Imagine you’re home and develop a sharp pain in your lower abdomen, and you think the best idea is to go to the hospital. A doctor stands, next to you, asking questions and explaining what comes next.
You don’t speak English or you have only a limited understanding of the language--enough to recognize a few words, maybe, but not enough to understand everything is happening, respond clearly and ask questions.
For millions of patients in the United States, this is their everyday reality. More than 25 million people who speak a language other than English face barriers to safe, high-quality medical care, including difficulty understanding discharge instructions after being release from the hospital and challenges with absorbing essential medical information.
At Mass General Brigham, researchers such as Jorge A. Rodriguez, MD, and Gezzer Ortega, MD, MPH, are studying how these communication gaps affect patient outcomes and how innovative tools, including artificial intelligence, can help close them.
Translation
The solution may seem simple at first glance—just find someone or something (like a computer program) that can translate the words of the doctor or nurse into a language that patients can understand. But communication in health care goes far beyond vocabulary.
For Gezzer Ortega, MD, MPH, a child of Spanish-speaking immigrants from the Dominican Republic, that understanding is deeply personal.
“Being in a healthcare setting and not being able to communicate clearly is pretty scary,” he says.
Ortega, has studied how emerging tools, including artificial intelligence, can help communication for patients whose primary language is not English.
In a recent study published in NEJM Catalyst of Spanish-speaking surgical patients, he and his team analyzed how patients experience--and benefit--from different types of language interpretation assistance in a clinical setting.
Participants were asked to watch standardized simulated postoperative encounters using both AI-based tools and remote video interpreters (RVI) to assist in interpretation. This design allowed patients to compare the two approaches under the same clinical conditions.
The study found that patients did not see AI and human interpreters as equally suitable tools for translation. Instead, their preferences shifted depending on the context.
- AI-based interpretation was valued for its speed, convenience, privacy, and ability to support more direct communication, especially in routine or time-sensitive interactions
- Remote video interpreters were preferred for more complex or emotionally sensitive conversations, where patients valued empathy, cultural nuance, and human connection
Patients described tradeoffs with both approaches. While many were open to AI, they questioned whether it could capture dialect, emotion or cultural context. At the same time, remote video interpreters were sometimes hampered by delays, interruptions, and concerns about accuracy.
Across interviews, the patients emphasized that control and flexibility matter most, with the ability to choose how they want to communicate.
The findings point toward establishing a more flexible, patient-centered model with options rather than entirely replacing remote human interpreters with computer models.
“AI appears most effective as part of a hybrid system, where tools are used in combination to support communication, trust, and patient autonomy,” adds Ortega.
Making Medical Information Clear and Usable
Jorge A. Rodriguez, MD, is focusing on another critical part of the problem: How patients understand written medical information.
His research has found that many patients never receive translated materials in a timely fashion, despite federal requirements that patients be provided with accessible materials. As a result, patients with non-English language preference are more likely to face higher emergency department use and hospital readmissions after discharge.
Even when translated materials are provided, they can be difficult to understand. If English is your native language, you've likely struggled to understand medical instructions at some point, too, just based on the complexity of the language that is often used in the clinical setting.
Traditional translation workflows are often slow, costly, and difficult to scale.
To address this, Rodriguez and colleagues are studying how computer-powered large language models can:
- Translate medical documents more quickly
- Simplify complex clinical language into plain terms
- Help patients better understand discharge instructions and next steps
Unlike earlier translation tools, newer AI systems can process entire passages and adapt the translated language for clarity, not just accuracy.
However, his research also identifies important limitations. Machine-generated translations can introduce errors, lose context, or perform poorly in less commonly represented languages. Because of this, federal guidance requires that AI-generated translations be reviewed by humans before reaching patients.
Rodriguez’s work points to a hybrid model, where technology improves speed and access, but human oversight ensures safety and trust.
“Being able to express yourself… goes a long way, especially in patient care,” he says.
Building a System That Works for Every Patient
Taken together, these research efforts point to a clear shift in how accessible medical information can be delivered in a way that values patient preferences.
Across Mass General Brigham, researchers and clinicians are working toward a more integrated approach that:
- Expands real-time access to interpreters
- Improves the quality and availability of translated materials
- Uses AI to support, not replace, human communication
- Centers patient preference and context in how language tools are used
The research from both Ortega and Rodriguez highlights the same conclusion: No single solution will work for every situation.
Instead, the future of language accessibility in healthcare will need to be flexible, patient-centered, and hybrid, combining technology and human expertise to ensure that every patient can understand their care, ask the right question and make informed decisions.
Subscribe to BenchMarks!
Subscribe to our monthly research newsletter and be the first to know what’s shaking in science at Mass General Brigham. From groundbreaking discoveries to the latest “you-heard-it-here-first” breakthroughs, we’ve got the updates that’ll keep your curiosity grooving.
Leave a Comment