23709 SpeechLinks: Developing a New Technology to Reduce Inequities in Health Care Delivery to Latino Patients with Limited English Proficiency

Win May, MD, PhD, Division of Medical Education, Educational Affairs, Keck School of Medicine of the University of Southern California, Los Angeles, CA, Panayiotis Georgiou, PhD, Ming Hsieh Department of Electrical Engineering-Systems, Viterbi School of Engineering, University of Southern California, Los Angeles, CA, Margaret McLaughlin, PhD, Annenberg School for Communication and Journalism, University of Southern California, Los Angeles, CA, Lourdes Baezconde-Garbanati, PhD, MPH, Department of Preventive Medicine, Keck School of Medicine, IInstitute for Healath Promotion and Disease Prevention Research, Alhambra, CA and Shrikanth Narayanan, PhD, Department of Electrical Engineering, University of Southern California, Los Angeles, CA

Background:  Numerous studies have documented inequities in delivery of health care to Latinos with limited English proficiency (LEP) .   Although LEP patients have the right to an interpreter, and research has shown that trained interpreters make fewer errors and have more satisfied patients, they are not readily available.  Despite system level efforts to improve the inter-cultural competency of health care institutions, the problem persists. Advances in technology, using automatic two-way speech translation offer new possibilities for cost-effective, targeted interpretation solutions for improved communication with monolingual Spanish speaking or Latino LEP patients and their families

Program background:  The National Science Foundation funded a transdisciplinary team of experts from engineering, communication, preventive medicine/ public health, and medical education to develop and evaluate the use, effectiveness and cultural appropriateness of SpeechLinks, a technology-based cross-lingual communication tool. Our target application focuses on communication between healthcare personnel who speak English only and monolingual Latino patients or patients with limited-English proficiency

Evaluation Methods and Results:  Data was compiled from 50 sessions of clinical encounters among triads of English speaking physicians, Spanish language simulated patients and bilingual Spanish/English interpreters. This data will be used in two ways: First, to develop the core SpeechLinks technologies including capturing the relevant dialectal and Spanish language information for the target interaction scenarios, and developing tools for easy end-user customization and updating; second, to assist and inform SpeechLinks to understand the most useful functionalities and form Findings reveal the importance of incorporating information on culturally bound syndromes, variations in the meaning of specific Spanish words, and factors impacting the communication of emotion in the medical environment.

Conclusions:  Interactions between simulated patients, interpreters and physicians not only provide the spoken language data needed for the development of SpeechLinks, but also assist in making the technology culturally intelligent.

Implications for research and/or practice: SpeechLinks has the potential to help bridge the language gap so that health care access and delivery to Latinos is not compromised by sociolinguistic barriers.