Plain Talk About Plain Language Summaries

At first glance, the creation of plain language summaries (PLS) appears rather straightforward. However, upon closer inspection, PLS require a multi-step process: planning, development, translation, and dissemination. While the use of the term “plain language summary” is generally preferred in the industry over “lay summary,” those four points are derived from the European Health Union’s Good Lay Summary Practice guidance issued in 2021.

Expert panelists at the annual TrialScope EXTRA customer event in New York City, which focuses on all aspects of clinical trial disclosure, discussed these considerations in depth.

Moderated by Sylvia Baedorf Kassis, Program Director, Multi-Regional Clinical Trials Center (MRCT) of Brigham and Women's Hospital and Harvard, the panel included:

• Julie Holtzople, President of Holtzople Consulting;
• Maureen Kashuba, Associate Director, CDD&T, Leading Health Literacy Projects and Plain Language Summaries at Merck; and
• Courtneay Parsons, Head of Plain Language Summaries at ClaritiDox.

“We have this idea of health literacy in clinical research and the overall need for clear communications … being the overall umbrella” that inputs and feeds into the PLS pipeline, Kassis said.

The first question she threw out to the panel, regarding US regulations, was, “What does it really mean for PLS to be non-promotional?”

Holtzople, former head of clinical transparency and data sharing at AstraZeneca, said being non-promotional goes beyond simply using the right language. It involves building a holistic planning process, plus a standard or template to ensure consistency. She noted that such a template could have variations, with guidelines governing them. Having a process in place provides a “safety net” for clinical trial sponsors.

Non-promotional also means “not picking and choosing which studies you would do this for,” Holtzople said. Failure to do so, she continued, might look like a sponsor is doing this for the “good” studies and not the “bad” studies, which is the complete opposite of transparency.

Kassis noted that secondary endpoints often present a challenge when authoring PLS. A best practice to address this, Parsons said, is to apply a predefined standard proactively, equitably, and systematically on how to include endpoints in PLS. Many sponsors defer to a primary-only approach and tend to get pushback from patients, patient advocates, and their study teams. Parsons said the danger is, “with the regulatory clock ticking,” when the medical and transparency teams are at an impasse over what to include. In short, she said the best approach is to have standard operating procedures (SOPs) in place and say, “These are the rules and we’re sticking to them.”

Shifting the focus, Kassis asked the panelists to discuss tools such as AI and large language models used in PLS authoring. “We’re not there,” Merck’s Kashuba said matter-of-factly. “We’re getting there, but it is emerging technology.

“There is this inherent bias that when things are written well, they’re very believable.” Kashuba pointed to the need for human involvement. Otherwise, she added, it’s a disservice to patients and could lead to more distrust in clinical research. “At Merck, we do all our plain language summary work in-house.”

Holtzople mentioned she is involved with a working group, formed by the Center for Information and Study on Clinical Research Participation (CISCRP), to look at best practices on using AI for writing trial summaries. CISCRP will release the group’s report shortly, but Holtzople saw a preview of the data, which found that “the general public trusts AI more than they trust our industry, which is scary.”

In terms of obtaining reviews from the general public, Kashuba said Merck has an internal process seeking feedback from employees with non-scientific backgrounds. “What we’ve found,” she said, “is that a lot of people have personal stories about being a caregiver or having to translate something for a family member.” Merck also has worked with patient reviews and has found that communication preferences vary across therapeutic areas and cultures.

Kashuba reminded attendees that PLS are for real people. “We’re not just checking a box” and fulfilling the regulatory obligation. However, there is a cost associated with patient panel reviews, which Kashuba said can cost around $3,000 to $3,500 for a single summary. “That’s a lot of money if you’re doing 70–80 summaries a year. That doesn’t mean it’s not worth it. … With the tight timelines and the increased volume, we have to operate in the real world.”

Holtzople noted that this cost is “a drop in the bucket” compared to the total cost of a clinical trial, and that it’s up to disclosure professionals to prioritize this with higher-ups in their organizations. “It’s the right thing to do.”

If it’s not feasible to obtain patient reviews, Parsons said sponsors can enlist the help of patient advocates. A caveat, she said, is that advocates themselves are often subject matter experts, so they don’t offer the same review of readability as actual patients.

Kassis said that while PLS translations are not required, they do add value. Holtzople said  sponsors translate informed consent forms (ICFs), “because we want to recruit patients. How is it fair to put the sales material in their language but not give them the output?”

A sponsor can be compliant by providing a PLS in English first, Kashuba said, and later providing translated versions. She added that sometimes simply translating is not enough. Sponsors must meet the cultural needs of audiences in different countries and regions.

The final step in the PLS process is dissemination, and Kassis asked the panelists why this is such a big stumbling block for study sponsors and how it can be addressed. Holtzople said there’s a theory that forced distribution of PLS may not be advised, as some participants or family members may not want to see the results. In addition, sponsors cannot effectively keep track of participants’ addresses in order to mail them copies of results.

Parsons made the distinction between direct vs. indirect dissemination of results. “The thing that appeals to me … about TrialSummaries.com … is you want to get results to participants. That’s the least you can do. … We’re also trying to get this information out to the general public. … I think it goes a long way to the transparency that we’ve been working toward.”

Francine Lane, Senior Director of Product Management at Citeline, asked the panelists about using alternative formats to deliver PLS, such as video. Holtzople said it might be a more appropriate format for pediatric trials.

“I do feel that this is sort of the next level and how to get there,” chimed in Parsons, noting that sometimes different graphics are used to help provide explanation to a pediatric audience. “The text is more for the caregivers, the pictures are more for the kids.”

Another pain point panelists cited is the regulation requiring a protocol synopsis with nine mandatory elements. But, Parsons said, this synopsis is limited to two pages.

“I think one of the biggest challenges we have with these summaries,” Kashuba said, “is that we really don’t know who the audience is.” She is part of the Drug Information Association (DIA) plain language summary working group, which deals with these challenges.

Kassis noted that UK requirements vary from those in the US. Parsons pointed out that the UK requests a plain English summary at the start of the trial and a PLS at the end, but this is currently a voluntary practice. She said that a plain language protocol synopsis developed for the EU can be adapted to a UK PLS version in about an hour by adding the types of sites to be used.

Additional guidance on PLS best practices can be found at ClinicalTrials.gov.