Dimensions of Trust: The Tensions and Challenges in Provider-Interpreter Trust

Abstract

This study problemtizes provider-interpreter trust, examining the challenges to providers' and interpreters' collaboration in bilingual health care. We conducted in-depth interviews and focus groups with 26 medical interpreters (from 17 languages) and 32 providers (from 4 specialties) in the United States to provide an empirically-based framework of provider-interpreter trust. Constant comparative analysis was used for data analysis. We identified four dimensions of trust, theoretical constructs that can strengthen or compromise provider-interpreter trust: interpreter's competence, shared goals, professional boundaries, and established patterns of collaboration. We further explore how these dimensions highlight tensions and challenges that are unique in provider-interpreter relationship. We concluded with practical guidelines that can enhance provider-interpreter trust and propose future research directions in bilingual health care.

Interpreters traditionally have been conceptualized as conduits, a mechanical role that transfers information from one language to another neutrally and faithfully (Dysart-Gale, 2005). Interpreters-as-conduits is found to be the dominant ideology in both interpreters' training programs and codes of ethics (Dysart-Gale, 2005). Health care providers also reported viewing interpreters as “a neutral ‘translating machine’ or neutral ally in the consultation” (Leanza, 2005, p. 177). From this perspective, successful bilingual health care is accomplished by an interpreter who does not influence the process or content of provider-patient interactions. However, researchers have raised concerns about the inadequacies of the communicative model of interpreters-as-conduits (Dysart-Gale, 2005; Kaufert & Koolage, 1984), noting that interpreters systematically adopt non-conduit behaviors to influence the process and content of provider-patient interactions (Bot, 2005; Davidson, 2000; Hsieh, 2008). Many researchers have concluded that the complexity of provider-patient interactions make the conduit model impractical, if not unrealistic (Angelelli, 2004; Dysart-Gale, 2007).

In the conduit model, interpreters are perceived as translation machines, which implies that (a) the ideal interpretation is the same for all interpreters, (b) interpreters' individuality is inconsequential to the communicative process, and (c) interpreters' understanding of the objectives or functions of provider-patient interactions is irrelevant. Researchers have demonstrated that interpreters do not assume a passive or neutral role as prescribed by the conduit model. Rather, they are active participants who adopt purposeful strategies to manage the contexts, problems, and even conflicts that emerge during dynamic provider-patient interactions (Hsieh, 2006a, 2007; Leanza, 2005; Rosenberg, Seller, & Leanza, 2008). Providers' ability to work with different types of interpreters appropriately and effectively can be valuable and critical to the efficiency, quality, and informal economy of bilingual health care (Green, Free, Bhavnani, & Newman, 2005; Rosenberg, et al., 2008). From this perspective, an interpreter-mediated provider-patient interaction is a communicative activity that is coordinated and collaborated between multiple individuals.

In bilingual health care, health care providers' and interpreters' communicative behaviors are interdependent with each other. For example, researchers found that health care providers' communicative practice and unrealistic role expectations might be potential factors that motivate interpreters to deviate from a conduit role (Angelelli, 2004; Hsieh, 2006a). Although the quality of interpreter-mediated medical encounters have traditionally been assessed by interpreters' performance, researchers recently have found that providers' expectations for interpreters, communicative style, and communicative goals might influence their coordination with interpreters (Hsieh, 2006a; Rosenberg, Leanza, & Seller, 2007).

This paper echoes the recent trend in highlighting provider-interpreter dynamics in bilingual health care. We problemtize provider-interpreter trust, examining the challenges to providers' and interpreters' collaboration in bilingual health care. Trust is an important element for interpersonal relationships in health care settings. Researchers have noted that health care providers' ability to demonstrate specific characteristics (e.g., competence and compassion) and to adopt certain communicative behaviors can be critical to the trust building process in provider-patient relationships (Pearson & Raeke, 2000). In addition, researchers have argued that trust is a complicated, multidimensional concept that might be influenced both by social trust (e.g., trust in institutional regulations and normative expectations) and interpersonal trust (e.g., repeated interactions and existing identities) (Pearson & Raeke, 2000). Trust is not only important for provider-patient relationships but also for members of health care teams in providing quality care (Pullon, 2008). When professionals understand each other's roles and adopt communicative strategies appropriately and effectively, they are more likely to successfully coordinate with each other and provide quality care (Keenan, Cooke, & Hillis, 1998).

Trust is fundamental to provider-interpreter relationships and bilingual health care (Greenhalgh, Robb, & Scambler, 2006; Robb & Greenhalgh, 2006). The provider-interpreter relationship is distinctive from any other interprofessional relationships in health care settings. Providers rely on interpreters to convey their voices, including their identities, emotions, and information to patients. In other words, interpreters have control over providers' identity and information management. Health care providers need to believe that interpreters can provide services without distorting their voice or compromising the quality of care. Despite the increasing theoretical discussion on trust as a construct in interpersonal relationships, researchers have noted that few studies have attempted to ground a conceptualization of trust in individuals' actual experience and perspectives (Pearson & Raeke, 2000). Rather than discussing trust as a general or abstract concept, we aim to identify and examine the specific dimensions of trust in the provider-interpreter relationship. Our objective is to present an empirically-based conceptualization of trust developed through providers' and interpreters' understanding, practice, and experience of bilingual health care. We examine the dimensions and meanings of trust in provider-interpreter relationships by exploring providers' evaluations on the interpreters' trustworthiness and interpreters' trust-building strategies. In addition, because providers often work with different types of interpreters (e.g., family members, bilingual colleagues, telephone interpreters, and on-site interpreters) (for a review of different types of interpreters, see Hsieh, 2006b), we aim to generate a framework of provider-interpreter trust that is applicable to various types of providers and interpreters. By identifying the elements that build or threaten provider-interpreter trust, we aim to provide both a theoretical framework and a set of practical guidelines that facilitate interpreter-mediated medical encounters.

Method

Participants and Procedure

This study is part of a larger study that examines the roles of medical interpreters. The data included in this study are in-depth interviews and focus groups with health care providers and interpreters. The first author recruited 26 interpreters (from 17 languages) and conducted 14 individual and 6 dyadic interviews (each lasting 1-1.5 hours). The demographic data of the interpreters are listed in Table 1. The interpreters are from two interpreting agencies in the Midwestern United States. Both agencies view medical interpreting as their primary task and have contractual relationships with local hospitals. Interpreters included in this study are all considered professional interpreters. Five of the interpreters also work as managers of interpreting agencies. The first author relied on her experience as a medical interpreter and prior data collected through participant observations of bilingual medical encounters to navigate through the design, preparation, and interview process. The research questions focused on exploring interpreters' understanding and practice of their roles.

Table 1. Interpreters' Demographic Data.

Category	Range	Number	%
Gender	Male	8	30.8
	Female	18	69.2
	Total	26	100.0
Age	18-30	2	7.7
	31-40	6	23.1
	41-50	12	46.2
	51-60	4	15.4
	61-70	2	7.7
	Total	26	100.0
Education	High school	1	3.8
	Undergraduate or vocational school	4	15.4
	Bachelor's degree	13	50
	Master's degree	2	7.7
	Doctorate degree	2	7.7
	M.D.	2	7.7
	Not reported	2	7.7
	Total	26	100.0
Experience	≤ 1 year	4	15.4
	2-5 years	6	23.1
	6-10 years	7	26.9
	11-15 years	7	26.9
	≥ 15 years	2	7.7
	Total	26	100.0

After the initial analysis of the interpreters' interview data, the first author and her research team recruited 32 health care providers from a major health care facility in the Southern United States as a part of funded research to examine the providers' views of the roles of medical interpreters. The health care providers are from four specialties: OB/GYN, nursing, mental health, and oncology. The demographic data of the providers are listed in Table 2. In total, the first author conducted 8 specialty-specific focus groups and 7 individual interviews (each lasting 1-1.5 hours). The research questions were designed to examine providers' perceptions, expectations, and evaluations of interpreters' roles and practices. The first author obtained informed consent from all interpreters and health care providers. All procedures of the study have been approved by the Institutional Review Boards involved.

Table 2. Providers' Demographic Data.

Category	Range	Number	%
Gender	Male	8	25.0
	Female	24	75.0
	Total	32	100.0
Age	18-30	7	21.9
	31-40	9	28.1
	41-50	4	12.5
	51-60	8	25.0
	61-70	4	12.5
	Total	32	100.0
Specialty	OB/GYN	8	25.0
	Nursing	6	18.8
	Mental Health	7	21.9
	Oncology	11	34.4
	Total	32	100.0
Experience	Work with interpreters
	Never	3	9.4
	1-5 times	2	6.3
	6-10 times	3	9.4
	>10 times	24	75.0
	Total	32	100.0

The differences in the data collection methods (i.e., individual interviews, dyadic interviews, and focus groups) are a result of the authors' effort to include a wide variety of participants. Because interpreters' schedules were variable and subject to change in the last minute, the first author conducted interviews at the first opportunity. All dyadic interviews consisted of two interpreters from different languages (except one interview that included two Spanish interpreters). The differences of interpreters' working languages allow them to compare or elaborate on their cultural practices and differences. On the other hand, the specialty-specific focus groups for health care providers often were conducted before or after departmental or clinic meetings. The similarities in the providers' specialty allowed participants to have lively discussions about their specific needs and expectations for interpreters. The first author also offered individual interviews to providers who were unable to attend the focus groups. Although there are differences in how the participants' narratives were solicited, we believe that the combination of data collection methods allow us to include more participants and obtain richer data.

Data Analysis

After the interviews were transcribed, all three authors of this study used constant comparative analysis for the data analysis (Charmaz, 2006), coding the data for dominant themes and categories. We independently reviewed all transcripts to identify interpreters' and providers' understanding of trust in provider-interpreter relationships. Due to the hierarchical nature of health care teams, providers typically discuss their trust for interpreters, evaluating an interpreter's trustworthiness. Providers often compare different types of interpreters to illustrate why interpreters' certain traits, behaviors, or characteristics can be trust-enhancing or trust-compromising. In contrast, interpreters do not talk about whether they trust the providers, but emphasize the various strategies utilized to earn the providers' trust.

In our initial analysis, we identified both providers' and interpreters' narratives that express the presence or absence of reassuring feelings of confidence or reliance toward the interpreter. We also identified narratives related to the interpreters' credibility, authority, reliability, dependability, competence, communicative/interpersonal skills, identity, and professional roles, all of which have been identified as dimensions of trust for interpersonal relationships in health care settings (Pearson & Raeke, 2000; Pullon, 2008). We then focused on the narratives that involve the conceptualization, evaluation, or construction of trust, exploring the possible dimensions of trust in provider-interpreter relationships. We view dimensions as theoretical constructs that can strengthen or weaken provider-interpreter trust. By juxtaposing providers' and interpreters' perspectives, we explored the complexity of the dimensions of trust and examined the tensions and challenges within and between these dimensions.

Each dimension of trust proposed by one of the investigators was then probed by the others in a second pass through the data. We then combined similar findings; however, claims proposed by one investigator but not corroborated by others were discussed in detail with further consultation of the data for evidence to support or contradict the claim. The authors had several meetings to ensure that these dimensions of trust are distinctive categories and provide a coherent framework on provider-interpreter trust. We then went back to the transcripts to examine (a) inconsistencies that would suggest modification or qualifications to our earlier typology and (b) examples consistent with our earlier typology that could amplify or elaborate the original findings. We then finalized a list of dimensions of trust.

The transcription is italicized to indicate participant emphasis. Each participant is assigned a pseudonym. We denote interpreters with a superscript I (i.e., ^I) and health care providers with a superscript H (i.e., ^H) after their pseudonyms.

Results

Trust is often referenced by our participants when discussing provider-interpreter collaboration. Trust is conceptualized and negotiated in the following four dimensions: (a) interpreter's competence, (b) shared goals, (c) professional boundaries, and (d) established patterns of collaboration. These dimensions are pathways to enhance or compromise provider-interpreter trust. Although they are distinctive categories, they are interrelated and interdependent in constructing provider-interpreter trust.

Interpreter's Competence

The emphasis on the literal, neutral, faithful relay of information is often expressed by providers as their initial response to what they expect from an interpreter. In fact, only one provider explicitly stated that a literal interpretation is not preferred. Cara^H offered a typical response that values interpreters' conduit role, “[An interpreter is] a literal person who is impartial and unemotionally involved, who translates as much as possible, word-for-word of what you said.” Providers' understanding of interpreters' competence often centered on their linguistic ability as opposed to cultural competence or other skills. In particular, providers often emphasized the necessity to find equivalent medical terms and transfer the exact information across different languages. Interpreters are conscientious about providers' role expectation, emphasizing their efforts to stay within the conduit role. In fact, 21 of the 26 interpreters claimed various forms of a conduit role. Stella^I, explained, “I would try to use the same vocabulary that the doctor is using and use the same words to put that across…. [I provide] as accurate as possible interpretation, word-for-word of what they are saying.” Claire^I stated, “When I went through the training, we have to interpret everything exactly as what the doctor said, even have to interpret exactly the same tone, and same expression, and the same use of words.” Both the interpreters and providers considered that an exact and literal interpretation builds the credibility of the interpreters' performance.

Providers are in a dilemma because despite their desires for a neutral conduit, they do not have the language skills to evaluate interpreters' performances. Candice^H explained, “Not knowing exactly what [interpreters] are saying is very frustrating [especially] if they are saying things that shouldn't be said…. You are always worried.” As a result, rather than directly evaluate the interpreters' linguistic skills, providers assess interpreters' competence through their assumptions about the interpreters' training, credentials, and official role. This is a type of social trust based on shared interests and common norms and values (Pearson & Raeke, 2000) as the providers demonstrate a general confidence in the collective institutions. For example, Gram^H said, “I know that the ones hired by the hospital to work in women's clinic go through our [human resources] people and so there's something about them in their background.” Cordell^H said that she prefers a paid interpreter because “for right or wrong somebody who works for the [hospital] is going to translate what I said word-for-word as opposed to family members who might tell their mother what they want to hear.” The trust invested in the institutional control over the quality of interpreters is so strong that providers demonstrate trust for interpreters despite their lack of knowledge about an interpreter's background, ability, or codes of ethic. For example,

Nora^H: I wouldn't have a clue what [the interpreters] were saying. So I have to trust them. I would assume that they also have certain policies. And that's where a lot of trust comes in too. It is because we see them as professionals and that they have guidelines and stuff.

Natelie^H: I don't know anything about their training, but it would just make sense to me that you would have that kind of protocol even policies in place that you only interpret what the nursing staff or what the doctor's saying.

Many providers believed that hospital interpreters are licensed or certified. However, although the medical interpreting industry has proposed some guidelines (Dysart-Gale, 2005), there is no official licensing or certification procedure for medical interpreters at the federal and state levels (in the states from which the participants are recruited). Many interpreters commented that the industry-standard 40-hour training is insufficient for the complexity of interpreting in medical contexts. In these cases, providers' trust for interpreters can be based on problematic assumptions, believing that professional interpreters have more credentials and trainings than they actually have.

Providers' trust for interpreters' competence is also extended to their colleagues (e.g., physicians and nurses) who are bilingual. All providers said that they would fully trust their bilingual colleagues' interpretation and consider it just as good, if not better, than professional interpreters. Professional interpreters and bilingual health professionals obtain the providers' trust for their competence differently. Whereas the providers' trust for professional interpreters centers on interpreters' linguistic competence, their trust for bilingual co-workers is based on their colleagues' knowledge of and familiarity with medical issues. As a result, several providers talked about how they have more confidence in interpreters stationed in their department than interpreters who provide services intermittently (e.g., hospital or telephone interpreters). Some providers discussed how family interpreters can be problematic due to their lack of medical knowledge even though the patients might have complete trust for them. From this perspective, interpreters' competence is understood through their medical expertise.

In fact, no providers questioned whether or not a bilingual colleague can be a competent interpreter. This assumption, however, is not without problem. Sherry^I noted, “Just because you say you are bilingual, it does not mean that you are an interpreter.” Sara^I commented, “You are not only interpreting the words, you are also interpreting the culture. You have to be also bicultural, not only bilingual.” In addition, Sharon^I, manager of an interpreting agency, noted that interpreters who have medical backgrounds often feel more liberty to modify the clients' narrative, which can pose risk to the conduit role. In short, although some providers realized that being bilingual does not equate to cultural competence or language proficiency in medical contexts, their narratives reflect little concerns for the interpreting competence of their colleagues.

An important aspect of providers' assessment of interpreters' competence is that providers often share stories about interpreters' problematic performances. One particular story that was mentioned repeatedly in different interviews involves an interpreter who failed to provide information neutrally. In a version of the story, Garner^H explained,

One of my partners had an interpreter who was putting way too much of her personal views into things. But the aunt [of the patient] in the room spoke English and Spanish. And the aunt goes, “You know, that she was saying things that you didn't say.” Like when they said, “The baby had stage 4 cancer,” the interpreter said, “Yes, stage 4 cancer. Say your prayers.” That [interpreter] needed to be fired. That one was fired…. I mean the parents didn't know that doctor wasn't saying “say your prayers.”

This is a story shared among providers in social settings and departmental meetings. In other versions of the story, the interpreter was reprimanded rather than fired or the providers of the same department were cautioned about the particular interpreter's tendency to add personal views. This story provides insights into an under-discussed counterforce to the social trust of interpreters' competence we discussed earlier. Although a professional interpreter might be given trust automatically due to their official role, a failure in performance might lead to erosion of trust in his or her competence. In addition, when the violation of expectation is significant, the lost of trust might be extended beyond the particular event and provider-interpreter pair, and become an erosion of trust for the particular interpreter or even interpreters in general.

Shared Goals

As a dimension of trust, interpreter's competence centers on providers' evaluation and assessment of the interpreter. In contrast, the second dimension, shared goals of the health care team, highlights the partnership between the provider and the interpreter. Provider-interpreter trust is enhanced when they believe that they have shared goals as a health care team. Many providers said that they view interpreters as professional colleagues, who are members of the health care team. For example,

Cleo^H: I consider the translator part of our team…. The translator is absolutely vital to get the [medical] history. I mean you have to be able to ask those questions in that language and get that answers back.

Gemma^H: I hope [interpreters] don't feel like that they are on the lower totem pole because I can't do my job and take care of a patient without them. So, I value them as an equal colleague. Because despite the fact they don't have the MD behind their name, I couldn't be an MD without their assistance.

Interpreters also recognized the notion of teamwork and their important role in the treatment process. For example, Shirley^I commented, “You are the person that the staff is depending on to provide that communication, to convey the same spirit that it is being given.” Sandra^I explained, “This is part of working as a team. Both of us are discovering what is going on.”

The sense of team made some providers expect interpreters' alliance. Camila^H argued, “The translators should work with [providers] in a way that best meet their needs. I don't think patients are really thinking about the translator. I think they are thinking about their health care.” Carmen^H explained, “[Interpreters are] pursuing [my] agenda. Their bias is towards us.” Providers' expectation of alliance, along with the implied trust, is particularly strong for professional interpreters and bilingual co-workers. Gram^H noted, “I rely on the fact that the professional interpreter is supposed to be working for me, as a go-between with the patient; whereas the family member might be working for themselves, or might be working for the patient, or who knows what. They are not there for me.” From this perspective, family members are distrusted due to their lack of relationship with the providers at both the organizational and interpersonal levels. The notion of alliance is interesting because this expectation contradicts the neutral performance that is emphasized in interpreters' competence discussed earlier (see also Fatahi, Hellstrom, Skott, & Mattsson, 2008).

Many interpreters viewed themselves as members of the health care team. However, both providers and interpreters noted that because interpreters often are not stationed in the clinics and thus, have a more peripheral membership and might feel like an outsider. For example, Steve^I noted,

Usually, [providers] don't see us as [professionals], a part of their team…. Instead of seeing that we are providing a service for them, they see us more on the patient's side. They sometimes treat us like relatives of the patient…. Like they are seeing me on the other side, like an outsider. Not part of the clinic.

The contrast between family members and professionals is important. Family members, regardless of their linguistic skills, are not subject to the control of the health care system; in contrast, providers and professional interpreters have a shared identity that grounds provider-interpreter trust in the shared goals of the team.

Providers also aligned themselves with interpreters, noting that they both shared similar goals. One of the goals mentioned repeatedly is patient care. For example, Michael^H noted, “I want to connect myself with the interpreter if that's going to give us some kind of benefit to that patient to improve their outcome or their capacity to work with us.” Monica^H noted, “We're all working as a team to make sure that person gets treated.” Because patient care is viewed as a shared goal of the team, interpreters' active participation in the provider-patient interactions is appreciated when their intervention facilitates the quality of care. Carmen^H noted,

I trust that [the interpreter in our department] is going to interpret what I say or at least maybe to interpret it into a culturally appropriate discussion…. You know she's gonna empathize with the patient but it's not gonna become a relationship between she and the patient that could be a barrier to I'm trying do.

It is important to note that with the second dimension of trust, interpreters are trusted for their ability to make active judgments to fulfill the goals of the team. Interpreters' deviation from the conduit role is acceptable and even valued when it accomplishes the team's objectives. Providers commented that they hope the interpreters would feel comfortable to interrupt them if the interpreters need further clarification, believe that the providers' care is not culturally appropriate, or think that the patients' care is compromised. Nacia^H noted, “We're a team. If a translator said, ‘What about this?’ I'd go, ‘Oh yeah, ask them anyway.’ Or if it wasn't an issue, I'd say, “Well, you don't need to worry about that.” I think that open dialogue is very important.”

Many providers noted that they do not mind interpreters developing rapport or providing emotional support to help patients feel more comfortable. Others, like Carmen^H, agreed that it is okay for interpreters to change the information in a culturally appropriate way to facilitate provider-patient communication. Cecil^H noted that as a pathologist, he cannot do his job without a sample of the patient's biopsy. Rather than having the interpreter to interpret back and forth between him and the patient, he allows the interpreter to “push” the patient, convincing the patient independently, to give consent to the procedure. He concluded, “Usually the interpreter understands that this is really necessary. [..] The bottom line is … interpreter does whatever on their side, and the patient put their name to the permission [for the biopsy].” From this perspective, the interpreters' ability to actively identify and meet the needs of the health care team is critical to provider-interpreter trust.

Two issues emerged when contrasting providers' and interpreters' understanding of shared goals of the team. First, interpreters seem to focus on the ever-present, umbrella goal of patient care, highlighting medically related issues. This focus might lead them to be less sensitive to the emergent, dynamic nature of provider-patient interactions: Various communicative goals might (a) emerge throughout various phases of a medical encounter and (b) be distinctive to the needs of the particular clinic/specialty. For example, some providers noted that their rapport-building strategies are important to the therapeutic relationship. Grace^H, an OB/GYN physician, explained,

If I walk in and I like my patient's shoes, I'd say, “Oh, I love your shoes! They are so cute” [high cherry tone]…. And some of [the interpreters] go like, “Yeah, haha.” I'm like, “No! Tell her!” “I like her shoes!” “Tell her I love her baby!” We are an emotional group. And I think that's a large part of what we do.

Providers in mental health also emphasized the importance of establishing a good provider-patient relationship in the therapeutic processes. It is important for interpreters to recognize that some non-medical talk might still serve therapeutic purposes. On the other hand, the same behaviors might have different functions and consequences in different specialty areas. Oncologists noted that interpreters are greatly appreciated to provide emotional support to their patients, who often experience emotional distress. In contrast, mental health professionals noted that interpreters are expected to refrain from chitchatting with their patients let along providing emotional support because any talk with their patients might have serious clinical and therapeutic consequences. From this perspective, it is important for interpreters to have a more fluid understanding of the shared goals of the health care team.

Second, interpreters might view providers' goals as the shared goal of the team due to the implicit hierarchy within the health care team. Past studies have found that interpreters often assume providers' communicative goals (Davidson, 2001; Hsieh, 2007) and are distressed when they believe that the providers' communicative practice might hinder the quality of care (Hsieh, 2006a). Interpreters talked about how the providers' time constraint might pose a threat to the quality of care. Rachel^I explained, “What I feel is that sometimes like, the situation that you are between the doctor who is in a hurry, who wants to leave, and the patient who wants to talk, who needs time. You know, that's a really difficult situation.” However, some providers differentiate their individual goals from the team goals. As discussed earlier, providers welcome interpreters' active intervention to ensure the quality of care, which might lengthen the time for or disrupt the flow of provider-patient interaction. Several providers argued that their personal agenda (e.g., time management) is secondary to the quality of care. They prioritized the shared team goals over their individual goals. Because providers' narratives are mediated through interpreters, interpreters are faced with the task of differentiating and prioritizing the shared team goals and the providers' individual goals. However, the differences between team goals and providers' individual goals are not always clear due to the hierarchical nature of the health care team.

The first two dimensions of trust create an inherent tension, requiring interpreters to balance the neutrality that is valued in interpreter's competence with active interventions to protect and facilitate the shared goals of the team. The third dimension, professional boundaries, provides further insights into the delicate balance of the first two dimensions.

Professional Boundaries

Interpreters often experience conflicts as they contemplate the balance between quality care and their role boundaries. Interpreters noted the difficulties to challenge the providers' opinions, which often reflect their lower status in the institutional hierarchy and their concerns for the conduit role. For example, Sophia^I explained, “When I am there, I cannot even tell the doctor, ‘I don't think you are saying the truth.’ Who is me? I am just there to be their voice and that's my role.” Roland^I echoed, “You don't want to overstep your duties. Nobody asks my opinion anyways. Doctors have their own opinions.” These concerns might be unfounded as many of our providers noted that they welcome interpreters' input when they enhance the shared goals of the team. However, providers' willingness to accept interpreters' active role is not without limits.

Providers emphasized that interpreters should not overstep their role boundaries and overtake providers' control over the health care services. Professional boundary as a dimension, thus, suggests that provider-interpreter trust is enhanced when interpreters stay within their professional boundaries. Trust is compromised when they don't. In other words, an interpreter's effort to gain trust through neutrality (i.e., interpreter's competence) or active intervention (i.e., shared goals) should be considered in relation to his or her ability to stay within the role boundaries. Several providers talked about the importance for interpreters to maintain their professional role. For example,

Norma^H: [The interpreter] was checking the arm bands with the mother and the baby. But see, that's not her responsibility. That's my responsibility. And I felt like she was overstepping her boundaries.

Mindy^H: If [interpreters] chat [with patients], it should never be anything besides “How are you today?” … I mean, they are not friends. This is a boundary issue.

Role boundary, however, is not always a clear-cut issue in bilingual health care and is often intertwined with issues of institutional structure, professional expertise, and control over medical encounters. First, an interpreter's status in the institutional structure can be unclear, which might lead to ambiguity in the interpreter's professional status (and boundaries). Because interpreters come from various sources (e.g., family interpreter, bilingual health care providers, and on-site interpreters) and often move between various clinics, providers might not always know which type of interpreter they are using. Some providers noted that interpreters sometimes claim to be the patient's relative or friend in an effort to advocate for the patient. One of our interpreters mentioned a similar strategy even though she is a professional interpreter. Stacey^I noted, “When I assume the role of advocate, I let the doctor know that I am assuming the role…. I said, ‘I am a friend of the family. And I know the situation. Let me explain to you the issues around this.’ So, they understand.” In later discussion, it was clear that she did not have any social interactions with her clients outside of the health care setting and only claimed the role for advocacy purposes. Part of the difficulties for interpreters to adopt an advocacy role is that their institutional role is often conceptualized as a neutral conduit, providing them little resources for other role alternatives within the institutional structure. By claiming roles outside of the institutional structure, interpreters are no longer subject to institutional control and might have more resources to mediate provider-patient conflicts. These strategies can be problematic because health care providers might become suspicious or uncertain about the role of an interpreter. Nacia^H explained, “I don't know who you are. And I might ask and you might say that you're her aunty. I don't even believe that half the time.”

Second, the boundaries of expertise in bilingual health care can be ambiguous. Ideally, the management and control over a bilingual medical encounter is best to have the linguistically and culturally knowledgeable interpreter and the medically knowledgeable physician exercise their expertise in the corresponding area. However, the boundaries between what is medical, social, cultural, and linguistic are not always obvious. Interpreters are in a difficult position as any talk can be perceived as medically meaningful in a medical encounter. Mindy^H, a mental health provider, explained that although the characteristics of a person's voice (e.g., volume, speech rate, and fluency) have diagnostic values, providers need to rely on interpreters to decipher the meanings of these signs. She explained, “If you hear people speaking Italian with their voice goes up or what if the interpersonal space is closer. To somebody who is only English dominant, it's ‘Oh, they are going to have an argument.’ But it's not. It's what's normal within the culture.” From this perspective, interpreters are responsible for diagnosing the medical meanings of these nonverbal behaviors. In addition, providers' understanding of what constitutes as medically meaningful talk can vary dramatically during the discursive process and their area of specialty. Gloria^H, an OB/GYN physician, talked about how she was frustrated about a patient who kept mentioning that her husband is going to be put on a ventilator that night. She concluded, “I got her into a psychiatrist after that too. But that's not a pap smear.” In this case, depression was considered medically irrelevant as it is not an OB/GYN issue. She added, “The interpreter needs to know how to keep the patient focused if the patient is not focusing so that the time management in the interaction is efficient as well.” In contrast, many mental health providers argued that they consider many issues (e.g., financial distress) medically meaningful even when those topics might be appear medically irrelevant to other specialties. For example, Michael^H, a psychologist, indicated that interpreters should refrain from having casual conversations with his patients, some of whom have paranoia or post-traumatic stress disorder. For those patients, even everyday casual interactions can have serious clinical consequences. He explained,

Cause I'm not there to participate in guiding that interaction. Should the family or the patient have a lot of angst or anxiety about seeing us, the interpreter won't be able to regulate that as well as if I was there. So that may impact patient care cause they are not gonna talk with us. Or they'll become too anxious and they'll kick us out or they're paranoid or afraid or angry.

Providers also noted that some interpreters were welcomed for a more active role because of their medical expertise. Nancy^H explained, “If I feel that they've been a nurse or they've been a doctor or whatever, I'm little bit more comfortable to let them step over the boundary a little bit.” Such an attitude was supported by some of our interpreters who noted that, due to their medical background (e.g., being a nurse or a physician in their home country), providers have asked them to perform tasks that are typically not associated with interpreters. From this perspective, not all interpreters are endowed with the same set of professional boundaries. In short, interpreters' role boundaries can vary drastically depending on the tasks, the clinical specialty, and the providers' knowledge of their background.

Finally, role boundaries imply that each individual is responsible and can be accounted for his or her behaviors and voices. For example, having control over the medical encounter is critical to providers as they feel the need to be in charge of interpreter-mediated interactions. Some providers were adamant that they have to be in charge of the medical encounter. For example, Nacia^H said, “I consider [interpreters] colleagues, but ancillary services to mine. I welcome [interpreters' input]. But I still get to call it. [laugh] I'm still the leader.” Candice^H also argued, “The [patients'] bonding and the feeling of ownership is to be with the doctor. This is coming from a doctor…. [Interpreters] are not the communicator. They are assisting the communication.” These comments reflect the providers' sense of hierarchy within the health care team. In addition, some providers noted that because interpreters are the voices of others, it is critical that interpreters are clear about whose voice (e.g., the provider's, the patient's, or their own voices) they are representing and do not hide behind others' voices. Role boundary is an important issue because all voices are conveyed through the interpreter, which creates difficulties for a patient to differentiate the opinions of the doctor and the interpreter. When an interpreters' opinion is blended into the providers' narrative, providers lose control over their own voices yet might still be held responsible for the information that is provided to the patient. Several providers are concerned about the blending of voices. They raised the issues of legal liability. Nancy^H explained,

I can tell that [the interpreter] tell them the extra stuff. You can just tell. And I was thinking, “I didn't say that, I didn't say all that.” … She forgets her boundaries and then she jumps into our boundaries. Then, what happens is legally we are responsible for that.

Having role boundaries in the provider-interpreter relationship allows both providers and interpreters to be held accountable for their voices in situations that might have legal or clinical consequences. Trust is compromised when providers believe that they have lost control of their voices (e.g., interpreters hiding behind their voices). Some researchers, however, have argued that because of interpreters' role in representing the voices of providers (and patients), the blending of voices from multiple participants are inevitable in bilingual health care (Hsieh & Kramer, in press). In other words, providers can never have full control over their voices, which presents an inherent challenge in an interpreter's ability to maintain his or her role boundaries.

Established Patterns of Collaboration

Trust is enhanced when providers and interpreters share an established pattern of collaboration. Providers noted that repeated interactions with a particular interpreter are desirable in the trust-building process. In particular, both providers and interpreters noted that repeated interactions build trust through established patterns of collaboration. An established pattern of collaboration is an emergent and evolving coordination between the provider-interpreter pair that is established over time (see also Goffman, 1959). Candice^H noted, “I know [some interpreters] personally. Often, I've worked with them more than once or twice. So, I know their style and they know mine. They know the words I use and all that. I mean, it's just better to have someone you know and trust.” Different pairs would have different dynamics and each pair's pattern of collaboration might be different from others. In addition, the established pattern of collaboration allows the pair to comfortably work with each other even when they deviate from normative expectations of the previous dimensions of trust.

Provider-interpreter trust is enhanced through their established pattern of collaboration in the following ways. First, both providers and interpreters can become more efficient in anticipating each other's communicative needs. For example, providers might become more sensitive to expressions that might be confusing when translated into different languages and cultures. Shirley^I mentioned that providers often are not aware of how their talk might cause confusion. She once asked a provider to clarify the term “giddy,” which might mean dizziness, lighthearted, or impulsive. Her questioning not only allowed her to interpret accurately but also made the provider become vigilant about possible confusions. Many providers said that they appreciated interpreters' suggestions to modify their narratives in a way that is culturally appropriate. For example, several oncologists talked about how using the concept of soup to discuss components of blood (e.g., red and white cells) might not be appropriate for cultures that have only purée soup. Interpreters have valuable knowledge to help providers to find the best metaphor in those situations. By learning each other's communicative needs and objectives, both providers and interpreters might help each other to become better communicators in future interactions.

Second, both providers and interpreters learn to adapt to and appreciate each other's communicative styles. Interpreters were trained to adopt a specialized style of speech, including first person interpreting style (e.g., talk as the original speaker), simultaneous interpreting (i.e., interpreting while other speakers are talking), and specific nonverbal strategies (e.g., avoiding of eye contact). For example, Stella^I explained, “I detach myself emotionally from many things that are going on there, and I look at the floor, and I look at the ceiling or something. And I make sure that they talk to each other.” Roger^I echoed, “I look at the floor, I look down, and I just interpret immediately what he said, or what she said.” Several interpreters talked about how they educate providers and patients to talk appropriately in interpreter-mediated interactions. For example, Sophia^I said that when she met clients for the first time, she would say, “I am gonna be just your voice. [..] Don't look at me, you have to look at [each other] because the conversation is with [him or her].” Interpreters' specialized style of speech aims to reinforce the provider-patient relationship and minimize their intrusion to the medical encounter. Many providers, however, commented on how interpreters' style of speech can be confusing and counter intuitive. Cory^H noted, “Before you finish your sentence, they are already speaking. That really bugs the tar out of me…. And not looking at the person and not looking at me…. It is distracting to everybody in the room.” As a result, some interpreters modified their interpreting style (e.g., changing from simultaneous to consecutive interpreting) to accommodate the providers' preference. Several providers also learned to adapt to and appreciate interpreters' strategies. For example, Candice^H said, “[The interpreter] would be talking as I was talking and there was no emotional reaction. Once I got used to that style, I kind of like it. Because the parents are looking at me and reading my nonverbal and my emotions.” Other providers made an effort to address patients directly, even when their intuition was to look at the interpreters. In short, interpreter-mediated interactions require individuals to adopt communicative styles that are different from the monolingual norms. As providers become more aware of the interpreters' communicative styles, they gain insights into the functions and values of these strategies and learn to modify their communicative behaviors accordingly. Once the providers and interpreters establish their pattern of collaboration, they no longer question each other's awkward way of speaking, but learn to collaborate in a way that best met their communicative goals.

Third, interpreters gain the opportunities to become more familiar with clinic-specific procedures. Many interpreters emphasized the importance of on-the-job training to complement their 40-hour training program. Some interpreters talked about their preferred clinic because they knew the providers and the procedures of those clinics better. Many providers echoed the notion of having the same interpreter for their everyday tasks to ensure and improve the quality of interpreting. Provider emphasized that such a practice would increase interpreters' familiarity to their routine talk, allowing them to provide better interpretation. In addition, providers argued that an interpreter who is familiar with their medical talk can be extremely valuable. For example, they can notice the variations across different patients, and be able to alert the provider if something warrants additional attention. Mindy^H explained, “It's like a baseline. If it's a skilled interpreter and they've seen many people coming with this problem, and this particular person is discussing it in a totally different way, then, I would want to know that.” Several providers talked about incidents that experienced interpreters reminded them about information that they forgot to ask or include in their routine procedure, arguing that the interpreters' active monitoring allowed them to provide better care. From this perspective, interpreters are trusted to have a certain level of medical expertise that is based on their familiarity with the clinic-specific procedures and medical dialogue.

It is important to note that it is not a common practice to consistently send the same interpreter to the same clinic, provider, or patient. In fact, several managers commented that they intentionally avoid keeping the same patient-interpreter pair too long to avoid the development of a patient-interpreter bond, fearing that such a bond might encourage the interpreter to become a patient advocate and neglect his or her professional boundaries. However, because the patient is likely to have the same provider throughout the course of an illness, changing the interpreter might also change the provider-interpreter pair. This practice might weaken the provider-interpreter trust because the new provider-interpreter pair needs to adapt to each other's communicative style again and develop new styles for collaboration.

Discussion

The objective of this study is to present an empirically-based conceptualization of provider-interpreter trust. These dimensions are the entry points to understanding how provider-interpreter trust can be compromised or strengthened. The dimensions identified in this study highlight inherent tensions and challenges that are unique in the provider-interpreter relationship.

First, these dimensions of trust are simultaneously complementary and competing constructs. The providers' and interpreters' understanding of these dimensions is best understood from a systems theory perspective (von Bertalanffy, 1976): The four dimensions are interdependent, functioning simultaneously to construct provider-interpreter trust. For example, the findings of this study demonstrate that although interpreters' competence is mainly evaluated through their neutrality, other dimensions suggest that deviation from the conduit role is acceptable and even appreciated when certain requirements are met. Nevertheless, interpreters' active interventions might concurrently reinforce certain dimensions (e.g., shared goals) while posing risks to others (e.g., role boundaries). From this perspective, trust building is not a zero-sum game for each individual dimension but a delicate balance between all four dimensions. For example, if an interpreter is working with a provider for the first time (i.e., lack of established communication patterns), he or she can still gain trust through neutrality (i.e., interpreter's competence). When deviating from a conduit role, interpreters can still secure trust by emphasizing their efforts to maintain shared goals or explaining why their interventions are still within the bounds of their professional responsibilities. When there is significant trust built through repeated interactions, an interpreter might even reshape their professional boundaries, being trusted with medically-related responsibilities.

Second, providers are not always consistent with their evaluation of an interpreter's trustworthiness through these dimensions. During the interviews, several providers noticed how their expectations for interpreters are self-contradictory. Different providers also might have different criteria and expectations in evaluating an interpreter's trustworthiness. Because providers have different expectations, the same communicative behavior (e.g., emotional support) might be viewed as trust-building by one provider and trust-compromising by another. Finally, these dimensions of trust might vary significantly due to the history of the provider-interpreter pair. An interpreter might be trusted with medical responsibilities by one provider but not another. In short, these dimensions represent a set of multi-dimensional and situational expectations that are placed on interpreters, requiring interpreters to be adaptive and responsive to providers' needs. They are best understood as situational and contextual guidelines, rather than fixed standards, to build trust.

Third, the hierarchy of health care teams present challenges to interpreters' trust-building strategies. When interpreters fail to foster provider-interpreter trust, they might risk losing their job; in contrast, providers face little consequences if there is little provider-interpreter trust. In addition, providers are the constant structure of the medical settings and the interpreters are only present as needed. A problematic performance by an interpreter can become an institutional cautionary tale, tainting not only the reputation of the particular interpreter but also interpreters in general. As a result, the structure of the team puts pressure on professional interpreters to be biased toward health care providers, risking the quality of care. Nevertheless, this study shows that providers do not necessary expect interpreters to side with them or to maintain neutrality. When certain behaviors can be justified, providers can be receptive to deviations from the norms. From this perspective, providers' knowledge about the underlying reasons for interpreters' communicative strategies is critical to their evaluation of interpreters' trustworthiness. Providing workshops to providers might be a valuable venue to strengthen provider-interpreter trust, allowing them (a) to be aware of the functions and meanings of interpreters' strategies and (b) to learn to coordinate with interpreters effectively and appropriately. On the other hand, interpreters' communicative strategies should be transparent so that providers are involved in deciding the best course of action. Interpreters need to be attentive in addressing providers' concerns and potential miscommunication to ensure that the integrity of the interpreter (and the interpreting services) is protected.

Finally, these dimensions highlight the importance of examining the impacts of various types of interpreters. Providers' assumptions about different types of interpreters might lead them to trust one type of interpreter more than others. At the same time, providers' trust is not always based on accurate assumptions. In addition, certain types of interpreters, by definition, have less trust in certain dimensions. For example, it is unlikely for a telephone interpreter to gain trust through repeated interactions because the service calls are typically directed to the interpreter who is available at the moment. On the other hand, the level of trust in different dimensions is subject to change. For example, as family members become more familiar with the patient's illness and his or her providers, they might gain trust in some dimensions that might be initially considered low (e.g., interpreter's competence and established patterns of collaboration). Consequently, different types of interpreters have variations in the initial baseline and the future potentials the dimensions of provider-interpreter trust. In future studies, researchers should examine (a) how these baselines might be different from one another and their corresponding impacts, (b) how different types of interpreters can effectively gain trust through these dimensions, and (c) how providers can influence interpreters to adopt trust-enhancing behaviors. Providers' and interpreters' abilities to negotiate and nurture trust through these dimensions are critical to the success of bilingual health care.