INTRODUCTION

Physicians underuse medical interpreters despite readily available interpreting services.1,2,3 They make decisions after weighing the benefits of accurate communication against competing demands such as “time constraints, alliances of care, therapeutic objectives and organizational-level considerations.”3, 4 Although physicians often resort to professional interpreters for interactions that are complex or involve “high-stakes,”1, 3 sometimes they find it easier to “get by” without one. Patients’ perspectives on interpreter use are rarely assessed.5 This paper sheds light on the challenges physicians and patients face with patients’ limited language proficiency (LP) in everyday clinical encounters and reflects upon the factors which drive the seemingly simple, unspectacular decision for—or against—the use of professional medical interpreters within a Swiss tertiary hospital.

The contemporary “speed, scale and spread”6 of international migration and the global refugee crisis place migration at the top of the political agenda in Switzerland and Europe. In the past 30 years in the German-speaking Swiss Canton of Basel-City, the foreign residential population has grown from 20.4 to 36% from 157 nations. These striking figures are reflected at the University Hospital Basel (USB), the canton’s main health care provider, where 36% of outpatients and 43% of hospital employees are “non-Swiss.” Therefore, the rapid and complex social change of contemporary societies is reproduced within health care institutions and individual clinical encounters, among and between patient populations and hospital staff.7,8,9 Basel-City and the USB, currently have the most diverse populations ever, culturally, ethnically, and linguistically.

Language forms a key tool to organize and navigate diversity, as is the use of interpreting services to provide appropriate health care across language barriers. Since the introduction of the interpreter service at the USB 30 years ago, the number of languages expanded from one to over one hundred languages available on-site and over-the-phone today. The use of these external language services increased decisively with Turkish interpreters consistently the most common. But languages like Tigrinya are on the rise representing 9.2% of all interpreter requests today, while Eritreans represent only 1% of all “non-Swiss” patients. This provides insight into the changes that the hospital’s patient populations have undergone and the inherent growing linguistic complexities. However, prior to using a medical interpreter, the need for one has to be acknowledged. This seemingly simple decision is complex and fraught with options and difficulties in the “real world” of clinical practice and may result in clinicians “getting by”1, 4, 5 without one.

While in the USA, patients with limited English proficiency have a legal right to access health care in their preferred language,12 this is not the case in Switzerland. Usually, advocates emphasize the need to address language barriers to avoid possible negative implications for quality of care and patient safety.1, 3,4,5, 10 Access to health care as a human right as defined by the United Nations and World Health Organization is inextricably linked to appropriate language services.11 Together with the recent statement by the Swiss National Advisory Commission on Biomedical EthicsFootnote 1 (NEK), they provide a valuable political instrument on a macro-level, although nonbinding. However, this “ideal” has to be transferred into the “real world” of clinical practice. Therefore, the micro-level of provider-patient interaction needs to be thoroughly considered and should be part of the inquiry. The micro-ethics approach by Komesaroff,12 Truog, and others13, 14 focuses on the unspectacular “choices that arise in everyday clinical encounters.”13 As Komesaroff states, “crucial ethical issues are involved […] in those clinical decisions which at first sight appear to be the simplest and most straightforward.”12 Data which relate the “ethical” with the “empirical”15, 16 are of fundamental importance to reveal the inconspicuous complexity of decision-making around patients’ limited LP.

We conducted a qualitative study to examine how “migrant patients” and hospital staff experience shared communication and interaction across language and culture throughout the entire hospital process from patients’ arrival to departure.

METHODS

The data presented are a subset of a larger study with a focus on how linguistic complexities shape the consultations of patients and physicians (primarily residents). The study was designed as hospital ethnography using semi-structured interviews, participant observation, and informal conversations conducted by the first author (KW). The research took place within the program on diversity management headed by the last author (SCH).

Patients were recruited from two main patient populations, 16 of TurkishFootnote 2 and 16 of AlbanianFootnote 3 origin, identified by their names on the clinic schedule and approached in the waiting area of two USB outpatient clinics. After introducing herself as a PhD student/Medical anthropologist (KW), patients were informed about the study and asked whether they agreed to participate if they had a follow-up appointment. The Ethics Committee of Northwestern and Central Switzerland (EKNZ) approved the study.

Ninety-four interviews were conducted between August 2012 and January 2015. Each patient interview was followed by interviews with all staff members that the patient had interacted with (front desk staff, nurses, physicians, and interpreters when required) (Table 1). The interviews were conducted face-to-face after informed consent was received, and privacy and confidentiality assured. Interviews lasted a maximum of 1.5 h, were tape-recorded, and transcribed verbatim.

Table 1 Data Sources
Full size table

The interview guide was developed based on literature17,18,19 and expertise of the co-authoring physicians (SCH, WL) with subsequent pilot testing and revision. The final version entailed general questions asked in each interview, and questions resulting from observations of the particular encounter to obtain the view of everyone involved. Each interview covered sections on demographics, cultural and social aspects, language, and communication. While patients, e.g., were asked: Were you able to understand everything in the encounter that was relevant to you? Physicians, e.g., were asked if they had the impression the patient understood everything that they perceived as relevant and how they noticed.

An observational grid was developed (KW, SCH) to pre-structure observation17, 20,21,22 of patients’ and staff members’ interactions. It included general conditions (e.g., attending persons), communication and interaction (e.g., verbal, non-verbal), and potential cultural components (e.g., discomfort due to gender discordance). The observations were recorded in field notes.

Content analysis according to Mayring23,24,25 was conducted (KW) using coding software (MAXQDA). After several cycles of analysis (inductive formation of categories by paraphrasing, generalization, and reduction), a category set entitled “insecure language” was formed with coded text sequences from 40 interviews. Codes within this category set were further analyzed (repeated review of interviews and category building). Open questions were discussed with the senior investigator (SCH). Patients’ and health care providers’ perspectives were compared and codes collated where appropriate (triangulationFootnote 4). Pseudonyms were used for cited actors and identifying participant details were omitted. For reporting, the COREQ guidelines were used.

The entire process was continuously accompanied by the co-authors’ insights and reviews, integrating their experience in communication, interpretation, and cultural competence. It was complemented by discussions with panels among medical anthropologists and clinical ethicists.

RESULTS

Three themes relevant to patients and physicians were identified: “Assessment of the language situation,” “use of interpreters,” and “dealing with conversational limits.”

Assessment of the Language Situation

Variations in Understanding and Communication Style

Physicians’ approaches to assessing their patients’ LP and comprehension included patients’ facial expression and body language, adequacy of their responses, presumptions about patients’ “intellectual capacity,” and their individual demeanor. When physicians assessed patients’ limited German proficiency (LGP), the relevant question was not whether patients understood, but rather how much. Physicians can judge a patient’s conversational understanding as inconsistent over the course of a consultation and be uncertain whether information came across or not. One physician described his difficulties assessing a patient’s capacity to understand as follows:

“Often, it isn’t easy […]. Partly, you sense that he understands quite a lot and that he is also able to answer […], but then there are situations where you get the impression that he understood nothing at all.”

Corresponding to physicians’ difficulties assessing patients’ level of comprehension, patients themselves reported how their capacity to understand clinical conversations depends on the difficulty level of the content, the speaking rate, or familiarity with the provider.

For example, Mrs. Arslan’s ability to understand depends on her capacity to cope with her “communicational needs.” To understand better, she interrupted the physician repeatedly to summarize or query contents she considered relevant and “checked back.” She explained:

“[…] I don’t understand much. Maybe [if he] just talks, I don’t understand everything. But [if] I ask by myself and he answers me, I do understand well.”

While helpful to Mrs. Arslan, Dr. Berg experienced her “strategy” as problematic. During the encounter, he interrupted her repeatedly and directed the conversation back towards his agenda because he wanted to make sure she understood. He guessed her conversational behavior was due to her impatience.

Physicians and patients applied and developed own techniques to address language barriers and found ways to “get by” with insufficient communication and limitations to understanding each other by making several minute decisions.

“Non-linguistic” Factors

Physicians’ assessment of patients’ LP was not based solely on patients’ capacity to express themselves. Additional “non-linguistic” factors were important. When patients appeared to be smart or self-confident, physicians frequently associated these features with a higher ability to understand. For example, Dr. Berg supposed that the patient was smart and able to understand his instructions on medical treatment despite LGP. He explained:

“[…] she isn’t stupid; it’s just the language barrier that hinders her.”

On the contrary, Dr. Mueller was unsure whether Mr. Begolli’s non-adherence was due to LGP or limited intellectual capacity:

“I (…) just don’t know whether it is a language problem or a problem of intelligence.”

Additionally, he suspected his patient had culturally driven convictions about the appropriate therapeutic approach. From the patient’s perspective, the difficulties were not rooted in the language barrier alone. In the interview, Mr. Begolli mentioned hearing loss, which he had not disclosed to the physician. This example shows how cultural and linguistic factors are not necessarily central but might hinder a more practical understanding.

Beyond that, patients’ self-confidence and resolute demeanor nurtured the impression that there was no language barrier. For example, Mrs. Pepshi firmly requested an iron infusion. Dr. Schmid perceived this as an indicator of her good linguistic skills:

“She was really self-confident, that’s how I also knew that there was no language barrier.”

Patients’ demeanor can influence how physicians rate their patients’ comprehension. However, physicians’ interpretation of a patient’s behavior can be misleading, as in Mr. Begolli’s case, who suffered from hearing loss. Again, minute decisions were made within the encounter with little distinction between patients’ LP, education, health literacy, and cognitive ability.

Use of Interpreters

Quality of Interpretation and Interpreter’s Reliability

Preference of direct communication, dissatisfaction with professional interpreters’ behavior towards patients, and the interplay and possible dynamics between a patient, interpreter, and provider were topics for physicians. Despite positive appraisals of interpreter services, for most physicians—regardless of professional or non-professional interpreters—the quality of interpretation was an issue. Physicians’ concernsFootnote 5 ranged from inaccurate interpretation, incomplete information to suspecting that the interpreter does not endorse their communicative agenda, as illustrated by the encounter between Dr. Wieden, Mrs. Abakay, and her husband.

When Dr. Wieden spoke about sexual intercourse, she felt deeply uncertain whether the husband interpreted properly to the patient, his wife:

“It was difficult for me to reconstruct whether he really interpreted it to his wife in the way I wanted him to.”

Another obstacle was the husband’s own limited LP. When it came to medical terms, Dr. Wieden used additional means:

“Fortunately, we have water bottles in the consultation room. So you can explain these [terms] ‘watery’ or ‘mucous’.”

Asked what she does when verbal communication is limited, she replied:

“If it is not possible at all to communicate with gestures or by a mix of English, French, Italian and German, at that point I would, depending on what is the matter, make an appointment with an interpreter as soon as possible.”

While an on-site interpreter is called for breaking bad news or when a patient consults the clinic for the first time, telephone interpreting is perceived valuable in case of clinical urgency (e.g., patient shows up in the emergency room with strong vaginal bleeding). In other words, the call for ensuring quality of communication by consulting available interpreting services is a decision determined by medical necessities. In less urgent situations, providers can rely on family members, multilingual hospital staff, own limited second language skills, or communication with gestures.

Hesitation

Both physicians and patients hesitated when making the decision for or against an interpreter. While physicians tended to doubt that an interpreter would be helpful, patients sometimes wished to manage the conversation themselves. Some patients feared that by asking for an interpreter, they would have to cover the costs or their request would be associated with negative connotations. Others were not informed about the availability of interpreter services at all.

Time constraints and concerns about disrupting their schedules can limit physicians’ use of interpreting services, although they simultaneously acknowledge the benefit of these services and their potential underuse. Dr. Mueller expressed:

“I never used telephone interpreting, but it is certainly a useful innovation although it takes an enormous amount of time. But yes, the interpreting service works out to some degree. One should probably use it much more.”

Dr. Mueller further reported that Mr. Begolli was “probably a model example” of a patient who might need an interpreter and that he “probably really should involve” one. When asked why he had not, he confessed:

“I don’t know.”

After the consultation, Mr. Begolli revealed that he would have appreciated an interpreter, but felt uncomfortable with the high cost of the service. Despite knowing that the hospital covers the costs, he feared difficulties with his health insurance because of it. In contrast, another female patient who admitted conversational limitations rejected consulting an interpreter:

“I will come also next time without interpreter. I want to manage it on my own.”

When asked how far she understood the conversation with the physician, she replied:

“I much understood, usual I understood not something missing. I understood. Only for speak I have little problems.”

Her wish to manage on her own is contrasted by her way of speaking, revealed in her answer when translated verbatim with grammatical errors and syntax.

Various decisions need to be made when dealing with a patient’s LGP by providers and patients. Although the language barrier can limit effective communication and influence patients’ capacity to understand physicians’ instructions and information, the decision to call for an interpreter is not necessarily made. Above, the physician concluded that an interpreter should have been involved only while reflecting retrospectively. Patients can consider it empowering to do it by themselves rather than with interpreters. Altogether, these decisions are fraught with uncertainties and hesitation.

Dealing with Conversational Limits

Being stretched to one’s limits to find common ground for mutual understanding was an issue for patients and physicians, which could sometimes result in feelings like “giving up.” For physicians, limitations became obvious, when repeated explaining and checking if a patient understood remained unsuccessful and resulted in accepting a patient’s lack of comprehension. As a consequence, “not having learned everything” (from a patient’s perspective) or “not having been able to communicate everything of importance” (as a provider) were common experiences. One physician explained:

“Some issues certainly remained open. You always try to ask and get to the point, but I don’t succeed every time. And then, you somehow give up and move on to the next point.”

Patients too accepted these limitations as a given “reality” and faced limits in their attempts to make themselves understood. Yet, for patients, this had a broader meaning including feelings of “being taken seriously” and “being understood.” For example, one patient felt the strong need to ask remaining questions about her condition. But, the physician cut her off. She wasn’t surprised that “he wouldn’t listen” to her and added resignedly:

“Many physicians do that.”

The acceptance of conversational limits—by physicians and patients—stands in contrast to the availability of interpreters. These conversational limits bear the risk of having a consultation and medical treatment of uncertain quality and presumably lower satisfaction for patients and providers. The uncertainty of this seemingly mundane decision to accept conversational limitations, instead of calling for an interpreter’s assistance, is not immediately apparent in everyday practice.

DISCUSSION

Our results reveal a troubling and heterogeneous range of factors shaping decision-making for or against an interpreter’s assistance by providers and patients alike. While patients often meet their need for interpretation by bringing someone along to the consultation, they rarely request professional medical interpreters. Providers categorize their patient conversations into ones of high or low medical significance. Physicians tend to use professional interpreters for “high-stakes” conversations, and to “get by” through “low-stakes” topics by resorting to bilingual family members or staff, their own second languages—even if incomplete—or by simply relying on gestures and mimicry.

“Getting by”1, 3, 4 describes a practice, which is rarely subject to closer examination. While the availability of language services is a necessary precondition for safe and effective communication across languages, physicians occupy a key position to initiate the process of medical interpreting.5 The seemingly simple decision to call for one or not, particularly when interpreters are available, is very complex and shaped by a range of small, frequently inconspicuous decisions in the “real world” of clinical practice. These decisions are often not part of an “actual process of clinical judgment,”12 and neither subject to critical scrutiny, nor of verbalized considerations or negotiations among the actors within an encounter. When language barriers exist, one crucial step often seems to be skipped—assessing a patient’s LP. Awareness about the decision-making potential of this very moment and its exploration hardly exists. Instead, research focuses on health care providers’ underuse of professional interpreters and risks though language barriers,5, 28, 29 the positive outcome when professional interpreters are used,30,31,32 or refers to clinicians’ second-language skills.30

Physicians respond with little uncertainty when patients’ LP is either high or absent. Their response is much more ambiguous when patients’ LP is somewhere in between. Guidance on how to assess a patient’s LP for the purposes of a medical encounter is rare33 aside from awareness for possible varieties in language proficiency and fluctuations in a conversation.34 While a patient’s limited LP can be sufficient for routine social demands and limited medical requirements,Footnote 6 proficiency is insufficient when the course of treatment, for example, reveals a poor medical outcome due to limited communication. Within a single consultation, the degree of problems in a conversation can vary due to fluctuations in difficulty level of content and uncertainties in comprehension by patients and providers. The mosaic-like character of these uncertainties is illustrated by multiple suboptimal circumstances, seemingly unspectacular when each is taken by itself. This reveals the risk of overlooking the significance of single routine actions, their potential importance for the course of a consultation, and the risk of probable subtle but adverse medical outcomes and suboptimal care.

Guidance is rare on how to balance this all to make an appropriate decision, for or against an interpreter’s assistance. This leaves physicians alone with uncertainties and places the responsibility of a decision solely on the individual physician.35 The conceptual framework of Schenker and colleagues36 is a valuable exception. It guides “physicians thinking through difficult choices about language services by four factors: ‘the clinical situation, degree of language gap, available resources, and patient preference.” But decision-making can be driven by factors more elusive than the objectifiable ones (e.g., available language services), for example, an environment which places higher value “on efficient completion of defined clinical tasks than on ensuring either effective communication or excellent care”37 or an individual provider’s level of engagement and moral commitment to ensure barrier-free communication. According to Komesaroff, “ethics is what happens in every interaction between every doctor and every patient.”12 A “doctor is involved in a constant stream of choices of an ethical kind, which are made at the local level of his or her interaction with the patient […]. The accumulation of these ‘micro-ethical’ decisions […] contributes importantly to the final qualitative and quantitative outcome of any particular medical encounter.”12 Presented results underline the National Advisory Commission on Biomedical Ethics (NEK) request for sensitizing providers to the challenges of cross-cultural communication and calls for teaching clinicians when and how to work with interpreters.38

Beyond that, the micro-decisions on the patient-provider level are closely intertwined with institutional conditions like facilitating language services. Within Switzerland, the institution under consideration is one of the advanced environments due to its long-standing availability of interpreter services, guidelines for choosing interpreters (e.g., on-site, bilingual staff), training courses, and crucially, readiness to cover the expenditures for medical interpreting. Nevertheless, huge intra-hospital variations still exist similar to the University Hospital Geneva.39 These variations range from organizational domains with high use of professional interpreters to domains with reluctant use, promoting bilingual staff instead. Ambiguities and uncertainties resulting in “getting by” are not only common on a patient-provider level but also on an institutional and even national level. Therefore, the micro-level reflects what takes places on the macro-level.

In Switzerland, the need to address the language barrier is widely acknowledged and the interpreting industry fast-growing. While in the USA, Medicaid has indicated that language services are eligible for federal matching funds;11 in Switzerland, no reimbursement is provided and a uniform solution for cost coverage (e.g., reimbursement by insurance companies) is not settled yet. Therefore, the financial burden to ensure barrier-free communication in health care rests on hospital decision-makers, or in private practices on the providers themselves, stressing their commitment. The commitments to human rights, to high quality care for all, and good communication are increasingly exposed to economic pressure. With all due respect to the moral imperative “it will be the financial equation that drives real change.”35

Research is needed on how LP can be assessed during hospital processing and to better understand physicians’ practice of “getting by” and the factors shaping their decision for—or against—an interpreter’s assistance. Physicians’ decision-making is a key element within the process of medical interpreting and demands more attention. The micro-ethics approach represents a vital concept for promoting awareness on the significance of “getting by” in the daily clinical practice of navigating language barriers. The concrete consequences of these decisions need to be examined in terms of their impact on the individual patient’s course of treatment and medical outcome. Patients have to be acknowledged as decision-makers in the interpreted encounter, and their attitudes, needs, and factors driving their decisions (e.g., rejection of professional interpreter use) need to be explored. The development and testing of a conceptual framework close to practice is crucial for guiding physicians’ decision-making on the use of interpreting services to ensure language equity and high quality care for all patients. However, these steps can only be successfully implemented with the support of institutional and political stakeholders in order to “get away” from “getting by.”