A new frontier?
Experts discuss the role of technology in clinical trials and what the future may or may not hold for dermatology
Feature
By Ruth Carol, contributing writer, July 1, 2021
The physical exam is at the heart of clinical trials in dermatology...or is it?
In the future, clinical trials in dermatology will likely be a hybrid of in-person and remote visits, noted Howard Sofen, MD, associate clinical professor of dermatology at Olive View-UCLA Medical Center. Many of the interim visits in select types of trials can be done virtually.
Using either the telephone or computer, remote visits within dermatology trials are becoming more common, stated Andrew Blauvelt, MD, MBA, president of Oregon Medical Research Center. Moving to completely remote trials in dermatology, however, is unlikely because they often require physical exams as well as blood tests. For dermatology trials to go entirely remote, the FDA would have to change its assessment requirements, he added.
In general, clinical trials are expected to move significantly toward including more remote components in the next three to five years, said Orlaith Burke, PhD, Life Sciences innovation portfolio lead at Accenture. Just how fast this move occurs will depend on the need and preferences for participants to interact in person with the physician, she said. Digital devices and artificial intelligence (AI) are already being used to enhance patient recruitment and augment enrollment in clinical trials. Eventually, there is a future vision where trials could be done by algorithm, which could significantly reduce the time it takes to run a clinical trial — from years to hours — and some of these are developments being tested already, she added (see sidebar). For now, Dr. Sofen would be happy to move beyond faxed papers that need to be signed, dated, verified, and filed each time he draws a lab.
Hybrid model will fill a niche
While palpation of the skin and blood tests rule out moving to entirely remote trials, interim visits that don’t require taking those measures can be done remotely, Dr. Blauvelt noted. Participants in a study for chronic urticaria, for example, could use an app or website to report the number of hives and/or how itchy the hives are as endpoints in a largely remote trial, Dr. Sofen said.
In theory, blood draws could be done by an office closer to the participant’s home, or by a home health nurse, reducing the burden on those who must travel long distances to the study site. But in Dr. Sofen’s experience, that option didn’t work well. “It sounds good but when it came down to it, it was a disaster,” he said. Dr. Sofen participated in one trial that relied on visiting nurses to do the blood draws; the primary outcome was the level of a medication in the patients’ system. Different nurses went out each time, making it difficult to gauge the quality and consistency of the work being done, and some were unable to go to where some participants lived, creating scheduling issues. “As an investigator, leaving those details up to someone else leaves room for errors,” he said. While Dr. Sofen is confident that his staff perform blood draws among other tasks correctly, signing off for other providers who he doesn’t know was uncomfortable from an accountability perspective.
On the other hand, qualitative research, including quality of life surveys, patient in-takes, and symptom assessments, could all be done completely remote, Dr. Blauvelt said. Participants could record patient-reported outcomes measures electronically from their homes.
Endgame: Trials by algorithm
The use of technology could revamp the future of clinical development and transform the trial process from nine years to mere hours, according to Accenture. This would occur in three successive waves.
The first wave will take advantage of new technologies to collect patient data in real time, Burke said. Taking wearable technology to the next level, built-in sensors in clothing, phones, and household devices could collect data from patients, effortlessly and continuously. Placing these small, wireless, self-powered, passive sensors in the participant’s body could increase the quality of the data being captured without impacting the user “wearing” them. Many of these tools are passive so the patient wouldn’t be burdened to fill out and submit multiple forms, she said. Temporary digital tattoos, which are the next generation of skin patches, are currently being developed to measure ECGs, detect falls, and even release drugs. How the data are tracked and sent to ensure authenticity will need to be addressed, Burke added.
In the second wave, physicians and patients will have access to AI-enhanced digital agents that will be used to direct them to appropriate clinical trials based on patient data, she continued. The agent will then check whether the patient is eligible for any of the suggested trials and provide additional information. Once a patient provides their informed consent, the agent will finish the onboarding process. This stage will require the standardization of all data collection and management across multiple systems. Paperless management systems will be needed to support the accurate collection of trial information, inclusion/exclusion criteria, and the patient’s electronic health record data. As AI technologies mature, agents will be trained to do complex cognitive tasks, such as determining patient eligibility.
In the final wave, AI-enhanced technology will direct patients to appropriate trials, determine eligibility, offer informed consent, and onboard them, Burke said. Decentralized data repositories will securely manage the data obtained from either existing data of patients who are currently receiving treatment or patients who contribute their data based on their medical condition, characteristics, lifestyle, etc. Traditional clinical trials will be replaced with trials by algorithm. As an example, AI-enhanced technology could be used to identify a target for a new medication, run synthetic cohort analyses, capture patient variability within the cohort, model likely interactions of a drug in those patients, and predict the likely outcome of a trial, all based on the data, Burke said. Such simulations are currently in the early stages of use for clinical trial design. HumMod, an integrative physiological model, has been used to create a synthetic control arm by replicating the findings of clinical trials evaluating how different human organs, such as the liver or kidney, interact with different drugs, she explained. This brave new trial world could boost patient participation as well as reduce patient risk and trial costs.
Real-time data capture
The emphasis on patient-reported outcomes measures in clinical trials within the past 10 years coupled with the growth in AI-enhanced wearable technology could enable participants to track and/or capture data electronically, Dr. Burke pointed out. Today, people use wearable technology to track exercise, heart rates, and sleep patterns as part of their daily routine. The Apple watch has been approved by the FDA as a Class II medical device capable of recording an electrocardiogram (ECG), detecting arrhythmias, and detecting falls, lending itself for possible use in some clinical trials, she said.
Capturing data in this manner could be beneficial for trials in some specialties. For example, collecting ECG and heart rate could be helpful in cardiology studies for primary or secondary endpoints, Dr. Sofen suggested. Patients could be sent a blood pressure device that sends the data electronically and could receive medications in the mail for a hypertension trial, Dr. Blauvelt said.
In dermatology, however, it’s unlikely that collecting data this way would be helpful for primary outcomes, Dr. Sofen said. He was involved in a clinical trial collecting data to determine if wrist movement correlated with nighttime itching and insomnia. Technical difficulties made data collection problematic and the results did not correlate with the patient itching, which was the primary outcome, Dr. Sofen explained. These digital methods, however, may be beneficial to collect secondary outcomes or provide interesting adjunct data in dermatology trials, he added.
Digital devices vs. AI-enhanced technology
Virtual props
Although teledermatology has assumed a place in practice, especially during the COVID-19 pandemic, it has limited usefulness for playing a major role in clinical trials, Dr. Blauvelt said. Teledermatology enables the U.S. military to have a dermatologist in Maryland evaluate a rash of a soldier in Iraq, Dr. Sofen said. But that setup requires technical people on both ends to ensure connectivity and quality images. “In the real world, people don’t know what to focus on,” he said.
It’s nearly impossible to get the quality of video and/or photography needed to do a true evaluation, and then have it validated remotely, Dr. Sofen said. He participated in one virtual acne trial, which was successful, but it was difficult to obtain consistent, high-quality photography for assessing skin lesions. The technology would have to improve to pick up skin differences on camera, Dr. Burke agreed.
Recruitment and enrollment
Digital devices and AI-enhanced technology could help cast a wider net for recruitment, Dr. Burke said. An estimated 86% of clinical trials fail to meet recruitment goals within their proposed timeframe, so any improvement would help. “The more AI or automation can identify, access, and recruit diverse sample groups for trials, the better,” she said. The caveat is that everyone must have access to baseline connectivity. “Assuming that access to baseline connectivity grows, so will the diversity of the patient population that can be accessed,” she added.
The pool of participants would grow, but only for individuals who are internet savvy, which is a fraction of the people out there, Dr. Sofen said. “Some people are very adept at logging into the internet, but the vast majority of the population either doesn’t have access to a computer or the ability to navigate the internet. They are lucky to have cell phones,” he added. The lack of quality internet access and optimal camera equipment complicates reaching potential candidates in rural areas as well.
In Dr. Blauvelt’s experience, social media advertising campaigns don’t increase enrollment in significant numbers. “They’re not very effective because the majority of people who express interest end up not being good candidates. It’s just a wider pool of people to wade through and reject,” he said.
“Virtual recruitment may sound good, but I’ve never found anything that replaces my sitting down and talking to a patient,” Dr. Sofen added.
“The more AI or automation can identify, access, and recruit diverse sample groups for trials, the better.”
Augmenting enrollment with automation could make the various steps more efficient and patient centric, Dr. Burke said. AI-enhanced technologies could help identify eligible patients, obtain informed consent, onboard them, and accurately collect trial information either from the patient or the patient’s electronic health record data. Informed consent could be simplified and accelerated by using an electronic consent platform, sometimes referred to as eConsent. “Core to the feasibility of digital control trials will be tenets around ownership of data, patient consent, and data security,” Dr. Burke emphasized. “The patient needs to know who has visibility of their data, what they are consenting to, and who has access to their data. The data has to be secure.”
Crowdsourcing — an open call for voluntary assistance from a large group of individuals — could potentially be harnessed to increase the speed and efficiency of data collection for clinical trials. Crowdsourcing was helpful in gathering information about COVID-19, especially in the beginning of the pandemic when information was rapidly evolving. As an example, the COVID-19 Dermatology Registry was launched in March 2020. There are also registries focused on psoriasis, atopic dermatitis, hidradenitis suppurativa, and alopecia, among others. Dr. Sofen believes that crowdsourcing would be most helpful for long-term, large, population studies, but not for dermatology trials because they tend to be small. Most Phase II dermatology trials seek to recruit only 100 participants and Phase III trials seek 1,000 participants, he said. Crowdsourcing can be helpful with the caveat of the inherent bias in people who volunteer, Dr. Burke added.
In theory, remote trials would be less expensive. However, investigators may have to provide patients with technology, such as a smartphone, broadband, or even basic WiFi, Dr. Burke said. Trials could be cheaper if lab tests are performed in a reliable manner and medications are shipped to the participant’s home, Dr. Sofen said. But remote trials put more burden on the site staff to coordinate them. Investigators will need a data strategy to manage and clean the data to ensure they are collecting only the data they need.
Ultimately, patient preference will have a lot to do with the progress of remote trials. “There is a large hands-on personal component for patients participating in a clinical trial,” Dr. Sofen said. They need to have in-depth conversations with the physician and/or coordinator. The potential downside of remote trials is an increased risk of depersonalization as interactions increasingly occur via technology, Dr. Burke said. Technology cannot entirely replace the human interaction, support, trust, and care provided by visiting a physician’s office, she concluded.
