Content Services
- Technical Writing
- Training & eLearning
- Financial Reports
- Digital Marketing
- SEO & Content Optimization
Translation Services
- Video Localization
- Software Localization
- Website Localization
- Translation for Regulated Companies
- Interpretation
- Instant Interpreter
- Live Events
- Language Quality Services
Testing Services
- Functional QA & Testing
- Compatibility Testing
- Interoperability Testing
- Performance Testing
- Accessibility Testing
- UX/CX Testing
Solutions
- Translation Service Models
- Machine Translation
- Smart Onboarding™
- Aurora AI Studio™
Our Knowledge Hubs
- Positive Patient Outcomes
- Modern Clinical Trial Solutions
- Future of Localization
- Innovation to Immunity
- COVID-19 Resource Center
- Disruption Series
- Patient Engagement
- Lionbridge Insights
Life Sciences
- Pharmaceutical
- Clinical
- Regulatory
- Post-Approval
- Corporate
- Medical Devices
- Validation and Clinical
- Regulatory
- Post-Authorization
- Corporate
Banking & Finance
Retail
Luxury
E-Commerce
Games
Automotive
Consumer Packaged Goods
Technology
Industrial Manufacturing
Legal Services
Travel & Hospitality
SELECT LANGUAGE:
This blog post is the latest in Lionbridge’s series on generative AI and what it means for the life sciences industry. In our last post, we shared an overview of the areas in which generative AI is likely to make an impact. This time, we’ll focus on patient literacy and patient communications. We’ll specifically discuss the challenges of making information about drugs and medical devices available to patients in plain language. Could emerging AI technologies help to overcome these challenges?
The Challenges of Patient Communications
Bringing a new drug or device to market is a lengthy and complex process. It generates enormous quantities of documentation. Much of that content is written by scientists and intended for a specialist audience. This isn’t a problem in itself. To ensure patient and public safety, regulators must impose a high burden of evidence for evaluating new therapies.
However, as regulators have realized, scientists and specialists are not the only stakeholders in the approval process. While protections for trial participants have been in place for many years, agencies like the European Medicines Agency (EMA) are increasingly concerned with making trials transparent to the wider public. As we outlined in a previous article, the EU Clinical Trials Regulation (EU CTR) now includes a requirement for plain language summaries.
Greater transparency is critical, but achieving it isn’t always easy. To begin with, writing in plain language for patient communications isn’t as simple as it sounds. For a more detailed look at these challenges, see our introductory article on plain language and healthcare.
Could AI Improve and Increase Patient Communications?
Generative AI systems are already skilled at summarizing complex content. But how well do they perform on dense medical texts? Here’s how Bing (using ChatGPT) responded when we asked it to summarize a BMJ article on immunomodulators in advanced multiple sclerosis.
The article discusses a study conducted by Lizak and colleagues, which aimed to evaluate the impact of highly active immunomodulatory therapy on the progression of moderately advanced and advanced multiple sclerosis (MS) 1. The study used data from over 4,000 patients with MS and found that the disability trajectories showed large variability. The probability of reaching the outcome was not associated with baseline variables, suggesting that previous disease activity before progress into advanced stages does not have a significant impact on later disease progression. However, higher relapsed rate was associated with disability. Additionally, highly active immunomodulatory therapy was associated with lower risk of reaching the outcome disability step 1. The authors conclude that disease progression in moderately advanced and advanced MS is highly variable and amnesic to prior disease activity. Lower relapse rates and greater time with active immunomodulatory therapy are associated with decreased risk of accumulating further disability. Highly active immunomodulatory therapy may delay disability in moderately advanced and advanced multiple sclerosis 1.
This summary isn’t a bad effort. It’s about half the original article’s length and reasonably well-structured. It begins by explaining the study’s objective in question and captures most of its key findings. It also accurately reports the authors’ conclusion.
However, much of the summary is made up of verbatim quotes from the original. The resulting language is dense, verbose, and full of specialist terminology. We asked the tool to try again using simpler wording.
The article talks about a study that looked at how a certain type of treatment affects the progression of multiple sclerosis (MS). The study found that the treatment can help delay disability in people with moderately advanced and advanced MS. The study also found that the risk of disability was lower in people who had fewer relapses and were on the treatment for a longer time. The authors suggest that this treatment may be helpful for people with moderately advanced and advanced MS.
This attempt is better, but achieves simplicity by omitting many significant details. For example, we can no longer tell what kind of treatment was being studied. More worryingly, eliminating nuances like “associated with” overstates the authors’ conclusions about risk.
Of course, this isn’t exactly a fair test. Bing’s AI chat tool is a preview product, and our methods here were far from rigorous. Generative AI technologies continue to improve. With appropriate training, we can expect much better outcomes. Still, some limitations are likely to persist. As we outlined in our previous post, adapting content for lay audiences requires skills encompassing multiple domains. Word choices and sentence length matter. So do layout, formatting, and adherence to regulatory guidelines. Crucially, too, layperson summaries must accurately represent the findings described in the original.
Lionbridge is actively exploring these potential applications. We’ve already implemented sophisticated techniques for quantifying complexity in source texts. As the technology advances, our advice to customers will continue to reflect our commitment to responsible innovation. We anticipate that generative AI will provide considerable productivity gains in supervised settings. We also believe such advances will enable Lionbridge and other Life Sciences organizations to make a wider pool of content accessible to patients and the public. However, our plain language service offerings will remain firmly expert-led.
Get in touch
Need assistance with your patient communications? Lionbridge has decades of experience and expertise in life sciences translation services. We know best practices for medical device translation services, clinical trial translation, and more. Contact us today to find out how Lionbridge can help as your language service provider.
