Trained medical interpreters in a face-to-face clinical setting for patients with low proficiency in the local language

  • Protocol
  • Intervention

Authors


Abstract

This is the protocol for a review and there is no abstract. The objectives are as follows:

To assess the effects of trained medical interpreters in face-to-face clinical settings for patients with low proficiency in the local language on:

  1. the quality of communication between patient and provider (as a precondition for the utilisation of professional knowledge to provide quality health care);

  2. the quality of health care, and health outcomes; and

  3. the cost benefit, cost effectiveness and cost utility of interventions by trained medical interpreters.

Background

Description of the condition

Population mobility is a global phenomenon, with about 214 million people, 3.1% of the world's population, living outside their country of birth (UN 2008). This number is increasing by almost 2% each year (UN 2008), creating various challenges for the countries of origin, host countries, and the migrants themselves (IOM 2010). Among these challenges is migrants' health. When they move, migrants can become vulnerable to disease and may face barriers to accessing appropriate health care due to poverty, marginal status, and/or limited access to social benefits (WHO 2003; IOM 2005; Ghent 2008). Although several studies have observed that the health of some populations improves after migration (Friis 1998), and that some populations are healthier than others, these positive effects may be lost over time (Ahmad 2004).

Because many migrants are not familiar with the local language, in face-to-face clinical settings they face language barriers that can diminish the quality of health care they receive. A number of studies have described the negative impact of language barriers on the quality of health services, on the utilisation of these services, and on patients’ health status as an outcome of service quality. These include excess hospitalisation, medical errors, and drug complications (Hampers 1999; Gandhi 2000; Bard 2004); poor access to medical care (Weinick 2000; Pippins 2007; DuBard 2008; Cruz-Flores 2011), and poor access to services promoting healthy behaviour change (Woloshin 1997; Jacobs 2005; Johnson-Kozlow 2009; Kaur 2009). Language barriers cause communication problems and misunderstanding of patients' explanations of their symptoms and health history. They also inhibit the health provider’s presentation of diagnosis, treatment and suggestions for healthful behavioural changes, and the development of a therapeutic patient-provider alliance. In the diagnosis and treatment process, and particularly for illnesses that cannot be identified by observable symptoms, this communication gap can lead to serious problems. The alleviation of language barriers may address these problems. One means of achieving this is by using trained medical interpreters.

Description of the intervention

A ‘trained medical interpreter’ works to overcome language and cultural barriers in a clinical setting (Hornberger 1997; Flores 2005; Bauer 2010; Leanza 2010) through oral restatement of words from one language into another language, simultaneously or consecutively. Trained medical interpretation is not simply any intervention involving an interpreter to provide a linguistic bridge between patient and health provider. There is no universal definition of the term 'trained medical interpreter', and different standards and training have been required by different institutions, agencies, and in various locations The International Medical Interpreters Association (IMIA) has defined standards of practice in the following three areas (IMIA 2007):

  1. clinical interpretation,

  2. cultural interface (understanding, attitudes and practices to reduce culturally-based dissimilarities of perception, presentation, course, and outcomes of illness, wellness and treatment as between providers and patients), and

  3. ethical behaviour.

Reflecting these standards, we define a trained medical interpreter as an internal (staff member employed in health facility in which a patient receives services) or external interpreter (staff member employed in different organisation from health facility in which a patient receives services), who has received training in clinical interpretation, particularly in some or all of these three areas of practice. It is reasonable to assume that trained medical interpreters provide superior and more accurate interpretation than untrained interpreters.

There is variation in how trained medical interpretation is delivered and utilised. For example, the quality of interpretation may vary depending on the professional interpreter's training. In addition, the cost of using professional interpreters is often regarded as a barrier to use, even though some studies have reported that the use of trained medical interpreters can offer cost benefits to the healthcare system, over other approaches or no interpretation (Hampers 2002; Jacobs 2004). Two obstacles to such positive utilisation, as pointed out in a recent study, are the availability of trained medical interpreters and accessibility to the agencies that provide them (Hadziabdic 2011).

Our review will compare the involvement of trained medical interpreters with other approaches which have similar goals but do not involve trained medical interpreters. These include ad hoc interpreters, bilingual health providers, and translated materials (Riddick 1998; Flores 2005). An ad hoc external interpreter is a friend, family member, relative, etc. who takes on the role of clinical interpreter, but has not received any training in interpretation. Ad hoc interpretation may be more convenient but also problematic, because an ad hoc interpreter may lack appropriate interpretation skills and knowledge of medical terminology. Also, the patient’s confidentiality may be compromised, and vital information may be distorted (Launer 1978; Flores 2005; Leanza 2010). A bilingual employee (ad hoc internal interpreter) is a health worker or support worker in a healthcare facility who takes on the role of clinical interpreter without having formal training in interpretation (Johnson 1998Elderkin-Thompson 2001Sevilla Matir 2004; Bischoff 2010). Finally, translated materials include documents and flip charts that offer written communication without an interpreter. Health providers and patients can communicate by pointing to an appropriate phrase in their respective languages, but optimal use requires the health provider to be trained to use them effectively as well as the patient to be literate in his/her mother tongue, which is not always the case. Each of the above modes of intervention may be best suited to different circumstances (Garcia-Castillo 2007; Vazquez Navarrete 2009; Manias 2010).

How the intervention might work

One aspect of the quality of health care for migrant patients is the degree to which their specific linguistic, cultural, and any other needs stemming from their migrant status are met in the process of healthcare delivery. Effectively meeting these needs increases the likelihood of achieving desired health outcomes consistent with the current state of professional knowledge (Lohr 1990). Trained medical interpretation can impact on various aspects of healthcare quality. Specifically, it can improve communication quality (Baker 1996; Flores 2003), and patient and healthcare provider satisfaction with communication (Lee 2002; Al-Khathami 2010). The quality of communication can have a substantial influence on: the suitability of clinical responses; diagnostic certainty and the likelihood of testing (Dodd 1984; Drennan 1996; Hampers 2002); timeliness in seeking medical care (Drennan 1996); visit duration (Kravitz 2000; Hampers 2002Fagan 2003); the utilisation of services including preventive screening (Bell 1999; Jacobs 2001; Bernstein 2002; Dang 2010); appointment keeping (Manson 1988; Sarver 2000); and the length of stay in hospital (Hampers 2002).

The context in which interpretation takes place can shape its effect, because medical interpretation is practised in different service settings and among different target groups. Our review will consider (in subgroup analyses) the following contextual factors, although we recognise that they may be poorly reported in studies:

  • Interpreters' training experience: Fulfilment or non-fulfilment of the three categories of standard practice recommended by IMIA, mentioned above, can influence the interpreter's competency. Interpreters' training experience can vary in terms of the content, duration and intensity of each of the three categories of standard practice (IMIA 2007).

  • Gender: The gender of the interpreter, or gender disharmony between the interpreter and patient, may influence their interaction (Roussos 2010).

  • Age of patient: Communication can differ between children, adolescents and adults due to differences in emotional development and cognitive ability. Quality of interpretation may influence the emotion and attitude of younger patients to health providers. For example, because paediatric patients may be intimidated in front of adults, they may not be able to verbalise their health condition (Purvis 2009).

  • Patient literacy: Information through interpretation for illiterate patients may be limited, since written materials in the patient's own language, for medication and for home follow-up or self-care, cannot be used as a supportive tool for medical interpretation.

  • Medical conditions that require sexual/cultural sensitivity: Some conditions such as reproductive illness, which are highly personal, call for sensitivity to sexual issues, which can influence the interaction between interpreter and patient.

Why it is important to do this review

Although some benefits of language interpretation are quite obvious, there is no systematic review of the effects of interpretation on the quality of health services. It is necessary to quantify the impact of interpretation on the quality of health care, in order to clarify its cost-effectiveness and the advantages it offers, as well as any disadvantages.

This review will provide such quantitative information on the impact of trained medical interpreters in face-to-face clinical settings, compared with other interpretation and translation measures. It will also present a subgroup analysis of the contexts in which interpretation takes place.

This information will offer essential assistance to policy makers, health facilities, and patients in the effective and efficient development of interpretation services, particularly in systems with a diversified context that serve patients with low proficiency in the local language.

Objectives

To assess the effects of trained medical interpreters in face-to-face clinical settings for patients with low proficiency in the local language on:

  1. the quality of communication between patient and provider (as a precondition for the utilisation of professional knowledge to provide quality health care);

  2. the quality of health care, and health outcomes; and

  3. the cost benefit, cost effectiveness and cost utility of interventions by trained medical interpreters.

Methods

Criteria for considering studies for this review

Types of studies

  1. Randomised controlled trials (RCTs)

  2. Cluster RCTs

  3. Randomised cross-over trials

  4. Quasi-RCTs

We will include quasi-RCTs as there are likely to be few RCTs available for inclusion in the review.

Types of participants

  1. Patients of any age with low proficiency in the local language, as determined by the study authors

  2. Health personnel who provide services for patients of any age with low proficiency in the local language

Each participant will be analysed separately.

Types of interventions

The main intervention to be considered is interpretation by a trained medical interpreter, in a face-to-face clinical setting.

The trained medical interpreter is an external or internal interpreter who has received training in medical interpretation, especially in all or some of the following areas:

  1. clinical interpretation,

  2. cultural interface, and

  3. ethical behavior (IMIA 2007).

An external interpreter is defined as an individual on dispatch from an outside organization such as a professional interpretation firm. An internal interpreter is defined as an individual employed in the health facility in which a patient receives care.

The intervention will be compared with one of the following control interventions:

  • Ad hoc external interpreter: a friend, family member, etc., who takes on the role of medical interpreter, but has not received any training in interpretation.

  • Bilingual employee (ad hoc internal interpreter): health worker, support worker at a health facility who takes on the role of medical interpreter, but has not received any training in interpretation.

  • Translated materials: document, flip chart, etc. for interpretation without an interpreter.

  • No interpretation.

Each comparison group will be analysed separately. We will not include comparison groups assessing remote interpretation via telephone or online.

Types of outcome measures

In accordance with the definition of quality of health care for patients with a low level of proficiency in the local language, we regard quality of communication as the primary outcome, which is a precondition for utilisation of professional knowledge, and can change the quality of health care either directly or indirectly. We regard the provision of health services, patients' health status and cost-benefit/effectiveness of medical interpretation as secondary outcomes. Primary outcomes can be determined by evaluating quality of interpretation, and secondary outcomes can be analysed using various measurements.

Primary outcomes
Quality of communication in medical interpretation

1) Quality of interpretation practice: omission of words or phrases, fluency of interpretation, substitution of words or phrases, editorialisation of words or phrases, addition of words or phrases

Measured by counting from audio or video record, self-report, or health provider report.

Interpretation of the results of analysis for these outcomes will be cautious. Omission, false fluency, substitution, editorialisation, and addition during interpretation are not always errors. They might be required to transform a patient or healthcare provider's discourse to make it understandable. In addition, the person assessing the audio or video record may be a trained interpreter who has a vested interested in showing that 'trained medical interpreters' have beneficial effects compared to no interpretation or other forms of interpretation (high risk of bias).

2) Quality of interpretation perceived by patient and/or health provider: patient understanding of diagnosis and treatment; patient satisfaction with information provided, decision made and interpretation; health provider satisfaction with information provided, decision made and interpretation; patient's sense of control over communication via interpretation; and health provider's sense of control over communication via interpretation.

Measured by self-report, health provider report, or standardised instruments.

In the table ' Characteristics of Included Studies' we will describe clearly the instruments used for outcome measurement, as well as the translation process to researchers, to achieve transparency of the data because some outcomes may be controversial. For example, patient satisfaction is widely used as an indicator to assess the quality of health care, but it is difficult to define this parameter because satisfaction can be influenced by various factors, such as patient/provider expectation, age, illness, previous experience, patient-health provider relationship, choice of provider, gender, ethnicity, and socio-economic status. In practice, there is no universally accepted method for the measurement of satisfaction (Saila 2008). In addition, articles may not report the details of the measurement and/or instruments used. Patient satisfaction can be low despite high quality interpretation.

Secondary outcomes

1) Patient engagement with health services

Delays in seeking medical care, visit duration, utilization of health services including preventive screening, missed appointments, length of hospitalisation.

Measured by medical records and the administrative databases of healthcare facilities.

2) Provision of health services

Diagnostic uncertainty, and the amount of medical testing. Communication problems can cause increased diagnostic uncertainty, which can then increase the amount of testing done.

Measured by medical records, and administrative databases of healthcare facilities.

3) Health outcomes (including health behavior, skills acquisition, medical errors, and drug complications).

Measured by medical records, administrative databases of healthcare facilities, standardized instruments, self-report, and provider report.

4) Cost and cost benefits, and effectiveness of medical interpretation

Measured or calculated by the cost of medical interpretation and the effects on health services (i.e. Impact on the cost of health services as well as the health outcomes achieved).

Search methods for identification of studies

Electronic searches

We will search for studies using the following databases:

  • Cochrane Central Register of Controlled Trials (CENTRAL, The Cochrane Library),

  • MEDLINE (OvidSP), 

  • PsycINFO (OvidSP),

  • Dissertations & Theses Database (Proquest),

  • ERIC (OvidSP),

  • Index to Theses,

  • Social Services Abstracts (CSA Illumina),

  • Sociological Abstracts (CSA Illumina).

  • Linguistic and Language Behavior Abstracts (CSA Illumina).

We present the strategy for MEDLINE in Appendix 1. Strategies will be tailored to other databases and reported in the review. There will be no language or date restrictions.

Searching other resources

Grey literature

We will search the reports and conference proceedings of IMIA and document resources linked in their web site (e.g. "Annotated Bibliography on Language Access and Interpretation" (http://www.imiaweb.org/resources/AnnotatedBibliography.asp)), and the reports and conference proceedings of Critical Link International (http://criticallink.org/)

Handsearching

We will manually search the following journals; Journal of Immigrant Minority Health (2006 to 2012), Social Science and Medicine (1967 to 2012) and the Journal of General Internal Medicine (1986 to 2012). We will also search reference lists of relevant studies.

Correspondence

We will contact experts in the field and authors of included studies for advice as to other relevant studies.

Data collection and analysis

Selection of studies

Two review authors will screen independently the titles and abstracts of the studies identified by the searches. We will retrieve full copies of all potentially-relevant articles selected by either of the authors. The two authors will then independently determine if studies meet the inclusion criteria mentioned above. We will list studies that initially appear to meet the inclusion criteria but are later excluded in the table 'Characteristics of excluded studies' with reasons for their exclusion. We will settle disagreements between the two authors through discussion with a third author. Potentially-relevant studies in languages other than English will be translated by collaborators within a group and/or translation agency in order to be considered for inclusion. We will provide citation details and any available information about ongoing studies and report details of duplicate publications. In addition, we will report the screening and selection process in an adapted PRISMA flowchart.

Data extraction and management

We will develop a 'data extraction' sheet (based on the Cochrane Consumer and Communication Review Group's data extraction template), pilot test it on ten randomly-selected included studies, and refine it accordingly. Independently, two authors will extract data from the included studies. Information extracted will include study design, information about the participants including patient's language proficiency, type of intervention, setting, and outcomes. We will settle disagreements between the two authors through discussion with a third author.

The patient's language proficiency is assessed by the study authors, who may ask questions in the local language and in the mother tongue of the patients to find whether there are concordant answers between the two languages. However, some studies might not describe how they identified patients with low proficiency in the local language. In this review, we will report on the method used to identify the language proficiency in the table 'Characteristics of Included Studies' and assess it as another source of bias (selection bias) in the assessment of the risk of bias. We will also report on the method used by the study authors to identify the health providers included in the studies.

All data will be entered into RevMan by one review author and checked for accuracy against the data extraction sheets by the other author working independently.

Assessment of risk of bias in included studies

Two review authors will assess the risk of bias of included studies using the criteria from the Cochrane Collaboration's tool (adapted to the Cochrane Consumer and Communication Review Group's data extraction template) as detailed in the Cochrane Consumer and Communication Review Group’s Study Quality Guide (CCRG 2011). These criteria consist of the following six domains:

  • Sequence generation: judged by the method used to generate the allocation sequence, reported in sufficient detail to allow an assessment of whether it should produce comparable groups. (Quasi-RCTs will be rated as 'high risk' of bias for sequence generation as the methods were not, by definition, truly random).

  • Allocation concealment: judged by the method used to conceal the allocation sequence, reported in sufficient detail to determine whether intervention allocation could have been foreseen in advance of, or during, enrolment.

  • Blinding of participants and personnel: judged by all measures used, if any, to blind study participants and personnel from knowledge of which intervention a participant received, and by any information relating whether the intended blinding was effective.

  • Blinding of outcome assessment: judged by all measures used, if any, to blind outcome assessors from knowledge of which intervention a participant received, and by any information relating whether the intended blinding was effective.

  • Incomplete outcome data: judged by the completeness of outcome data for each main outcome, including attrition and exclusions from the analysis. If 80% or more of the data are complete, it will be rated as 'low risk' of bias. Otherwise, it will be rated as 'high risk' of bias. If it cannot be identified in a study, it will be rated as being at 'unclear' risk of bias.

  • Selective outcome reporting: judged by the review authors' findings about the possibility of selective outcome reporting. If a study protocol is available and all outcomes in the study method are reported by the study report it will be rated as 'low risk' of bias; if no protocol is available and not all outcomes in the method are reported it will be rated as 'high risk' of bias; if no protocol is available but all outcomes in the method are reported it will be rated as being at 'unclear' risk of bias.

  • Other sources of bias: the authors will identify any important concerns about bias not addressed in the other domains. For example, we will assess the method used to identify patients' language proficiency as a source of potential selection bias. In addition, we will assess the baseline discordance between groups. If cluster RCTs are included in the review we will also assess and report the risk of bias associated with selective recruitment of cluster participants (CCRG 2011).

Further, as outlined in the Cochrane Handbook (Higgins 2011), we will categorise the risk of bias of included studies as: low risk of bias (plausible bias unlikely to seriously alter the results), unclear risk of bias (plausible bias that raises some doubt about the results), and high risk of bias (plausible bias that seriously weakens confidence in the results).

Two authors will conduct the risk of bias assessment independently. Disagreements will be resolved by discussion between the two review authors; if agreement cannot be reached, a third review author will decide.

The risk of bias of included studies will be used to inform the discussion of the review's findings.

Measures of treatment effect

Continuous data

We anticipate that the majority of outcomes will be measured and reported as continuous data. We will compute a standardised mean difference (SMD) for continuous outcome variables, with a 95% confidence interval (CI). For this review, a corrected Hedges' g will be computed by dividing the difference between intervention means (trained interpreters versus other interpretation) by the pooled and weighted standard deviation of the intervention. Specifically, Hedges' g corrects for a bias (overestimation) that occurs when the uncorrected SMD effect size is used on small samples. The combined effect size for each outcome will be computed as a weighted mean of the effect size for each study, with the weight being the inverse of the square of the standard error. Thus, a study will be given greater weight for a larger sample size and more precise measurement, both of which reduce standard error.

Dichotomous data

We will compute odds ratios (ORs) for dichotomous outcomes with a 95% CI. Based on the assumption of proportional odds, ORs can be compared between variables with different distributions, including very rare and more frequent occurrences.

Unit of analysis issues

Cluster-randomised trials

If the unit of allocation (e.g. hospital) is different from the unit of analysis (e.g. people with low proficiency in the local language), we will seek statistical advice to determine whether appropriate methods were used to avoid unit-of-analysis errors. When suitable cluster analysis is used, effect estimates and their standard errors will be meta-analysed. Otherwise they will be excluded from the meta-analyses unless the review authors can control for the clustering from the available information.

Crossover trials

If studies are conducted in crossover design, we will use the results from the first intervention period.

Multiple intervention groups

Within the intervention and/or control groups mentioned in the section Types of interventions, if multiple groups with different individuals are presented in studies, all relevant intervention and/or control groups will be combined into a single group to create a single pair-wise comparison.

If the multiple intervention and/or control group include the same individuals, only one intervention group will be compared to the control group to avoid overlapping samples. Data from studies comparing different type of interventions/ comparisons will be analysed separately.

Dealing with missing data

Although studies with incomplete outcome data (e.g. missing means, standard deviations, sample sizes) will be included in the review, we will exclude them from the meta-analyses unless we can calculate an effect size from the available information. When outcome data may be missing from an article or report, we will make reasonable attempts to retrieve these data from the original researchers. We will account for overall and differential attrition in the risk of bias and sensitivity analyses.

Assessment of heterogeneity

Clinical heterogeneity may be a prevalent issue. The contents of training, and the experience and ability of medical interpreters may differ among individuals. Health providers' cultural competence, clinical skills and medical specialty may be diverse. Further, patients' background, illness and previous experiences in the clinical setting may be diverse.

If studies are considered homogeneous based on consideration of the intervention's features (the interpreter's training experience, their gender and/or gender discordance between interpreter and patient) and patient factors (the patient's age, literacy or their medical condition), to allow pooling of data using meta-analysis, we will assess the degree of heterogeneity by visual inspection of forest plots and by examining the Chi2 test for heterogeneity. Heterogeneity will be quantified using the I2 statistic. An I2 value of 50% or more with a P value from the test < 0.1 will be considered to represent substantial levels of heterogeneity.

If we detect substantial clinical, methodological or statistical heterogeneity across included studies, we will not report pooled results from meta-analysis, but will instead use a narrative approach to data synthesis. In this case we will attempt to explore possible clinical or methodological reasons for this variation by grouping studies that are similar in terms of training experience of trained interpreter, interpreter's gender and/or gender discordance between interpreter and patient, age of patient, literacy of patient, or medical condition.

Assessment of reporting biases

We will investigate possible reporting biases by means of funnel plots and testing for funnel plot asymmetry using Egger's test (Higgins 2011). If important asymmetries are found, possibly due to small-study effects (the tendency for effect sizes from small studies to deviate from effect sizes from large studies), we will undertake a sensitivity analysis using the 'trim and fill' method to correct for the asymmetry.

Data synthesis

The decision to meta-analyse data or not will be based on an assessment of whether the interventions in the included study are similar enough in terms of participants intervention, comparison and outcome measures to ensure meaningful conclusions if data are statistically pooled. Due to the anticipated variability in training experience of trained interpreter, gender of interpreter and/or gender discordance between interpreter and patient, age of patient, literacy of patient, or medical condition, we will use a random-effects model for meta-analysis.

If a study reports multiple effect sizes, the results will be included in the meta-analysis for each outcome. If a study reports effect sizes for multiple samples, the results will be aggregated for the main effects meta-analysis before being used for the subgroup meta-analyses.

If we are unable to pool the data statistically using meta-analysis, we will group the data based on the category that best explores the heterogeneity of studies and makes most sense to the reader (i.e. by interventions, populations or outcomes). Within each category we will present the data in the tables and summarise the results narratively .

Subgroup analysis and investigation of heterogeneity

We will consider subgroup analyses to examine different effects of the intervention or subgroup characteristics as follows:

  • Training experience of trained interpreter based on the three categories of standard practice recommended by IMIA: (1) interpretation, (2) cultural interface, and (3) ethical behavior, as well as (4) duration or intensity of training.

  • Gender of interpreter and/or gender discordance between interpreter and participant.

  • Age of patient (child, adolescent and adult). Children, adolescent, adults are defined as young people less than 10 years old, young people between 10 and 19 years old and people more than 19 years old, respectively.

  • Literate and illiterate patients.

  • Medical conditions that need sexual/cultural sensitivity.

This subgroup analysis will take comparison groups separately into analyses, because the variability among comparison intervention groups across studies might impact on the outcomes more than any of the characteristics mentioned above.

Sensitivity analysis

Sensitivity analyses will be considered to explore the impact of the risk of bias dimensions on the outcomes of the review. Specifically, we will consider the following comparisons:

  • RCT and quasi-RCT.

  • Studies with low overall or differential completeness of outcome data, and studies with high overall or differential completeness of outcome data. If 20% or more of outcome data are incomplete, it will be rated as 'low' completeness. Otherwise, it will rated as 'high' completeness.

Consumer views and participation

In order to obtain consumer views and feedback on the interpretation of meta-analysis results, we will conduct the following:

  • a descriptive (non-systematic) review of qualitative research on the quality of medical interpretation and its impact;

  • after the completion of the meta-analysis, an in-depth interview or focus group discussion in Japan; and

  • a similar interview/discussion in Bangladesh.

The descriptive review can provide supplemental information on consumer views, especially where accessibility to consumer views or the variety of accessible consumer views is limited. In this study, ‘consumers’ include patients with low proficiency in the local language, healthcare providers, medical interpreters, administrative staff of health facilities, and public employees.

The results of these methods will be included in the Cochrane review's 'Discussion' to assist us to interpret the results of any meta-analysis.

1) Descriptive (non-systematic) review of qualitative research into the quality of medical interpretation and its impact

We will conduct a descriptive review of qualitative research. Because qualitative research is generally used to gain in-depth understanding of and insight into human practices and the reasons that govern such practices, it will be useful in this review to understand consumer views on medical interpretation. In addition, the review will be able to cover a greater variety of populations and clinical settings, in comparison to direct contact with different consumers mentioned below in methods 2) and 3). We will identify articles for this descriptive review through the process of selection of literature for meta-analysis, particularly through the process of searching other resources. We will select studies with in-depth interviews, focus group discussions, and other participatory methods analysed with grounded theory, for this descriptive review.

2)  In-depth interview and/or focus group discussion with consumers or supporting organizations (in Japan)

We will conduct in-depth interviews and/or focus group discussions with different consumer groups or supporting organizations to obtain feedback on meta-analysis results. With this method, diversity by ethnicity or native country may be looked into more deeply, particularly in comparison to method 3). The in-depth interview/focus group discussion will take place at least three times with three different consumers, to enable triangulation of data. The information may be gained mainly from supporting organizations for patients with low proficiency in the local language, rather than patient themselves.

3)  In-depth interview and/or focus group discussion with returnees (in Bangladesh) who have experience as patients with low local-language proficiency in a host country

In-depth interviews and/or focus group discussions will be conducted with returnees who went abroad to work and/or live and then returned to Bangladesh, and who have experience as patients with low local-language proficiency in a host country. By involving such returnees, more in-depth information can be gleaned to better understand the context of medical interpretation. In addition, the diversity of the host country may be looked into more deeply, in comparison to method 2). In migrant host countries such as Japan, it is sometimes difficult to access patients with low local-language proficiency or, sometimes, they hesitate to participate in research. In addition, the analysis on the data gathered can be complicated, time consuming, and expensive because the process of qualitative analysis has to include language interpretation of in-depth interviews and focus group discussions, translation of transcripts of interviews and discussions for analysis, and translation of the analysis results for verification.

Acknowledgements

We thank Dr. Brian McKinstry of the University of Edinburgh, the contact editor of this review, and Dr. Megan Prictor, the Managing Editor of Cochrane Consumers and Communication Review Group, for their continuous suggestions and support since the title registration.

Appendices

Appendix 1. MEDLINE (OvidSP) Search Strategy

1. translating/

2. communication/

3. communication barriers/

4. ((communication or language or linguistic) adj barrier*).tw.

5. exp professional patient relations/

6. professional-family relations/

7. hospital-patient relations/

8. community-institutional relations/

9. interviews as topic/

10. "referral and consultation"/

11. minority groups/

12. (second language* or ((language or low or limited or english) and proficien*)).tw.

13. (language adj (concordan* or discordan*)).tw.

14. health services accessibility/

15. cultural competency/

16. cultural diversity/

17. transcultural nursing/

18. ((cultural* or multicultural or transcultural or cross-cultural or ethnic*) adj3 (service* or care)).tw.

19. or/2-18

20. 1 and 19

21. ((medical or healthcare or health care or professional or trained) adj (interpret* or translat*)).tw.

22. (interpreter? or translator or translators).tw.

23. ((interpret* or translat* or language or linguistic*) adj2 (service* or assist*)).tw.

24. language support.tw.

25. interpreted consult*.tw.

26. or/20-25

27. randomized controlled trial.pt.

28. controlled clinical trial.pt.

29. clinical trial.pt.

30. evaluation studies.pt.

31. comparative study.pt.

32. random*.tw.

33. placebo*.tw.

34. trial.tw.

35. research design/

36. follow up studies/

37. prospective studies/

38. cross over studies/

39. (experiment* or intervention*).tw.

40. (pre test or pretest or post test or posttest).tw.

41. (preintervention or postintervention).tw.

42. time series.tw.

43. (cross over or crossover or factorial* or latin square).tw.

44. (assign* or allocat* or volunteer*).tw.

45. (control* or compar* or prospectiv*).tw.

46. (impact* or effect? or chang* or evaluat*).tw.

47. or/27-46

48. exp animals/ not humans.sh.

49. 47 not 48

50. 26 and 49

Contributions of authors

HT conceived, designed, drafted and coordinated the protocol. RM contributed to the writing of the protocol. RM, DK, and TS commented upon and approved the final version of the protocol.

Declarations of interest

None known

Sources of support

Internal sources

  • Medical Library, University of Tokyo, Japan.

    In-kind support for searching articles

External sources

  • No sources of support supplied

Ancillary