Mobile Video Interpretation to Optimize Communication Across Language Barriers

Overview

Patients and families with limited English proficiency (LEP) and/or who use a language other than English for medical care (LOE) experience substantial barriers to communication in the healthcare setting, leading to medical care that is less safe, less efficient, less effective, and inequitable. Professional interpreter use has repeatedly been shown to improve outcomes and reduce disparities, yet it remains widely underused. This study will test two discrete, scalable implementation strategies for improving professional interpreter use in primary care, while collecting detailed information about mechanisms of action and costs that will be applicable to many sites of care and contribute meaningfully to the goal of reducing communication-based disparities. The investigators hypothesize that improving provider knowledge, confidence, and interpreter access will change behavior under real-world conditions.

This study will be conducted over a series of 4 phases. Phase 0 (15 months): The study team will enroll 55 primary care providers from 4-8 enrolled primary care clinics within the WWAMI region. Enrolled providers will complete a baseline survey and a baseline in-depth qualitative interview, exploring baseline barriers, facilitators, moderators, necessary prerequisites, and proximal outcomes of professional interpreter use in their current clinic setting. Providers will receive a gift card for completing the survey and a gift card for completing the 30-60 minute interview. For participating providers who cannot receive direct incentives per local policy, lunch or equivalent incentive will be provided. During this phase we will calculate baseline interpreter use for each enrolled provider by matching interpreter vendor invoice data to visits with patients listed as preferring a language other than English for medical care over the past 6 months. Phase 1 (9 months): The study team will randomize enrolled providers 1:1 to either web-based educational modules or mobile video interpreting (mVI), stratified by clinic and baseline interpreter use. For education-assigned providers, they will receive access to the online modules and begin receiving weekly reminder emails. They will also receive a tip sheet for making the best use of the current interpreter services they have access to in their clinic. mVI-assigned providers will have mVI downloaded onto their preferred device (smartphone or tablet), or be issued a study smartphone if they do not want to use their own. They will be shown how to use it and receive a tip sheet for optimal mVI use in clinic. The study team will then begin 9 months of data collection for phase 1. In addition to tracking interpreter use, the study team will also begin recruiting patients or parents of pediatric patients with LOE to participate in one of 3 data collection activities. The study team will enroll patients or parents of pediatric patients who are being seen by an enrolled provider for an acute concern and who prefer one of the top 7 non-English languages in the area (Spanish, Vietnamese, Somali, Amharic, Russian, Ukrainian, Marshallese) to complete a post-visit survey (n=662 total; 331 during phase 1). Patients will be approached by telephone before the visit or in-person in the waiting room prior to the clinic visit, then the survey will be administered immediately after the visit via tablet, or up to 7 days after the visit, online or over the telephone. The survey will collect patient reported communication quality, interpretation satisfaction, demographic information, and patient reported diagnosis. The survey will be administered via REDCap, using an audio feature that will allow participants to read or listen to the survey questions in their preferred language for care. Patients will receive a gift card or cash for completing the survey. Patients who report a concern about how communication occurred on the survey, along with a randomly-selected subset of others, will be invited to complete a 20 to 30 minutes interview to provide greater detail about how communication and interpretation occurred and what their experience of it was like (n=75 total, 38 during Phase 1). Whenever possible, the interview will be completed immediately after the survey, while the patient or family is still in the clinic. If the patient wishes to participate but they are unable to stay or their language is different than the bilingual research coordinator's, the interview will be conducted over the telephone within 7 days of the clinic visit. The interview will be conducted by a research coordinator bilingual in the patient's preferred language. Patients will receive a gift card or cash for completing the interview. The investigators will also be requesting permission to video or audio record a subset of patient visits (up to n=200 total, 100 during Phase 1). Patients or parents of pediatric patients who prefer one of the top of 7 non-English language, as well as a set of patients who indicate English as their language for care and are being seen by an enrolled provider for any visit reason will be eligible and approached in the waiting room prior to the clinic visit. Patients will have the opportunity to choose video or audio-only recordings. Those who consent to recording will have a video camera mounted in the corner of the patient room and will have their encounter recorded. For patients who choose audio only recording, the camera will be returned to the wall and covered to avoid capturing any video. During Phase 1, providers that will complete a 2nd in-depth qualitative interview, revisiting some of the barriers, facilitators, moderators, necessary prerequisites, and proximal outcomes from the 1st interview, and assessing experiences and perceived change in behavior with the assigned strategy. They will receive a gift card or other equivalent incentive per local policy for completing the 30-60 minute interview. Phase 2 (9 months): At the beginning of phase 2, the investigators will calculate interpreter use for enrolled providers during Phase 1. Providers will be rank-ordered by interpreter use within assigned strategy. Those with interpreter use in the top tertile will receive an email of positive feedback and be encouraged to continue with their originally assigned strategy. Providers with interpreter use in the lowest two tertiles will undergo a 2nd round of randomization, either to continue with the originally assigned strategy or to add the 2nd strategy on top of the 1st. Those assigned to continue with the initially assigned strategy will receive a brief check-in from study staff to be sure they know how to access and interact with their assigned strategy. Those assigned to the 2nd strategy will receive access to it, as described above. Patient data collection will continue as described for Phase 1, with ongoing recruitment for patient surveys (n=331 for Phase 2), interviews (n=37 for Phase 2), and recording (up to n=100 for Phase 2). Providers will complete a 3rd in-depth interview during Phase 2, again assessing barriers, facilitators, moderators, necessary prerequisites, and proximal outcomes from the 1st interview, and eliciting experiences and perceived change in behavior with the assigned strategy or strategies. They will receive a gift card or equivalent incentive per local policy for completing the 30-60 minute interview. During Phase 2, the study team will begin abstracting provider-reported diagnosis from clinic visit chart notes for patients who completed the survey. Abstraction will be completed by a trained abstractor, blinded to provider assigned strategy, and continue into Phase 3. Diagnosis concordance coding will begin at the end of Phase 2. Two trained coders will compare patient-reported diagnosis (collected through the survey and professionally translated into English) to provider-documented diagnosis (abstracted from the chart). Each pair will be coded as yes, concordant; no, not concordant; or unclear/insufficient information, using the standard of whether a follow-up provider would be likely to understand what the patient was seen or treated for based on the information provided by the patient. Disagreements will be resolved through discussion, with input from the PI as needed. All coders will be blinded to provider assigned strategy and patient information during the coding process. Coding of qualitative interviews and video/audio recordings will also begin at the end of Phase 2. These will be conducted by trained research coordinators, under the supervision of study investigators. Phase 3 (9 months): During phase 3, the investigators will provide access to mVI and the educational modules for any enrolled providers who would like them. The investigators will continue to measure interpreter use and will track uptake of both strategies by providers who were not initially assigned to them. The study team will not continue patient surveys or interviews during this phase. Providers will be asked to complete a final survey, re-assessing constructs from the Theoretical Domains Framework that were measured at baseline, and eliciting time and out-of-pocket costs related to each strategy. During phase 3, diagnosis chart abstraction will be completed, as will diagnosis concordance coding. Qualitative coding and analysis, and video/audio recording coding and analysis will continue throughout phase 3. During this phase, the investigators will work with each clinic to develop a sustainment plan, informed by a review of the site-specific barriers to interpreter use identified in the final survey and interviews. Phase 4 (18 months): This final study phase will be dedicated to analysis and results dissemination. Provider participation will be complete, and access to mVI and the educational modules for study purposes will be discontinued, although investigators will work with clinics to establish methods for ongoing access should the clinics wish to pursue this option. Provider interpreter use during the 6 month follow-up period (Phase 3) will be calculated and shared with clinics. Study staff will continue qualitative and video/audio-recording coding and analysis, along with our quantitative analyses related to overall interpreter use, patient comprehension, and the cost-effectiveness of each strategy (mVI, education, or both) for improving interpreter use and patient comprehension. During this time, the investigators will also refine our preliminary causal pathway models, exploring potential mechanisms by which mVI and education may be operating to improve interpreter use and/or patient comprehension.