Background

Emergency care systems, which include prehospital care and emergency departments (EDs), provide accessible medical care to all individuals, regardless of their background. However, patients with preferred languages other than English, particularly those labeled as limited English proficiency (LEP), often face barriers to communication and access to care in English-speaking emergency care systems. This has been found to lead to increased rates of ED investigations [1–5] and healthcare utilization [1, 2]. In addition, patients with non-English preferred languages have decreased comprehension of medical instructions and reduced comfort with postdischarge care plans [6], likely leading to higher rates of unplanned ED returns and readmission [7, 8]. They also report decreased satisfaction with care and communication in the ED [9].

Professional interpretation improves outcomes for patients with LEP compared to no translation or ad hoc, untrained interpreters such as family members or members of the healthcare team [10–12]. Despite the demonstrated effectiveness of interpreters, uptake has been poor [13–16] and we have yet to understand the key barriers and facilitators of implementing different interpretation/translation services and tools. With emerging technology and increasing availability of online tools, new modes of interpretation/translation may also be able to help overcome historical barriers and improve uptake, such as machine translation tools [17, 18] and apps [19, 20]. Indeed, a recent consensus study found language translation to be one of the most desired applications of artificial intelligence by emergency physicians [21].

We are therefore seeking to map the existing literature on interpretation/translation modalities for patients in emergency care, report what key barriers/facilitators to implementation exist, and how their effectiveness and implementation is studied.

Research questions

Our research question is: What language interpretation/translation modalities have been studied in emergency care?

Sub-questions include:

1. What type of language interpretation/translation modalities have been studied in emergency care?
2. What are the key barriers/facilitators to implementation of each type of language interpretation/translation modality in emergency care?
3. How is the effectiveness and implementation of language interpretation/translation in emergency care assessed?

Information sources & search strategy

The search strategy (example in S2 Appendix) was developed and refined by a Health Sciences librarian (LD) in consultation with content and methodology experts. Peer review of the search strategy was done by a second expert librarian using the Peer Review of Electronic Search Strategies (PRESS) checklist [25].

We searched the following electronic databases from inception until May 15, 2024: MEDLINE [Ovid], Embase [Ovid], PsycINFO [Ovid], CINAHL [EBSCO], Scopus, iPortal (https://iportal.usask.ca/), Native Health Database (https://nativehealthdatabase.net/), and Cochrane Library CENTRAL [Wiley]. The search includes subject headings and free text terms for two concepts: language translation services and emergency care. For the emergency care concept, we used the filter by Campbell and Dennett, 2024 [26]. The search was optimized for use in each database. No language restrictions were included in the search. Opinion pieces, letters, editorials, and news articles were removed if possible. Results were then imported into Covidence and de-duplication was performed.

References of all included studies will be searched for additional studies. Any relevant articles/studies encountered during our work that are not captured by the systematic search will also be considered and included if they meet the eligibility criteria, following a snowballing approach.

Article selection

Two independent reviewers will screen titles and abstracts in Covidence to identify studies relevant to this review. The two reviewers involved in any disagreement will meet to discuss and resolve through consensus. If consensus is not achieved through this process, then a third reviewer will be engaged for resolution. Subsequently, full texts of all relevant records will be retrieved and independently reviewed for their adherence to the a priori-defined inclusion criteria. These criteria will be finalized with team input and pilot tested prior to implementation. Disagreements at this stage will again be resolved through consensus or by a third reviewer.

Non-English texts will be included for consideration, however, due to financial limitations we will not be able to employ professional translation. As a result, we will use the team’s internal resources alongside online language translation tools to translate non-English texts for selection and data extraction. Members of our study team are familiar with over 10 different languages. We will use both Google Translate and ChatGPT to translate non-English texts and compare the resulting translations. Machine translation tools provide a practical solution, however their accuracy may vary, especially in technical or medical contexts. Where discrepancies arise, they will be verified by a study team member familiar in the language of the text to ensure translation fidelity. Where no study team members are familiar with the language, we will ask a colleague who is familiar with the language to verify. In the event this is not possible we will report this as such in the limitations of the analysis.

Inclusion criteria

Peer-reviewed articles reporting original research will be included according to the PCC (population, concept, context) inclusion criteria in Table 1.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Inclusion and exclusion criteria for study selection.

https://doi.org/10.1371/journal.pone.0314049.t001

We use the terms translation and interpretation according to Allen et al’s definitions [27]. Interpretation is defined as "the rendering of live utterances in one language to live utterances in another language" which includes both spoken and signed languages. Translation is defined as "movement between a recorded form of one language into a recorded form of another language" which includes written text or audio recordings of spoken language [27].

We will include descriptive, observational, quantitative, qualitative, and mixed methods studies. Editorials, letters, opinion pieces, policy recommendations, case reports, and case series will be excluded.

Data charting

The following data will be extracted from each article:

1. Study details (e.g., first author, year of publication, country, sample size)
2. Study design (e.g., interventional, observational)
3. Setting characteristics (e.g. hospital ED, stand-alone ED/urgent care, pre-hospital setting)
4. Characteristics of interpretation/translation modality (e.g. modality, resources/technology needed, unidirectional vs bidirectional, languages involved, relevant stage of clinical care, personnel utilizing tool)
5. Participant characteristics (e.g. inclusion/exclusion, age, gender, languages spoken)
6. Outcomes including effectiveness (e.g. accuracy) and implementation (e.g. acceptability, adoption, appropriateness, cost, feasibility, fidelity, penetration, sustainability)
7. Barriers/facilitators to implementation

The data charting form will be pilot-tested using a sample of 10 full-text articles and then revised accordingly (current version in S3 Appendix). Data charting will be completed by two independent reviewers. Disagreements will be resolved through consensus or a third reviewer.

Data synthesis

We will summarize and report the quantity, content, and coverage of the evidence. This will include 1) a summary of the language interpretation/translation modalities studied in emergency care; 2) a list of outcomes used to evaluate language interpretation/translation in emergency care; and 3) a narrative summary of key barriers and facilitators to implementation specific to each identified interpretation/translation modality, mapped to the Consolidated Framework for Implementation Research [28]. Results of included studies such as the impact of interpretation/translation on outcomes will not be extracted or assessed. Descriptive statistics will be used where appropriate, including means and standard deviations for continuous data and proportions for categorical data.

Stakeholder engagement & knowledge dissemination

Our team includes the lead for interpretation services of our provincial health authority, Alberta Health Services (KG), who reviewed and informed the research questions for this scoping review. They also bring lived experience of navigating language barriers in healthcare and will be engaged according to the “collaborate” goals of the International Association for Public Participation spectrum for engagement [29]. Furthermore, we will be engaging additional patient partners to help inform and validate our ScR results and help identify future areas of needed research. In addition to academic knowledge dissemination venues such as publication in a peer-reviewed journal and conference presentations, we will also develop a graphical abstract and social media campaign to help share our findings more broadly. Our included stakeholders will help us identify key messages and appropriate methods for dissemination to maximize knowledge translation to patients/families, local policy/clinical practice, and funding agencies.