Professional interpreter: the visit in another language

An 82-year-old man who speaks only Spanish is discharged after a hospital stay for heart failure. The person who helps the nurse review his medications and explain the instructions, which are written in English, is an aide who speaks a little of the language but is not an interpreter. Ten days later, he comes back decompensated and needs ventilatory support: he had stopped the diuretic because of the diuresis and barely understood what the new medications were for (Medrano and colleagues, 2023). The case, described by Medrano and colleagues, sums up a problem that appears in any service that cares for people who do not speak the team's language (Medrano and colleagues, 2023).

A consultation in another language is not just a matter of courtesy. The way communication is done changes what the doctor understands of the history, what the patient understands of the plan and, in some scenarios, the clinical outcome. This text reviews what studies show about professional interpreters, family members and bilingual professionals, the risks of improvising and the practices that make an interpreted consultation safer.

Key points

• A professional interpreter improves the quality of care more than an improvised interpreter.
• Family members make more interpretation errors than professional interpreters, and omissions are the most common error.
• Consent, discharge, medication reconciliation and conversations about goals of care are high-risk scenarios.
• Video, telephone and in-person interpreters get similar ratings from patients, but in-person tends to generate more satisfaction.
• Speaking to the patient in short sentences and in the first person reduces losses in interpretation.

The size of the problem

The topic has been studied mainly in the United States, where the numbers are large. According to the 2000 census, 47 million people in the country spoke another language at home, and half of them said they spoke English less than very well (Karliner and colleagues, 2007). In 2013, there were about 61.6 million, of whom approximately 25 million, or 41 out of every 100, had limited English proficiency (Yang and colleagues, 2023). In 2015, about 70 million residents spoke another language at home, and 25.1 million had limited proficiency (Medrano and colleagues, 2023).

The phenomenon is not exclusive to countries with traditional immigration. Any health service cares for tourists, migrants, refugees, deaf people who use sign language and speakers of indigenous languages. The practical question is always the same: who will ensure that the patient and the doctor are talking about the same thing?

Medrano and colleagues call "making do" the practice of using non-professional interpreters, bilingual employees without certification or family members and friends (Medrano and colleagues, 2023). This improvisation is associated with lower patient satisfaction, lower self-efficacy, worse medication adherence, worse self-care and less understanding of their own illness (Medrano and colleagues, 2023). The authors also note that poor verbal communication increases the risk of safety errors and is associated with suboptimal clinical decisions, diagnostic errors, delays in care, medication errors and adverse events (Medrano and colleagues, 2023). For those who study clinical reasoning, this is the central point: a poorly taken history generates bad hypotheses, however careful the rest of the process may be. With patients who speak another language, it is also worth looking at one's own implicit bias.

Professional interpreter and improvised interpreter

The systematic review by Karliner and colleagues is the classic reference. Among 3698 references, 28 studies met the criteria, and 21 of them assessed the professional interpreter separately from the ad hoc interpreter (Karliner and colleagues, 2007). In the four outcomes examined, communication, use of services, clinical outcomes and satisfaction, the professional interpreter was associated with better care than the ad hoc interpreter (Karliner and colleagues, 2007). More than that, it seems to bring the quality of care close to that offered to patients without a language barrier (Karliner and colleagues, 2007).

Some numbers from this review help to size the difference. In the study by Flores, cited by Karliner, the proportion of errors with potential clinical consequence was 53 per 100 with professional interpreters and 77 per 100 with ad hoc interpreters, such as family members and staff (Karliner and colleagues, 2007). In a small Australian study, full understanding of a terminal diagnosis was 67 per 100 with an interpreter and 30 per 100 without an interpreter, but the difference was not significant (Karliner and colleagues, 2007).

The review included studies published between 1966 and 2005 that compared at least two language groups (Karliner and colleagues, 2007). The interpreters' training ranged from a few hours to more than a year across the studies, and half of them did not even report the training (Karliner and colleagues, 2007). In other words, "professional" does not always mean the same thing, and this heterogeneity limits the conclusions.

The review by Heath and colleagues, with 29 reports, reached a similar conclusion: the in-person professional interpreter had the best results in satisfaction and communication, any interpreter was better than none, and interpreters from the patient's circle can be a valuable resource in private practice (Heath and colleagues, 2023). In one of the included studies, satisfaction was 92.4 per 100 with the hospital's professional interpreter, 85.1 with an interpreter from the patient's circle, 53.3 with a telephone interpreter and 40 with an ad hoc interpreter (Heath and colleagues, 2023). Hearing the story in the patient's own words is the heart of narrative medicine.

The errors that happen in interpretation

Interpretation errors were studied in detail by Lauridsen and colleagues, who found only 7 studies on the topic (Lauridsen and colleagues, 2024). There were fewer errors with the in-person professional interpreter than with family members (Lauridsen and colleagues, 2024). The comparison with health team members without training in interpretation, however, showed no significant difference (Lauridsen and colleagues, 2024).

The most common error is omission. In three studies that separated the types of error, omissions accounted for 40, 42 and 63 per 100 of errors with a professional interpreter, and for 51, 46 and 61 per 100 with an ad hoc interpreter (Lauridsen and colleagues, 2024). Two studies found that the longer and more complex the sentences, the higher the risk of mistaken interpretation (Lauridsen and colleagues, 2024).

From this come the authors' three recommendations (Lauridsen and colleagues, 2024). First, the professional interpreter should be the first choice. Second, if one is not available, a health professional without training in interpretation should be preferred to family members and friends. Third, everyone in the consultation should use short, direct sentences, which reduces the risk of omissions. When an interpretation error causes harm, the conversation with the patient follows the principles of how to communicate a diagnostic error.

There is a type of error that directly affects diagnostic reasoning. Bauer and Alegría reviewed 26 articles on mental health assessment and described that an assessment conducted in a language that is not the patient's primary language can lead to an incomplete or distorted mental status examination (Bauer and Alegría, 2010). Untrained interpreters can make errors that cause the clinician to fail to identify disorganized thinking or delusional content (Bauer and Alegría, 2010). The language barrier can mask changes in speech, in the course and content of thought, and in perception (Bauer and Alegría, 2010). The authors note that a professional interpreter can improve the disclosure of information and lessen some of these difficulties, but that diagnostic agreement, joint treatment planning and referral to specialized care can be compromised (Bauer and Alegría, 2010). A hallucination that is not reported, or disorganized speech that is "tidied up" in interpretation, changes the list of hypotheses. Since much of the reasoning starts from the clinical history, a poorly translated history compromises what comes next.

When the doctor speaks the patient's language

A common way out is for the doctor to conduct the conversation in the patient's language. This can be great, but it depends on the real level of fluency. Patel and colleagues conducted a survey with surgeons at a university hospital, with a response rate of 33.1 per 100, or 158 of 477 (Patel and colleagues, 2016). About one third, 32 per 100, reported using their own knowledge of another language to obtain surgical consent (Patel and colleagues, 2016). Among those who considered themselves fluent, 67 per 100 rarely or never used an interpreter in that language (Patel and colleagues, 2016).

The same survey showed that many surgeons turned to bilingual staff, family members and even minors as interpreters for consent, although hospital policy expressly prohibited the use of minors (Patel and colleagues, 2016). When a professional interpreter was not available in time, improvisation became more frequent (Patel and colleagues, 2016). Overestimating one's own fluency is, at bottom, a problem of calibration of confidence.

When fluency is real, language concordance is advantageous. According to Medrano and colleagues, care provided by professionals who speak the patient's language is associated with greater satisfaction, better control of chronic diseases and fewer emergency department visits, compared with interpreter-mediated care (Medrano and colleagues, 2023). The authors note that even the best care with an interpreter is still inferior to care in the same language in comprehension and satisfaction, and that patients who switched to a doctor who spoke their language had a significant improvement in glycemic control (Medrano and colleagues, 2023). They therefore recommend that professionals who consider themselves fluent formally assess their own proficiency and seek certification (Medrano and colleagues, 2023). Fluency to chat about the weekend is not the same as what is needed to explain the risks of surgery.

Consent is precisely one of the scenarios that Medrano and colleagues classify as high risk without professional interpretation, alongside discharge, medication reconciliation and conversations about goals of care (Medrano and colleagues, 2023). These are moments when a misunderstanding has a direct consequence, as the case at the beginning shows.

The impact on outcomes

Beyond communication and satisfaction, there are data on clinical outcomes. Bertram and colleagues reviewed 1133 trauma activations at a referral center (Bertram and colleagues, 2024). Of the 70 patients with limited English proficiency, 62, or 88.6 per 100, needed an interpreter, and 41 of these, or 66.1 per 100, had a professional interpreter in the first 24 hours (Bertram and colleagues, 2024).

Patients without early interpretation stayed in the hospital longer than English-proficient patients, with a hazard ratio of 0.59 for discharge, and longer than patients with limited proficiency who had an early interpreter, with a hazard ratio of 0.51 (Bertram et al., 2024). Among those who had an early interpreter, the length of stay was the same as for proficient patients (Bertram et al., 2024). The study is retrospective and from a single center, but the signal is consistent with the rest of the literature.

Time with the interpreter also seems to matter. Torresdey et al. studied 338 patients seen with a professional interpreter (Torresdey et al., 2024). The consultation lasted an average of 47.7 minutes, the interpreter's auxiliary time before and after the visit was 43.8 minutes, and the total time reached 91.1 minutes (Torresdey et al., 2024). Patients gave an average score of 97.4 out of 100 to the interpreter's support and 93.7 out of 100 to communication with the clinician (Torresdey et al., 2024). More total time with the interpreter was associated with a better rating of communication, with a β of 7.23, but only when auxiliary time was included (Torresdey et al., 2024). The interpreter who accompanies the patient to the laboratory or explains the next step is also part of the care.

In person, by phone or by video

There is not always an interpreter in the room. Pathak et al. compared 326 patients' ratings of interpretation quality by video, in person and by phone (Pathak et al., 2021). The adjusted means were 3.91 for video, 3.86 for in person and 3.73 for phone, with no significant differences between the modalities (Pathak et al., 2021). The scale ranged from 1, poor, to 5, excellent, and most patients rated quality as very good or excellent in all three modalities (Pathak et al., 2021). Medrano et al., in turn, suggest preferring video over phone when interpretation is remote (Medrano et al., 2023).

The covid-19 pandemic tested these arrangements. In the review by Yang et al., concern about contamination and the lack of protective equipment reduced the availability of in-person interpreters at the bedside, and there was even reluctance to use phone and video devices (Yang et al., 2023). Hospitals responded with solutions such as a tablet mounted on a wheeled stand for audio and video interpretation in the emergency department, which logged 477 virtual visits, totaling 4123 minutes, in about two months (Yang et al., 2023). Another strategy programmed the phone interpretation number into an electronic badge that clinicians already wore, which allowed them to call the interpreter to the bedside by voice command (Yang et al., 2023). Phone interpretation minutes doubled between the first and second quarters of 2020, and clinicians reported a greater sense of connection with patients, because the interpreter's voice seemed to come from their own chest (Yang et al., 2023). The authors note that end-of-life and goals-of-care conversations are especially vulnerable to misunderstandings without professional interpretation (Yang et al., 2023).

In the ICU, interpretation has an extra layer. Suarez and colleagues interviewed 16 physicians, 12 nurses and 12 interpreters and described three roles for the professional interpreter (Suarez and colleagues, 2021). In the first, the interpreter interprets literally. In the second, the interpreter acts as a guardian of health literacy, taking on some patient advocacy. In the third, the interpreter works as a cultural mediator, conveying information with its cultural nuances (Suarez and colleagues, 2021). In all, there were 40 participants from 3 ICUs (Suarez and colleagues, 2021). Some physicians preferred literal interpretation, others felt less comfortable with it (Suarez and colleagues, 2021). There were reports of side conversations between interpreter and family from which the physician felt excluded, and of situations in which the information conveyed seemed incomplete (Suarez and colleagues, 2021). The guardian of literacy role was valued for alerting the team when understanding was at risk and for encouraging patients and families to ask questions (Suarez and colleagues, 2021). Clinicians also saw advantages and disadvantages in each mode of interpretation for the performance of these roles (Suarez and colleagues, 2021). In this setting, the interpreter becomes part of the team, and the problems described in interprofessional clinical reasoning also appear.

How to conduct an interpreted consultation

The good practices gathered by Medrano and colleagues form a simple script (Medrano and colleagues, 2023):

Do	Avoid
Agree with the interpreter before the consultation, when possible	Referring to the patient in the third person, such as "tell her"
Using first-person sentences	Addressing the interpreter instead of the patient
Using short sentences and asking for sentence-by-sentence interpretation	Side conversations with the interpreter
Speaking directly to the patient and keeping eye contact	Leaving the phone, the screen or the interpreter between you and the patient
Prioritizing the key points and addressing complex topics one at a time
Actively seeking the patient's perspective

The idea behind the list is simple: the interpreter gives voice to the conversation, but the relationship remains between the physician and the patient. Speaking in the third person or talking with the interpreter separately makes the patient a spectator of their own consultation. At the end, it is worth asking the patient to explain in their own words what they understood, a technique that the participants of the workshop described below cited as a learning point (Jones and colleagues, 2020). Attention to the patient's cues, discussed in how to train empathy in the consultation, also applies with an interpreter.

Training helps turn these practices into habit. Jones and colleagues described a 90-minute workshop on working with professional interpreters (Jones and colleagues, 2020). Among the 53 participants who answered the evaluation, 98 out of 100 said the workshop met its objectives, 98 out of 100 that it was a valuable use of time and 92 out of 100 that the supporting material was useful (Jones and colleagues, 2020). In addition, 90 out of 100 said they would apply the content in practice (Jones and colleagues, 2020). Participants came from various backgrounds, including medical students, residents and physicians (Jones and colleagues, 2020). It is a satisfaction evaluation right after the workshop, not an evaluation of behavior change, but it shows that the topic fits in a short teaching session. Participants cited as learning points their positioning in the room, eye contact and the method of asking the patient to repeat what they understood (Jones and colleagues, 2020).

In practice, a few questions organize the reasoning before starting:

• What is the patient's preferred language for talking about health, and not just for everyday conversation?
• Does the conversation involve consent, discharge, medications or goals of care, the highest-risk scenarios (Medrano and colleagues, 2023)?
• Is a professional interpreter available now, in person or remote, or is it possible to wait for one?
• If not, who is the least risky alternative, remembering that a health professional is preferable to a family member (Lauridsen and colleagues, 2024)?
• How much time will the consultation need, counting the interpretation and the time before and after the appointment (Torresdey and colleagues, 2024)?

These questions are part of clinical decision-making, just as much as the choice of test. The answers, including who interpreted and how, must be recorded in the medical record.

Difficult conversations call for even more care. When delivering bad news with an interpreter, the steps of the SPIKES protocol still apply, and it is worth agreeing with the interpreter beforehand on the content and tone of the conversation.

Frequently asked questions

Can a family member serve as interpreter?

In exceptional situations, yes, but it is not ideal. Family members make more interpretation errors than professional interpreters, and the recommendation is to prefer a professional interpreter and, failing that, a health professional over family members or friends (Lauridsen et al., 2024).

Can a child interpret for their parents?

They should not. In the hospital studied by Patel et al., policy expressly prohibited the use of minors as interpreters, yet some surgeons reported resorting to them (Patel et al., 2016).

Is video interpretation as good as in-person interpretation?

In patients' assessment, video, in-person and telephone interpretation received similar scores, with no significant difference (Pathak et al., 2021). Another review found greater satisfaction with a professional interpreter in person (Heath et al., 2023).

What is the most common interpretation error?

Omission, which accounted for a good share of the errors with both professional and ad hoc interpreters (Lauridsen et al., 2024).

If I speak a little of the patient's language, can I skip the interpreter?

It depends on fluency and on what is at stake. In consent, discharge and goals of care, which are high-risk scenarios, a professional interpreter is the safest choice (Medrano et al., 2023).

Should the interpreter just translate word for word?

Not always. In the ICU, interpreters describe roles that range from literal interpretation to cultural mediation and protecting the patient's health literacy (Suarez et al., 2021).

Can the language barrier get in the way of diagnosis?

It can. In mental health, for example, it can distort the mental status examination and hide disorganized thinking or delusions (Bauer and Alegría, 2010).

Take-home messages

• Prefer a professional interpreter; any interpreter is better than none, but family members make more errors.
• Omissions are the most common error, and short sentences reduce that risk.
• Consent, discharge, medications and goals of care are high-risk scenarios.
• Speak to the patient in the first person, looking at them.
• Early interpretation and adequate time with the interpreter are associated with better outcomes.

References

1. Bauer AM, Alegría M. Impact of patient language proficiency and interpreter service use on the quality of psychiatric care: a systematic review. Psychiatr Serv. 2010. PMID 20675834.
2. Bertram SC, Nichols FR, Cox LE, et al. Early use of professional interpreters improves trauma outcomes: Results of a single-center retrospective study. Surg Open Sci. 2024. PMID 39399161.
3. Heath M, Hvass AMF, Wejse CM. Interpreter services and effect on healthcare - a systematic review of the impact of different types of interpreters on patient outcome. J Migr Health. 2023. PMID 36816444.
4. Jones J, Rice K, Cueto V, et al. Increasing Health Care Workers' Proficiency With Using Professional Medical Interpretation: A Workshop. MedEdPORTAL. 2020. PMID 33204841.
5. Karliner LS, Jacobs EA, Chen AH, et al. Do professional interpreters improve clinical care for patients with limited English proficiency? A systematic review of the literature. Health Serv Res. 2007. PMID 17362215.
6. Lauridsen IG, Terkildsen MD, Sørensen LU. A systematic review of whether the number of linguistic errors in medical interpretation is associated with the use of professional vs ad hoc interpreters. Arch Public Health. 2024. PMID 39696623.
7. Medrano FJ, Fernandez AJ, Sudore RL, et al. Limited English Proficiency in Older Adults Referred to the Cardiovascular Team. Am J Med. 2023. PMID 36822259.
8. Patel DN, Wakeam E, Genoff M, et al. Preoperative consent for patients with limited English proficiency. J Surg Res. 2016. PMID 26541685.
9. Pathak S, Gregorich SE, Diamond LC, et al. Patient Perspectives on the Quality of Professional Interpretation: Results from LASI Study. J Gen Intern Med. 2021. PMID 33515189.
10. Suarez NRE, Urtecho M, Jubran S, et al. The Roles of medical interpreters in intensive care unit communication: A qualitative study. Patient Educ Couns. 2021. PMID 33168459.
11. Torresdey P, Chen J, Rodriguez HP. Patient Time Spent With Professional Medical Interpreters and the Care Experiences of Patients With Limited English Proficiency. J Prim Care Community Health. 2024. PMID 38912573.
12. Yang C, Prokop L, Barwise A. Strategies Used by Healthcare Systems to Communicate with Hospitalized Patients and Families with Limited English Proficiency During the COVID-19 Pandemic: A Narrative Review. J Immigr Minor Health. 2023. PMID 36821068.