There’s no question that decentralized trial (DCT) methodologies have cemented themselves as a central part of the future of clinical research, but sponsors aren’t on the same page regarding their potential risks to good clinical practice (GCP) and how to mitigate them, a new survey shows, highlighting a critical need for greater dialogue, research and education.

The survey, which was conducted by ZS and garnered responses from quality leaders at 15 household-name sponsors, such as AstraZeneca, Eli Lilly, Johnson & Johnson and Roche, brings a crucial insight to light as DCT methods continue to increase in use: there’s significant variance among top sponsors in how they perceive the possible quality risks of DCT approaches.

ZS asked the 15 pharma titans to rank the level of risk they felt nine DCT methods present to patient safety/welfare, patient rights/privacy, data integrity/credibility, oversight (sponsor, vendor and site) and documentation of regulatory compliance from zero risk to very high risk. The findings demonstrate the current thinking on quality risks for telemedicine/virtual visits; home health services/home health nursing; eConsent; direct-to-patient shipment; remote monitoring; digital communication for patients and sites; local labs and imaging; direct data capture; and on-demand tech support for participants.

In particular, the survey’s findings on eConsent serve as a strong example of the differing perceptions among these change-driving sponsors when it comes to potential risks to GCP quality. And it’s this variability that should raise critical questions in future dialogues about DCTs and be explored more deeply through research, Jonathan Rowe, principal and head of clinical development quality, operations and risk management, ZS, told CenterWatch Weekly.

“The question really becomes, did one sponsor actually mitigate effectively, or is another sponsor unaware of the risk? We didn’t have that discussion, and I think that’s the next important thing to do,” he said. “This should be a noncompetitive area, where folks are going after the same modalities, and some do know how to control the risk and others aren’t aware of the risks.”

Specifically, while most sponsors (11) believed eConsent has a medium level of risk when it comes to patient rights/privacy considerations, two sponsors felt it bore low risk in this area, one sponsor felt it bore high risk and another felt it posed no risk at all. Similarly, nine sponsors felt it had no risk when it comes to patient safety/welfare, a single sponsor believed it had a high risk level and the remaining five felt it came with medium risks.

Sponsors also varied on their perception of eConsent’s risk to data integrity and credibility: one felt it presented very low risk, two felt it presented low risk and seven felt it posed no risk at all, while three perceived medium risk and two felt its risks were potentially high. Similar results were seen for eConsent’s potential impact on oversight at the sponsor, vendor and site level.

In addition to the differing views on its potential risks to GCP quality, eConsent has seen varied rates of use in the respondents’ trials and has thus far been kept from uniform adoption globally due to current regulations, added Arnab Roy, manager of ZS and lead of development and commercialization at the company’s DCT office. This murkiness around regulations and their implementation also creates potential risks, respondents said.

Some methodologies have clearer GCP quality risks to sponsors than others, according to the 15 respondents, but none were fully clear cut. For instance, nearly all felt that direct data capture poses no risk to patient safety/welfare, save for a single sponsor, who felt the methodology presents low levels of risk. Similarly, two felt it could present a low risk to patient rights and privacy and one even perceived a medium risk in this area. Conversely, sponsors varied significantly across the board in the level of risk they felt telemedicine presented to patient safety and welfare, patient rights and privacy, data integrity and credibility, oversight and regulatory compliance documentation.

While no approach was considered risky enough to bar its use in trials, the survey results suggest that DCT methodologies require more careful consideration when it comes to their possible impact on quality. Future discussions should certainly involve sponsors and their experiences with DCT, but it’s also important that investigators aren’t left out of these talks, Rowe says.

Sponsors employing decentralized approaches could, for example, unknowingly be adding to an investigator and site’s plates or taking away an important intimacy with participants instead of reducing burden. And while there may be situations in which DCT methodologies can actually cut down on risk and quality concerns compared to traditional approaches, they could also bring new risks into play, such as through potential software errors or the need to integrate technology.

Patients, too, must obviously be considered as the risks of DCTs are discussed and assessed at greater depth. The use of telemedicine could make it more difficult to catch adverse events (AE) in a trial of a drug that causes bruising as an AE, for instance, when a bruise on their body may not be visible on camera. And the question of encouraging and tracking drug adherence when a patient gets the medication shipped to their home is also one that needs answers.

“It is becoming very clear in industry right now that DCTs do not mean that sites are not going to be involved. It has to be a mixture of old and new when it comes to the SOPs, the processes and how we actually design and execute some of these trials,” Roy said.

“These risks will become more and more apparent and, going forward, what I’m looking forward to is … parsing out the perceived risks from the real risks. There seem to be a lot of perceived risks out there at this point. To what extent are those real risks something that we need to start looking into as an industry?”

Read the survey authors’ write-up here: https://bit.ly/3QjEEeb.