Developing Discrete Choice Experiments for Populations with Non-Dominant Language Preference: A Methodological Framework and Case Example

Developing Discrete Choice Experiments for Populations with Non-Dominant Language Preference: A Methodological Framework and Case Example | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Developing Discrete Choice Experiments for Populations with Non-Dominant Language Preference: A Methodological Framework and Case Example Kirsten Austad, Noelia Lugo, Diana Bautista-Hurtado, Stephane Labossiere, and 6 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6976834/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Background: Discrete choice experiments (DCEs) are widely used to elicit patient preferences for health interventions and implementation strategies. However, individuals with non-dominant language preference (NDLP) are frequently excluded from DCE studies due to language and literacy needs. As health systems increasingly serve linguistically diverse populations, there is an critical need to adapt DCE methods for inclusion of NDLP populations. Objective: To develop and pilot a practical methodology for designing and conducting DCEs in high-diversity, multilingual settings. We use a case example focused on tailoring a hospital discharge intervention for Spanish and Haitian Creole speakers with NDLP to illustrate the methodology. Methods: We adapted a five-step DCE development framework, including evidence synthesis, end-user input, expert review, pretesting and pilot testing, for use with NDLP populations. Modifications focused on pretesting using cognitive interviews with patients and professional interpreters, and a rapid-cycle approach to iterative revision in two languages. The DCE focused on hospital discharge preferences and included four attributes with visual and text-based levels. Pilot testing assessed comprehension, cognitive burden, and usability of the final DCE instrument. Results: We identified multiple barriers to DCE participation for patients with NDLP, including difficulty grasping the critical task of elicitation (making tradeoffs) in a DCE choice set, misinterpretation of abstract attributes, and inconsistent understanding of visual materials. Tailored solutions included bilingual interviewer-led administration, integration of a “warm-up” elicitation task, use of labeled photos to describe levels, and iterative revision of wording and decisional context to improve comprehension. Pilot testing showed high acceptability and comprehension, with no evidence of cognitive fatigue. Participants were able to articulate tradeoffs and report decision-making strategies aligned with stated preferences. Conclusions: This study provides a practical, equity-centered methodology for conducting DCEs with linguistically diverse populations. Our findings demonstrate the feasibility of rigorous DCE administration in NDLP populations and offer a replicable framework for implementing DCEs high diversity contexts. Future work should explore adaptation across additional languages, cultures, and settings and evaluate the impact of preference-informed tailoring on intervention outcome. non-dominant language preference limited English proficiency implementation science discrete choice experiment intervention adaptation transitions of care hospital discharge professional interpreter Figures Figure 1 Figure 2 Figure 3 Figure 4 BACKGROUND Discrete choice experiments (DCEs) are a powerful tool for eliciting the preferences of individuals and are increasingly applied to evaluate health care decision-making.( 1 , 2 ) Participants are presented with hypothetical scenarios and asked to select their preferred options from two or more alternatives, simulating real world decision-making. Participants must weigh competing factors and make deliberate trade-offs between different features. Through probabilistic models grounded in consumer and random utility theory, it is possible to quantify the importance of different attributes, the utility participants assign to various choices, and the trade-offs they are willing to make between attribute levels.( 3 , 4 ) Within the field of implementation science, DCEs have been used to make implementation efforts more responsive to end-users by refining and tailoring implementation strategies to the preferences of a given population.( 5 , 6 ) A promising yet underutilized application of DCEs involves engaging minoritized populations in adapting interventions to better meet their needs and preferences.( 7 ) Many health interventions are not tailored to the needs of linguistically and culturally diverse populations, leading to poor fit, limited uptake, and reduced real-world effectiveness.( 8 – 10 ) DCEs offer a systematic method for capturing the specific needs and preferences of diverse populations that can be used to tailor interventions, a strategy shown to reduce inequities in health outcomes.( 11 ) As linguistic and cultural diversity continues to grow throughout the world, developing interventions that fit the needs of a single demographic does not align with real world health care delivery, where multiple linguistic and cultural groups are often served simultaneously in one setting.( 10 , 12 ) In these high diversity contexts, a DCE needs to be designed and deployed in multiple languages, but this introduces methodological challenges that threaten the feasibility of conducting a DCE and reliability of the results.( 13 , 14 ) Currently there is a lack of up-to-date guidance for conducting DCEs in high diversity contexts.( 15 ) Without rigorous development of cross-linguistic research processes, DCEs risk introducing biases that reduce generalizability of the findings, and compromise the validity of insights, ultimately undermining efforts to ensure equity for individuals with a non-dominant language preference (NDLP). The goals of this manuscript are to provide practical solutions to address the challenges of conducting DCEs in linguistically and culturally diverse populations. We first summarize these challenges and propose solutions to address them. We then propose modifications to the existing approach to DCE development, including a novel approach to pretesting the DCE instrument in a linguistically and culturally diverse patient population. Finally, we present a case to illustrate our methodology. Designing and Conducting DCEs in Populations with NDLP: Anticipated Challenges To illustrate how a DCE is operationalized, Fig. 1 depicts an example choice set, a presentation of two hypothetical scenarios requiring participants to select their preferred option based on trade-offs between different characteristics. In this example, the choice set measures the importance of characteristics weighed when choosing a grocery store. Two hypothetical grocery stores are defined according to discrete characteristics (or attributes), such as hours of operation, transportation to the store, and available food types. Each of these attributes has multiple options (or levels). For example, for the attribute of store hours the levels shown are “open only in the morning” and “open only in the afternoon.” Participants are asked to select option A or B as a set, which requires making tradeoffs between the levels displayed as a package. An experimental design is used to systematically vary the scenarios shown to each participant, since the total number of attribute-level permutations are too great to test.( 16 ) Given the complexity of the task, designing and conducting a DCE with culturally and linguistically diverse populations presents unique challenges. Specific challenges, which are summarized in Table 1 , include: ( 1 ) effectively communicating the DCE content and task, ( 2 ) developing culturally and linguistically appropriate content that maintains equivalent meaning, ( 3 ) providing equitable access to completing the DCE regardless of internet connectivity and electronic device availability. Table 1 Considerations in planning a discrete choice experiment (DCE) in a high linguistic diversity contexts Consideration Rationale Assemble a diverse research team Qualitative methods, such as cognitive interviewing, used to design the DCE instrument are better accomplished via language concordant data collection and data analysis and interpreter by a multidisciplinary research team Involve professional interpreters Professional interpreters can provide valuable insights to word choice and phrasing to ensure the content and description of elicitation are understood by a wide range of speakers of language is critical Utilize strategies to overcome linguistic, literacy, and health literacy barriers DCEs most often convey attributes and levels with text only; use of visuals, such as photos, pictographs, or even videos may improve comprehension Ensure cultural appropriateness of attributes and levels If key determinants of choice in real life are not represented in DCE the results will not represent real-world decision making Engage with end-users early There may be little qualitative literature to inform attribute and level selection in specific group of interest; to ensure content is culturally appropriate and inclusive DCE design should ensure the voices of those with NDLP are included Include language concordance as an attribute or level in DCE Patients with NDLP have a strong preference for language concordant care. Understanding how much they are willing to tradeoff other preferences to achieve this can directly inform where to invest resources in adapting interventions. Develop and refine the description of decisional context Participants may come from different contexts where health systems work differently and may need more explanation to understand trade offs Sample size Recruiting a large sample of the target population with NDLP may be challenging. Limiting the number of attributes and levels tested may be necessary. How to administer DCE DCEs are most often administered via a survey link to be completed independently on a computer or phone by the participation; having trained research staff administer the DCE in person may improve comprehension and allow recruitment of populations otherwise excluded Cognitive burden on participants Those who face linguistic and literacy barriers may be more cognitively stressed by a DCE. To minimize cognitive burden, consider pairwise comparisons only (A vs B), limit number of choice sets shown to each participant, and consider administration and amount of text (as described above) Anticipated Challenge 1: Understanding Typical approaches to orienting participants to a DCE may not adequately address the communication needs of those with NDLP due to barriers posed by limited fluency in the dominant language. While professional translation of written DCE content is one solution, it may fail to capture the intended meaning or use phrasing that is not understood by the average speaker.( 17 ) For widely spoken languages such as Spanish, difference in word meaning and colloquial usage across Spanish-speaking countries may compromise understanding. In a high diversity context where the DCE will be completed by speakers of different languages, the translation of attributes and levels must result in uniform understanding across linguistic groups. Additional barriers to understanding for individuals with NDLP include low literacy, the inability to read written text at a level needed for everyday tasks, and low health literacy, the limited ability to understand and use health information. In the United States, individuals with a non-English language preference are more than three times as likely to have low health literacy compared to their English fluent counterparts.( 18 ) Because many DCEs are self-administered via an written survey, literacy barriers may preclude participation or impede full understanding for those with NDLP. DCEs conducted in settings where low literacy is common have addressed this by using alternatives to traditional text heavy DCEs, such as use of visual depictions of attributes and levels.( 19 , 20 ) Individuals with NDLP may also face challenges during the elicitation of tradeoffs in the DCE. While everyone considers tradeoffs in everyday decisions, explicitly asking participants to make these tradeoffs in an abstract task may create confusion. Introductory materials must explain the task clearly. For some individuals with NDLP, this task may be affected by a cultural tendency to avoid direct expression of opinions. Anticipated Challenge 2: Content One challenge is ensuring that the content tested in the DCE adequately represents the preferences of the participating groups. If the most important attributes and levels are not included in a DCE, the results become uninterpretable. For example, physician gender may be unimportant for some cultural groups and critically important for others. Before being presented with choice sets, DCE participants should receive an introduction to the topic that frames the relevant decisions and tradeoffs. This so-called “decisional context” is critical to ensure that participants are imagining the scenario as intended, so that choices truly represent preferences. Some individuals with NDLP that were born in other countries may require different or more extensive explanation of the context within the local health care system. The attribute of language concordance is likely to be relevant in DCEs for those with NDLP. Research has shown a strong preference for a language concordant provider over the use of an interpreter. This strong preference may drive participants to make decisions solely based on this attribute, limiting the ability to draw conclusions about other preferences from a DCE. Anticipated Challenge 3: Administration Administering a DCE in a high linguistic diversity context may also pose challenges. As some DCEs recruit participants via an online panel (pre-recruited individuals who have agreed to participate in research), certain populations with NDLP who have limited access to digital devices may be systematically excluded. While DCEs are often self-administered electronically, comfort with using a phone or tablet to complete a DCE survey without dedicated support may be low among individuals with NDLP. Lastly, reliable answers to a DCE depend on the participant not being cognitive overburdened. Therefore, approaches to reduce cognitive burden should be considered, such as limiting the number of choice sets shown to each participant and only presenting pairwise comparisons. Attention to these considerations is essential for maintaining the validity and reliability of DCE results in multilingual contexts. METHODS Description of Case Example We will illustrate our methodology by applying it to the development of a DCE to improve the hospital discharge experience for patients with NDLP. The transition from hospital to home is notoriously difficult for both patients and caregivers.( 21 ) It is prone to errors that compromise patient safety and costs the health care system billions of dollars every year. Patients with NDLP face unique challenges due to multifactorial communication barriers and structural inequities.( 22 ) Most trials that test evidence-based interventions to improve the hospital discharge experience have categorically excluded individuals with NDLP. The goal of this DCE is to inform a multidisciplinary team tasked with tailoring a hospital discharge intervention to populations with NDLP. The study was conducted at Boston Medical Center (BMC), a tertiary academic medical center and New England’s largest safety net hospital. About 80% of BMC patients are covered by public insurance. Its patient population is diverse both in terms of race and ethnicity, with 50% of patients identifying as Black and 25% as Latino, and linguistically, with over one-third of patients reporting a NDLP. The top two non-English languages of those discharged from medical hospitalizations are Spanish (54%) and Haitian Creole (18%). Because our primary focus is on the development of the DCE instrument for this high diversity context, other aspects of the DCE development, such as experimental design and quantitative data analysis of DCE results, are not addressed. Overview of Approach to DCE Design A five-step process for designing a DCE instrument has emerged as a best practice as described by Campoamor and colleagues.( 23 ) The steps include: ( 1 ) evidence synthesis; ( 2 ) expert input; ( 3 ) end-user engagement; ( 4 ) pretesting; and ( 5 ) pilot testing. Figure 2 a depicts these steps along with the methods that can be used in each step. To address the challenges of developing a DCE for participants with NDLP, we propose changes to all five phases as described below and depicted in Fig. 2 b. Our design for the DCE was a two option (A versus B) choice set with an opt out option. We administered it in person to patients currently admitted to the hospital who spoke Spanish or Haitian Creole by a bilingual researcher. Given the intensity of the data collection approach, we sought to limit it to 4 to 5 attributes with 2 to 4 levels each. Our goal was to ensure that the instrument captured all strongly held preferences across the two linguistic groups. Procedures Phase 1: Evidence synthesis Phase 1 involves review of existing literature to identify gaps in understanding that could be addressed using a DCE. For a DCE focused on populations with NDLP, there may be limited data. In our example, we followed the best practices and synthesized findings from studies identified in two recent systematic reviews on hospital discharge interventions and a non-systematic PubMed search for additional articles on challenges of patients with NDLP at hospital discharge.( 21 , 24 – 26 ) The first author used this information to create a preliminary list of attributes and levels. Phase 2 and 3: End user and Expert input Best practice recommends first seeking input from experts (phase 2) followed by end-users (phase 3), which may include patients and community members or any others who are expected to be the users of the intervention. Both steps may be formally conducted using qualitative methods and participatory approaches, such as community advisory boards. We reversed these steps and first engaged end-users with the goal to center the voices of the target population and ensure that culturally relevant attributes and levels emerged as early in the process as possible, which we did in our study. In our phase 2, we utilized existing data from a separate study involving structured qualitative interviews with Spanish and Haitian Creole speakers recently discharged from the hospital to understand needs and challenges during the transition from hospital to home. The first author reviewed the interview transcripts to assess for alignment with attributes and levels included in the draft DCE instrument and identify additional attributes and levels. In our phase 3, three types of experts provided input on the draft DCE: two with knowledge of hospital discharge interventions, one with expertise in populations with NDLP, and one DCE expert. Phase 4: Pretesting Phase 4 involves pretesting the draft DCE instrument, often utilizing cognitive interviewing, cognitive debriefing, or co-creation—a collaborative and iterative process of researchers creating materials with participants.( 27 ) The goals of pretesting are to evaluate: ( 1 ) comprehension of attributes and levels as well as the introductory script; ( 2 ) presentation including the images used and how choice sets are displayed; ( 3 ) elicitation, or whether participants understand the hypothetical scenarios and the task of making tradeoffs; and ( 4 ) content, specifically if any attributes or levels should be added or removed. We modified the typical approach to DCE pretesting to meet the unique needs of our culturally and linguistically diverse patient population. We developed an innovative rapid cycle analytic approach to collect, analyze, and revise DCE materials (Fig. 3 ). We initially developed a first draft of materials in Spanish and them to conducted cognitive interviews with Spanish speakers, iteratively modifying materials after each pair of interviews. Once few new themes emerged, we translated the DCE materials into Haitian Creole and conducted cognitive interviews with this population, again making iterative changes to improve clarity and content when needed. We translated the materials into Spanish once again and revised both sets with insights from cognitive interviews with a professional interpreter from each language group, which allowed us to assess the comprehension of formal translation of study materials—akin to the approach of forward and backward translation—and to solicit input on whether the DCE task, attributes, and levels would be understood by a diverse group of Spanish and Haitian Creole speakers.( 28 ) Finally, we conduct additional interviews with these revised materials in both languages in parallel until saturation was reached in both groups. This approach allowed us to test solutions quickly and achieving uniform understanding by participants and to overcome barriers arising from the need to analyze data and revise materials in two different non-English languages. Sample size was determined by saturation of the data at each step. Phase 4: Pretesting We used cognitive interviews to pretest draft DCE materials, including an introductory script explaining the decisional context, an in-depth explanation of attributes and levels using both text and pictures, and three example choice sets. Interviews were conducted by native speakers of Spanish and Haitian Creole. The interview guide first assessed understanding of the concept of hospital discharge and of DCE with a focus on the participant’s grasp of hypothetical comparisons and elicitation. Second, the interviewer assessed understanding of each attribute and its levels and solicited feedback on how to improve clarity of text descriptions and pictorial depictions, including whether they liked the photograph or pictograph better for each level (and any input on revisions). Third, the interviewer administered three rounds of DCE on paper followed by a cognitive autopsy to evaluate the participant’s understanding, decision making processes, and alignment with preferences stated earlier. We also collected demographic data from patient participants, including age, gender, duration of time in the U.S., language preferences, self-reported ability to read and write in their preferred language, and self-assessed fluency in English. We asked about personal experience with forms of language facilitation, including professional interpreters by phone, video, and in-person and language concordant providers, and asked participant to rank each from most to least preferred. In phase 4b, in-depth cognitive interviews were conducted with professional Spanish and Haitian Creole interpreters using the same process as with patient participants. The interviewer solicited feedback using a semi-structured interview guide addressing the same domains used in the patient participant interviews. Additional probes aimed to assess if terms would be understood by diverse speakers of the language (i.e. colloquial differences based on country or rurality) and identify ways to improve the description of the DCE task and address potential areas of confusion. Phase 5: Pilot testing Prior to launching the DCE, best practice includes pilot testing the instrument under actual study conditions. For example, if DCE will collect data from participants via a self-administered survey then pilot testing should follow this same procedure. Piloting can also assess cognitive fatigue, which may require reducing the number of choice sets shown to each participant. In our case, we elected to conduct pilot testing in person with a language concordant administrator. Each participant was asked to complete 20 choice sets via Sawtooth, a commercial software for conducting DCEs.( 29 ) Participants were given the option of holding the tablet and selecting each choice themselves, or for the researcher to read each choice set aloud and record their preferences. The researcher took structured notes on signs of cognitive fatigue and on the points of clarification needed. The participants were asked to rate how difficult and cognitively taxing they found the task on a five-point Likert scale. They were then asked to explain their decision-making strategy and discuss opportunities to improve clarity of the DCE materials. Participants Participants in phases 4a and 5 included patients aged 18 or older who were admitted to a medicine team at BMC from June to July 2024 and spoke Spanish and Haitian Creole. They were approached in person by bilingual research staff, unless they were on infectious or suicide precautions or if their nurse reported they were unable to provide informed consent due to cognitive impairment. For phase 4b, all Spanish and Haitian Creole BMC interpreters were eligible to participate and were recruited via email. The participants were required to provide informed consent and pass a teach back. The project was approved by the Boston University Medical Center Institutional Review Board (IRB). Data Analysis To analyze the patient cognitive interviews conducted in Spanish, the PI (who is fluent in Spanish) reviewed audio recordings and created a summary table in English (Appendix 3). Much like the summary tables utilized in rapid qualitative analysis, each table was structured around the topic subheadings of the semi-structured interview guide. Since the PI was not fluent in Haitian Creole, the interviewer took notes in English during these interviews for the PI to review, which she used to create a summary table. To avoid misinterpretation, another researcher fluent in Haitian Creole reviewed the audio recordings and independently created summary tables, which the PI then merged with her own. Any discrepancies were resolved by group discussion. Summary tables were reviewed by the entire research team, including native speakers of both Spanish and Haitian Creole, within 48 hours of the interview to refine findings and determine revisions to DCE materials for the next cycle of cognitive interviews. When suggestions for new attributes or levels were suggested by participants, the PI conferred with content experts to determine whether the content of the DCE should be modified. The professional interpreter cognitive interviews were also audio recorded and analyzed using summary tables as described above. Insights were used to modify DCE materials before a final round of patient participant cognitive interviews. During pilot testing, research staff systematically documented participant interactions using REDCap. They recorded total completion time, frequency and types of clarification questions, and time comparison between early choice sets (second) and final choice sets to detect potential cognitive fatigue (whether participants began answering without deliberation as the DCE progressed). The PI reviewed all observation data within 24 hours of collection, followed by a team debriefing to evaluate necessary modifications before full implementation. Reflexivity All members of the data analysis team are multilingual and have lived experience in high diversity contexts. All are first or second-generation immigrants to the United States who have assisted family members with NDLP and/or low health literacy. Nonetheless, the study team did not represent all perspectives represented by the participants in this study. We engaged in regular team debriefings and reflexivity exercises to critically examine how our own experiences and biases might shape (or miss) the themes we identified. RESULTS Phase 1: Evidence synthesis In reviewing the literature, we identified numerous hospital discharge interventions that included speakers of the dominant language, however, there are few interventions including patients with NDLP and limited research on their preferences during discharge.( 21 , 24 , 30 ) Based on the minimal evidence base, we selected four attributes: ( 1 ) “Form of interpreter”; ( 2 ) “Family involvement in discharge teaching”; ( 3 ) “Form of written instructions”; and ( 4 ) “Deliverer of intervention” (Table 2 ). This preliminary draft of the DCE instrument included an opt out, meaning the participant could state that they would not elect either option A or B. Table 2 Overview of content changes to discrete choice experiment attributes and levels across the five phases of development Attribute: Brief Title Phase 1 Evidence synthesis Phases 2/3 End user input/expert review Phase 4a and 4b Pretesting Phase 5 Pilot testing Attribute 1: Language Form of interpreter level 1) telephone level 2) video level 3) in person level 4) family member ▪ Revised attribute to Way to overcome language barrier ▪ Drop level 4 “family member” ▪ Add level 4 “language concordant provider” ▪ Revise attribute description to Form of verbal communication ▪ Shorten level descriptions ▪ Reorder attribute to be last in choice set to avoid priming No changes Attribute 2: Family Family involvement in discharge teaching level 1) solo level 2) family/caregiver present in person ▪ Add level 3 “family/caregiver present by phone” ▪ Shorten level descriptions ▪ Revise attribute description to Family participation when nurse reviews discharge instructions No changes Attribute 3: Written instructions Form of written instructions level 1) written in preferred language level 2) written in English ▪ Add level 3 “audio recording” ▪ Revise wording of attribute to Written communication ▪ Shorten level descriptions ▪ Reword level 3 to “audio recording in my language” No changes Attribute 4 Deliverer of intervention : level 1) nurse level 2) physician level 3) community health worker Dropped Form of follow up after discharge: level 1) nurse from primary care office level 2) nurse from hospital level 3) number I can call ▪ Revise wording of attribute to Phone follow up after discharge ▪ Shorten level descriptions No changes Opt Out Included Dropped Appendix 1 provides more details about specific changes per iterative cycle and Appendix 2 provides word for word change to attribute titles and level descriptions Phase 2: End user input Our review of post-discharge qualitative interviews with Spanish and Haitian Creole speaking BMC patients (n = 37) validated the importance of attributes 1, 2, and 3. Revisions to the levels were made based on insights gained from the interviews. The title of attribute 2 (“Family involvement in discharge teaching”) was modified to incorporate the finding that family members often cannot be present at discharge, so the level of “family/caregiver present by phone” was added. For attribute 3 (“Form of written discharge instructions”), “audio recording” was added because some participants expressed satisfaction with audio recording of discharge instructions in their preferred language. One theme that was not addressed by the DCE was post-discharge care. Participants articulated significant challenges with systems navigation, including strategies for getting their questions answered and resolving issues post-discharge. We added attribute 4,“Form of follow up after discharge,” to address this theme. Phase 3: Expert input The DCE draft was reviewed with the expert team. They recommended eliminating the attribute “Deliverer of intervention” for two reasons. First, a paraprofessional such as a community health worker would not be qualified to deliver the intervention because it requires clinical expertise. Second, due to the intensive nature of collecting DCE data in person and with a language concordant administrator, adequate sample size was a major concern and maintaining a maximum of four attributes was important. Experts also recommended removing the opt out option. This was both to minimize missing data (selecting the opt out option reduces the number of data points for a participant) and because it did not reflect a real-world option. They also suggested removing the family member option for attribute 1 (“Form of interpreter”) because it is not legal unless a professional interpreter is also present to correct misunderstandings. Finally, due to evidence of a very strong preference for language concordant interactions, the option of a language-concordant provider was added, which required rephrasing attribute 1 to “Way to overcome language barrier.” A noted risk of adding this option is its potential to drive preference so strongly that it “crowd out” preferences for other attributes. Phase 4: Pretesting with patients and professional interpreters Demographics of the pretesting participants are summarized in Table 3 . Of the 21 participants, 62% were male and 62% were Spanish speaking. Almost all participants (95%) reported speaking English less than very well. Only 24% reported reading in their preferred language less than very well. Most reported a preference for language concordant health care providers (76%) while 14% preferred a phone interpreter, and 9.5% preferred a family member. We summarize our findings using the four domains assessed in pretesting: comprehension, presentation, elicitation and content. Table 3 Demographics of patient participants in discrete choice experiment pretesting and pilot testing Characteristic Phase 4a N = 21 Phase 5 N = 8 Age (years): mean (SD) 47.2 (12.3) 53.5 (15.8) Gender Male 13 (61.9%) 7 (87.5%) Female 8 (38.1%) 1 (12.5%) Preferred language Spanish 13 (61.9%) 4 (50.0%) Haitian Creole 8 (38.1%) 4 (50.0%) Language spoken most often at home Spanish 13 (61.9%) 4 (50.0%) Haitian Creole 8 (38.1%) 4 (50.0%) English 0 (0%) 0 (0%) Confidence in filling out forms Extremely confident 6 (28.6%) 1 (12.5%) Quite a bit 3 (14.3%) 1 (12.5%) Somewhat / A little bit / Not at all 12 (57.1%) 6 (75.0%) Self-reported fluency in English Understands less than “very well” 20 (95.2%) 8 (100%) Speaks less than “very well” 20 (95.2%) 8 (100%) Reads less than “very well” 20 (95.2%) 8 (100%) Self- reported ability to read in preferred language “Very well” 16 (76.2%) 4 (50%) Less than “very well” 5 (23.8%) 4 (50%) Prior experience with language facilitation Language concordant provider 19 (90.5%) 7 (87.5%) In person professional interpreter 15 (71.4%) 4 (50%) Phone professional interpreter 20 (95.2%) 8 (100%) Video professional interpreter 8 (38.1%) 1 (12.5%) Preference for language facilitation Language concordant provider 16 (76.2%) 4 (50.0%) In person professional interpreter 0 (0%) 0 (0%) Phone professional interpreter 3 (14.3%) 3 (37.5%) Video professional interpreter 0 (0%) 0 (0%) Family 2 (9.5%) 1 (12.5%) Table 4 summarizes the challenges faced in pretesting. In the initial interviews, many participants had difficulty comprehending the attributes and levels as intended. For example, many participants thought that attribute 4 (“Follow up after discharge ”) referred to an in-person appointment with their primary care doctor instead of a phone call from a nurse. Rephrasing this attribute to “Phone call after discharge” addressed this misinterpretation. Similar modifications were needed to clarify that the audio recording was in the patient’s preferred language (Appendix 2 and 3). Another comprehension error was related to the use of a professional interpreter. In the early draft of materials, the first attribute presented in the choice set was “Verbal communication”. This order primed some participants to interpret subsequent attributes 2 to 4 as asking about the type of professional interpreter used for that task. We remedied this by presenting “Verbal communication” last in the choice set. We also moved survey questions about their explicit preferences for type of interpreter to after the DCE was completed. Table 4 Domains assessed in pretesting and strategies used in high linguistic diversity context to achieve understanding Domain Strategies Comprehension Understanding of attributes, levels, and decisional context Use native speaker research assistants Rapid cycle iterative changes to materials using sequential group analysis Presentation Text descriptions and visuals Use visuals to augment descriptions of attributes and/or levels Strike balance in length to adequately explain but not promote disengagement and only use of visuals Elicitation Process of making a choice Emphasize choice as a “set” Warm up choice card for everyday decision Content Relevance of topic, attributes, and levels Explicitly ask about “missing” levels or attributes that are important to participant Ensure aspects of health system not universally understood are addressed in introductory script The presentation of attributes, levels, and choice sets also required significant revision. When attribute and level descriptions were too long, participants looked only at the images, which alone did not communicate the full meaning. Shortening attribute and level labels and placing them directly under images improved understanding. Overall, participants reported that images were helpful and had a slight preference for photos over pictographs. However, some participants were not able to discern the roles of different people in the photos. In attribute 1 “Form of verbal communication”, one photo is meant to depict a patient, a nurse, and an in-person interpreter but these roles were not obvious to some participants. Photos were revised to include written labels for each person which improved understanding (Fig. 4 ). In addition, our initial photos aimed to capture diversity in skin color of the patient and nurse. However, this led to confusion when comparing photos in which these individuals represented different roles. As such, participants preferred that the same cast of characters be used across all photos. Most participants did not grasp the task of elicitation during the pretest with the initial version of DCE materials. Instead of selecting Option A or B as a set, participants often stated a preference for a single level instead of considering the tradeoffs of all levels displayed as a package. Revising the introductory script and placing a large rectangular box around Option A and B (Fig. 4 ) improved understanding, but not for all participants. In response, we designed a “warm up” DCE choice card (Appendix 3) that used the everyday example of selecting a grocery store. The warmup was integrated into the introductory script, and participants uniformly grasped the concept of tradeoffs and task of selecting A or B in its entirety from each choice set. Querying the participant’s process of elicitation also detected some cases in which personal preference was not driving decision making. For example, two participants reported that they did not select certain levels because they did not believe the hospital would offer them. We addressed this by revising the introductory script to explicitly ask for personal preference and not what was felt likely to be offered by the hospital. The pretesting identified preferences regarding the DCE content. In both Spanish and Haitian Creole groups, some individuals requested the intervention be delivered by a physician, instead of a nurse. After discussion the study team decided not to revise the levels as it would not be a feasible adaptation in real world implementation. Instead, the introductory script was modified to specifically describe this as a nurse-delivered intervention. The second preference that emerged was to have family serve as an interpreter for discharge teaching. Ultimately the consensus in discussion with experts was not to include this option as it could not be legally offered as part of care. During the interpreter interviews, both Spanish and Haitian Creole interpreters felt that all word choices for levels and attributes were understandable and accurate. They provided feedback on how to revise the introductory script for improved clarity, including suggestions for explaining “theoretical” and “elicitation”. Step 5: Pilot testing Demographics of participants were similar to those including in pretesting and are summarized in Table 2 . During the pilot tests, none of the participants expressed confusion after the introductory script. While completing the DCE on the tablet, there was a mix of participants who elected to hold the tablet and select their preferred option themselves after reviewing options silently, while others requested that the administrator read each choice set aloud and stated aloud their preferred option in each pairwise comparison. Observations did not reveal signs of cognitive fatigue with nearly equal time needed to complete the second and last choice set. Participants reported that both cognitive demand and task difficulty were low. When questioned about decision making, all participants were able to articulate the tradeoffs behind their decisions. As such, the DCE content was not modified after the pilot testing. The initial and final version of the DCE (translated into English) are presented in Fig. 4 for comparison. DISCUSSION This study aimed to address the methodological challenges of conducting DCEs with individuals who have a NDLP, a population often excluded from health services research. We designed a rigorous, culturally and linguistically tailored approach to DCE development and applied it to a study of preferences for hospital discharge supports. We found that participants with NDLP faced challenges in understanding trade-offs, abstract attributes, and initial draft of visual materials in the DCE. Through features of our methodology—namely, bilingual interviewer-led administration, cognitive interviews with both patients and interpreters, warm-up elicitation tasks, heavy use of labeled visual descriptions of attributes and levels, and rapid-cycle analysis for iterative refinement—that demonstrate that it is possible to achieve good comprehension of DCE content through this tailored approach suitable for high diversity contexts. Pilot testing of the DCE instrument designed using our proposed methodology demonstrated high comprehension with no cognitive fatigue, as participants successfully articulated decision-making strategies consistent with their stated preferences. Most of the literature on administering DCEs to populations who face linguistic and literacy barriers has been conducted with health workers in low- and middle-income countries.( 31 – 36 ) For instance, Hanson and colleagues conducted a DCE with community members in Zambia. To improve understanding, they used a picture board with symbolic representations of the attributes and levels. Similarly, Brunie and colleagues conducted a DCE with community health workers in Uganda and chose to develop the DCE instrument in English and translate the final versions into three local languages. However, no cognitive testing was done to ensure equivalent understanding in these three languages. Our results can inform the development of DCEs in these contexts by offering a methodology for careful adaptation and testing of DCE instruments across multiple languages to ensure content validity and equity. The existing DCE literature provides limited guidance on effective approaches to collecting and analyzing data for pretesting in phase 4 (which includes developing introductory scripts, attribute descriptions, and visual materials to support comprehension). Given the lack of existing models, including those that are feasible across linguistic and cultural communication barriers, we suggest and demonstrate a rapid cycle approach to cognitive interviews involving data collection, analysis, and iterative changes, utilizing summary tables from rapid qualitative analysis.( 37 ) Our iterative pretesting process revealed consistent misunderstandings of key attributes such as “follow-up after discharge” and “verbal communication,” which were only resolved through revisions grounded in real-time participant feedback. These findings underscore the importance of combining cultural competence with methodological rigor when engaging diverse populations. The clinical significance of our study lies in its potential to improve health equity. Transitions from hospital to home are a well-documented point of vulnerability for patients, and individuals with NDLP face disproportionate risk due to language barriers and systemic inequities.( 38 – 40 ) By identifying preferences for communication modalities, instructional formats, and family involvement in discharge processes, our approach supports the development of more responsive, person-centered interventions. Incorporating these preferences into intervention design is a strategy to improve patient comprehension, satisfaction, and safety during care transitions.( 22 ) Even when responding to all preferences is not feasible—for example, a physician-led hospital discharge intervention would be significantly less scalable—our findings highlight the importance of explaining the value and expertise of nursing staff among populations whose experiences with healthcare systems in other countries may differ. Innovative next steps include exploring the feasibility and relative advantage of using a video introduction to the DCE or testing remote DCE administration with Research Assistant support via phone or video to reduce the barriers presented by in-person completion. This study has several limitations. First, the methodology was applied to a DCE including two languages. Expanding to three or more languages would increase complexity and should be explored in future studies as this could require expanding the number of levels offered, necessitating a larger sample size. Second, only one interpreter from each linguistic group was asked to review materials as the final step in pretesting. Engaging multiple interpreters or a more traditional forward and back translation approach could be tested, but would slow down the rapid cycle analytic approach. Third, this methodology was demonstrated at a single site, which may limit generalizability. Further testing across more contexts and with different subject matters would help validate and refine the proposed approach. CONCLUSIONS We provide a practical framework for designing and conducting DCEs with populations who speak non-dominant languages. Our approach addresses longstanding methodological gaps by integrating bilingual data collection, culturally informed instrument refinement, and pilot testing in real-world conditions. These findings support the feasibility of conducting rigorous DCEs in high-diversity settings and highlight the value of inclusive design to inform equitable implementation. Future research should expand this work by testing the generalizability of our approach across additional languages, healthcare contexts, and decision-making scenarios. Efforts to quantify how tailoring based on DCE findings affects intervention outcomes will also be critical to advancing equity in intervention implementation. Abbreviations AVS After Visit Summary BMC Boston Medical Center LEP limited English proficiency NDLP non-English language preference Declarations Ethics approval and consent to participate This study was approved by the Boston University Medical Center IRB. Consent for publication Not applicable Clinical trial number Not applicable Availability of data and materials Datasets will be available on reasonable request from the corresponding author (KA). Competing interests The authors declare that they have no competing interests. Funding Dr. Austad is supported by a K23 from the National Institute for Minority Health and Health Disparities (K23 MD019068). Authors’ contributions KA conceived of the study. SM, BJ, AF, RS provided input on study design. KA, NL, DBH, SL, and KP collected data and conducted the analysis. SM, BJ, AF, MLD, and RS provided input on the results. KA led the drafting of the manuscript with significant contributions from all co-authors. All authors read and approved the final manuscript. Acknowledgements We would like to thank Director of Interpreter Services Alegna Zavatti for her collaboration as well as the interpreters who participated in the cognitive interviews. We would also like to thank Amy Ries for her input as a professional writer on the organization and prose of the manuscript. Authors’ information KA is a Family Medicine hospitalist and implementation science researcher at Boston Medical Center. References Zhang Y, Anh Ho TQ, Terris-Prestholt F, Quaife M, De Bekker-Grob E, Vickerman P, et al. Prediction accuracy of discrete choice experiments in health-related research: a systematic review and meta-analysis. eClinicalMedicine. 2025;79:102965. Ryan M, Gerard K, Amaya-Amaya M. Using Discrete Choice Experiments to Value Health and Health Care. 1st ed. Springer Dordrecht; 2007. p. 256. Wang H, Rowen DL, Brazier JE, Jiang L. Discrete Choice Experiments in Health State Valuation: A Systematic Review of Progress and New Trends. Appl Health Econ Health Policy. 2023;21(3):405–18. Quaife M, Terris-Prestholt F, Di Tanna GL, Vickerman P. How well do discrete choice experiments predict health choices? A systematic review and meta-analysis of external validity. Eur J Health Econ. 2018;19(8):1053–66. Irie WC, Kerkhoff A, Kim HY, Geng E, Eshun-Wilson I. Using stated preference methods to facilitate knowledge translation in implementation science. Implement Sci Commun. 2024;5(1):32. s43058-024-00554–3. Salloum RG, Shenkman EA, Louviere JJ, Chambers DA. Application of discrete choice experiments to enhance stakeholder engagement as a strategy for advancing implementation: A systematic review. Implement Sci. 2017;12(1):1–12. Ostermann J, Njau B, Hobbie A, Mtuy T, Masaki ML, Shayo A, et al. Using discrete choice experiments to design interventions for heterogeneous preferences: protocol for a pragmatic randomised controlled trial of a preference-informed, heterogeneity-focused, HIV testing offer for high-risk populations. BMJ Open. 2020;10(11):e039313. Lau AS, Huey SJ, Baumann AA. Advances in the adaptation and implementation of evidence-based interventions for historically marginalized groups. Behav Res Ther. 2023;168:104377. Pineros-Leano M, Desrosiers A, Piñeros-Leaño N, Moya A, Canizares-Escobar C, Tam L, et al. Cultural adaptation of an evidence-based intervention to address mental health among youth affected by armed conflict in Colombia: An application of the ADAPT-ITT approach and FRAME-IS reporting protocols. Camb prisms Glob ment health. 2024;11:e114. Healey P, Stager ML, Woodmass K, Dettlaff AJ, Vergara A, Janke R, et al. Cultural adaptations to augment health and mental health services: a systematic review. BMC Health Serv Res. 2017;17(1):1–26. Chin MH, Walters AE, Cook SC, Huang ES. Interventions to reduce racial and ethnic disparities in health care. Med Care Res Rev. 2007;64(5 SUPPL.). Austad K, Cordova-Ramos EG, Fernandez A, Drainoni ML. Lost in translation: advancing intervention adaptation for populations with non-dominant language preference in high diversity settings. Implement Sci Commun. 2025;6(1):66. Squires A, Gerchow L, Ma C, Liang E, Trachtenberg M, Miner S. A multi-language qualitative study of limited English proficiency patient experiences in the United States. PEC Innov. 2023;2:100177. Liamputtong P. Doing Research in a Cross-Cultural Context: Methodological and Ethical Challenges. In: Diener E, Glatzer W, Moum T, Sprangers M, Vogel J, Veenhoven R, editors. Springer; 2008. pp. 3–20. Available from: http://link.springer.com/ 10.1007/978-1-4020-8567-3_1 Mangham LJ, Hanson K, McPake B. How to do (or not to do)… Designing a discrete choice experiment for application in a low-income country. Health Policy Plann. 2009;24(2):151–8. Wang Y, Wang Z, Wang Z, Li X, Pang X, Wang S. Application of Discrete Choice Experiment in Health Care: A Bibliometric Analysis. Front Public Health. 2021;9. Naveen P, Trojovský P. Overview and challenges of machine translation for contextually appropriate translations. iScience. 2024;27(10):110878. Sentell T, Braun KL. Low health literacy, limited English proficiency, and health status in Asians, Latinos, and other racial/ethnic groups in California. In: J Health Communication. 2012. pp. 82–99. Moor S, Tusubira AK, Akiteng AR, Hsieh E, Ngaruiya C, Rabin TL et al. Developing a Discrete Choice Experiment to Understand Patient Preferences in Resource-Limited Settings: a Six-Step Guide. Available from: https://doi.org/10.1101/2020.10.05.20206557 Kohler RE, Lee CN, Gopal S, Reeve BB, Weiner BJ, Wheeler SB. Developing a discrete choice experiment in Malawi: Eliciting preferences for breast cancer early detection services. Patient Prefer Adherence. 2015;9:1459–72. Tyler N, Hodkinson A, Planner C, Angelakis I, Keyworth C, Hall A, et al. Transitional Care Interventions from Hospital to Community to Reduce Health Care Use and Improve Patient Outcomes: A Systematic Review and Network Meta-Analysis. JAMA Netw Open. 2023;6(11):E2344825. Austad K, Jack BW. Linguistic and Cultural Competence at Hospital Discharge. J Healthc Manage Stand. 2023;3(1):1–16. Campoamor NB, Guerrini CJ, Brooks WB, Bridges JFP, Crossnohere NL. Pretesting Discrete-Choice Experiments: A Guide for Researchers. Patient. 2024;17(2):109–20. Becker C, Zumbrunn S, Beck K, Vincent A, Loretz N, Müller J et al. Interventions to Improve Communication at Hospital Discharge and Rates of Readmission: A Systematic Review and Meta-analysis. JAMA Netw Open. 2021;4(8). Balaban RB, Galbraith AA, Burns ME, Vialle-Valentin CE, Larochelle MR, Ross-Degnan D. A Patient Navigator Intervention to Reduce Hospital Readmissions among High-Risk Safety-Net Patients: A Randomized Controlled Trial. J Gen Intern Med. 2015;30(7):907–15. Balaban RB, Weissman JS, Samuel PA, Woolhandler S. Redefining and redesigning hospital discharge to enhance patient care: A randomized controlled study. J Gen Intern Med. 2008;23(8):1228–33. Agnello DM, Anand-Kumar V, An Q, De Boer J, Delfmann LR, Longworth GR et al. Co-creation methods for public health research — characteristics, benefits, and challenges: a Health CASCADE scoping review. BMC Med Res Methodol [Internet]. 2025 Mar 6 [cited 2025 May 23];25(1). Available from: https://bmcmedresmethodol.biomedcentral.com/articles/ 10.1186/s12874-025-02514-4 Lee WL, Chinna K, Lim Abdullah K, Zainal Abidin I. The forward-backward and dual‐panel translation methods are comparable in producing semantic equivalent versions of a heart quality of life questionnaire. Int J of Nursing Practice [Internet]. 2019 Feb [cited 2025 May 23];25(1). Available from: https://onlinelibrary.wiley.com/doi/ 10.1111/ijn.12715 Lighthouse Studio. Choice-Based Conjoint Module. Provo, UT: Sawtooth Software, Inc.; 2023. Kansagara D, Chiovaro JC, Kagen D, Jencks S, Rhyne K, O’Neil M et al. Transitions of care from hospital to home: An overview of systematic reviews and recommendations for improving transitional care in the Veteral Health Administration. Transitions of Care from Hospital to Home: An Overview of Systematic Reviews and Recommendations for Improving Transitional Care in the Veterans Health Administration [Internet]. 2015; Available from: http://www.ncbi.nlm.nih.gov/pubmed/26312362 Brunie A, Chen M, Akol A. Qualitative Assessment of the Application of a Discrete Choice Experiment With Community Health Workers in Uganda: Aligning Incentives With Preferences. Glob Health Sci Pract. 2016;4(4):684–93. Abdel-All M, Angell B, Jan S, Howell M, Howard K, Abimbola S, et al. What do community health workers want? Findings of a discrete choice experiment among Accredited Social Health Activists (ASHAs) in India. BMJ Glob Health. 2019;4(3):e001509. Honda A, Vio F. Incentives for non-physician health professionals to work in the rural and remote areas of Mozambique—a discrete choice experiment for eliciting job preferences. Hum Resour Health [Internet]. 2015 Dec [cited 2025 May 23];13(1). Available from: https://human-resources-health.biomedcentral.com/articles/ 10.1186/s12960-015-0015-5 Blaauw D, Erasmus E, Pagaiya N, Tangcharoensathein V, Mullei K, Mudhune S, et al. Policy interventions that attract nurses to rural areas: a multicountry discrete choice experiment. Bull World Health Organ. 2010;88(5):350–6. vujicic marko. alfano marco, ryan mandy, wesseh CS, Brown-Annan J. Policy options to attract nurses to rural Liberia: Evidence from a discrete choice experiment. HNP; 2010 Nov. Shiratori S, Agyekum EO, Shibanuma A, Oduro A, Okawa S, Enuameh Y et al. Motivation and incentive preferences of community health officers in Ghana: an economic behavioral experiment approach. Hum Resour Health [Internet]. 2016 Dec [cited 2025 May 23];14(1). Available from: http://human-resources-health.biomedcentral.com/articles/ 10.1186/s12960-016-0148-1 Kowalski CP, Nevedal AL, Finley EP, Young JP, Lewinski AA, Midboe AM, et al. Planning for and Assessing Rigor in Rapid Qualitative Analysis (PARRQA): a consensus-based framework for designing, conducting, and reporting. Implement Sci. 2024;19(1):71. Khoong EC, Sherwin EB, Harrison JD, Wheeler M, Shah SJ, Mourad M et al. Impact of standardized, language-concordant hospital discharge instructions on postdischarge medication questions. J Hosp Med. 2023. Woods AP, Alonso A, Duraiswamy S, Ceraolo C, Feeney T, Gunn CM, et al. Limited English Proficiency and Clinical Outcomes After Hospital-Based Care in English-Speaking Countries: a Systematic Review. J Gen Intern Med. 2022;37(8):2050–61. Chauhan A, Walton M, Manias E, Walpola RL, Seale H, Latanik M, et al. The safety of health care for ethnic minority patients: A systematic review. Int J Equity Health. 2020;19(1):1–25. Additional Declarations No competing interests reported. Supplementary Files Appendix5.23.2025.docx ADDITIONAL FILES Appendix 1: Detailed description of changes to DCE instrument and introductory script during pretesting Appendix 2: Evolution of working of attribute and level descriptions Appendix 3: Comparison of version 1 (A) to version 8 (B) level descriptions for “Verbal communication” attribute Appendix 4: Example of warm up discrete choice experiment to help participants understand elicitation Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-6976834","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":484169423,"identity":"5b2865a6-c2c8-42b3-88e0-cbfd7680cd6b","order_by":0,"name":"Kirsten Austad","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA90lEQVRIiWNgGAWjYBACCWYULptNAnFaDkDYjA0MbGlEaGFA1XKYsBbJdt7Hnz9U3GHgn5H+/MGPsvN5uu0HGB9X/MKtRZqZ3UziwJlnDBI3cgwbe87dLjY7k8BseLYPtxY5ZjY2hoNthxkYbuQwNvC23U7cdoOBTbKxB68W5g8H/x1mkL+R/rDxb9s5wlqkmdkYJA42HGYwuJFg2MzbdgCipeEHHu83s7FJnDl2mMfwzBvD2TLnkoF+SWw2bGzArUXi/DHmDxU1h+Xkjqc/+PimzC7P7Pjhgw8b/uDWAgM8SGxg/DC2EdaCDoiwZRSMglEwCkYMAACJzVbDPVZ0+QAAAABJRU5ErkJggg==","orcid":"","institution":"Boston University Chobanian and Avedisian School of Medicine","correspondingAuthor":true,"prefix":"","firstName":"Kirsten","middleName":"","lastName":"Austad","suffix":""},{"id":484169428,"identity":"a40775fe-343a-447e-a227-f9a74bbd7647","order_by":1,"name":"Noelia Lugo","email":"","orcid":"","institution":"Boston University Chobanian and Avedisian School of Medicine","correspondingAuthor":false,"prefix":"","firstName":"Noelia","middleName":"","lastName":"Lugo","suffix":""},{"id":484169431,"identity":"d30b9c3a-d392-4f77-b9b3-e5bea4f92094","order_by":2,"name":"Diana Bautista-Hurtado","email":"","orcid":"","institution":"Boston University Chobanian and Avedisian School of Medicine","correspondingAuthor":false,"prefix":"","firstName":"Diana","middleName":"","lastName":"Bautista-Hurtado","suffix":""},{"id":484169434,"identity":"76df6125-ec6b-47e7-b1b5-7889eda1f491","order_by":3,"name":"Stephane Labossiere","email":"","orcid":"","institution":"Northeastern University","correspondingAuthor":false,"prefix":"","firstName":"Stephane","middleName":"","lastName":"Labossiere","suffix":""},{"id":484169438,"identity":"5ecd4630-4b3f-4f6c-8163-e9a1731d12fa","order_by":4,"name":"Kathryn Pluta","email":"","orcid":"","institution":"University of Florida","correspondingAuthor":false,"prefix":"","firstName":"Kathryn","middleName":"","lastName":"Pluta","suffix":""},{"id":484169440,"identity":"ed9668c6-280c-4602-b5c3-49199d73f5ce","order_by":5,"name":"Suzanne Mitchell","email":"","orcid":"","institution":"UMass Chan Medical School","correspondingAuthor":false,"prefix":"","firstName":"Suzanne","middleName":"","lastName":"Mitchell","suffix":""},{"id":484169446,"identity":"413e65b2-ba45-44d1-873c-45b8afd71607","order_by":6,"name":"Brian W. Jack","email":"","orcid":"","institution":"Boston University Chobanian and Avedisian School of Medicine","correspondingAuthor":false,"prefix":"","firstName":"Brian","middleName":"W.","lastName":"Jack","suffix":""},{"id":484169452,"identity":"5da4231b-eb5d-4fab-b879-9e5538ca979b","order_by":7,"name":"Alicia Fernandez","email":"","orcid":"","institution":"University of California San Francisco","correspondingAuthor":false,"prefix":"","firstName":"Alicia","middleName":"","lastName":"Fernandez","suffix":""},{"id":484169457,"identity":"77717e1d-7afe-44a7-94dd-745b57db2155","order_by":8,"name":"Ramzi G. Salloum","email":"","orcid":"","institution":"University of Florida","correspondingAuthor":false,"prefix":"","firstName":"Ramzi","middleName":"G.","lastName":"Salloum","suffix":""},{"id":484169460,"identity":"f726b4f0-8ee9-4cdb-b6f8-d07fc7007142","order_by":9,"name":"Mari-Lynn Drainoni","email":"","orcid":"","institution":"Boston University Chobanian and Avedisian School of Medicine","correspondingAuthor":false,"prefix":"","firstName":"Mari-Lynn","middleName":"","lastName":"Drainoni","suffix":""}],"badges":[],"createdAt":"2025-06-25 16:53:23","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-6976834/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6976834/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":87214201,"identity":"c334c688-256c-49fa-ba28-986be048d032","added_by":"auto","created_at":"2025-07-21 15:09:51","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":421367,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eDiscrete choice experiment choice set for exploring decision-making about selecting a grocery store\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-6976834/v1/22090ff22033c507c961fcbd.png"},{"id":87214778,"identity":"8aaeea35-4d78-4f36-a57a-5a9a5b0a5aaf","added_by":"auto","created_at":"2025-07-21 15:17:51","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":319518,"visible":true,"origin":"","legend":"\u003cp\u003eProposed process for developing a discrete choice experiment to understand preferences of patients with a non-English language preference\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-6976834/v1/37a87bfbe62b1138443f429b.png"},{"id":87214203,"identity":"6aa29125-bb18-438a-ab11-61710a2dd8fd","added_by":"auto","created_at":"2025-07-21 15:09:51","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":136495,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eApproach to discrete choice experiment pretesting for a multi-lingual patient population: Cognitive interview data collection, analysis, and iterative revision\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-6976834/v1/7f3635f831e13ca0c754c723.png"},{"id":87214779,"identity":"50d5e401-03ac-4bc7-b0c3-2ed02df92c60","added_by":"auto","created_at":"2025-07-21 15:17:51","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":388983,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eExample scenario card for the pilot tested discrete choice experiment\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"4.png","url":"https://assets-eu.researchsquare.com/files/rs-6976834/v1/d4757cf2ef118503ae42a969.png"},{"id":104429977,"identity":"1bbbc206-3b65-4eb3-a0d1-d3aea6532af7","added_by":"auto","created_at":"2026-03-11 15:27:29","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2095578,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6976834/v1/87e1676e-ada8-436c-acbe-2fba0c2d30c7.pdf"},{"id":87214208,"identity":"63987acd-e879-4f85-ba00-a94d56d03a2e","added_by":"auto","created_at":"2025-07-21 15:09:51","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":3050319,"visible":true,"origin":"","legend":"\u003cp\u003eADDITIONAL FILES\u003c/p\u003e\n\u003cp\u003eAppendix 1: Detailed description of changes to DCE instrument and introductory script during pretesting\u003c/p\u003e\n\u003cp\u003eAppendix 2: Evolution of working of attribute and level descriptions\u003c/p\u003e\n\u003cp\u003eAppendix 3: Comparison of version 1 (A) to version 8 (B) level descriptions for “Verbal communication” attribute\u003c/p\u003e\n\u003cp\u003eAppendix 4: Example of warm up discrete choice experiment to help participants understand elicitation\u003c/p\u003e","description":"","filename":"Appendix5.23.2025.docx","url":"https://assets-eu.researchsquare.com/files/rs-6976834/v1/5f270a3f2268a35a4a5c6211.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Developing Discrete Choice Experiments for Populations with Non-Dominant Language Preference: A Methodological Framework and Case Example","fulltext":[{"header":"BACKGROUND","content":"\u003cp\u003eDiscrete choice experiments (DCEs) are a powerful tool for eliciting the preferences of individuals and are increasingly applied to evaluate health care decision-making.(\u003cspan class=\"CitationRef\"\u003e1\u003c/span\u003e, \u003cspan class=\"CitationRef\"\u003e2\u003c/span\u003e) Participants are presented with hypothetical scenarios and asked to select their preferred options from two or more alternatives, simulating real world decision-making. Participants must weigh competing factors and make deliberate trade-offs between different features. Through probabilistic models grounded in consumer and random utility theory, it is possible to quantify the importance of different attributes, the utility participants assign to various choices, and the trade-offs they are willing to make between attribute levels.(\u003cspan class=\"CitationRef\"\u003e3\u003c/span\u003e, \u003cspan class=\"CitationRef\"\u003e4\u003c/span\u003e)\u003c/p\u003e\n\u003cp\u003eWithin the field of implementation science, DCEs have been used to make implementation efforts more responsive to end-users by refining and tailoring implementation strategies to the preferences of a given population.(\u003cspan class=\"CitationRef\"\u003e5\u003c/span\u003e, \u003cspan class=\"CitationRef\"\u003e6\u003c/span\u003e) A promising yet underutilized application of DCEs involves engaging minoritized populations in adapting interventions to better meet their needs and preferences.(\u003cspan class=\"CitationRef\"\u003e7\u003c/span\u003e) Many health interventions are not tailored to the needs of linguistically and culturally diverse populations, leading to poor fit, limited uptake, and reduced real-world effectiveness.(\u003cspan class=\"CitationRef\"\u003e8\u003c/span\u003e\u0026ndash;\u003cspan class=\"CitationRef\"\u003e10\u003c/span\u003e) DCEs offer a systematic method for capturing the specific needs and preferences of diverse populations that can be used to tailor interventions, a strategy shown to reduce inequities in health outcomes.(\u003cspan class=\"CitationRef\"\u003e11\u003c/span\u003e)\u003c/p\u003e\n\u003cp\u003eAs linguistic and cultural diversity continues to grow throughout the world, developing interventions that fit the needs of a single demographic does not align with real world health care delivery, where multiple linguistic and cultural groups are often served simultaneously in one setting.(\u003cspan class=\"CitationRef\"\u003e10\u003c/span\u003e, \u003cspan class=\"CitationRef\"\u003e12\u003c/span\u003e) In these high diversity contexts, a DCE needs to be designed and deployed in multiple languages, but this introduces methodological challenges that threaten the feasibility of conducting a DCE and reliability of the results.(\u003cspan class=\"CitationRef\"\u003e13\u003c/span\u003e, \u003cspan class=\"CitationRef\"\u003e14\u003c/span\u003e) Currently there is a lack of up-to-date guidance for conducting DCEs in high diversity contexts.(\u003cspan class=\"CitationRef\"\u003e15\u003c/span\u003e) Without rigorous development of cross-linguistic research processes, DCEs risk introducing biases that reduce generalizability of the findings, and compromise the validity of insights, ultimately undermining efforts to ensure equity for individuals with a non-dominant language preference (NDLP).\u003c/p\u003e\n\u003cp\u003eThe goals of this manuscript are to provide practical solutions to address the challenges of conducting DCEs in linguistically and culturally diverse populations. We first summarize these challenges and propose solutions to address them. We then propose modifications to the existing approach to DCE development, including a novel approach to pretesting the DCE instrument in a linguistically and culturally diverse patient population. Finally, we present a case to illustrate our methodology.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eDesigning and Conducting DCEs in Populations with NDLP: Anticipated Challenges\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eTo illustrate how a DCE is operationalized, Fig. \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e depicts an example choice set, a presentation of two hypothetical scenarios requiring participants to select their preferred option based on trade-offs between different characteristics. In this example, the choice set measures the importance of characteristics weighed when choosing a grocery store. Two hypothetical grocery stores are defined according to discrete characteristics (or attributes), such as hours of operation, transportation to the store, and available food types. Each of these attributes has multiple options (or levels). For example, for the attribute of store hours the levels shown are \u0026ldquo;open only in the morning\u0026rdquo; and \u0026ldquo;open only in the afternoon.\u0026rdquo; Participants are asked to select option A or B as a set, which requires making tradeoffs between the levels displayed as a package. An experimental design is used to systematically vary the scenarios shown to each participant, since the total number of attribute-level permutations are too great to test.(\u003cspan class=\"CitationRef\"\u003e16\u003c/span\u003e)\u003c/p\u003e\n\u003cp\u003eGiven the complexity of the task, designing and conducting a DCE with culturally and linguistically diverse populations presents unique challenges. Specific challenges, which are summarized in Table \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e, include: (\u003cspan class=\"CitationRef\"\u003e1\u003c/span\u003e) effectively communicating the DCE content and task, (\u003cspan class=\"CitationRef\"\u003e2\u003c/span\u003e) developing culturally and linguistically appropriate content that maintains equivalent meaning, (\u003cspan class=\"CitationRef\"\u003e3\u003c/span\u003e) providing equitable access to completing the DCE regardless of internet connectivity and electronic device availability.\u003c/p\u003e\n\u003cp\u003e\u003cbr\u003e\u003c/p\u003e\n\u003cdiv class=\"gridtable\"\u003e\n \u003ctable id=\"Tab1\" border=\"1\"\u003e\n \u003ccaption language=\"En\"\u003e\n \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e\n \u003cdiv class=\"CaptionContent\"\u003e\n \u003cp\u003eConsiderations in planning a discrete choice experiment (DCE) in a high linguistic diversity contexts\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 294px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eConsideration\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 561px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eRationale\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 294px;\"\u003e\n \u003cp\u003eAssemble a diverse research team\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 561px;\"\u003e\n \u003cp\u003eQualitative methods, such as cognitive interviewing, used to design the DCE instrument are better accomplished via language concordant data collection and data analysis and interpreter by a multidisciplinary research team\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 294px;\"\u003e\n \u003cp\u003eInvolve professional interpreters\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 561px;\"\u003e\n \u003cp\u003eProfessional interpreters can provide valuable insights to word choice and phrasing to ensure the content and description of elicitation are understood by a wide range of speakers of language is critical\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 294px;\"\u003e\n \u003cp\u003eUtilize strategies to overcome linguistic, literacy, and health literacy barriers\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 561px;\"\u003e\n \u003cp\u003eDCEs most often convey attributes and levels with text only; use of visuals, such as photos, pictographs, or even videos may improve comprehension\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 294px;\"\u003e\n \u003cp\u003eEnsure cultural appropriateness of attributes and levels\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 561px;\"\u003e\n \u003cp\u003eIf key determinants of choice in real life are not represented in DCE the results will not represent real-world decision making\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 294px;\"\u003e\n \u003cp\u003eEngage with end-users early\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 561px;\"\u003e\n \u003cp\u003eThere may be little qualitative literature to inform attribute and level selection in specific group of interest; to ensure content is culturally appropriate and inclusive DCE design should ensure the voices of those with NDLP are included\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 294px;\"\u003e\n \u003cp\u003eInclude language concordance as an attribute or level in DCE\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 561px;\"\u003e\n \u003cp\u003ePatients with NDLP have a strong preference for language concordant care. Understanding how much they are willing to tradeoff other preferences to achieve this can directly inform where to invest resources in adapting interventions.\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 294px;\"\u003e\n \u003cp\u003eDevelop and refine the description of decisional context\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 561px;\"\u003e\n \u003cp\u003eParticipants may come from different contexts where health systems work differently and may need more explanation to understand trade offs\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 294px;\"\u003e\n \u003cp\u003eSample size\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 561px;\"\u003e\n \u003cp\u003eRecruiting a large sample of the target population with NDLP may be challenging. Limiting the number of attributes and levels tested may be necessary.\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 294px;\"\u003e\n \u003cp\u003eHow to administer DCE\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 561px;\"\u003e\n \u003cp\u003eDCEs are most often administered via a survey link to be completed independently on a computer or phone by the participation; having trained research staff administer the DCE in person may improve comprehension and allow recruitment of populations otherwise excluded\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 294px;\"\u003e\n \u003cp\u003eCognitive burden on participants\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 561px;\"\u003e\n \u003cp\u003eThose who face linguistic and literacy barriers may be more cognitively stressed by a DCE. To minimize cognitive burden, consider pairwise comparisons only (A vs B), limit number of choice sets shown to each participant, and consider administration and amount of text (as described above)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n \u003cp\u003e\u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003eAnticipated Challenge 1: Understanding\u003c/span\u003e\u003c/p\u003e\n \u003cp\u003eTypical approaches to orienting participants to a DCE may not adequately address the communication needs of those with NDLP due to barriers posed by limited fluency in the dominant language. While professional translation of written DCE content is one solution, it may fail to capture the intended meaning or use phrasing that is not understood by the average speaker.(\u003cspan class=\"CitationRef\"\u003e17\u003c/span\u003e) For widely spoken languages such as Spanish, difference in word meaning and colloquial usage across Spanish-speaking countries may compromise understanding. In a high diversity context where the DCE will be completed by speakers of different languages, the translation of attributes and levels must result in uniform understanding across linguistic groups.\u003c/p\u003e\n \u003cp\u003eAdditional barriers to understanding for individuals with NDLP include low literacy, the inability to read written text at a level needed for everyday tasks, and low health literacy, the limited ability to understand and use health information. In the United States, individuals with a non-English language preference are more than three times as likely to have low health literacy compared to their English fluent counterparts.(\u003cspan class=\"CitationRef\"\u003e18\u003c/span\u003e) Because many DCEs are self-administered via an written survey, literacy barriers may preclude participation or impede full understanding for those with NDLP. DCEs conducted in settings where low literacy is common have addressed this by using alternatives to traditional text heavy DCEs, such as use of visual depictions of attributes and levels.(\u003cspan class=\"CitationRef\"\u003e19\u003c/span\u003e, \u003cspan class=\"CitationRef\"\u003e20\u003c/span\u003e)\u003c/p\u003e\n \u003cp\u003eIndividuals with NDLP may also face challenges during the elicitation of tradeoffs in the DCE. While everyone considers tradeoffs in everyday decisions, explicitly asking participants to make these tradeoffs in an abstract task may create confusion. Introductory materials must explain the task clearly. For some individuals with NDLP, this task may be affected by a cultural tendency to avoid direct expression of opinions.\u003c/p\u003e\n \u003cp\u003e\u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003eAnticipated Challenge 2: Content\u003c/span\u003e\u003c/p\u003e\n \u003cp\u003eOne challenge is ensuring that the content tested in the DCE adequately represents the preferences of the participating groups. If the most important attributes and levels are not included in a DCE, the results become uninterpretable. For example, physician gender may be unimportant for some cultural groups and critically important for others.\u003c/p\u003e\n \u003cp\u003eBefore being presented with choice sets, DCE participants should receive an introduction to the topic that frames the relevant decisions and tradeoffs. This so-called \u0026ldquo;decisional context\u0026rdquo; is critical to ensure that participants are imagining the scenario as intended, so that choices truly represent preferences. Some individuals with NDLP that were born in other countries may require different or more extensive explanation of the context within the local health care system.\u003c/p\u003e\n \u003cp\u003eThe attribute of language concordance is likely to be relevant in DCEs for those with NDLP. Research has shown a strong preference for a language concordant provider over the use of an interpreter. This strong preference may drive participants to make decisions solely based on this attribute, limiting the ability to draw conclusions about other preferences from a DCE.\u003c/p\u003e\n \u003cp\u003e\u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003eAnticipated Challenge 3: Administration\u003c/span\u003e\u003c/p\u003e\n \u003cp\u003eAdministering a DCE in a high linguistic diversity context may also pose challenges. As some DCEs recruit participants via an online panel (pre-recruited individuals who have agreed to participate in research), certain populations with NDLP who have limited access to digital devices may be systematically excluded. While DCEs are often self-administered electronically, comfort with using a phone or tablet to complete a DCE survey without dedicated support may be low among individuals with NDLP. Lastly, reliable answers to a DCE depend on the participant not being cognitive overburdened. Therefore, approaches to reduce cognitive burden should be considered, such as limiting the number of choice sets shown to each participant and only presenting pairwise comparisons.\u003c/p\u003e\n \u003cp\u003eAttention to these considerations is essential for maintaining the validity and reliability of DCE results in multilingual contexts.\u003c/p\u003e\n\u003c/div\u003e"},{"header":"METHODS","content":"\u003cp\u003e\u003cem\u003eDescription of Case Example\u003c/em\u003e\u003c/p\u003e\u003cp\u003eWe will illustrate our methodology by applying it to the development of a DCE to improve the hospital discharge experience for patients with NDLP. The transition from hospital to home is notoriously difficult for both patients and caregivers.(\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e) It is prone to errors that compromise patient safety and costs the health care system billions of dollars every year. Patients with NDLP face unique challenges due to multifactorial communication barriers and structural inequities.(\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e) Most trials that test evidence-based interventions to improve the hospital discharge experience have categorically excluded individuals with NDLP.\u003c/p\u003e\u003cp\u003eThe goal of this DCE is to inform a multidisciplinary team tasked with tailoring a hospital discharge intervention to populations with NDLP. The study was conducted at Boston Medical Center (BMC), a tertiary academic medical center and New England’s largest safety net hospital. About 80% of BMC patients are covered by public insurance. Its patient population is diverse both in terms of race and ethnicity, with 50% of patients identifying as Black and 25% as Latino, and linguistically, with over one-third of patients reporting a NDLP. The top two non-English languages of those discharged from medical hospitalizations are Spanish (54%) and Haitian Creole (18%). Because our primary focus is on the development of the DCE instrument for this high diversity context, other aspects of the DCE development, such as experimental design and quantitative data analysis of DCE results, are not addressed.\u003c/p\u003e\u003cp\u003e\u003cem\u003eOverview of Approach to DCE Design\u003c/em\u003e\u003c/p\u003e\u003cp\u003eA five-step process for designing a DCE instrument has emerged as a best practice as described by Campoamor and colleagues.(\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e) The steps include: (\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e) evidence synthesis; (\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e) expert input; (\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e) end-user engagement; (\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e) pretesting; and (\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e) pilot testing. Figure\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ea depicts these steps along with the methods that can be used in each step. To address the challenges of developing a DCE for participants with NDLP, we propose changes to all five phases as described below and depicted in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eb.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eOur design for the DCE was a two option (A versus B) choice set with an opt out option. We administered it in person to patients currently admitted to the hospital who spoke Spanish or Haitian Creole by a bilingual researcher. Given the intensity of the data collection approach, we sought to limit it to 4 to 5 attributes with 2 to 4 levels each. Our goal was to ensure that the instrument captured all strongly held preferences across the two linguistic groups.\u003c/p\u003e\u003cp\u003e\u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003eProcedures\u003c/span\u003e\u003c/p\u003e\u003cp\u003e\u003cem\u003ePhase 1: Evidence synthesis\u003c/em\u003e\u003c/p\u003e\u003cp\u003ePhase 1 involves review of existing literature to identify gaps in understanding that could be addressed using a DCE. For a DCE focused on populations with NDLP, there may be limited data. In our example, we followed the best practices and synthesized findings from studies identified in two recent systematic reviews on hospital discharge interventions and a non-systematic PubMed search for additional articles on challenges of patients with NDLP at hospital discharge.(\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e, \u003cspan additionalcitationids=\"CR25\" citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e–\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e) The first author used this information to create a preliminary list of attributes and levels.\u003c/p\u003e\u003cp\u003e\u003cem\u003ePhase 2 and 3: End user and Expert input\u003c/em\u003e\u003c/p\u003e\u003cp\u003eBest practice recommends first seeking input from experts (phase 2) followed by end-users (phase 3), which may include patients and community members or any others who are expected to be the users of the intervention. Both steps may be formally conducted using qualitative methods and participatory approaches, such as community advisory boards. We reversed these steps and first engaged end-users with the goal to center the voices of the target population and ensure that culturally relevant attributes and levels emerged as early in the process as possible, which we did in our study.\u003c/p\u003e\u003cp\u003eIn our phase 2, we utilized existing data from a separate study involving structured qualitative interviews with Spanish and Haitian Creole speakers recently discharged from the hospital to understand needs and challenges during the transition from hospital to home. The first author reviewed the interview transcripts to assess for alignment with attributes and levels included in the draft DCE instrument and identify additional attributes and levels.\u003c/p\u003e\u003cp\u003eIn our phase 3, three types of experts provided input on the draft DCE: two with knowledge of hospital discharge interventions, one with expertise in populations with NDLP, and one DCE expert.\u003c/p\u003e\u003cp\u003e\u003cem\u003ePhase 4: Pretesting\u003c/em\u003e\u003c/p\u003e\u003cp\u003ePhase 4 involves pretesting the draft DCE instrument, often utilizing cognitive interviewing, cognitive debriefing, or co-creation—a collaborative and iterative process of researchers creating materials with participants.(\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e) The goals of pretesting are to evaluate: (\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e) comprehension of attributes and levels as well as the introductory script; (\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e) presentation including the images used and how choice sets are displayed; (\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e) elicitation, or whether participants understand the hypothetical scenarios and the task of making tradeoffs; and (\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e) content, specifically if any attributes or levels should be added or removed.\u003c/p\u003e\u003cp\u003eWe modified the typical approach to DCE pretesting to meet the unique needs of our culturally and linguistically diverse patient population. We developed an innovative rapid cycle analytic approach to collect, analyze, and revise DCE materials (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e). We initially developed a first draft of materials in Spanish and them to conducted cognitive interviews with Spanish speakers, iteratively modifying materials after each pair of interviews. Once few new themes emerged, we translated the DCE materials into Haitian Creole and conducted cognitive interviews with this population, again making iterative changes to improve clarity and content when needed. We translated the materials into Spanish once again and revised both sets with insights from cognitive interviews with a professional interpreter from each language group, which allowed us to assess the comprehension of formal translation of study materials—akin to the approach of forward and backward translation—and to solicit input on whether the DCE task, attributes, and levels would be understood by a diverse group of Spanish and Haitian Creole speakers.(\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e) Finally, we conduct additional interviews with these revised materials in both languages in parallel until saturation was reached in both groups. This approach allowed us to test solutions quickly and achieving uniform understanding by participants and to overcome barriers arising from the need to analyze data and revise materials in two different non-English languages. Sample size was determined by saturation of the data at each step.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003e\u003cem\u003ePhase 4: Pretesting\u003c/em\u003e\u003c/p\u003e\u003cp\u003eWe used cognitive interviews to pretest draft DCE materials, including an introductory script explaining the decisional context, an in-depth explanation of attributes and levels using both text and pictures, and three example choice sets. Interviews were conducted by native speakers of Spanish and Haitian Creole. The interview guide first assessed understanding of the concept of hospital discharge and of DCE with a focus on the participant’s grasp of hypothetical comparisons and elicitation. Second, the interviewer assessed understanding of each attribute and its levels and solicited feedback on how to improve clarity of text descriptions and pictorial depictions, including whether they liked the photograph or pictograph better for each level (and any input on revisions). Third, the interviewer administered three rounds of DCE on paper followed by a cognitive autopsy to evaluate the participant’s understanding, decision making processes, and alignment with preferences stated earlier.\u003c/p\u003e\u003cp\u003eWe also collected demographic data from patient participants, including age, gender, duration of time in the U.S., language preferences, self-reported ability to read and write in their preferred language, and self-assessed fluency in English. We asked about personal experience with forms of language facilitation, including professional interpreters by phone, video, and in-person and language concordant providers, and asked participant to rank each from most to least preferred.\u003c/p\u003e\u003cp\u003e In phase 4b, in-depth cognitive interviews were conducted with professional Spanish and Haitian Creole interpreters using the same process as with patient participants. The interviewer solicited feedback using a semi-structured interview guide addressing the same domains used in the patient participant interviews. Additional probes aimed to assess if terms would be understood by diverse speakers of the language (i.e. colloquial differences based on country or rurality) and identify ways to improve the description of the DCE task and address potential areas of confusion.\u003c/p\u003e\u003cp\u003e\u003cem\u003ePhase 5: Pilot testing\u003c/em\u003e\u003c/p\u003e\u003cp\u003ePrior to launching the DCE, best practice includes pilot testing the instrument under actual study conditions. For example, if DCE will collect data from participants via a self-administered survey then pilot testing should follow this same procedure. Piloting can also assess cognitive fatigue, which may require reducing the number of choice sets shown to each participant.\u003c/p\u003e\u003cp\u003eIn our case, we elected to conduct pilot testing in person with a language concordant administrator. Each participant was asked to complete 20 choice sets via Sawtooth, a commercial software for conducting DCEs.(\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e) Participants were given the option of holding the tablet and selecting each choice themselves, or for the researcher to read each choice set aloud and record their preferences. The researcher took structured notes on signs of cognitive fatigue and on the points of clarification needed. The participants were asked to rate how difficult and cognitively taxing they found the task on a five-point Likert scale. They were then asked to explain their decision-making strategy and discuss opportunities to improve clarity of the DCE materials.\u003c/p\u003e\u003cp\u003e\u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003eParticipants\u003c/span\u003e\u003c/p\u003e\u003cp\u003eParticipants in phases 4a and 5 included patients aged 18 or older who were admitted to a medicine team at BMC from June to July 2024 and spoke Spanish and Haitian Creole. They were approached in person by bilingual research staff, unless they were on infectious or suicide precautions or if their nurse reported they were unable to provide informed consent due to cognitive impairment. For phase 4b, all Spanish and Haitian Creole BMC interpreters were eligible to participate and were recruited via email. The participants were required to provide informed consent and pass a teach back. The project was approved by the Boston University Medical Center Institutional Review Board (IRB).\u003c/p\u003e\u003ch2\u003eData Analysis\u003c/h2\u003e\u003cp\u003eTo analyze the patient cognitive interviews conducted in Spanish, the PI (who is fluent in Spanish) reviewed audio recordings and created a summary table in English (Appendix 3). Much like the summary tables utilized in rapid qualitative analysis, each table was structured around the topic subheadings of the semi-structured interview guide. Since the PI was not fluent in Haitian Creole, the interviewer took notes in English during these interviews for the PI to review, which she used to create a summary table. To avoid misinterpretation, another researcher fluent in Haitian Creole reviewed the audio recordings and independently created summary tables, which the PI then merged with her own. Any discrepancies were resolved by group discussion.\u003c/p\u003e\u003cp\u003e Summary tables were reviewed by the entire research team, including native speakers of both Spanish and Haitian Creole, within 48 hours of the interview to refine findings and determine revisions to DCE materials for the next cycle of cognitive interviews. When suggestions for new attributes or levels were suggested by participants, the PI conferred with content experts to determine whether the content of the DCE should be modified.\u003c/p\u003e\u003cp\u003eThe professional interpreter cognitive interviews were also audio recorded and analyzed using summary tables as described above. Insights were used to modify DCE materials before a final round of patient participant cognitive interviews.\u003c/p\u003e\u003cp\u003eDuring pilot testing, research staff systematically documented participant interactions using REDCap. They recorded total completion time, frequency and types of clarification questions, and time comparison between early choice sets (second) and final choice sets to detect potential cognitive fatigue (whether participants began answering without deliberation as the DCE progressed). The PI reviewed all observation data within 24 hours of collection, followed by a team debriefing to evaluate necessary modifications before full implementation.\u003c/p\u003e\u003cp\u003e\u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003eReflexivity\u003c/span\u003e\u003c/p\u003e\u003cp\u003eAll members of the data analysis team are multilingual and have lived experience in high diversity contexts. All are first or second-generation immigrants to the United States who have assisted family members with NDLP and/or low health literacy. Nonetheless, the study team did not represent all perspectives represented by the participants in this study. We engaged in regular team debriefings and reflexivity exercises to critically examine how our own experiences and biases might shape (or miss) the themes we identified.\u003c/p\u003e"},{"header":"RESULTS","content":"\u003cp\u003e\u003cem\u003ePhase 1: Evidence synthesis\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eIn reviewing the literature, we identified numerous hospital discharge interventions that included speakers of the dominant language, however, there are few interventions including patients with NDLP and limited research on their preferences during discharge.(\u003cspan class=\"CitationRef\"\u003e21\u003c/span\u003e, \u003cspan class=\"CitationRef\"\u003e24\u003c/span\u003e, \u003cspan class=\"CitationRef\"\u003e30\u003c/span\u003e) Based on the minimal evidence base, we selected four attributes: (\u003cspan class=\"CitationRef\"\u003e1\u003c/span\u003e) \u0026ldquo;Form of interpreter\u0026rdquo;; (\u003cspan class=\"CitationRef\"\u003e2\u003c/span\u003e) \u0026ldquo;Family involvement in discharge teaching\u0026rdquo;; (\u003cspan class=\"CitationRef\"\u003e3\u003c/span\u003e) \u0026ldquo;Form of written instructions\u0026rdquo;; and (\u003cspan class=\"CitationRef\"\u003e4\u003c/span\u003e) \u0026ldquo;Deliverer of intervention\u0026rdquo; (Table \u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003e). This preliminary draft of the DCE instrument included an opt out, meaning the participant could state that they would not elect either option A or B.\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" width=\"896\" class=\"fr-table-selection-hover\"\u003e\n \u003ccaption language=\"En\"\u003e\n \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e\n \u003cdiv class=\"CaptionContent\"\u003e\n \u003cp\u003eOverview of content changes to discrete choice experiment attributes and levels across the five phases of development\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 5.422%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eAttribute:\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003eBrief Title\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 12.049%;\"\u003e\n \u003cp\u003e\u003cstrong\u003ePhase 1\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003eEvidence synthesis\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 13.1334%;\"\u003e\n \u003cp\u003e\u003cstrong\u003ePhases 2/3\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003eEnd user input/expert review\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 13.0732%;\"\u003e\n \u003cp\u003e\u003cstrong\u003ePhase 4a and 4b\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003ePretesting\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 9.1572%;\"\u003e\n \u003cp\u003e\u003cstrong\u003ePhase 5\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003ePilot testing\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 5.422%;\"\u003e\n \u003cp\u003eAttribute 1:\u003c/p\u003e\n \u003cp\u003eLanguage\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 12.049%;\"\u003e\n \u003cp\u003e\u003cem\u003eForm of interpreter\u003c/em\u003e\u003c/p\u003e\n \u003cp\u003elevel 1) telephone\u003c/p\u003e\n \u003cp\u003elevel 2) video\u003c/p\u003e\n \u003cp\u003elevel 3) in person\u003c/p\u003e\n \u003cp\u003elevel 4) family member\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 13.1334%;\"\u003e\n \u003cp\u003e▪ Revised attribute to \u003cem\u003eWay to overcome language barrier\u003c/em\u003e\u003c/p\u003e\n \u003cp\u003e▪ Drop level 4 \u0026ldquo;family member\u0026rdquo;\u003c/p\u003e\n \u003cp\u003e▪ Add level 4 \u0026ldquo;language concordant provider\u0026rdquo;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 13.0732%;\"\u003e\n \u003cp\u003e▪ Revise attribute description to \u003cem\u003eForm of verbal communication\u003c/em\u003e\u003c/p\u003e\n \u003cp\u003e▪ Shorten level descriptions\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e▪ Reorder attribute to be last in choice set to avoid priming\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 9.1572%;\"\u003e\n \u003cp\u003eNo changes\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 5.422%;\"\u003e\n \u003cp\u003eAttribute 2:\u003c/p\u003e\n \u003cp\u003eFamily\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 12.049%;\"\u003e\n \u003cp\u003e\u003cem\u003eFamily involvement in discharge teaching\u003c/em\u003e\u003c/p\u003e\n \u003cp\u003elevel 1) solo\u003c/p\u003e\n \u003cp\u003elevel 2) family/caregiver present in person\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 13.1334%;\"\u003e\n \u003cp\u003e▪ Add level 3 \u0026ldquo;family/caregiver present by phone\u0026rdquo;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 13.0732%;\"\u003e\n \u003cp\u003e▪ Shorten level descriptions\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e▪ Revise attribute description to \u003cem\u003eFamily participation when nurse reviews discharge instructions\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 9.1572%;\"\u003e\n \u003cp\u003eNo changes\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 5.422%;\"\u003e\n \u003cp\u003eAttribute 3:\u003c/p\u003e\n \u003cp\u003eWritten instructions\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 12.049%;\"\u003e\n \u003cp\u003e\u003cem\u003eForm of written instructions\u003c/em\u003e\u003c/p\u003e\n \u003cp\u003elevel 1) written in preferred language\u003c/p\u003e\n \u003cp\u003elevel 2) written in English\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 13.1334%;\"\u003e\n \u003cp\u003e▪ Add level 3 \u0026ldquo;audio recording\u0026rdquo;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 13.0732%;\"\u003e\n \u003cp\u003e▪ Revise wording of attribute to \u003cem\u003eWritten communication\u003c/em\u003e\u003c/p\u003e\n \u003cp\u003e▪ Shorten level descriptions\u003c/p\u003e\n \u003cp\u003e▪ Reword level 3 to \u0026ldquo;audio recording in my language\u0026rdquo;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 9.1572%;\"\u003e\n \u003cp\u003eNo changes\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"2\" valign=\"top\" style=\"width: 5.422%;\"\u003e\n \u003cp\u003eAttribute 4\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 12.049%;\"\u003e\n \u003cp\u003e\u003cem\u003eDeliverer of intervention\u003c/em\u003e:\u0026nbsp;\u003c/p\u003e\n \u003cp\u003elevel 1) nurse\u003c/p\u003e\n \u003cp\u003elevel 2) physician\u003c/p\u003e\n \u003cp\u003elevel 3) community health worker\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 13.1334%;\"\u003e\n \u003cp\u003eDropped\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 13.0732%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 9.1572%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 12.049%;\"\u003e\n \u003cp\u003e\u003cem\u003e\u0026nbsp;\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 13.1334%;\"\u003e\n \u003cp\u003e\u003cem\u003eForm of follow up after discharge:\u0026nbsp;\u003c/em\u003e\u003c/p\u003e\n \u003cp\u003elevel 1) nurse from primary care office\u003c/p\u003e\n \u003cp\u003elevel 2) nurse from hospital\u003c/p\u003e\n \u003cp\u003elevel 3) number I can call\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 13.0732%;\"\u003e\n \u003cp\u003e▪ Revise wording of attribute to \u003cem\u003ePhone follow up after discharge\u003c/em\u003e\u003c/p\u003e\n \u003cp\u003e▪ Shorten level descriptions\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 9.1572%;\"\u003e\n \u003cp\u003eNo changes\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 5.422%;\"\u003e\n \u003cp\u003eOpt Out\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 12.049%;\"\u003e\n \u003cp\u003eIncluded\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 13.1334%;\"\u003e\n \u003cp\u003eDropped\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 13.0732%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 9.1572%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"5\" valign=\"top\" style=\"width: 52.8349%;\"\u003e\n \u003cp\u003e\u003cem\u003eAppendix 1 provides more details about specific changes per iterative cycle and Appendix 2 provides word for word change to attribute titles and level descriptions\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003cem\u003ePhase 2: End user input\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eOur review of post-discharge qualitative interviews with Spanish and Haitian Creole speaking BMC patients (n\u0026thinsp;=\u0026thinsp;37) validated the importance of attributes 1, 2, and 3. Revisions to the levels were made based on insights gained from the interviews. The title of attribute 2 (\u0026ldquo;Family involvement in discharge teaching\u0026rdquo;) was modified to incorporate the finding that family members often cannot be present at discharge, so the level of \u0026ldquo;family/caregiver present by phone\u0026rdquo; was added. For attribute 3 (\u0026ldquo;Form of written discharge instructions\u0026rdquo;), \u0026ldquo;audio recording\u0026rdquo; was added because some participants expressed satisfaction with audio recording of discharge instructions in their preferred language. One theme that was not addressed by the DCE was post-discharge care. Participants articulated significant challenges with systems navigation, including strategies for getting their questions answered and resolving issues post-discharge. We added attribute 4,\u0026ldquo;Form of follow up after discharge,\u0026rdquo; to address this theme.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003ePhase 3: Expert input\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eThe DCE draft was reviewed with the expert team. They recommended eliminating the attribute \u0026ldquo;Deliverer of intervention\u0026rdquo; for two reasons. First, a paraprofessional such as a community health worker would not be qualified to deliver the intervention because it requires clinical expertise. Second, due to the intensive nature of collecting DCE data in person and with a language concordant administrator, adequate sample size was a major concern and maintaining a maximum of four attributes was important. Experts also recommended removing the opt out option. This was both to minimize missing data (selecting the opt out option reduces the number of data points for a participant) and because it did not reflect a real-world option. They also suggested removing the family member option for attribute 1 (\u0026ldquo;Form of interpreter\u0026rdquo;) because it is not legal unless a professional interpreter is also present to correct misunderstandings. Finally, due to evidence of a very strong preference for language concordant interactions, the option of a language-concordant provider was added, which required rephrasing attribute 1 to \u0026ldquo;Way to overcome language barrier.\u0026rdquo; A noted risk of adding this option is its potential to drive preference so strongly that it \u0026ldquo;crowd out\u0026rdquo; preferences for other attributes.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003ePhase 4: Pretesting with patients and professional interpreters\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eDemographics of the pretesting participants are summarized in Table\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003e. Of the 21 participants, 62% were male and 62% were Spanish speaking. Almost all participants (95%) reported speaking English less than very well. Only 24% reported reading in their preferred language less than very well. Most reported a preference for language concordant health care providers (76%) while 14% preferred a phone interpreter, and 9.5% preferred a family member. We summarize our findings using the four domains assessed in pretesting: comprehension, presentation, elicitation and content.\u003c/p\u003e\n\u003cdiv class=\"gridtable\"\u003e\n \u003cdiv align=\"left\" class=\"colspec\"\u003e\u003cbr\u003e\u003c/div\u003e\n\u003c/div\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" width=\"540\"\u003e\n \u003ccaption language=\"En\"\u003e\n \u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e\n \u003cdiv class=\"CaptionContent\"\u003e\n \u003cp\u003eDemographics of patient participants in discrete choice experiment pretesting and pilot testing\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 258px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eCharacteristic\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 156px;\"\u003e\n \u003cp\u003e\u003cstrong\u003ePhase 4a\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003eN = 21\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 126px;\"\u003e\n \u003cp\u003e\u003cstrong\u003ePhase 5\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003eN = 8\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 258px;\"\u003e\n \u003cp\u003eAge (years): mean (SD)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 156px;\"\u003e\n \u003cp\u003e47.2 (12.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 126px;\"\u003e\n \u003cp\u003e53.5 (15.8)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 258px;\"\u003e\n \u003cp\u003eGender\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 156px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 126px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 258px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp; Male\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 156px;\"\u003e\n \u003cp\u003e13 (61.9%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 126px;\"\u003e\n \u003cp\u003e7 (87.5%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 258px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp; Female\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 156px;\"\u003e\n \u003cp\u003e8 (38.1%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 126px;\"\u003e\n \u003cp\u003e1 (12.5%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 258px;\"\u003e\n \u003cp\u003ePreferred language\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 156px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 126px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 258px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;Spanish\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 156px;\"\u003e\n \u003cp\u003e13 (61.9%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 126px;\"\u003e\n \u003cp\u003e4 (50.0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 258px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;Haitian Creole\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 156px;\"\u003e\n \u003cp\u003e8 (38.1%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 126px;\"\u003e\n \u003cp\u003e4 (50.0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 258px;\"\u003e\n \u003cp\u003eLanguage spoken most often at home\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 156px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 126px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 258px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;Spanish\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 156px;\"\u003e\n \u003cp\u003e13 (61.9%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 126px;\"\u003e\n \u003cp\u003e4 (50.0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 258px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;Haitian Creole\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 156px;\"\u003e\n \u003cp\u003e8 (38.1%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 126px;\"\u003e\n \u003cp\u003e4 (50.0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 258px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;English\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 156px;\"\u003e\n \u003cp\u003e0 (0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 126px;\"\u003e\n \u003cp\u003e0 (0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 258px;\"\u003e\n \u003cp\u003eConfidence in filling out forms\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 156px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 126px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 258px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;Extremely confident\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 156px;\"\u003e\n \u003cp\u003e6 (28.6%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 126px;\"\u003e\n \u003cp\u003e1 (12.5%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 258px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;Quite a bit\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 156px;\"\u003e\n \u003cp\u003e3 (14.3%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 126px;\"\u003e\n \u003cp\u003e1 (12.5%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 258px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;Somewhat / A little bit / Not at all\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 156px;\"\u003e\n \u003cp\u003e12 (57.1%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 126px;\"\u003e\n \u003cp\u003e6 (75.0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 258px;\"\u003e\n \u003cp\u003eSelf-reported fluency in English\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 156px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 126px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 258px;\"\u003e\n \u003cp\u003e\u0026nbsp; Understands less than \u0026ldquo;very well\u0026rdquo;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 156px;\"\u003e\n \u003cp\u003e20 (95.2%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 126px;\"\u003e\n \u003cp\u003e8 (100%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 258px;\"\u003e\n \u003cp\u003e\u0026nbsp; Speaks less than \u0026ldquo;very well\u0026rdquo;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 156px;\"\u003e\n \u003cp\u003e20 (95.2%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 126px;\"\u003e\n \u003cp\u003e8 (100%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 258px;\"\u003e\n \u003cp\u003e\u0026nbsp; Reads less than \u0026ldquo;very well\u0026rdquo;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 156px;\"\u003e\n \u003cp\u003e20 (95.2%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 126px;\"\u003e\n \u003cp\u003e8 (100%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 258px;\"\u003e\n \u003cp\u003eSelf- reported ability to read in preferred language\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 156px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 126px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 258px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026ldquo;Very well\u0026rdquo;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 156px;\"\u003e\n \u003cp\u003e16 (76.2%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 126px;\"\u003e\n \u003cp\u003e4 (50%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 258px;\"\u003e\n \u003cp\u003e\u0026nbsp; Less than \u0026ldquo;very well\u0026rdquo;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 156px;\"\u003e\n \u003cp\u003e5 (23.8%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 126px;\"\u003e\n \u003cp\u003e4 (50%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 258px;\"\u003e\n \u003cp\u003ePrior experience with language facilitation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 156px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 126px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 258px;\"\u003e\n \u003cp\u003e\u0026nbsp; Language concordant provider\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 156px;\"\u003e\n \u003cp\u003e19 (90.5%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 126px;\"\u003e\n \u003cp\u003e7 (87.5%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 258px;\"\u003e\n \u003cp\u003e\u0026nbsp; In person professional interpreter\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 156px;\"\u003e\n \u003cp\u003e15 (71.4%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 126px;\"\u003e\n \u003cp\u003e4 (50%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 258px;\"\u003e\n \u003cp\u003e\u0026nbsp; Phone professional interpreter\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 156px;\"\u003e\n \u003cp\u003e20 (95.2%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 126px;\"\u003e\n \u003cp\u003e8 (100%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 258px;\"\u003e\n \u003cp\u003e\u0026nbsp; Video professional interpreter\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 156px;\"\u003e\n \u003cp\u003e8 (38.1%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 126px;\"\u003e\n \u003cp\u003e1 (12.5%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 258px;\"\u003e\n \u003cp\u003ePreference for language facilitation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 156px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 126px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 258px;\"\u003e\n \u003cp\u003e\u0026nbsp; Language concordant provider\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 156px;\"\u003e\n \u003cp\u003e16 (76.2%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 126px;\"\u003e\n \u003cp\u003e4 (50.0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 258px;\"\u003e\n \u003cp\u003e\u0026nbsp; In person professional interpreter\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 156px;\"\u003e\n \u003cp\u003e0 (0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 126px;\"\u003e\n \u003cp\u003e0 (0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 258px;\"\u003e\n \u003cp\u003e\u0026nbsp; Phone professional interpreter\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 156px;\"\u003e\n \u003cp\u003e3 (14.3%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 126px;\"\u003e\n \u003cp\u003e3 (37.5%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 258px;\"\u003e\n \u003cp\u003e\u0026nbsp; Video professional interpreter\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 156px;\"\u003e\n \u003cp\u003e0 (0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 126px;\"\u003e\n \u003cp\u003e0 (0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 258px;\"\u003e\n \u003cp\u003e\u0026nbsp; Family\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 156px;\"\u003e\n \u003cp\u003e2 (9.5%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 126px;\"\u003e\n \u003cp\u003e1 (12.5%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003eTable \u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003e summarizes the challenges faced in pretesting. In the initial interviews, many participants had difficulty comprehending the attributes and levels as intended. For example, many participants thought that attribute 4 (\u0026ldquo;Follow up after discharge\u003cem\u003e\u0026rdquo;)\u003c/em\u003e referred to an in-person appointment with their primary care doctor instead of a phone call from a nurse. Rephrasing this attribute to \u0026ldquo;Phone call after discharge\u0026rdquo; addressed this misinterpretation. Similar modifications were needed to clarify that the audio recording was in the patient\u0026rsquo;s preferred language (Appendix 2 and 3). Another comprehension error was related to the use of a professional interpreter. In the early draft of materials, the first attribute presented in the choice set was \u0026ldquo;Verbal communication\u0026rdquo;. This order primed some participants to interpret subsequent attributes 2 to 4 as asking about the type of professional interpreter used for that task. We remedied this by presenting \u0026ldquo;Verbal communication\u0026rdquo; last in the choice set. We also moved survey questions about their explicit preferences for type of interpreter to after the DCE was completed.\u003c/p\u003e\n\u003cdiv class=\"gridtable\"\u003e\n \u003cdiv align=\"left\" class=\"colspec\"\u003e\u003cbr\u003e\u003c/div\u003e\n \u003ctable id=\"Taba\" border=\"1\"\u003e\n \u003ccaption language=\"En\"\u003e\n \u003cdiv class=\"CaptionNumber\"\u003eTable 4\u003c/div\u003e\n \u003cdiv class=\"CaptionContent\"\u003e\n \u003cp\u003eDomains assessed in pretesting and strategies used in high linguistic diversity context to achieve understanding\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eDomain\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eStrategies\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eComprehension\u003c/p\u003e\n \u003cp\u003e\u003cem\u003eUnderstanding of attributes, levels, and decisional context\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eUse native speaker research assistants\u003c/p\u003e\n \u003cp\u003eRapid cycle iterative changes to materials using sequential group analysis\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003ePresentation\u003c/p\u003e\n \u003cp\u003e\u003cem\u003eText descriptions and visuals\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eUse visuals to augment descriptions of attributes and/or levels\u003c/p\u003e\n \u003cp\u003eStrike balance in length to adequately explain but not promote disengagement and only use of visuals\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eElicitation\u003c/p\u003e\n \u003cp\u003e\u003cem\u003eProcess of making a choice\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eEmphasize choice as a \u0026ldquo;set\u0026rdquo;\u003c/p\u003e\n \u003cp\u003eWarm up choice card for everyday decision\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eContent\u003c/p\u003e\n \u003cp\u003e\u003cem\u003eRelevance of topic, attributes, and levels\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eExplicitly ask about \u0026ldquo;missing\u0026rdquo; levels or attributes that are important to participant\u003c/p\u003e\n \u003cp\u003eEnsure aspects of health system not universally understood are addressed in introductory script\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n\u003c/div\u003e\n\u003cp\u003e\u003cbr\u003e\u003c/p\u003e\n\u003cp\u003eThe presentation of attributes, levels, and choice sets also required significant revision. When attribute and level descriptions were too long, participants looked only at the images, which alone did not communicate the full meaning. Shortening attribute and level labels and placing them directly under images improved understanding. Overall, participants reported that images were helpful and had a slight preference for photos over pictographs. However, some participants were not able to discern the roles of different people in the photos. In attribute 1 \u0026ldquo;Form of verbal communication\u0026rdquo;, one photo is meant to depict a patient, a nurse, and an in-person interpreter but these roles were not obvious to some participants. Photos were revised to include written labels for each person which improved understanding (Fig. \u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003e). In addition, our initial photos aimed to capture diversity in skin color of the patient and nurse. However, this led to confusion when comparing photos in which these individuals represented different roles. As such, participants preferred that the same cast of characters be used across all photos.\u003c/p\u003e\n\u003cp\u003eMost participants did not grasp the task of elicitation during the pretest with the initial version of DCE materials. Instead of selecting Option A or B as a set, participants often stated a preference for a single level instead of considering the tradeoffs of all levels displayed as a package. Revising the introductory script and placing a large rectangular box around Option A and B (Fig. \u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003e) improved understanding, but not for all participants. In response, we designed a \u0026ldquo;warm up\u0026rdquo; DCE choice card (Appendix 3) that used the everyday example of selecting a grocery store. The warmup was integrated into the introductory script, and participants uniformly grasped the concept of tradeoffs and task of selecting A or B in its entirety from each choice set. Querying the participant\u0026rsquo;s process of elicitation also detected some cases in which personal preference was not driving decision making. For example, two participants reported that they did not select certain levels because they did not believe the hospital would offer them. We addressed this by revising the introductory script to explicitly ask for personal preference and not what was felt likely to be offered by the hospital.\u003c/p\u003e\n\u003cp\u003eThe pretesting identified preferences regarding the DCE content. In both Spanish and Haitian Creole groups, some individuals requested the intervention be delivered by a physician, instead of a nurse. After discussion the study team decided not to revise the levels as it would not be a feasible adaptation in real world implementation. Instead, the introductory script was modified to specifically describe this as a nurse-delivered intervention. The second preference that emerged was to have family serve as an interpreter for discharge teaching. Ultimately the consensus in discussion with experts was not to include this option as it could not be legally offered as part of care.\u003c/p\u003e\n\u003cp\u003eDuring the interpreter interviews, both Spanish and Haitian Creole interpreters felt that all word choices for levels and attributes were understandable and accurate. They provided feedback on how to revise the introductory script for improved clarity, including suggestions for explaining \u0026ldquo;theoretical\u0026rdquo; and \u0026ldquo;elicitation\u0026rdquo;.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eStep 5: Pilot testing\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eDemographics of participants were similar to those including in pretesting and are summarized in Table \u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003e. During the pilot tests, none of the participants expressed confusion after the introductory script. While completing the DCE on the tablet, there was a mix of participants who elected to hold the tablet and select their preferred option themselves after reviewing options silently, while others requested that the administrator read each choice set aloud and stated aloud their preferred option in each pairwise comparison.\u003c/p\u003e\n\u003cp\u003eObservations did not reveal signs of cognitive fatigue with nearly equal time needed to complete the second and last choice set. Participants reported that both cognitive demand and task difficulty were low. When questioned about decision making, all participants were able to articulate the tradeoffs behind their decisions. As such, the DCE content was not modified after the pilot testing. The initial and final version of the DCE (translated into English) are presented in Fig. \u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003e for comparison.\u003c/p\u003e"},{"header":"DISCUSSION","content":"\u003cp\u003eThis study aimed to address the methodological challenges of conducting DCEs with individuals who have a NDLP, a population often excluded from health services research. We designed a rigorous, culturally and linguistically tailored approach to DCE development and applied it to a study of preferences for hospital discharge supports. We found that participants with NDLP faced challenges in understanding trade-offs, abstract attributes, and initial draft of visual materials in the DCE. Through features of our methodology\u0026mdash;namely, bilingual interviewer-led administration, cognitive interviews with both patients and interpreters, warm-up elicitation tasks, heavy use of labeled visual descriptions of attributes and levels, and rapid-cycle analysis for iterative refinement\u0026mdash;that demonstrate that it is possible to achieve good comprehension of DCE content through this tailored approach suitable for high diversity contexts. Pilot testing of the DCE instrument designed using our proposed methodology demonstrated high comprehension with no cognitive fatigue, as participants successfully articulated decision-making strategies consistent with their stated preferences.\u003c/p\u003e\u003cp\u003eMost of the literature on administering DCEs to populations who face linguistic and literacy barriers has been conducted with health workers in low- and middle-income countries.(\u003cspan additionalcitationids=\"CR32 CR33 CR34 CR35\" citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e) For instance, Hanson and colleagues conducted a DCE with community members in Zambia. To improve understanding, they used a picture board with symbolic representations of the attributes and levels. Similarly, Brunie and colleagues conducted a DCE with community health workers in Uganda and chose to develop the DCE instrument in English and translate the final versions into three local languages. However, no cognitive testing was done to ensure equivalent understanding in these three languages. Our results can inform the development of DCEs in these contexts by offering a methodology for careful adaptation and testing of DCE instruments across multiple languages to ensure content validity and equity.\u003c/p\u003e\u003cp\u003eThe existing DCE literature provides limited guidance on effective approaches to collecting and analyzing data for pretesting in phase 4 (which includes developing introductory scripts, attribute descriptions, and visual materials to support comprehension). Given the lack of existing models, including those that are feasible across linguistic and cultural communication barriers, we suggest and demonstrate a rapid cycle approach to cognitive interviews involving data collection, analysis, and iterative changes, utilizing summary tables from rapid qualitative analysis.(\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e) Our iterative pretesting process revealed consistent misunderstandings of key attributes such as \u0026ldquo;follow-up after discharge\u0026rdquo; and \u0026ldquo;verbal communication,\u0026rdquo; which were only resolved through revisions grounded in real-time participant feedback. These findings underscore the importance of combining cultural competence with methodological rigor when engaging diverse populations.\u003c/p\u003e\u003cp\u003eThe clinical significance of our study lies in its potential to improve health equity. Transitions from hospital to home are a well-documented point of vulnerability for patients, and individuals with NDLP face disproportionate risk due to language barriers and systemic inequities.(\u003cspan additionalcitationids=\"CR39\" citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e) By identifying preferences for communication modalities, instructional formats, and family involvement in discharge processes, our approach supports the development of more responsive, person-centered interventions. Incorporating these preferences into intervention design is a strategy to improve patient comprehension, satisfaction, and safety during care transitions.(\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e) Even when responding to all preferences is not feasible\u0026mdash;for example, a physician-led hospital discharge intervention would be significantly less scalable\u0026mdash;our findings highlight the importance of explaining the value and expertise of nursing staff among populations whose experiences with healthcare systems in other countries may differ. Innovative next steps include exploring the feasibility and relative advantage of using a video introduction to the DCE or testing remote DCE administration with Research Assistant support via phone or video to reduce the barriers presented by in-person completion.\u003c/p\u003e\u003cp\u003eThis study has several limitations. First, the methodology was applied to a DCE including two languages. Expanding to three or more languages would increase complexity and should be explored in future studies as this could require expanding the number of levels offered, necessitating a larger sample size. Second, only one interpreter from each linguistic group was asked to review materials as the final step in pretesting. Engaging multiple interpreters or a more traditional forward and back translation approach could be tested, but would slow down the rapid cycle analytic approach. Third, this methodology was demonstrated at a single site, which may limit generalizability. Further testing across more contexts and with different subject matters would help validate and refine the proposed approach.\u003c/p\u003e"},{"header":"CONCLUSIONS","content":"\u003cp\u003eWe provide a practical framework for designing and conducting DCEs with populations who speak non-dominant languages. Our approach addresses longstanding methodological gaps by integrating bilingual data collection, culturally informed instrument refinement, and pilot testing in real-world conditions. These findings support the feasibility of conducting rigorous DCEs in high-diversity settings and highlight the value of inclusive design to inform equitable implementation. Future research should expand this work by testing the generalizability of our approach across additional languages, healthcare contexts, and decision-making scenarios. Efforts to quantify how tailoring based on DCE findings affects intervention outcomes will also be critical to advancing equity in intervention implementation.\u003c/p\u003e"},{"header":"Abbreviations","content":"\u003cdiv class=\"DefinitionList\"\u003e\u003cdiv class=\"DefinitionListEntry\"\u003e\u003cdiv class=\"Term\"\u003eAVS\u003c/div\u003e\u003cdiv class=\"Description\"\u003e\u003cp\u003eAfter Visit Summary\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e\u003cdiv class=\"DefinitionListEntry\"\u003e\u003cdiv class=\"Term\"\u003eBMC\u003c/div\u003e\u003cdiv class=\"Description\"\u003e\u003cp\u003eBoston Medical Center\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e\u003cdiv class=\"DefinitionListEntry\"\u003e\u003cdiv class=\"Term\"\u003eLEP\u003c/div\u003e\u003cdiv class=\"Description\"\u003e\u003cp\u003elimited English proficiency\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e\u003cdiv class=\"DefinitionListEntry\"\u003e\u003cdiv class=\"Term\"\u003eNDLP\u003c/div\u003e\u003cdiv class=\"Description\"\u003e\u003cp\u003enon-English language preference\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e\u003c/div\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eEthics approval and consent to participate\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study was approved by the Boston University Medical Center IRB.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for publication\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eClinical trial number\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAvailability of data and materials\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eDatasets will be available on reasonable request from the corresponding author (KA).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare that they have no competing interests.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eDr. Austad is supported by a K23 from the National Institute for Minority Health and Health Disparities (K23 MD019068). \u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthors’ contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eKA conceived of the study. SM, BJ, AF, RS provided input on study design. KA, NL, DBH, SL, and KP collected data and conducted the analysis. SM, BJ, AF, MLD, and RS provided input on the results. KA led the drafting of the manuscript with significant contributions from all co-authors. All authors read and approved the final manuscript.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgements\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe would like to thank Director of Interpreter Services Alegna Zavatti for her collaboration as well as the interpreters who participated in the cognitive interviews. We would also like to thank Amy Ries for her input as a professional writer on the organization and prose of the manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthors’ information\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eKA is a Family Medicine hospitalist and implementation science researcher at Boston Medical Center.\u0026nbsp;\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eZhang Y, Anh Ho TQ, Terris-Prestholt F, Quaife M, De Bekker-Grob E, Vickerman P, et al. Prediction accuracy of discrete choice experiments in health-related research: a systematic review and meta-analysis. eClinicalMedicine. 2025;79:102965.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eRyan M, Gerard K, Amaya-Amaya M. Using Discrete Choice Experiments to Value Health and Health Care. 1st ed. Springer Dordrecht; 2007. p. 256.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWang H, Rowen DL, Brazier JE, Jiang L. Discrete Choice Experiments in Health State Valuation: A Systematic Review of Progress and New Trends. Appl Health Econ Health Policy. 2023;21(3):405\u0026ndash;18.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eQuaife M, Terris-Prestholt F, Di Tanna GL, Vickerman P. How well do discrete choice experiments predict health choices? A systematic review and meta-analysis of external validity. Eur J Health Econ. 2018;19(8):1053\u0026ndash;66.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eIrie WC, Kerkhoff A, Kim HY, Geng E, Eshun-Wilson I. Using stated preference methods to facilitate knowledge translation in implementation science. Implement Sci Commun. 2024;5(1):32. s43058-024-00554\u0026ndash;3.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSalloum RG, Shenkman EA, Louviere JJ, Chambers DA. Application of discrete choice experiments to enhance stakeholder engagement as a strategy for advancing implementation: A systematic review. Implement Sci. 2017;12(1):1\u0026ndash;12.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eOstermann J, Njau B, Hobbie A, Mtuy T, Masaki ML, Shayo A, et al. Using discrete choice experiments to design interventions for heterogeneous preferences: protocol for a pragmatic randomised controlled trial of a preference-informed, heterogeneity-focused, HIV testing offer for high-risk populations. BMJ Open. 2020;10(11):e039313.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLau AS, Huey SJ, Baumann AA. Advances in the adaptation and implementation of evidence-based interventions for historically marginalized groups. Behav Res Ther. 2023;168:104377.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003ePineros-Leano M, Desrosiers A, Pi\u0026ntilde;eros-Lea\u0026ntilde;o N, Moya A, Canizares-Escobar C, Tam L, et al. Cultural adaptation of an evidence-based intervention to address mental health among youth affected by armed conflict in Colombia: An application of the ADAPT-ITT approach and FRAME-IS reporting protocols. Camb prisms Glob ment health. 2024;11:e114.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHealey P, Stager ML, Woodmass K, Dettlaff AJ, Vergara A, Janke R, et al. Cultural adaptations to augment health and mental health services: a systematic review. BMC Health Serv Res. 2017;17(1):1\u0026ndash;26.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eChin MH, Walters AE, Cook SC, Huang ES. Interventions to reduce racial and ethnic disparities in health care. Med Care Res Rev. 2007;64(5 SUPPL.).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eAustad K, Cordova-Ramos EG, Fernandez A, Drainoni ML. Lost in translation: advancing intervention adaptation for populations with non-dominant language preference in high diversity settings. Implement Sci Commun. 2025;6(1):66.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSquires A, Gerchow L, Ma C, Liang E, Trachtenberg M, Miner S. A multi-language qualitative study of limited English proficiency patient experiences in the United States. PEC Innov. 2023;2:100177.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLiamputtong P. Doing Research in a Cross-Cultural Context: Methodological and Ethical Challenges. In: Diener E, Glatzer W, Moum T, Sprangers M, Vogel J, Veenhoven R, editors. Springer; 2008. pp. 3\u0026ndash;20. Available from: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://link.springer.com/\u003c/span\u003e\u003cspan address=\"http://link.springer.com/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/978-1-4020-8567-3_1\u003c/span\u003e\u003cspan address=\"10.1007/978-1-4020-8567-3_1\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMangham LJ, Hanson K, McPake B. How to do (or not to do)\u0026hellip; Designing a discrete choice experiment for application in a low-income country. Health Policy Plann. 2009;24(2):151\u0026ndash;8.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWang Y, Wang Z, Wang Z, Li X, Pang X, Wang S. Application of Discrete Choice Experiment in Health Care: A Bibliometric Analysis. Front Public Health. 2021;9.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eNaveen P, Trojovsk\u0026yacute; P. Overview and challenges of machine translation for contextually appropriate translations. iScience. 2024;27(10):110878.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSentell T, Braun KL. Low health literacy, limited English proficiency, and health status in Asians, Latinos, and other racial/ethnic groups in California. In: J Health Communication. 2012. pp. 82\u0026ndash;99.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMoor S, Tusubira AK, Akiteng AR, Hsieh E, Ngaruiya C, Rabin TL et al. Developing a Discrete Choice Experiment to Understand Patient Preferences in Resource-Limited Settings: a Six-Step Guide. Available from: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1101/2020.10.05.20206557\u003c/span\u003e\u003cspan address=\"10.1101/2020.10.05.20206557\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKohler RE, Lee CN, Gopal S, Reeve BB, Weiner BJ, Wheeler SB. Developing a discrete choice experiment in Malawi: Eliciting preferences for breast cancer early detection services. Patient Prefer Adherence. 2015;9:1459\u0026ndash;72.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eTyler N, Hodkinson A, Planner C, Angelakis I, Keyworth C, Hall A, et al. Transitional Care Interventions from Hospital to Community to Reduce Health Care Use and Improve Patient Outcomes: A Systematic Review and Network Meta-Analysis. JAMA Netw Open. 2023;6(11):E2344825.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eAustad K, Jack BW. Linguistic and Cultural Competence at Hospital Discharge. J Healthc Manage Stand. 2023;3(1):1\u0026ndash;16.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eCampoamor NB, Guerrini CJ, Brooks WB, Bridges JFP, Crossnohere NL. Pretesting Discrete-Choice Experiments: A Guide for Researchers. Patient. 2024;17(2):109\u0026ndash;20.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBecker C, Zumbrunn S, Beck K, Vincent A, Loretz N, M\u0026uuml;ller J et al. Interventions to Improve Communication at Hospital Discharge and Rates of Readmission: A Systematic Review and Meta-analysis. JAMA Netw Open. 2021;4(8).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBalaban RB, Galbraith AA, Burns ME, Vialle-Valentin CE, Larochelle MR, Ross-Degnan D. A Patient Navigator Intervention to Reduce Hospital Readmissions among High-Risk Safety-Net Patients: A Randomized Controlled Trial. J Gen Intern Med. 2015;30(7):907\u0026ndash;15.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBalaban RB, Weissman JS, Samuel PA, Woolhandler S. Redefining and redesigning hospital discharge to enhance patient care: A randomized controlled study. J Gen Intern Med. 2008;23(8):1228\u0026ndash;33.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eAgnello DM, Anand-Kumar V, An Q, De Boer J, Delfmann LR, Longworth GR et al. Co-creation methods for public health research \u0026mdash; characteristics, benefits, and challenges: a Health CASCADE scoping review. BMC Med Res Methodol [Internet]. 2025 Mar 6 [cited 2025 May 23];25(1). Available from: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://bmcmedresmethodol.biomedcentral.com/articles/\u003c/span\u003e\u003cspan address=\"https://bmcmedresmethodol.biomedcentral.com/articles/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1186/s12874-025-02514-4\u003c/span\u003e\u003cspan address=\"10.1186/s12874-025-02514-4\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLee WL, Chinna K, Lim Abdullah K, Zainal Abidin I. The forward-backward and dual‐panel translation methods are comparable in producing semantic equivalent versions of a heart quality of life questionnaire. Int J of Nursing Practice [Internet]. 2019 Feb [cited 2025 May 23];25(1). Available from: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://onlinelibrary.wiley.com/doi/\u003c/span\u003e\u003cspan address=\"https://onlinelibrary.wiley.com/doi/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1111/ijn.12715\u003c/span\u003e\u003cspan address=\"10.1111/ijn.12715\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLighthouse Studio. Choice-Based Conjoint Module. Provo, UT: Sawtooth Software, Inc.; 2023.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKansagara D, Chiovaro JC, Kagen D, Jencks S, Rhyne K, O\u0026rsquo;Neil M et al. Transitions of care from hospital to home: An overview of systematic reviews and recommendations for improving transitional care in the Veteral Health Administration. Transitions of Care from Hospital to Home: An Overview of Systematic Reviews and Recommendations for Improving Transitional Care in the Veterans Health Administration [Internet]. 2015; Available from: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://www.ncbi.nlm.nih.gov/pubmed/26312362\u003c/span\u003e\u003cspan address=\"http://www.ncbi.nlm.nih.gov/pubmed/26312362\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBrunie A, Chen M, Akol A. Qualitative Assessment of the Application of a Discrete Choice Experiment With Community Health Workers in Uganda: Aligning Incentives With Preferences. Glob Health Sci Pract. 2016;4(4):684\u0026ndash;93.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eAbdel-All M, Angell B, Jan S, Howell M, Howard K, Abimbola S, et al. What do community health workers want? Findings of a discrete choice experiment among Accredited Social Health Activists (ASHAs) in India. BMJ Glob Health. 2019;4(3):e001509.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHonda A, Vio F. Incentives for non-physician health professionals to work in the rural and remote areas of Mozambique\u0026mdash;a discrete choice experiment for eliciting job preferences. Hum Resour Health [Internet]. 2015 Dec [cited 2025 May 23];13(1). Available from: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://human-resources-health.biomedcentral.com/articles/\u003c/span\u003e\u003cspan address=\"https://human-resources-health.biomedcentral.com/articles/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1186/s12960-015-0015-5\u003c/span\u003e\u003cspan address=\"10.1186/s12960-015-0015-5\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBlaauw D, Erasmus E, Pagaiya N, Tangcharoensathein V, Mullei K, Mudhune S, et al. Policy interventions that attract nurses to rural areas: a multicountry discrete choice experiment. Bull World Health Organ. 2010;88(5):350\u0026ndash;6.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003evujicic marko. alfano marco, ryan mandy, wesseh CS, Brown-Annan J. Policy options to attract nurses to rural Liberia: Evidence from a discrete choice experiment. HNP; 2010 Nov.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eShiratori S, Agyekum EO, Shibanuma A, Oduro A, Okawa S, Enuameh Y et al. Motivation and incentive preferences of community health officers in Ghana: an economic behavioral experiment approach. Hum Resour Health [Internet]. 2016 Dec [cited 2025 May 23];14(1). Available from: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://human-resources-health.biomedcentral.com/articles/\u003c/span\u003e\u003cspan address=\"http://human-resources-health.biomedcentral.com/articles/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1186/s12960-016-0148-1\u003c/span\u003e\u003cspan address=\"10.1186/s12960-016-0148-1\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKowalski CP, Nevedal AL, Finley EP, Young JP, Lewinski AA, Midboe AM, et al. Planning for and Assessing Rigor in Rapid Qualitative Analysis (PARRQA): a consensus-based framework for designing, conducting, and reporting. Implement Sci. 2024;19(1):71.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKhoong EC, Sherwin EB, Harrison JD, Wheeler M, Shah SJ, Mourad M et al. Impact of standardized, language-concordant hospital discharge instructions on postdischarge medication questions. J Hosp Med. 2023.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWoods AP, Alonso A, Duraiswamy S, Ceraolo C, Feeney T, Gunn CM, et al. Limited English Proficiency and Clinical Outcomes After Hospital-Based Care in English-Speaking Countries: a Systematic Review. J Gen Intern Med. 2022;37(8):2050\u0026ndash;61.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eChauhan A, Walton M, Manias E, Walpola RL, Seale H, Latanik M, et al. The safety of health care for ethnic minority patients: A systematic review. Int J Equity Health. 2020;19(1):1\u0026ndash;25.\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"non-dominant language preference, limited English proficiency, implementation science, discrete choice experiment, intervention adaptation, transitions of care, hospital discharge, professional interpreter","lastPublishedDoi":"10.21203/rs.3.rs-6976834/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6976834/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cb\u003eBackground:\u003c/b\u003e\u003c/p\u003e\u003cp\u003eDiscrete choice experiments (DCEs) are widely used to elicit patient preferences for health interventions and implementation strategies. However, individuals with non-dominant language preference (NDLP) are frequently excluded from DCE studies due to language and literacy needs. As health systems increasingly serve linguistically diverse populations, there is an critical need to adapt DCE methods for inclusion of NDLP populations.\u003c/p\u003e\u003cp\u003e\u003cb\u003eObjective:\u003c/b\u003e\u003c/p\u003e\u003cp\u003eTo develop and pilot a practical methodology for designing and conducting DCEs in high-diversity, multilingual settings. We use a case example focused on tailoring a hospital discharge intervention for Spanish and Haitian Creole speakers with NDLP to illustrate the methodology.\u003c/p\u003e\u003cp\u003e\u003cb\u003eMethods:\u003c/b\u003e\u003c/p\u003e\u003cp\u003e We adapted a five-step DCE development framework, including evidence synthesis, end-user input, expert review, pretesting and pilot testing, for use with NDLP populations. Modifications focused on pretesting using cognitive interviews with patients and professional interpreters, and a rapid-cycle approach to iterative revision in two languages. The DCE focused on hospital discharge preferences and included four attributes with visual and text-based levels. Pilot testing assessed comprehension, cognitive burden, and usability of the final DCE instrument.\u003c/p\u003e\u003cp\u003e\u003cb\u003eResults:\u003c/b\u003e\u003c/p\u003e\u003cp\u003eWe identified multiple barriers to DCE participation for patients with NDLP, including difficulty grasping the critical task of elicitation (making tradeoffs) in a DCE choice set, misinterpretation of abstract attributes, and inconsistent understanding of visual materials. Tailored solutions included bilingual interviewer-led administration, integration of a \u0026ldquo;warm-up\u0026rdquo; elicitation task, use of labeled photos to describe levels, and iterative revision of wording and decisional context to improve comprehension. Pilot testing showed high acceptability and comprehension, with no evidence of cognitive fatigue. Participants were able to articulate tradeoffs and report decision-making strategies aligned with stated preferences.\u003c/p\u003e\u003cp\u003e\u003cb\u003eConclusions:\u003c/b\u003e\u003c/p\u003e\u003cp\u003eThis study provides a practical, equity-centered methodology for conducting DCEs with linguistically diverse populations. Our findings demonstrate the feasibility of rigorous DCE administration in NDLP populations and offer a replicable framework for implementing DCEs high diversity contexts. Future work should explore adaptation across additional languages, cultures, and settings and evaluate the impact of preference-informed tailoring on intervention outcome.\u003c/p\u003e","manuscriptTitle":"Developing Discrete Choice Experiments for Populations with Non-Dominant Language Preference: A Methodological Framework and Case Example","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-07-21 15:09:46","doi":"10.21203/rs.3.rs-6976834/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"9b6e9d7d-2617-48cd-a541-e53a8f399c9d","owner":[],"postedDate":"July 21st, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2026-03-11T15:27:09+00:00","versionOfRecord":[],"versionCreatedAt":"2025-07-21 15:09:46","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-6976834","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6976834","identity":"rs-6976834","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}