Department of Language and Culture Studies, University of Modena and Reggio Emilia, Modena, Italy
Received: 17 April 2012
Accepted: 13 July 2012
This article presents an analysis of 55 conversations between doctors and patients, who did not share a common language, in the Emilia Romagna region of Italy. Interpreters took part in each of these conversations to interpret Italian and Arabic. The findings suggest that whether or not the patient’s voice, wishes and feelings are heard by the doctor is largely dependent on the interpreter’s actions. In some instances, the interpreter may exclude the patient or the doctor from relevant healthcare information. Conversely, the emotional rapport between the patient and the doctor may improve if the interpreter conveys implicit content to the doctor, thus creating opportunities for him or her to respond. These findings indicate that interpreters are expected not only to interpret what is said but also to act as mediators to enable both parties to communicate effectively.
Italy, Northern African migrants, office visits, personal communication, women’s health
Studies of interpreting from the field of applied linguistics have shown that interpreters are active participants in the interaction between the other individuals in the conversation (Davidson, 2000, 2001, 2002; Bolden, 2000; Mason, 2001; Wadensjo¨, 1998).
Wadensjo¨ (1998) has suggested that the interpreter’s activities can be seen as primarily oriented towards the joint construction of meaning by interpreting both what is said and what is implied. Interpreters therefore need to consider the meanings and purposes that are achieved through a conversation, and thus they play a double role, as they both interpret and coordinate communication. For this reason, interpreting may be understood as a form of mediation. The interpreter– mediator is the only participant in the interaction who is able to understand everything that the others in the conversation say. Therefore he or she can define the context of the encounter, draw attention to the production of shared topics, and manage misunderstandings. This paper focuses on this dual role as it emerged from analysis of conversations between interpreters, patients and doctors in an Italian healthcare setting.
Mediation is a strategy that is used to enable individuals who are in conflict to resolve their differences and modify their relationship (Carnevale and Pruitt, 1992). The intervention of a third party, namely the mediator, does not seek to establish who is right or wrong (Bush and Folger, 1994; Mulcahy, 2001). Rather, the mediator intervenes as a provider of opportunities to talk, inducing both parties to introduce and deal with particular issues (Fisher and Shapiro, 2005; Katz Jameson et al, 2006; Schulz, 2006) and construct acceptable solutions (Picard and Melchin, 2007; Winslade, 2006; Winslade and Monk, 2000). However, doubts have been raised about the effectiveness of mediation in promoting balanced power relationships. Welsh and Coleman (2002, pp. 345–6) have observed that mediators may ‘become de facto agents of the status quo invested in maintaining the stability of the current social system and stopping the conflict before it moves beyond the affected institutions.’ Therefore an important question concerns the effectiveness of mediation in challenging the status quo and in empowering the less influential (Gwartney et al, 2002), for example, in medical systems (Bolden, 2000; Davidson 2000, 2001).
Interpreter-mediated interaction is a triadic interaction involving an interpreter as the third party in a communication process between individuals who do not share a common language. Interpreter-mediated interactions help to create an effective form of intercultural communication, giving voice to cultural diversity in the interaction (Baraldi, 2006). Situations that require interpreters are increasingly common in western medical systems where healthcare providers encounter migrant patients (Angelelli, 2004; Baraldi and Gavioli, 2007; Cambridge, 1999; Bu¨hrig and Meyer, 2004; Po¨chhacker and Kadric, 1999; Tebble, 1999). The development of rapport between doctors and patients is considered very important for the successful outcome of treatment and care (Barry et al, 2001; Charles et al, 1999; Epstein et al, 2005; Heritage and Maynard, 2005, 2006; Mead and Bower, 2000; Zandbelt et al, 2005, 2006). The patient’s feelings and wishes and the doctor’s affective involvement in the interaction are considered of primary importance (Barry et al, 2001) in helping the patient to comply with treatment (Kiesler and Auerbach, 2003; Mangione- Smith et al, 2003; Robinson and Heritage, 2005; Stivers, 2002). In this respect, healthcare providers are invited to observe illness through the patient’s lens and to ‘treat the patient, rather than just the disease’ (Heritage and Maynard, 2006, p. 355).
These western expectations mean that migrant patients may encounter severe difficulties in presenting their case histories, concerns and worries. The difficulties involved in handling emotional expressions in interpreter-mediated patient–doctor interactions have been reported in several studies (Bolden, 2000; Cambridge, 1999; Davidson, 2000; Hsieh, 2010; Po¨chhacker and Kadric, 1999). Instead of relaying the patient’s concerns in full to the doctor, interpreters tend to summarise what the patient has said, focusing on medical problems and treatments. Consequently, emotional expressions may be overlooked or omitted. Thus interpreters become gatekeepers, controlling what is passed between doctor and patient, and fuelling asymmetrical power relationships between the two parties (Davidson, 2000, 2001; Bu¨hrig and Meyer, 2004).
Recent data (from 2011) indicate that there are 89 346 immigrants (12.7% of the resident population) in the Modena district, and 69 060 immigrants in the Reggio Emilia district (13% of the resident population). In both cases, the majority of migrants originate from Morocco and Albania. Modena also has a population of Tunisian migrants, and Reggio Emilia has quite large Indian and Chinese communities.
The major driver for institutional change in healthcare systems is the requirement to provide appropriate services for migrant women who are pregnant or who already have children, When using healthcare services, migrant women encounter different and unfamiliar cultural constructions of health, disease, therapy, sexuality and motherhood which their husbands and fathers may not understand or approve of, and which may therefore be a source of conflict. As a result, healthcare providers are being encouraged to reorganise their services in innovative and creative ways based on migrant-friendly models, which may potentially be extended to all patients. For example, intercultural mediators have been appointed by the General Hospital Board and Local Health Board in Modena to help in reception, obstetrics, nursery, paediatrics, gynaecology, neonatology and the family advice bureau. Reggio Emilia Local Health Board uses intercultural mediators in the outpatients’ departments and specialised units for the care of women and children.
This paper is based on a project undertaken in Emilia Romagna, Italy, entitled ‘Interlinguistic and intercultural communication: analysis of interpretation as a form of mediation for the bilingual dialogue between foreign citizens and institutions.’ The aims of this project were as follows:
1. to create a method of analysing healthcare practices
2. to draw up specific criteria to identify good practices
3. to develop criteria for selecting models of good practice for evaluative analysis, pointing out the indicators of effectiveness concerning their functionality, correspondence to patients’ needs, and opportunities for access
4. to develop instruments to monitor these models, with reference to the reduction of inequalities and barriers
5. to develop guidelines to be used in personnel training.
The project was reviewed and approved by a Management Coordination Committee consisting of the research coordinator and the coordinators of the healthcare service. The Committee was in charge of decision making on knowledge protection, ethical and legal issues.
Written information about the project was provided for doctors, interpreters and patients. This included details of the aim of the project, a request for permission to audio-record each conversation, and details of how the results would be used. Written permission was requested from patients, interpreters and doctors. The privacy of participantswas preserved according to the Italian Data Protection Act 675 (31 December 1996).
Before each recording, the participants were reminded about the aims of the research, what taking part involved, and their right to withdraw. Assurances about anonymity were important to avoid anyone being blamed or stigmatised as a result of taking part in the research. Removing or changing names was not always enough to ensure anonymity. In such cases the ethical need for anonymity was prioritised over scientific considerations about documentation. These ethical considerations are not and cannot possibly be exhaustive. Ethical research practice requires continuous reflexivity and coping with ethical problems as they arise. This requires dialogue at two levels, between researchers as a means of collectively sharing experience, and between researchers and participants in the ongoing research project.
Four doctors, four nurses and four interpreters took part in the research. All of the healthcare professionals were of Italian origin and were native speakers of Italian. The interpreter originated from Tunisia and Jordan, spoke Arabic, and had been living in Italy for at least 6 years at the time of this research. Other interpreters who took part in the project originated from Ghana and Nigeria and spoke English. All of these professionals had undergone formal training to enable them to work as intercultural mediators. Resolution 265 of the Regional Government of Emilia- Romagna (2005) establishes training standards for intercultural mediators. In order to be qualified as intercultural mediators in public services, it is necessary to follow courses organised by training centres approved by the regional authorities. The minimum duration of such a training course is 200 hours, including at least 40 hours of traineeship. The interpreters involved in this research were professional mediators who were officially recognised by theEmilia Romagna regional authorities. Although the research involved patients who spoke various languages other than Italian, this paper is based on Arabic–Italian interactions. Thus the 55 patients who were involved in the interactions that we discuss in this paper are Arabic speakers, most of them fromMorocco, Tunisia and Egypt.
As the mediation service is used predominantly in the areas of nursery, infant care and women’s health, most of the patients involved in the research were women. With regard to the present paper, 51 patients (92.72%) were women, and only 5 patients (7.28%) were men.
The data consisted of 55 conversations, in Arabic and Italian, in two public healthcare services, namely the Centro per la Salute Delle Famiglie Straniere (Healthcare Support Centre for Foreign Families) in Reggio Emilia, and the Consultorio (Local Centre for Health and Social Services) in Vignola (Province of Modena). In most cases (47 cases, 85.4%) these conversations concerned issues related to obstetrics, nursery, paediatrics, gynaecology and neonatology. The conversations took place in general hospital settings, between Italian doctors who did not speak Arabic, Arabic-speaking patients who did not speak Italian, and interpreters, acting as intercultural mediators, who spoke both languages and who facilitated communication between the other two parties. Each conversation was audio-recorded, transcribed and analysed using two socio-linguistic methods of analysis. The first was based on conversation analysis (CA) and the ways in which participants in a conversation talk according to a coordinated system of turn-taking (Sacks et al, 1974). CA looks at the mechanisms that invite participants in a conversation to talk, and at the acceptance or rejection of their contributions in subsequent interaction (Schegloff, 1980; Pomerantz, 1984). CA suggests that responses to contributions are very important in explaining how each participant reacts and how they achieve an understanding of what is going on (Mason, 2006, p. 364). What we gained from CAwas an understanding of the system through which the speakers achieved their understanding.
The second analytical approach was derived from studies on intercultural communication (e.g.Gudykunst, 2005; Samovar and Porter, 1997; Ting-Toomey and Kurogi, 1998), as we considered interpreter-mediated interaction to be a form of intercultural communication. We analysed the use of language and language diversity from the perspective of intercultural communication, observing whether the features of bilingual or multilingual talk in the interaction reproduced and/or tackled particular cultural aspects of the interactions.
In the following section we discuss three types of interaction, namely those which exclude or inhibit the patient from communicating with the doctor, those which include the patient but not the doctor, and three-way interactions that include the patient, the doctor and the interpreter/intercultural mediator. In each instance we present extracts from our data using transcription conventions set out by Jefferson (2004) (see Table 1). In each extract the following code applies: D = Italian healthcare provider of the institution; IM = Arabic-speaking interpreter/intercultural mediator; P = Arabic-speaking patient.
The most common types of exclusion were reduced renditions, a term that refers to situations in which the IM cut out some of the patient’s and doctor’s talk (Wadensjo¨, 1998). Reduced renditions usually occurred when the interpreter was passing information from the patient to the doctor, and vice versa. In excerpt CS5 (see Appendix) the patient asks two questions (turns 3 and 5) to try to find out whether the doctor is going to treat her leg in the office. Instead of translating the patient’s questions, the interpreter responds directly, hindering patient–doctor communication.
In turn 2, the interpreter produces a reduced rendition of the doctor’s contribution in the previous turn (‘she is giving you the ointment’), leaving out the information about the drug not being available at the doctor’s office. This reduced rendition indicates a doctor-centred culture (Barry et al, 2001) in which the patient is expected to follow instructions and the doctor does not have to account for his or her decisions. In this excerpt, reduced rendition creates some concern for the patient, who is told that the doctor is going to treat her leg with the ointment and to buy the drug at the pharmacy in the same turn (turn 3). The interpreter uses the word ‘give’ instead of ‘prescribe’, so the patient understands ‘giving you’ as ‘treating you with the ointment.’
Consequently, the patient is uncertain about the doctor’s intentions. Moreover, she does not know that the drug is not available because the interpreter did not tell her this in turn 2, and she has no reason to believe that the doctor will not treat her leg in the office. In order to solve this problem, the patient initiates a repair sequence in turn 3: ‘Is she giving it to me?’ The repair is completed by the IM, who responds to the patient without translating the request to the doctor: ‘They do not have it’ (turn 4).
The doctor has not said anything to justify the new piece of information which arrives too late in the interaction, and the patient understands it as a way to cover the fact that the doctor does not want to treat her. The patient’s reiteration of the question in a different format (turn 5) is evidence of her dissatisfaction. The IM’s actions create distance between the two parties, making the creation of common ground between them very unlikely. The IM notices the patient’s increasing dissatisfaction and tries to mitigate it.However, she does not interpret the question to the doctor but provides a direct response (turn 6), thus increasing the distance between them. We can imagine the attitude that the patient will have towards medical prescriptions if she believes that the doctor is not interested in her health.
Reduced renditions speed up encounters between doctors and patients and so appear to support the functionality of the healthcare system. However, we may ask what kind of system functionality is supported by these actions. Recent research (Leanza et al, 2010; Schouten et al, 2007) has confirmed the efficacy of reduced renditions in keeping the interaction coherent by, for example, censoring a part of the medical discourse that might not be comprehensible to the patient or a part of the patient’s discourse that might not be relevant to healthcare treatment. However, the same research shows that reduced renditions also hinder the trust-building process between patient and healthcare provider because they create more distance between the principal participants. Thus reduced renditions, interruptions and substitutions pose a threat to the therapeutic process and, paradoxically, compromise the core values (i.e. selfdeterminism and informed decision making) of the western medical system (Hsieh, 2010).
The IM’s actions could promote the patient’s active involvement in the medical encounter by encouraging their self-expression, and giving voice to their concerns, doubts, needs and requests. This was achieved through two types of interactional practices, namely dyadic (patient–interpreter) and triadic (patient–interpreter– doctor).
In dyadic interactions, the IM supported the voice of the patient through backchannelling (Schegloff, 1982; Schiffrin, 1999), using short conversational markers such as acknowledgement tokens (e.g. ‘Sı`’, ‘Yeah’, ‘OK’), which convey that the stated information has been received, continuers (e.g. ‘hmmm’, ‘mmh’), which maintain the flow of conversation and the desire of the current speaker to continue his or her turn, and echoing, which serves as a floor-yielding device that provides feedback to demonstrate that attention has been paid to what is being said. These conversational markers provide space for the speakerin- turn to go on, adding or completing their contribution with the help of the IM. Thus the IM and the speaker are both involved in the interpreting process.
In excerpt CO1 (see Appendix), the IM expresses her attentiveness and understanding through feedback tokens (‘Ah’, ‘mmh’, ‘Ah, I understand you’). In turn 122, the IM encourages the patient to express her concerns, making her participation relevant as a person with specific needs and worries rather than a generic sick person who is expected to provide a list of current symptoms. In this excerpt, the IM systematically encourages the patient to express her doubts about the therapy, thus promoting her participation in the conversation. Being empowered as an active participant, the patient is confident enough to finally advance a request for clarification (turn 123).
In triadic interactions, all three participants (patient, doctor and IM) are involved in the conversation. The IM’s actions bring the doctor into the conversation through the use of formulations of the patient’s contributions. Formulations are summaries or the gist of what someone has said (Heritage, 1985). They provide directions for subsequent turns by inviting responses in so far as they ‘advance the prior report by finding a point in the prior utterance and thus shifting its focus, redeveloping its gist, making something explicit that was previously implicit in the prior utterance, or by making inferences about its presuppositions or implications’ (Heritage, 1985, p. 104). In our study, the IM’s formulations consisted of interpretations which followed patient–IM dyadic sequences, but with adaptations to accommodate the doctor. Formulations enabled IMs to build, expand and recreate the meanings of previous dyadic sequences according to presuppositions and orientations for which they were responsible. Formulations are not word-for-word interpretations of contributions in previous dyadic sequences, but rather they rely on the IM’s discursive initiative and willingness to create common ground between the patient and the doctor. Thus the IM acts as both interpreter and mediator.
Specifically, formulations are conversational resources available to the IM in order firstly to provide an interpretation that highlights content from previous sequences, secondly, to make explicit what was thought in previous turns of talk to be implicit or unclear, and thirdly, to propose inferences about presuppositions or implications of the participants’ contributions (Baraldi and Gavioli, 2008). Formulations are informational when they elicit an explanation from the doctor which the patient is somehow inhibited from requesting, and affective when they bring the patient’s emotions, doubts and concerns into the conversation.
Excerpt CO23 (see Appendix) is an example of an informational formulation. In the course of the examination of a pregnant woman, it becomes apparent that the fetus is not yet in the appropriate position. In turn 59 the doctor reassures the patient about this issue, and in turn 60 the IM offers reassurance and further suggestions to the patient. The doctor is then re-involved in the interaction in turn 63, through an informational formulation which is introduced by the IM to obtain therapeutic recommendations for the patient’s benefit. In CO23, the informational formulation is an initiative motivated by the IM’s interpretation of the patient’s situation.
Affective formulations may be understood as discursive initiatives undertaken by the IM to give voice to the patient’s emotions, which in most cases manifest themselves implicitly. Patients rarely talk about their emotions directly and without prompting. Instead they provide clues about their feelings, thus providing health professionals and IMs with ‘potential emphatic opportunities’ (Beach and Dixson, 2001, p. 39). Affective formulations focus on the emotional point of the patient’s utterances, giving the doctor the chance to share and become involved in the affective dimension of the interaction. In this way the doctor is made aware of the patient’s concerns, and the patient assumes an identity that goes beyond the generic social role of being sick.
In excerpt CO11 (see Appendix), the patient reports a delay in her menstrual period, but mitigates the relevance of this information by assuming that her period will start within the next few days. Through affective formulations, the IM brings the patient’s emotions, which have remained implicit up to that moment, to the fore, making them a topic for communication and concern. The IM’s formulation, ‘She’s a bit worried’ (in turn 65), is affective because, while making current symptoms available to the doctor, it highlights the patient’s emotional situation which could otherwise have gone unnoticed in previous turns. The IM’s formulation of affective understanding involves the doctor in the affective exchange, and promotes a shift from a dyadic to a triadic interaction.
The IM’s affective formulation offers the doctor the opportunity to tune into the emotional status of the patient, and to reassure her as necessary. Affective formulations are inclusive because, while highlighting the emotions of the patient, they also involve the doctor in the formation of affective relations. By producing an affective formulation, the IM develops and emphasises an implicit emotional expression as a basis for subsequent interaction. Affective formulation reveals the IM not as a neutral conduit, but rather as an active mediator of the preceding talk. In particular, the IM’s active participation concerns the patient’s implicit, difficult and embarrassed emotional expressions, providing a way to include such expressions in the triadic sequence, and for it to be treated in a patient-centred interaction involving the doctor (Baraldi and Gavioli, 2007).
The dual function of interpreter and mediator can make positive contributions to patient-centred care and treatment. This paper has focused on how these two functions are intertwined and how they affect doctor–patient communication. When the interpreter acts effectively as a mediator, otherwise hidden factors, such as the patient’s emotional expressions, can be relayed to the doctor, thus creating opportunities for him or her to respond. If the interpreter does not act in this way, the patient’s feelings may be neglected.
Analysis of emergency visits in two large paediatric departments in the USA (Flores et al, 2012) suggests an association between the number of previous hours of interpreter training and error numbers, types and potential consequences in English–Spanish mediated interactions. Well-trained professional interpreters demonstrated a significantly lower likelihood of errors than ad-hoc interpreters such as family members or other hospital staff. The study suggests that training for interpreters might have a major impact on reducing interpreter errors and their consequences in healthcare, while improving quality of care and patient safety. Although we agree with the importance of professional training for interpreters, we also argue that the complexity of the mediator’s task needs to be acknowledged. According to the literature on dialogue interpreting (Angelelli, 2004; Baker, 2006; Bolden, 2000; Davidson, 2000, 2001; Mason, 1999, 2006; Wadensjo¨, 1998), in triadic interactions the interpreters are the only participants who can effectively understand all of the content and the intentions of the other participants. This implies that interpreters are never neutral conduits, and that errors are not the only issue. Interpreter–mediators necessarily coordinate the contingent and changeable construction of a difference between cultural presuppositions and the corresponding distribution of communicative resources, through their translation activity in intercultural contexts.
Our data suggest that the dual roles of interpreter– mediators are crucial in enabling patients to make their voices and their wishes heard during medical encounters. We have observed how reduced renditions may exclude the patient or the doctor from the conversation and from relevant healthcare information. On the other hand, we have seen how the use of formulations improves the emotional rapport between patient and doctor, taking the medical encounter well beyond a mere exchange based on standardised roles. In our analysis of patient–doctor conversations, interpreter- mediators contribute to dialogue management in at least two ways, first, as responders, affiliating with the patient in dyadic interactions, and secondly, as coordinators, affiliating with the patient and then involving the doctor. As responders, interpretermediators have an opportunity to check and echo the patient’s perceptions and emotions, actively listen to and appreciate their expressions, provide positive feedback, and express personal concern for them. In a more complex interaction, this can form the basis of an important step in patient care.