REVIEW

Translation of interviews from a source language to a target language: examining issues in cross-cultural health care research Rasmieh Al-Amer, Lucie Ramjan, Paul Glew, Maram Darwish and Yenna Salamonson

Aims and objectives. To illuminate translation practice in cross-language interview in health care research and its impact on the construction of the data. Background. Globalisation and changing patterns of migration have created changes to the world’s demography; this has presented challenges for overarching social domains, specifically, in the health sector. Providing ethno-cultural health services is a timely and central facet in an ever-increasingly diverse world. Nursing and other health sectors employ cross-language research to provide knowledge and understanding of the needs of minority groups, which underpins cultural-sensitive care services. However, when cultural and linguistic differences exist, they pose unique complexities for cross-cultural health care research; particularly in qualitative research where narrative data are central for communication as most participants prefer to tell their story in their native language. Consequently, translation is often unavoidable in order to make a respondent’s narrative vivid and comprehensible, yet, there is no consensus about how researchers should address this vital issue. Design. An integrative literature review. Methods. PubMed and CINAHL databases were searched for relevant studies published before January 2014, and hand searched reference lists of studies were selected. Results. This review of cross-language health care studies highlighted three major themes, which identify factors often reported to affect the translation and production of data in cross-language research: (1) translation style; (2) translators; and (3) trustworthiness of the data. Conclusion. A plan detailing the translation process and analysis of health care data must be determined from the study outset to ensure credibility is maintained. A transparent and systematic approach in reporting the translation process not only enhances the integrity of the findings but also provides overall rigour and auditability. Relevance to clinical practice. It is important that minority groups have a voice in health care research which, if accurately translated, will enable nurses to improve culturally relevant care. Authors: Rasmieh Al-Amer, MSc, RN, PhD Candidate, School of Nursing and Midwifery, University of Western Sydney, Penrith, NSW; Lucie Ramjan, PhD, RN, BN, Senior Lecturer, School of Nursing and Midwifery, University of Western Sydney, Penrith, NSW; Paul Glew, PhD, RN, MN, Senior Lecturer, School of Nursing and Midwifery, University of Western Sydney, Penrith, NSW Maram Darwish, MD, General Surgery Resident, Department of General Surgery, Jordanian Royal Medical Services, Amman, Jor-

© 2014 John Wiley & Sons Ltd Journal of Clinical Nursing, 24, 1151–1162, doi: 10.1111/jocn.12681

What does this paper contribute to the wider global clinical community?

• Highlights •

different strategies used in translation Recommends a framework for optimal practice using standardised guidelines

dan; Yenna Salamonson, PhD, RN, Associate Professor, School of Nursing and Midwifery, University of Western Sydney, Penrith, NSW, Australia Correspondence: Rasmieh Al-Amer, PhD Candidate, School of Nursing and Midwifery, University of Western Sydney, Locked Bag 1797, Penrith, NSW 2751, Australia. Telephone: +61 2 46203355. E-mail: [email protected]

1151

R Al-Amer et al.

Key words: cross-cultural research, cultural issues, health, interviews, language barriers, qualitative approach, review, translation Accepted for publication: 15 July 2014

Introduction In conducting cross-language qualitative health care research, the idiosyncratic use of language is often unique, for language is a medium that conveys culture as a ‘set of distinctive spiritual, material, intellectual and emotional features of a society or a social group, that encompasses, in addition to art and literature, lifestyles, ways of living together, value systems, traditions and beliefs’(UNESCO 2001). Language conveys understandings and assumptions commonly held by a socio-cultural group; and often determines acceptance and participation in events, rituals and social phenomena (Temple et al. 2006, Lopez et al. 2008). Therefore, throughout this review, the term of cross-language research will be used; this is mainly because the focus of this review addresses the effect of translation from a source language to another language ‘a target language’ in health care research. In cross-language qualitative health research, translation remains invisible for most of the time, and it is often regarded as a straightforward process; however, translation is often problematic and this may threaten the integrity of the data unless it is reviewed in a systematic way (Temple 2002). Researchers, who undertake qualitative studies often recognise multiple realities and utilise different structural methods to access such realities, this includes the restrictions posed by linguistic and ethnic barriers when engaging participants from different ethno-cultural backgrounds (Temple et al. 2006). Therefore, rigorous translation in qualitative health care research is indispensable in the production of valid data in cross-language research in order to minimise discrepancies, offer insights into health beliefs and behaviours among different ethno-cultural groups and to construct knowledge on health care systems. There are clear standardised guidelines for the translation of research surveys from one language to another; however, when it comes to qualitative data translation, it usually differs in relation to the language discourse. Namely, the former involves the translation of professional source language to professional target language; whereas, the latter entails translating the participant’s narrative data, which could yield colloquial source language and nuances, to a professional style in a

1152

different language; this may impose difficulties when trying to reduce the gap between different languages. To illustrate; interviews, which allow participants to narrate their stories, have been extensively used within the health and humanities disciplines as a method for data collection (Randall et al. 2012). Narration allows participants to relate their experiences and stories in their own words; however sometimes these stories are constructed using clich
es, and expressions containing jargon or slang. Therefore, not paying attention to these idiosyncrasies and nuances, can lead to language discrepancies during the translation phase of qualitative research, particularly when interview data are collected in one language, whereas the results are presented in another language. It is worth stating that; the way words are often presented in published cross-language studies can give the impression that the participant’s speak proficiently in the target language, regardless of their mother tongue, which is in fact not the case (Temple et al. 2006).Thus, researchers must be cautious when conducting cross-language research, particularly in fields such as health care, as misrepresentation of meaning due to language or cultural differences can jeopardise the quality of translated data (Temple et al. 2006). These misinterpretations may unintentionally contribute to poor health care outcomes or even result in harm.

Objectives Strategies to manage the translation of interviews in cross-language qualitative health research have not been adequately identified. This paper is an integrative review of the literature that seeks to address the following questions in relation to translation in cross-language interviews: • What is the accepted translation practice in cross-language qualitative health care research? • What are the factors that influence the data in cross-language qualitative health care research? • What is the impact of translation on the study outcomes in cross-language qualitative health care research? The current study defined translation as a multifaceted activity that requires processing of information, and involves written text being transformed into another language in a written form (Squires 2009). © 2014 John Wiley & Sons Ltd Journal of Clinical Nursing, 24, 1151–1162

Review

Interview translation between languages

PubMed and CINAHL electronic databases were searched for articles published prior to first of January 2014. The authors developed the search syntaxes with the help of a qualified research librarian, taking into consideration a wide range of terms used in defining qualitative research, translation, and cross-cultural context (Appendix S1). Reference lists of the retrieved articles were searched by hand and two experts in the field were personally contacted on 5 February 2014 to identify datasets not captured by the search strategy. A subsequent assessment was performed on the full text of retrieved articles to identify articles dealing with the following concepts; ‘qualitative research’; ‘cross-language studies’; ‘addressed the translationprocess’;’ healthfocuses’; and ‘non–English speaking participants’. Only articles published in English were included. Studies identified for exclusion were: non-English studies, nonclinical studies, studies that used an interpreter as a translator in a real life translation event, and studies that did not report the translation process. Two reviewers independently screened titles and abstracts of all citations retrieved from the electronic search to identify potentially relevant studies. Another two reviewers verified the appraisal process, and discrepancies were resolved through discussion.

et al. 2001, Lee et al. 2007, Lopez et al. 2008, Lee & Weiss 2009), United Kingdom (Temple et al. 2006, Irvine et al. 2007), China (Twinn 1997), Japan (Irvine et al. 2007), Canada (Karwalajtys et al. 2010), Australia (Kwok & White 2011), and Spain (Esposito 2001). The sample sizes ranged from 6–60; and in two studies (Temple et al. 2006, Karwalajtys et al. 2010) the number of focus groups was reported instead of the number of participants. At least 14 different languages were represented across these studies, including: Spanish (Elderkin-Thompson et al. 2001, Esposito 2001, Lopez et al. 2008, Karwalajtys et al. 2010); Arabic (Temple et al. 2006, Karwalajtys et al. 2010); Cantonese (Twinn 1997, Temple et al. 2006, Karwalajtys et al. 2010, Kwok & White 2011); Somali (Karwalajtys et al. 2010) Dari (Karwalajtys et al. 2010); Mandarin (Temple et al. 2006, Lee & Weiss 2009, Kwok & White 2011); Polish (Temple et al. 2006); Gujarati (Temple et al. 2006), Bangladeshi, Sylheti, Turkish and Kurdish (Temple et al. 2006); Japanese (Irvine et al. 2007); and Korean (Lee et al. 2007). The health focus of the included studies were: (1) cancer screening; (2) beliefs, treatment and perceptions among the studied group (Twinn 1997, Lee et al. 2007, Lopez et al. 2008, Karwalajtys et al. 2010, Kwok & White 2011); (3) patients’ families experiences (Lee & Weiss 2009);(4) menopause (Esposito 2001); (5) nursing educators’ roles (Irvine et al. 2007); and (6) interpretation in the healthcare settings (Elderkin-Thompson et al. 2001, Temple et al. 2006).

Search outcome

Results

Methods Search methods

The databases search yielded 1356 potentially relevant records. Of these, 29 potentially eligible articles were considered. Additional hand searching identified two further potentially relevant citations. Thirty-two potentially eligible articles were considered, of which, 10 satisfied the inclusion criteria. Fifteen studies did not report or discuss the translation challenges, three studies translated the interview guide only (i.e. the proposed questions), two studies translated and interpreted sign language, one study used an interpreter, and one study was not health focused. In total, 22 studies were excluded; Table 1 provides an overview of these studies. Ten studies were included in the review (Table 2). Figure 1 illustrates the overall selection process.

Descriptive data synthesis Study characteristics of the 10 papers depicted in Table 2 included studies published in the period 1997–2011 that originated from the United States (Elderkin-Thompson © 2014 John Wiley & Sons Ltd Journal of Clinical Nursing, 24, 1151–1162

Theme identification For identifications of the themes, an open-coding approach was used. Firstly, the authors read each scholarly article independently and identified the relevant themes. Similarities and discrepancies were then identified to group main themes. Afterwards, coding was evaluated to comprise key themes related to translation. The authors reviewed every scholarly paper thoroughly until agreement was reached on the main concepts in relation to challenges facing the translation process. Of the 10 papers included, three themes emerged and were clustered into subcategories as summarised in Table 3. Theme 1: Translation Methods of translation. Different methods of translation have been identified across the reviewed articles. Temple et al. (2006), for instance, emphasised the importance of maintaining semantic and content equivalence to produce a

1153

R Al-Amer et al. Table 1 Characteristics of excluded studies Study identification

Country

Language

Reasons for exclusion

Bradby (2002) Craig et al.(2010) Hill et al.(2003) Jagosh and Boudreau (2009)

United Kingdom Nepal United States Canada

Punjabi Nepali and Tibetan Spanish French

Lazear et al.(2008)

United States

Lin et al.(2007)

Canada

Long et al.(2012) Merrell et al.(2005) Nachtigall et al.(2009) Poureslami et al.(2011)

United States Bangladesh United States Canada

Quach et al.(2012) Sheppard (2011)

United States United States

Hmong, English, Spanish, Haitian-Creole; Nuer; Somali; Khmer (Cambodian), Vietnamese, Arabic Chinese/Mandarin and Cantonese Spanish Bengali Spanish and English Persian, Chinese, Punjabi, and Spanish Mandarin, Cantonese, Spanish Sign language to English

Insufficient data (translation challenges not discussed) Insufficient data (translation challenges not discussed) Translated the interview guide not the interview The translation was for the interview guide not the interview Insufficient data (translation challenges not discussed)

Sneesby et al.(2011) Stone and West (2012)

Australia United Kingdom

Dinka and English British Sign language to English

Upvall et al.(2002) Wadensten et al.(2008)

Pakistan Sweden and China

Urdu Chinese, Swedish and English

Wallin and Ahlstr€ om (2010) Watson-Johnson et al.(2011) Stone et al. (2005) Woo and Twinn (2004) Yin and Twinn (2004) Wong and Poon (2010)

Sweden United States United Kingdom Hong Kong, China Hong Kong, China Canada

Somali English and Japanese English, Gujarati, and Punjabi Cantonese Cantonese Cantonese

translated text that is as similar as possible to the source text. Conversely, Esposito (2001) and Twinn (1997)stated that semantic equivalence alone is not enough to produce accurate findings, and different approaches are needed, including textual and content equivalence in translation along with semantic equivalence in the target language. Other studies applied forward–backward translation while maintaining the semantic equivalence, and considered this method of translation to produce the best possible rhetorical translation (Elderkin-Thompson et al. 2001, Irvine et al. 2007, Lopez et al. 2008, Karwalajtys et al. 2010, Kwok & White 2011). Nevertheless, Lopez et al. (2008) pointed out that a negative aspect of this approach was in it being a demanding, laborious and lengthy process. Additionally, this approach, which can be tedious and complex, was found to give word-for-word translation a considerable weight, and this can be problematic when there is no word equivalence in the target language (Twinn 1997). To

1154

Insufficient data (translation challenges not discussed) Insufficient data (translation challenges not discussed) Insufficient data (translation challenges not discussed) Insufficient data (translation challenges not discussed) Direct interpretation in health care system Insufficient data (translation challenges not discussed) The authors used translation/interpretation of Sign language Insufficient data (translation challenges not discussed) The authors used translation/interpretation of Sign language Insufficient data (translation challenges not discussed) The translation was for the interview guide not the interview Insufficient data (translation challenges not discussed) Insufficient data (translation challenges not discussed) Insufficient data (translation challenges not discussed) Insufficient data (translation challenges not discussed) Insufficient data (translation challenges not discussed) No health focus

illustrate; Lopez et al. (2008) reported that a term or an expression may be present in one language, yet not exist in another; thus, regional variations in words and interpretations are a concern, also word-for-word translation may jeopardise the readability of the data. Quality of translation. The quality of translation was reported in five of the reviewed studies; with cohesion, congruence, and clarity, being the overarching attributes of translation’s quality (Twinn 1997, Temple et al. 2006). Namely, cohesion which is the holding together of a sentence or a script using linguistic and verbal links (Halliday & Hasan 1976), may not always be feasible, as grammatical errors or awkward syntax from the source language can be challenging and require correction or alteration during translation (Esposito 2001, Temple et al. 2006). Therefore, it is highly recommended to recruit members of the participating local community to work with the researchers, as © 2014 John Wiley & Sons Ltd Journal of Clinical Nursing, 24, 1151–1162

21

United States

Elderkin-

Spain

Japan/United Kingdom

Canada

Australia

Esposito (2001)

Irvine et al. (2007)

© 2014 John Wiley & Sons Ltd Journal of Clinical Nursing, 24, 1151–1162

Karwalajtys et al. (2010)

Kwok & White (2011)

23

11*

27

40

size

identification

Thompson et al. (2001)

Sample

Country

Study

Breast cancer

Cervical cancer screening

Roles of that nurse educators

Menopause

interpretation

Medical

focus

Health

Table 2 Characteristics of included studies

Cantonese or Mandarin

and English English

Cantonese or Mandarin

first language

(five ethno linguisticgroups),

Focus groups in English

English

and in the participants’

Japanese/ English

Spanish

Spanish

language

Interview

English

English

English

language

Analysis

Somali. Dari, Spanish

Arabic, Cantonese,

Japanese/ English

Spanish

Spanish

language

Source

local syntax (b) limited depth of analysis (c) dialect differences

the study (b) using two different translators and bilingual reviewers

Not addressed

students

verbatim translation

Forward and backward

Forward translation

through discussion

(b) discrepancies were resolved

(a) translation verification

Not addressed

resolved through discussion

were

categories and concepts

(a) Disagreements in relation to

(c) dialect differences

translation services,

(b) High cost of professional

(a) Data collection delays,

the translators and the equivalence

faculty member

Field workers/ university

(b) discrepancies in translation

(b) discussions between translation/ semantic

service and a nursing

researchers (c) discrepancies

a nursing background,

(a) Translation verification

(a) The translator did not have

(a) popular idioms and

Differences in translation:

analysis from the beginning of

(a) Developing a plan for

Forward and backward

and semantic equivalence

Verbatim translation

transcripts to the tapes

in translations

Discrepancies

verification; (2) comparing

(1) translation

Forward and backward translation

challenges

Translation

control

Translation

method

Translation

University translation

service

A professional translation

assistant

-interpreters and a research

Nurse

personnel

Translation

Review Interview translation between languages

1155

1156

United States

United

Lee et al. (2007)

Lopez

© 2014 John Wiley & Sons Ltd Journal of Clinical Nursing, 24, 1151–1162

China

Kingdom

United

*Number of focus groups.

Twinn (1997)

et al. (2006)

Temple

States

United States

Lee & Weiss (2009)

et al. (2008)

Country

Study identification

Table 2 (continued)

6

10*

60

33

47

Sample size

Pap smears

interpreters

User views of

treatment side-effects

Prostate cancer

cancer screening

Breast and cervical

Hospitalised infants in ICU

Health focus

Cantonese

Cantonese

Kurdish, and Turkish

Sylheti, Bengali,

Gujarati, Arabic,

and Kurdish

Polish, Cantonese, Mandarin,

English & Chinese

English

Spanish

Korean

Mandarin

Interview language

Gujarati Indian, Bangladeshi,

Polish, Chinese,

English

English

Korean

Spanish

Chinese and English

Analysis language

Mandarin

Source language

(a) Interviews were not tape-recorded;

Translation verification

assistants

Independent research

researchers

Verbatim and semantic

equivalence

language (c) The influence research assistants independently

(d) dross rate

of the grammatical style on the analysis

equivalent word existed in the target

(a) Themes differences, (b) sometimes, no

presentation of possible tension

validity Contextual issues of

participants, (g) reliability and

limited contact with

has been conducted by two

Translation and transcription

sessions

Debriefing

(e) More appropriate replacements in some, (b) Translation verification

Semantic

(d) low education level of participants, slang terms (translation log).

Field

(c) translation process delays, for translation variations and

translation

research staff

(f) bilingual/ bicultural staff had

(b) lack of efficient translators, of words was created to control

backward verbatim

bilingual/ bicultural

(a) Dialect differences,

disagreements

(b) disagreements were resolved by discussion (a) Compendium

Transcription and translation

(a) The transcripts were checked

(b) incongruence in code assignments

Translation challenges

Translation control

Forward and

Forward translation

translation

Forward and backward

Translation method

Independent

A professional translator

Not addressed

Translation personnel

R Al-Amer et al.

Review

Interview translation between languages

1356 Citations identified from literature search PubMed and CINAHL electronic databases

1326 Citations excluded after deduplication based on screening titles and, or abstracts for

32 Potentially relevant citations for inclusion

2 Citations identified from reference lists

22 Citations excluded 15 Studies did not report or discuss translation challenges 3 Studies translated the interview guide only

10 Citations included for review

2 Studies were about translating and interpreting sign-language 1 direct interpretation in health care system 1 not health focus

Figure 1 Flow-chart of study selection process. Table 3 Themes and related articles Themes identified in the included studies Theme 1: Translation Methods of translation

Quality of translation Theme 2: Translators Number of Translators Translators’ qualifications and characteristics Theme 3: Trustworthiness of the data

Related articles

(Twinn 1997, Elderkin-Thompson et al. 2001, Esposito 2001, Temple et al. 2006, Irvine et al. 2007, Lee et al. 2007, Lopez et al. 2008, Karwalajtys et al. 2010, Kwok & White 2011) (Twinn 1997, Esposito 2001, Temple et al. 2006, Irvine et al. 2007, Lopez et al. 2008) (Twinn 1997, Esposito 2001, Woo & Twinn 2004, Temple et al. 2006, Lopez et al. 2008, Lee & Weiss 2009, Karwalajtys et al. 2010) (Twinn 1997, Elderkin-Thompson et al. 2001, Esposito 2001, Yin & Twinn 2004, Irvine et al. 2007, Lee et al. 2007, Lopez et al. 2008, Karwalajtys et al. 2010) (Twinn 1997, Elderkin-Thompson et al. 2001, Esposito 2001, Woo & Twinn 2004, Yin & Twinn 2004, Stone et al. 2005, Temple et al. 2006, Irvine et al. 2007, Lee et al. 2007, 2007, Lopez et al. 2008, Lee & Weiss 2009, Kwok & White 2011)

these individuals are crucial for verifying the translation (Karwalajtys et al. 2010). Congruence, which is the agreement and consistency of the text between the source and target languages corresponding to one another without changing the meaning (Irvine et al. 2007), is often difficult to achieve during translation due to a range of possible interpretations, particularly when cultural context differs, hence, inter-lingual translation is often advised (Temple et al. 2006, Irvine et al. 2007). Generally speaking, there may be no singular meaning or translation possible for an expression, particularly when © 2014 John Wiley & Sons Ltd Journal of Clinical Nursing, 24, 1151–1162

the semantics and syntactic structures of the target language do not match the source language (Twinn 1997, Lopez et al. 2008). Therefore, it is crucial to comprehend the ways in which the source language is tied into the specific culture and social contexts in order to clearly understand its nuances and differences to the target language (Temple et al. 2006). For instance, the verbatim quotes used by researchers in the writing up of their reports may reflect not only the participants’ voices, but also, in some way, the translators’ interpretation as well. In line with this debate, several studies showed that the use of a board of experts including those with linguistic, cultural, topic

1157

R Al-Amer et al.

and methodological proficiency is essential to verify the accuracy in translation (Esposito 2001, Lee et al. 2007, Karwalajtys et al. 2010). Theme 2: Translators In this review, the term ‘translator’ refers to:

possible translation errors (Temple et al. 2006, Lopez et al. 2008, Lee & Weiss 2009, Karwalajtys et al. 2010); moreover, translation by committee allows for meaningful judgment of the translation, limits the potential loss of word meaning and reduces differences in translation (Twinn 1997, Esposito 2001, Lee et al. 2007, Lopez et al. 2008).

an individual who transforms the research data (i.e. audio-taped interviews or documents) from one language to another. (Wong & Poon 2010, p. 152)

Number of translators. Up until now, there have not been clear guidelines regarding the appropriate number of translators needed to achieve good translation. Twinn (1997) reported that using one translator maximises the reliability of the data. The author also argued that multiple translators may create inaccurate data during translations. In particular, multiple dialects could emerge across translated texts, which in turn may greatly influence the analysis process and the findings of a study and might leave readers sceptical about the quality of the translated texts. This is illustrated in a sample of two translated versions of one Cantonese script from two independent translators in a study that explored the perceptions of Chinese women in relation to the influences on the uptake of Pap smears (Twinn 1997). Transcript I Even if the doctor is female, the check-up is very painful. Do try to avoid it. But sometimes. . . is scared. When it [last check translation, up] was a long time ago. . . then have to do it again Transcript II Even it comes with female doctors; it is very painful when having the check-up. So, it is avoided if it is possible. But some participant times it will be worrying if the check-up has not been done for a long time. So I just force myself to go for the check-up. (Twinn 1997, p. 421)

Although both transcripts seem similar, there is a subtle difference in the interpretation of the underlined statements, which imply that the discourse competence in combining grammatical form, local syntax and the cultural context is not congruent across the two translated scripts. However, it was argued that, when applying such an approach, there is the issue of language dialects in relation to the respondent’s accounts; the contexts of phrases are diverse within each language (Esposito 2001), and it is almost impossible for one translator to be fully versed with all language dialects. Thus, using multiple translators as a committee can not only solve this problem, but also limit

1158

Translators’ qualifications and characteristics. Translators weave documents with threads drawn from their own culture along with their linguistic abilities to produce the original text in another language; they are not merely bilingual persons, they also offer the researcher further insight and understanding of a studied phenomenon. Therefore, translators have often been viewed as authors of new texts. Eight of the studies included highlighted the importance of translators’ qualifications as their education and training influenced the quality of translations (Twinn 1997). Vigilant selection of translators and the whole translation practice often guarantees the integrity of the data (Twinn 1997, Esposito 2001, Lee et al. 2007, Lopez et al. 2008), for translators walk a fine line between the source and target languages. Three of the reviewed articles stated that translators’ demographic characteristics should be reported and it would be of great benefit if they match the participants’ characteristics for they can integrate the nuances of meaning in the language that are embedded in the cultural issue among and within different groups of participants (Twinn 1997, Esposito 2001, Irvine et al. 2007). This would add authenticity and contextual data that assign the meaning of the narrative in two different languages. Language proficiency is another core issue to be considered; this could be demonstrated through qualifications from an accredited translators’ association (i.e. the American Translators Association 2005) or another comparable regulation body (Esposito 2001). There is also a need for criteria to include the level of certification and experience of the translators to be set in advance to gain the maximum excellence of translation. Disconcertingly; the cost of professional translation services varies according to how uncommon is a specific language in a particular context, it could reach up to $12000 in some cases (Karwalajtys et al. 2010). It is acknowledged that securing the services of efficient translators can be complex and may take more time than researchers anticipate (Lopez et al. 2008); yet, that should not prohibit researchers from pursing rigorous translation when facing such difficulties. Another core argument in favour of translator’s qualifications reported that, in health research in particular, when one of the researchers takes over the translation, medical

© 2014 John Wiley & Sons Ltd Journal of Clinical Nursing, 24, 1151–1162

Review

terminology is not jeopardised (Twinn 1997, Lopez et al. 2008); specialised healthcare translators were used in these studies due to their familiarity with both common jargon and medical terminology (Twinn 1997, Lopez et al. 2008). Thus, a translator who is bilingual, and acquainted with the studied phenomenon could facilitate producing comparable versions of meaning in both languages. Theme 3: Trustworthiness of the data within cross-cultural studies Language differences can pose considerable barriers for cross-language research such as when the intended meaning of a participant’s words in one language cannot be accurately conveyed in another, which in turn can render the data useless (Esposito 2001). This raises important issues for the trustworthiness of the data as this must take into account methodological and epistemological concerns. Twinn (1997)argued that in using an ethnographic approach, managing translation difficulties is feasible as this research approach seeks to reveal collectivistic experiences within a specific culture. However, in phenomenological studies, where exact word usage is paramount, the use of translation may hinder the trustworthiness of the data as understanding the subjective, lived experiences and perspectives of participants is the cornerstone to maintaining the authenticity of the data. In addition, the author argued that in the analysis of cross-language interviews in phenomenology, studies should be held in the source language. Esposito (2001) acknowledged that immersion in the data in crosslanguage research was unachievable due to the language barrier, analysis and interpretation of the data depth being limited; however, the author argued that ensuring the integrity of the data could be achieved through different means, namely by using multiple translators, triangulation of the respondents, triangulation of different methods, and engaging independent bilingual reviewers. Whereas, Temple et al. (2006) maintained that the validity of their data could be achieved through conducting debriefing sessions after each interview. This involved researchers in the field highlighting the translation choices that were made during translation process and justifying their choices as well (Temple et al. 2006).Finally, several of the studies recommended using cultural, bilingual, linguistic and methodological experts to ensure satisfactory trustworthiness of the data (Esposito 2001, Irvine et al. 2007, Lee et al. 2007, Lopez et al. 2008, Lee & Weiss 2009, Kwok & White 2011). In summary, recommendations regarding the means of ensuring trustworthiness were: holding debriefing sessions (Temple et al. 2006), independently verifying the accuracy of the translation (Twinn 1997, Elderkin-Thompson et al. 2001, Lopez © 2014 John Wiley & Sons Ltd Journal of Clinical Nursing, 24, 1151–1162

Interview translation between languages

et al. 2008), and constructing independent categories to assess equivalence with the researcher’s categories.

Discussion Translation is of principal importance in cross-language research. Unfortunately, there are no standardised procedures for translation in qualitative healthcare research when involving non-English speaking participants. This review identifies some key issues in cross-language qualitative health care research and provides recommendations about the practice of translation. It clarifies crucial issues related to methodological translation choices made by researchers that might affect the results of cross-language studies. In addition, it illuminates a touch stone in the translation process and its related issues, including translation methods, translators, and maintaining trustworthiness in the data. This review may help the researcher(s) to be able to reflect on their competence in handling language issues in crosslanguage qualitative research. Reporting the reason behind choosing a target language instead of the source language in disseminating the results of multi-language studies can help reviewers and readers obtain a clearer picture of the translation (Esposito 2001, Lopez et al. 2008, B. Temple et al. 2006). Also, it could reduce readers’ assumptions about the effects on the data of the selected translation methods, translators and their credentials. Consequently, providing detailed information allows others to reflect on the identity of cross-cultural research team members, and determine their competence in the source and target language. In relation to phenomenological studies, it is paramount to analyse the data in the source language and then employ translation at the themes and categories levels so as to enhance the trustworthiness of the data(Twinn 1997). In addition, this review shows that translation methods can significantly impact the data in cross-language research. Therefore, translators should transfer the meaning of the words rather than re-writing them; so considering content equivalence in translation while maintaining semantic equivalence as this is necessary to produce informative text (Esposito 2001, Temple et al. 2006, Lopez et al. 2008). Therefore, attention should be given to the number of translators and their credentials and skills. Using one translator can maintain translation’s consistency and enhances the data analysis’ reliability (Twinn 1997). In addition to that, using a qualified translator who has a good command of the target language and the source language is indispensable (Lopez et al. 2008, Squires 2009). Translators who are acquainted with the culture and values of the studied group can help to contextualise the narrations. However, the

1159

R Al-Amer et al.

Scholarly reviews are known to contain some publication and selection bias. The reviewed articles for this paper were chosen after meeting predefined criteria. Limiting the review to English language articles might have limited findings in this matter.

ation of a variety of measures that take into account the source and target languages, planning and allocation of resources and translator’s numbers and credentials. Moreover, developing a plan for analysis from the outset of the study using multiple means of assuring credibility can provide researchers an opportunity to budget appropriately. It also recommends that data should be collected in the participants’ native language, and have a clear rationale and process for the translation of the qualitative data. Transparent reporting of the translation process will not only enhance the integrity of the findings but also the overall rigour and auditability. Several steps could facilitate the translation of the data and help in protecting the credibility of the data. It is critical to assemble a qualified team of bilingual and bicultural members who possess the required expertise in qualitative research, translators should be chosen based on clear criteria; as well as sharing the origin, language, culture and religious beliefs of the participants and they should have the ability to deal with the medical terminology. Semantic and content equivalence should be given a priority over word-for-word translation, specifically, when there is no equivalent meaning for words, as this takes into account the overall meaning in the translation process (Larkin et al. 2007). In addition to that, after each interview, debriefing sessions between the translators and the researchers should be held to check the accuracy and credibility of the data (Temple et al. 2006). Also, translators should be asked to show the researchers examples about the choices they made during the translation process. To enhance consistency and reduce variation in the translation process, this review suggests that one translator should be selected to avoid discrepancies in the translated data (Twinn 1997). This limits the risks of misinterpretation inherent in having multiple translators involved in validating data and conceptual equivalence (Squires 2009). To handle slang, jargon and regional differences, a trail of ‘atypical ‘words that are not documented in the dictionary should be created, and to manage dross data, nonverbal sounds and cues should be reported when quoting patients. Finally, as a framework for optimal practice has not yet been devised, this review calls for further standardisation of guidelines concerning the use of translation in qualitative health care research to enhance the quality of the data.

Conclusion

Relevance to clinical practice

This review shows that cross-language qualitative health care researchers face methodological challenges in conducting research in culturally diverse and different communities in terms of the translation process. Hence, these difficulties have to be recognised, and such difficulties require the consider-

The findings of this review highlight the importance of transparency in how language and translation of it is handled in cross-language research. As societies are increasingly made up of people with different ethnicities and cultures, hearing their voices in relation to health care issues is

challenge in dealing with qualitative data and the likelihood of translation errors can also justify consulting experts to assess translation differences and reduce errors. Translating interviews can involve considerable time, effort (Lopez et al. 2008)and finances (Esposito 2001). Therefore, the authors recommend planning for translation in cross-cultural health care research for effective allocation of financial and human resources. Needless to say, a higher degree of transparency and visibility in the conduct and writing up of the translation process in cross-language research is required to assure the integrity and reliability of the findings. This also mirrors the researchers’ competence in handling language issues between the researchers themselves and the studied group (Larkin et al. 2007).

Strengths This is the first review to discuss the translation challenges inherent in cross-cultural qualitative research with a focus on health issues among non–English speaking participants. The review recognises that translation is a time-consuming task for the research team. However, this investment promotes rigour and trustworthiness of the results. The strength of this review lies in its attempt to assess the available reported patterns of translation in cross-cultural qualitative health care research in a systematic way. It illustrates the differences in the practice of translation across languages using various research processes. We acknowledge that this review could only assess what was stated in the scholarly articles. These often did not discuss the translation challenges, as this issue has been largely overlooked in qualitative research. It is foreseeable that the lack of the latter might compromise one’s ability to independently assess the rigour and value of the findings for a given healthcare setting.

Limitations

1160

© 2014 John Wiley & Sons Ltd Journal of Clinical Nursing, 24, 1151–1162

Review

Interview translation between languages

important in promoting culturally sensitive nursing care. Understanding important issues in cross-language research such as how translation occurred, the numbers and qualifications of translators will add to trustworthiness and ensure that nurses are providing care in line with best evidence.

or revising it critically for important intellectual content, and (3) final approval of the version to be published.

Funding No funding was necessary for the undertaking of this integrative literature review.

Acknowledgements The authors thank Dr. Sue Randall and Ms. Marian Martin for their valuable comments.

Conflict of interest No conflict of interest has been declared by the authors.

Disclosure The authors have confirmed that all authors meet the ICMJE criteria for authorship credit (www.icmje.org/ethical_1author.html), as follows: (1) substantial contributions to conception and design of, or acquisition of data or analysis and interpretation of data, (2) drafting the article

Supporting information Additional Supporting information may be found in the online version of this article: Appendix S1. PubMed Search syntax. Appendix S2. CINAHL Search Syntax.

References American Translators Association (2005) Minimum qualifications for translator certification. Available at: http://www. atanet.org/ (accessed 1 July 2005). Bradby H (2002) Translating culture and language: a research note on multilingual settings. Sociology of Health & Illness 24, 842–855. Craig SR, Chase L & Lama TN (2010) Taking the MINI to Mustang, Nepal: methodological and epistemological translations of an illness narrative interview tool. Anthropology and Medicine 17, 1–26. Elderkin-Thompson V, Silver RC & Waitzkin H (2001) When nurses double as interpreters: a study of Spanish-speaking patients in a US primary care setting. Social Science & Medicine 52, 1343–1358. Esposito N (2001) From meaning to meaning: the influence of translation techniques on non-English focus group research. Qualitative Health Research 11, 568–579. Halliday M & Hasan R (1976) “Cohesion in” English. Longman, London. Hill NE, Ramirez C & Dumka LE (2003) Early adolescents’ career aspirations: a qualitative study of perceived barriers and family support among low-income, ethnically diverse adolescents. Journal of Family Issues 24, 934–959. Irvine FE, Lloyd D, Jones PR, Allsup DM, Kakehashi C, Ogi A & Okuyama M

© 2014 John Wiley & Sons Ltd Journal of Clinical Nursing, 24, 1151–1162

(2007) Lost in translation? Undertaking transcultural qualitative research Nurse Researcher 14, 46–59. Jagosh J & Boudreau D (2009) Lost and found in translation: an ecological approach to bilingual research methodology. International Journal of Qualitative Methods 8, 102–114. Karwalajtys TL, Redwood-Campbell LJ, Fowler NC, Lohfeld LH, Howard M, Kaczorowski JA & Lytwyn A (2010) Conducting qualitative research on cervical cancer screening among diverse groups of immigrant women: research reflections: challenges and solutions. Canadian Family Physician 56, e130–e135. Kwok C & White K (2011) Cultural and linguistic isolation: the breast cancer experience of Chinese-Australian women – a qualitative study. Contemporary Nurse: A Journal for the Australian Nursing Profession 39, 85–94. Larkin PJ, Dierckx de Casterle B & Schotsmans P (2007) Multilingual translation issues in qualitative research: reflections on a metaphorical process. Qualitative Health Research, 17, 468–476. Lazear KJ, Pires SA, Isaacs MR, Chaulk P & Huang L (2008) Depression among low-income women of color: qualitative findings from cross-cultural focus groups. Journal of Immigrant & Minority Health 10, 127–133.

Lee S & Weiss SJ (2009) When east meets west: intensive care unit experiences among first-generation Chinese American parents. Journal of Nursing Scholarship, 41, 268–275. Lee EE, Tripp-Reimer T, Miller AM, Sadler GR & Lee S-Y (2007) Korean American Women’s Beliefs about Breast and Cervical Cancer and Associated Symbolic Meanings. Paper presented at the Oncology nursing forum. Lin C, Williams RM, Shannon HS & Wilkins S (2007) Work organization and its effects on the health of Chinese workers with English as a second language: a qualitative approach. Work 28, 379–390. Long JM, Sowell R, Bairan A, Holtz C, Curtis AB & Fogarty KJ (2012) Exploration of commonalities and variations in health beliefs across four Latino subgroups using focus group methodology: implications in care for Latinos with Type 2 diabetes. Journal of Cultural Diversity 19, 133–142. Lopez GI, Figueroa M, Connor SE & Maliski SL (2008) Translation barriers in conducting qualitative research with Spanish speakers. Qualitative Health Research 18, 1729–1737. Merrell J, Kinsella F, Murphy F, Philpin S & Ali A (2005) Support needs of carers of dependent adults from a Bangladeshi community. Journal of Advanced Nursing 51, 549–557.

1161

R Al-Amer et al. Nachtigall RD, Castrillo M, Shah N, Turner D, Harrington J & Jackson R (2009) The challenge of providing infertility services to a low-income immigrant Latino population. Fertility and Sterility 92, 116–123. Poureslami I, Rootman I, Doyle-Waters M, Nimmon L & FitzGerald J (2011) Health literacy, language, and ethnicity-related factors in newcomer asthma patients to Canada: a qualitative study. Journal of Immigrant & Minority Health 13, 315–322. Quach T, Nuru-Jeter A, Morris P, Allen L, Shema SJ, Winters JK, Le GM, Gomez SL (2012) Experiences and perceptions of medical discrimination among a multiethnic sample of breast cancer patients in the Greater San Francisco Bay Area, California. Am J Public Health, 102(5), 1027–1034. doi: 10. 2105/ajph.2011.300554 Randall E, Crooks VA & Goldsmith LJ (2012) In search of attachment: a qualitative study of chronically ill women transitioning between family physicians in rural Ontario, Canada. BioMed Central Family Practice 13, 125. Sheppard K (2011) Using American sign language interpreters to facilitate research among deaf adults: lessons learned. Journal of Transcultural Nursing 22, 129–134. Sneesby L, Satchell R, Good P & van der Riet P (2011) Death and dying in Australia: perceptions of a Sudanese community. Journal of Advanced Nursing, 67, 2696–2702.

Squires A (2009) Methodological challenges in cross-language qualitative research: a research review. International Journal of Nursing Studies 46, 277–287. Stone C & West D (2012) Translation, representation and the Deaf ‘voice’. Qualitative Research 12, 645–665. Stone M, Pound E, Pancholi A, Farooqi A & Khunti K (2005) Empowering patients with diabetes: a qualitative primary care study focusing on South Asians in Leicester, UK. Family Practice 22, 647–652. Temple B (2002) Crossed wires: interpreters, translators, and bilingual workers in cross-language research. Qualitative Health Research 12, 844–854. Temple B, Edwards R & Alexander C (2006) Grasping at context: Cross language qualitative research as secondary qualitative data analysis. Forum: Qualitative Social Research, 7(4). Retrieved December 10, 2008, from http://www.qualitative-research.net/ index.php/fqs/article/view/176/393 Twinn S (1997) An exploratory study examining the influence of translation on the validity and reliability of qualitative data in nursing research. Journal of Advanced Nursing 26, 418–423. UNESCO (2001, November 2). Universal Declaration on Cultural Diversity. Adopted by the 31st Session of the General Conference of UNESCO. Available at: http://unesdoc.unesco. org/images/0012/001271/127160m.pdf (accessed 12 December 2013).

Upvall MJ, Kanji Z, Jaffer S, Khowaja K, Barolia R & Amarsi Y (2002) Defining the cultural milieu for implementing faculty practice in Pakistan. International Journal of Nursing Practice 8, 315–323. Wadensten B, Wenneberg S, Sil
en M & Ahlstr€ om G (2008) A cross-cultural comparison of nurses’ ethical concerns. Nursing Ethics 15, 745–760. Wallin A & Ahlstr€ om G (2010) From diagnosis to health: a cross-cultural interview study with immigrants from Somalia. Scandinavian Journal of Caring Sciences, 24, 357–365. Watson-Johnson LC, DeGroff A, Steele CB, Revels M, Smith JL, Justen E, Barron-Simpson R, Sanders L, Richardson LC (2011) Mammography adher- ence: A qualitative study. Journal of Womens Health 20(12), 1887–1894. doi: http://dx.doi.org/10. 1089/jwh.2010.2724 Wong JP-H & Poon MK-L (2010) Bringing translation out of the shadows: translation as an issue of methodological significance in cross-cultural qualitative research. Journal of Transcultural Nursing 21, 151–158. Woo H & Twinn S (2004) Health needs of Hong Kong Chinese pregnant adolescents. Journal of Advanced Nursing 45, 595–602. Yin LK & Twinn S (2004) The effect of childhood cancer on Hong Kong Chinese families at different stages of the disease. Cancer Nursing 27, 17–24.

The Journal of Clinical Nursing (JCN) is an international, peer reviewed journal that aims to promote a high standard of clinically related scholarship which supports the practice and discipline of nursing. For further information and full author guidelines, please visit JCN on the Wiley Online Library website: http:// wileyonlinelibrary.com/journal/jocn

Reasons to submit your paper to JCN: High-impact forum: one of the world’s most cited nursing journals, with an impact factor of 1
316 – ranked 21/101 (Nursing (Social Science)) and 25/103 Nursing (Science) in the 2012 Journal Citation Reportsâ (Thomson Reuters, 2012). One of the most read nursing journals in the world: over 1
9 million full text accesses in 2011 and accessible in over 8000 libraries worldwide (including over 3500 in developing countries with free or low cost access). Early View: fully citable online publication ahead of inclusion in an issue. Fast and easy online submission: online submission at http://mc.manuscriptcentral.com/jcnur. Positive publishing experience: rapid double-blind peer review with constructive feedback. Online Open: the option to make your article freely and openly accessible to non-subscribers upon publication in Wiley Online Library, as well as the option to deposit the article in your preferred archive.

1162

© 2014 John Wiley & Sons Ltd Journal of Clinical Nursing, 24, 1151–1162

This document is a scanned copy of a printed document. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material.