”
 
One of the dimensions McKendrick points to as having the greatest potential to benefit from AI is research and development of new drugs and therapies. He reported that currently the clinical studies phase of development for the average drug takes nearly a decade and costs in excess of $1.3 billion…figures that could be reduced through the efficiencies of AI.
 
McKendrick cautions that because of their impacts on human life, certain portions of development — particularly around decision-making — will require 100 percent accuracy from AI before they are widely adopted. Nevertheless, there are aspects of the clinical studies process — particularly around patient recruitment and engagement — that already may benefit from AI such as:
 
Developing target populations: Using algorithms, data and direction input by the study sponsor, AI efficiently replaces the otherwise time-consuming and labor-intensive manual tasks of reviewing patient data, interpreting patient behavior patterns and incorporating data from experts to develop an accurate, targeted, potential patient population.
 
Improving diversity, equity and inclusion (DEI): Not only can AI identify target patient populations for a clinical study, but also it can pinpoint those within the target population from typically underrepresented segments. AI’s ability to analyze available socio-economic, race, lifestyle, genetic data and other factors within a target patient population quickly and intuitively should provide organizations with the information they need to increase the DEI of study participants, which could improve the likelihood they will better represent those whom they hope to help with their investigational drugs, devices or therapies.
 
Prescreening efficiency: AI can add efficiency to the prescreening of clinical study candidates by reducing the time and resources spent trying to do them manually, including:

• Establishing comprehensive inclusion and exclusion criteria
• Collecting preliminary data from a targeted population
• Reviewing screening questions and electronically (e.g., via email) interview candidates for the clinical study
• Identifying and recommending the best possible participants for a study based upon the established inclusion and exclusion criteria

Managing decentralized clinical trials (DCTs): Whether fully remote (no visits to the study site) or hybrid (limited visits to the study site), DCTs help increase participation among patients for whom travel to a study site would create a burden; however, management of those patients can be a time consuming and costly task. Organizations could minimize the labor and cost of managing DCTs and even improve the utilization of the information by enabling AI systems to monitor, collect and analyze data from remote DCT patients (e.g., vital signs, adverse reactions, medication adherence, et al), and transmit that data and analysis in real time to researchers.
 
Improving patient safety: From identifying a target population through the end of a study, AI systems can help make studies safer. AI’s ability to analyze candidate information can identify potential challenges or risks and forecast which patients may be at higher risks of experiencing adverse reactions to the investigational study drug. For those already participating in a study, AI’s real-time monitoring, collection, analyzing and transmission of data can identify early health warning signals, which could allow for quicker medical intervention if necessary.
 
Engaging with patients: AI systems can field questions from clinical study participants and — using algorithms that identify keywords within a patient inquiry — develop a human-like response that either replies to the inquiry, provides information to the patient about where to find the answer, or forwards the inquiry to a person who can assist the patient. This enhanced patient engagement expands a sponsor’s efforts to customize and scale solutions to meet patients’ unique and diverse needs associated with their study involvement such as:

• Travel & reimbursement resources

• Study alerts

• Appointment and medication reminders

• Visit-by-visit guides

• Educational resources and videos

• Gamification

   

The Future Has Not Been Written

With more data available than ever before and the need for more powerful tools to analyze it, AI use will only grow throughout the drug development process. Its potential to deliver almost limitless improvements in clinical study processes and outcomes will also grow as organizations learn how to take advantage of its ability to centralize knowledge, unlock sharing between subject matter experts and analyze all available data. 
 
Increased use of AI’s ability to enhance the clinical study process should enable developers of new drugs and therapies to increase medical advances while significantly reducing time and cost…goals that should ultimately improve human lives rather than “terminating” them.