Improving interpretation for inpatients
Experts offer tips for hospitalists on providing care to patients speaking any language amid an ongoing shortage of interpreters.
It's well established that patients with non-English-language preference (NELP) have worse health care outcomes in the United States than their English-speaking counterparts, and census data show the U.S. is home to more than 25 million people who speak English “less than very well.”
Thus, it's imperative for hospitals and hospitalists to do their best to overcome these language barriers and provide culturally competent care. In-person interpreters are the gold standard, as they're better positioned to build relationships with patients and understand cultural nuances, compared with phone or video interpreters. However, high demand for these professionals has led to shortages, and even when they are available, interpreters are underutilized.
“Studies indicate that interpreters are utilized in less than 20% of encounters with eligible patients,” said Taru Saigal, MD, a clinical assistant professor of internal medicine at The Ohio State University's Wexner Medical Center in Columbus. That's despite the proven benefits. “We have ample evidence that when hospitals utilize interpreter services, readmission rates decrease and the length of stay is reduced,” she said.
To improve care for patients with NELP, researchers are working to better understand why uptake is low and how that can be corrected.
Understanding underutilization
Title VI of the Civil Rights Act mandates that health care institutions receiving federal funding offer interpretation services. While in person is the preferred option, many hospitals meet those legal obligations using video interpretation—which has the benefit of displaying facial expressions—and phone services, considered to be the least preferable method.
“The video and telephone interpreter[s] are much easier to get because it's on demand any time you need it,” said Lev Malevanchik, MD, a hospitalist and assistant professor of medicine at University of California, San Francisco. “But an in-person interpreter, although not as possible to get, is essential, especially if there are patients who can't hear well, patients who can't see well, patients who have any sort of confusion, dementia, anything like that.”
In-person interpretation is also the most expensive, which can lead to disparities in care. The “best cost the most money and the worst [telephone services] cost the least money. So it is really a matter of resources,” Dr. Malevanchik said.
A 2021 survey by AMN Healthcare, a staffing company, found that among their current clients, use of in-person interpretation fell by 15% after the COVID-19 pandemic and 52% of respondents said they were unlikely to use in-person interpreting in the future due mainly to cost concerns.
It can also be difficult to find an in-person interpreter for languages that are less common in the U.S., explained Amelia K. Barwise, MB, BCh, BAO, PhD, an associate professor of medicine and biomedical ethics at Mayo Clinic in Rochester, Minn.
From the physicians' perspective, logistical barriers may prevent uptake of interpreter services even when they are available. Clinicians may think it takes too long to contact and coordinate the service, they're too busy, it's inconvenient, or the patient speaks English well enough to understand or “get by,” Dr. Barwise said.
Assessing patients' English proficiency can be challenging. Institutions often lack standardized screening questions. Typically, patients are either asked “What is your preferred language for medical communication?” or “Do you need an interpreter today?” Others respond to a census-style question at admission that asks if they speak any language other than English at home, Dr. Saigal explained. If a patient answers yes, they're asked, “How well do you speak English?”
“Any person who speaks English less than ‘very well’ is categorized as limited English proficiency,” said Dr. Saigal, adding that this is a very subjective measure. “I may perceive my English proficiency very different, and it may vary, depending on the level of complexity of the conversation and how the stress levels are.” One workaround to this is to reassess patients' proficiency throughout their hospital stay through check-in questions, teach-back methods, and ongoing observations, she suggested.
Technological solutions
Given the difficulties of accurately determining patients' English proficiency at admission and shortages of interpretation services, physicians “may end up sending an interpreter to somebody who actually wouldn't need it as much as someone else,” Dr. Barwise said.
To address this and other barriers, she and colleagues recently developed and tested an artificial intelligence (AI) algorithm to promote equitable care for inpatients with language barriers and complex medical needs.
“If patients have a particularly complicated illness, they really benefit from the in-person interpreter who can understand the dynamics in the room, the nuances, the nonverbal communication,” Dr. Barwise said. “Remote or video or telephone interpreters can't pick up all that.”
Between May 2023 and June 2024, researchers carried out the randomized trial at Saint Marys Hospital & Methodist Hospital, part of Mayo Clinic in Rochester, Minn. Using an AI algorithm, the study team generated a daily list of adult inpatients with language barriers and ranked them based on complexity scores. They then shared the list with the hospital's interpreter services, which sent a reminder of the need for an interpreter to the bedside nurse for any patients judged to be high priority.
“We had several levels of human oversight of the AI algorithm,” Dr. Barwise noted. In the study's control group, who received standard care, it was up to clinician initiative to request interpreter services.
Results of the trial have not been released yet, but in December, the team published a report on clinical stakeholders' perspectives of the intervention in the Journal of the American Medical Informatics Association.
Findings suggested the AI intervention “would help overcome clinician bias, because clinicians avoid conversation sometimes with patients with language barriers, just because it is so onerous and tricky to get interpreters at times,” Dr. Barwise said.
Several common concerns about AI were raised by stakeholders, including its reliability and accuracy, but less frequently than potential benefits. Overall, “this specific application of AI may remediate gaps in clinician knowledge and training about when to use in-person interpreters; may effectively support prioritization for interpreters if shortages; can highlight an important human resource to optimize face to face communication and may support best practice to address health disparities, quality, and safety,” the authors concluded.
In another study, researchers at an academic medical center used automated early warning scores and integrated interpreters into their rapid response system to improve care for patients with NELP.
That intervention was associated with a decrease in average monthly mortality rates and expedited the process of getting an interpreter to the bedside, explained lead author Evan Raff, MD, ACP Member, professor of hospital medicine at the University of North Carolina School of Medicine in Chapel Hill. Findings were published by the Journal of General Internal Medicine in February, and the intervention is ongoing.
The team created a dashboard that displayed every adult inpatient who reported speaking a language other than English at admission. “With every rapid [response] event with a Spanish-speaking patient, the interpreter was paged overhead directly to bedside, come urgently,” Dr. Raff said. For patients who spoke languages other than Spanish, the rapid response nurses utilized virtual interpreter services.
Of the 222 patients with NELP included, 135 had rapid response system activations before the intervention and 167 afterward. There was an overall decrease in average monthly mortality rate from 7.42% (8 of 107 patients) to 6.09% (7 of 115 patients).
An alternative, low-tech solution could be to match physicians with patients based on language concordance. “Many hospitals have multilingual physicians or bilingual physicians, it's just we're not utilizing them like we should,” said Dr. Saigal.
This requires having an ample, linguistically diverse workforce to meet the demand, Dr. Saigal added. She's currently working to compile data on the linguistic diversity of hospitalists at her own institution “so that we can come up with some kind of workflow for patients and physicians who speak the same language.”
Matching hospitalists to patients would be ideal, but Dr. Malevanchik cautioned that if a hospital employs a strategy of relying on staff who aren't assigned to the patient but speak the same language to help with interpretation, “you're taking them away from their job, whatever that job is, to do a job that is not their job. … You're just asking them to do more work for free without paying them.”
Do's and don'ts
Busy physicians may resort to using hand signals or gesturing when interacting with patients with NELP, but experts recommend against this, noting it can be dangerous and result in inaccuracies.
Crunched for time, some may also turn to tools like Google Translate. This may be permissible in situations where you're asking the patient simple questions like “Do you need to use the bathroom?” Dr. Malevanchik said. But for any more substantial conversation, such apps should be avoided.
“If patients speak a language that doesn't have a big digital footprint, often a language not widely spoken, I think it's going to be very difficult for some of those translation apps to work effectively,” said Dr. Barwise.
She pointed out that communicating empathy and compassion is vital and something at which AI translations often fall short. It's important to also keep in mind that patients with language barriers can have health literacy barriers too, Dr. Barwise said.
In addition, experts cautioned against using family members as interpreters for a variety of reasons, including that the family member may omit details when interpreting for the patient, that the information they do relay may be unclear, and that by adopting the role of interpreter, they are no longer fully present to emotionally support the patient.
Whether using in-person or remote interpreters, ideally, hospitalists should briefly meet with interpreters before the visit to get them up to speed or give them time to prepare for delivering sensitive news, Dr. Barwise suggested. A debrief after the meeting can be useful too.
According to Dr. Saigal, there's a myth that “we should only utilize either video remote interpreting or in-person interpreting, and [physicians] may forgo the over-the-phone interpreting.” But any interpretation is more useful than none, she stressed.
Dr. Saigal also underscored the importance of cultural competency when caring for patients with NELP. “The staff has to use some kind of cultural engagement tools to understand where these patients are coming from, what their beliefs are, what they are trying to say. Everybody's expression can be different, and based on their health beliefs, what they're saying may not have a literal meaning.”
She acknowledged it would be impossible to be well versed in every culture, “but you have to be naturally curious and sensitive and be open.” The U.S. Department of Health and Human Services' Office of Minority Health offers resources for physicians on culturally competent care.
Modeling best practices for students and trainees is also key, Dr. Saigal said. If attendings don't use interpreter services, students are learning to do the same. “Research shows that when you actually train people on how to use interpreter services, our medical students and our residents do better at it. They are more inclined to use it and they use it properly,” she said.
Ongoing research
Building on his study integrating interpreters into rapid response alerts, Dr. Raff has continued to research assessments of English proficiency and health literacy among hospitalized patients, finding “significant” gaps.
“If we are going to make a change in how we provide language-concordant care to limited English proficiency patients, then we really need to have a better way of assessing and documenting this stuff at the outset,” he said, adding it's not currently feasible to perform a comprehensive language assessment on every patient who comes through the door.
Additional research by Dr. Raff and colleagues, published in Hospital Practice, found that compared with primary English speakers, primary Spanish speakers who had a rapid response had significantly higher disease severity scores and an 18.5% increase in length of stay afterward that was not mitigated after adjusting for disease severity.
That suggests that their care may already have been less than optimal in some way, he noted. “Let's intervene sooner with language-concordant care and see if we can make improvements in their disease severity when we get to them,” he said.
Down the line, one potential alternative to comprehensive language assessments at admission could be using AI to assess English proficiency based on documents or responses from a patient, Dr. Barwise said. But interventions like this are in nascent stages. More research to better understand where interpreters are running short or underused and the reasons why is also needed, Dr. Saigal added.
In the meantime, “we should be advertising our language services better, promoting them, letting patients know their rights to have interpreter services, and educating staff,” said Dr. Barwise. “We know if we use interpreters, they improve the quality of care for patients. It improves outcomes, it reduces readmission … all of those things that are really important to both patients families and institutions.”