Research organizations and IRBs continue to face challenges and make tough decisions based on the best available information about a pandemic that changes daily as it spreads across the world.
Early on, most colleges and universities closed campuses and started online classes. This change left researchers in limbo. By April, most hospitals had begun shifting resources to clinical care for nonelective procedures and more serious conditions, including COVID-19 patients. Biomedical researchers and their study sponsors have had to decide what to do about ongoing clinical trials.
The Food and Drug Administration (FDA) provided some guidance in March, but left the decision to each investigator, sponsor, and institution.
The FDA chiefly said that researchers need to document the changes they make for each participant enrolled in research, said James Riddle, MCSE, CIP, CPIA, CRQM, vice president of institutional services and strategic consulting with Advarra in Columbia, MD.
But how do they make the best decision? Other than COVID-19 research, which clearly is the top priority during the pandemic, what type and phases of studies take priority over others? Riddle and other experts offered these suggestions:
• Think of clinical trials in terms of their lifespan. “Its lifespan could be as short as a week or two — or for 10 years in cardiovascular outcomes trials,” said Janet Wittes, PhD, founder and president of WIRB-Copernicus Group (WCG) Statistics Collaborative. Wittes spoke at a WCG COVID-19 web conference on April 1.
Studies in the design and screening stages can be halted until the pandemic is over or winding down in the trial site’s region.
“If your trial is in the design phase or in the screening phase, the decision is clear,” Wittes said. “This is not the time to start recruiting; sit back, look at the protocol, make sure it’s clean, and take time to fix it up.” Investigators can complete their database so the trial can start quickly, she added.
Trials that are in their last stage, where study visits are complete or nearly complete, also can stop any remaining participant visits — so long as this is feasible, scientifically and for safety, Wittes said. Researchers should limit queries only to those central to interpretation of the study. “Think harder about how to collect that query,” she explained. “Centers will be very busy with COVID-19 patients. For queries that are not essential to the central interpretation of the study, just sit back and don’t collect them.”
Efficacy data, especially for primary and secondary outcomes, are what are important.
• For middle stages of studies, make hard decisions. Studies that have collected fewer than 20% of endpoints or have collected 20-80% of endpoints are the ones for which decisions are the most challenging, Wittes noted.
“Think specifically about that particular trial and what the nature of the trial is,” she explained. “Is it a trial with imaging studies? Then the decision is quite different.”
Other questions to consider include:
- What is the study’s design? Is it randomized, adaptive, etc.?
- Would the study design impinge on how COVID-19 affects operations?
- Where does the study take place?
- Are there study sites where the pandemic currently is raging, or are the proposed study sites in locations where the pandemic has peaked?
- What axis should the investigator consider? For instance, are patients healthy, or are they seriously ill?
- Is the study drug well studied, or a new drug?
- Is the study’s purpose to assess symptoms, or is it curative?
If study participants do not require medical treatment, investigators may pause such trials during the pandemic, Wittes said. If a trial is taking place in a hospital, perhaps because participants already are patients there, investigators would be less likely to stop or pause, she added.
Some well-studied drugs have been prescribed to thousands of people; investigators may be testing these therapeutics for other indications. In these cases, investigators likely do not need to collect as much information on safety, she noted.
“[Investigators] can reduce data collection,” Wittes said. “If it’s a new chemical entity, you may have to think very hard about consequence of continuing the trial.”
• Know the best way to modify trials. Research organizations should consider the study sites, environment, sponsor information, and data collection challenges before deciding to modify a clinical trial because of COVID-19, said Jonathan Seltzer, MD, MBA, MA, FACC, chief scientific officer with WCG, and president of WCG ACI Clinical. Seltzer also spoke at the April 1 web conference.
“Is it a risky environment for patients because of risk of infection?” Seltzer asked. “Will the site have [clinical staff] available, because a lot of researchers are not allowed to come into [offices] because of no more outpatient visits?”
Here are some additional questions to consider before modifying a trial:
- How can the site minimize risk to participants?
- Are there benefits for participants remaining in the trial?
- What are the risks of staying in the trial?
- Can fewer data be collected?
- Can the number of visits be reduced?
- Is it possible for study staff to make home visits to participants?
- Are phone or videoconferencing visits possible?
- What types of digital technologies are available that could help reduce in-clinic visits?
Understanding the risks of having participants pull out of a study and developing COVID-19 is crucial to decision-making, Seltzer says.
“You don’t want to send people home to die of heart failure vs. taking a 10-20% risk of symptomatic COVID-19,” he says. “Do a risk-benefit analysis, and if you decide the trial goes on, think about how you can minimize that risk.”
For instance, investigators might be able to conduct fewer visits, and have two data points.
• Make practical decisions, but maintain study integrity. During the pandemic era, researchers and IRBs might need to consider various trade-offs. For example, would it be better to keep participant monitoring the same as pre-pandemic, or is there a way to obtain the same or similar data using a remote device?
“Say you’re looking at cardiac [outcomes]. Can you get [participants] an ECG at home? Or, do you need to get them into the hospital, or maybe wearing an Apple watch is good enough for what you’re looking at?” Seltzer asked. “If you have X-rays on the schedule, think about whether you really need them.”
Sites might be able to use a local lab for lab data, or forgo standard vitals if the data are not necessary, he added.
“These are all very specific to the trial,” Seltzer said. “Balance safety for the patient with preserving trial integrity. At the end of the day, if you can’t use data, then we wasted everyone’s time.”
Researchers also might decide to take their foot off the gas pedal of gathering so much data, Seltzer observed. “This is something that has to be reported to the IRB, discussing modifications.”
• Develop a statistical analysis plan (SAP). One way to balance the risks and benefits of trial modifications is through review and modification of an SAP, Wittes said.
Research organizations might consider the following:
- Carefully read SAP;
- Examine how operational changes due to COVID-19 affect the SAP;
- Ask statisticians who wrote the SAP speak with operational people;
- Amend the SAP as needed to conform to the operational changes;
- Explain why suggested changes are necessary;
- Provide thoughtfully designed sensitivity analyses.
Major changes to the SAP would include changes in how data are collected, how much data are collected, who collects data, and the primary endpoint.
“These have direct implications for a statistical analysis,” Wittes explained. “One thing we think hard about is making sure the people who are making these operational changes, clinical changes, and statistical analysis plans are talking.”
Here is an example of a change: Before COVID-19, researchers scheduled patients visit at weeks 22, 23, and 24. The operations people say participants could come in for just one of those three visits during the pandemic, Wittes explained. The statistical professionals say they will define weeks 22, 23, and 24 as the same point as just week 24.
“That’s an example of the intersection of operational and statistical changes,” she said. “It’s important to explain in that analysis plan why those changes are being made [for] when that [information] goes to regulators.”
Changes that are not explained to regulators are difficult to defend. “We have studies with a pre-COVID part, a during part, and an after part,” Wittes said. “There must be a standard analysis plan to address that. Think very hard about plans’ missing data.”