Case Report: Benefits of LSVT LOUD in a multilingual patient with hypokinetic-hyperkinetic dysarthria and suspected progressive supranuclear palsy.

Purpose: This case study measured how well the Lee Silverman Voice Treatment (LSVT) improved vocal features, intelligibility, and communicative effectiveness for a multilingual participant with hypokinetic-hyperkinetic dysarthria secondary to suspected progressive supranuclear palsy. LSVT treatment was chosen for the participant due to the strengths and deficits he presented with prior to treatment, and for the anticipated challenges in treatment that may arise from the presence of multilingualism and impaired cognitive functioning. Methods: A multilingual patient in their 60's (English, Spanish, and French) with hypokinetic-hyperkinetic dysarthria secondary to suspected progressive supranuclear palsy completed the standard treatment sessions for LSVT. Assessment measures were taken at baseline, immediately post-treatment, and three-months post-treatment. Results: Improvements were measured in vocal quality, vocal loudness, intelligibility, and communicative effectiveness immediately post-treatment. Three months post-treatment, improvements in vocal quality and intelligibility were maintained. Conclusion: This case study illustrates that LSVT may be a beneficial treatment for complex clients who are multilingual and present with complex comorbidities and cognitive deficits. LSVT resulted in some meaningful changes in vocal quality, intelligibility, and communicative effectiveness for this individual. Clinicians who work with complex patients may wish to consider the theoretical underpinnings of LSVT, client profile, areas of client need, and ability and desire to complete an intensive treatment program to determine if trialing LSVT is appropriate. The use of LSVT with complex clients may yield positive outcomes.

Challenges of the Calgary-Cambridge Consultation Guide in Veterinary Multicultural and Multilingual Scenarios and the Role of Veterinary Translators.

The Calgary-Cambridge Guide is a widely recognised framework for teaching communication skills to healthcare professionals that has become a cornerstone of communication training programs in medicine and other healthcare fields. In the context of veterinary medicine, its integration into communication training programs has become an asset improving communication, education, interaction, and quality of service, enhancing the veterinary-client-patient relationship (VCPR). In veterinary medicine, however, a more challenging consultation dynamic involves the veterinarian, the owner, and the animal. The addition of a veterinary assistant that acts as an interpreter or translator is common in Hong Kong where the native language (Cantonese) coexists with English when consultations are led by non-native language speakers. This addition converts this commonly dyadic model into a triadic communication model. The addition of an assistant interpreter influences the way consultations are conducted, how information is conveyed, and how interpersonal cues and empathy are delivered. In this report we depict challenges applying the Calgary-Cambridge Guide in multicultural and multilingual veterinary medical centres in Hong Kong and highlight the role of veterinary supporting staff in these scenarios, specifically veterinary assistant interpreters.

Federated learning for privacy-preserving depression detection with multilingual language models in social media posts.

The incidences of mental health illnesses, such as suicidal ideation and depression, are increasing, which highlights the urgent need for early detection methods. There is a growing interest in using natural language processing (NLP) models to analyze textual data from patients, but accessing patients' data for research purposes can be challenging due to privacy concerns. Federated learning (FL) is a promising approach that can balance the need for centralized learning with data ownership sensitivity. In this study, we examine the effectiveness of FL models in detecting depression by using a simulated multilingual dataset. We analyzed social media posts in five different languages with varying sample sizes. Our findings indicate that FL achieves strong performance in most cases while maintaining clients' privacy for both independent and non-independent client partitioning.

Quebec-based parents' concerns regarding their children's multilingual development.

Many parents express concerns for their children's multilingual development, yet little is known about the nature and strength of these concerns - especially among parents in multilingual societies. This pre-registered, questionnaire-based study addresses this gap by examining the concerns of 821 Quebec-based parents raising infants and toddlers aged 0-4 years with multiple languages in the home. Factor analysis of parents' Likert-scale responses revealed that parents had (1) concerns regarding the effect of children's multilingualism on their cognition, and (2) concerns regarding children's exposure to and attainment of fluency in their languages. Concern strength was moderate to weak, and cognition concerns were weaker than exposure-fluency concerns. Transmission of a heritage language, transmission of three or more languages, presence of developmental issues, and less positive parental attitudes towards childhood multilingualism were associated with stronger concerns. These findings have both theoretical and practical implications: they advance our understanding of parental concerns and facilitate the development of support for multilingual families.

Annotated dataset for sentiment analysis and sarcasm detection: Bilingual code-mixed English-Malay social media data in the public security domain.

Sentiment analysis in the public security domain involves analysing public sentiment, emotions, opinions, and attitudes toward events, phenomena, and crises. However, the complexity of sarcasm, which tends to alter the intended meaning, combined with the use of bilingual code-mixed content, hampers sentiment analysis systems. Currently, limited datasets are available that focus on these issues. This paper introduces a comprehensive dataset constructed through a systematic data acquisition and annotation process. The acquisition process includes collecting data from social media platforms, starting with keyword searching, querying, and scraping, resulting in an acquired dataset. The subsequent annotation process involves refining and labelling, starting with data merging, selection, and annotation, ending in an annotated dataset. Three expert annotators from different fields were appointed for the labelling tasks, which produced determinations of sentiment and sarcasm in the content. Additionally, an annotator specialized in literature was appointed for language identification of each content. This dataset represents a valuable contribution to the field of natural language processing and machine learning, especially within the public security domain and for multilingual countries in Southeast Asia.

An Overview of the ADReSS-M Signal Processing Grand Challenge on Multilingual Alzheimer's Dementia Recognition Through Spontaneous Speech.

The ADReSS-M Signal Processing Grand Challenge was held at the 2023 IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP 2023. The challenge targeted difficult automatic prediction problems of great societal and medical relevance, namely, the detection of Alzheimer's Dementia (AD) and the estimation of cognitive test scoress. Participants were invited to create models for the assessment of cognitive function based on spontaneous speech data. Most of these models employed signal processing and machine learning methods. The ADReSS-M challenge was designed to assess the extent to which predictive models built based on speech in one language generalise to another language. The language data compiled and made available for ADReSS-M comprised English, for model training, and Greek, for model testing and validation. To the best of our knowledge no previous shared research task investigated acoustic features of the speech signal or linguistic characteristics in the context of multilingual AD detection. This paper describes the context of the ADReSS-M challenge, its data sets, its predictive tasks, the evaluation methodology we employed, our baseline models and results, and the top five submissions. The paper concludes with a summary discussion of the ADReSS-M results, and our critical assessment of the future outlook in this field.

Naming in a multilingual context: Norms for the ICMR-Manipal colour picture corpus in Kannada from the Indian context.

There have been many published picture corpora. However, more than half of the world's population speaks more than one language and, as language and culture are intertwined, some of the items from a picture corpus designed for a given language in a particular culture may not fit another culture (with the same or different language). There is also an awareness that language research can gain from the study of bi-/multilingual individuals who are immersed in multilingual contexts that foster inter-language interactions. Consequently, we developed a relatively large corpus of pictures (663 nouns, 96 verbs) and collected normative data from multilingual speakers of Kannada (a southern Indian language) on two picture-related measures (name agreement, image agreement) and three word-related measures (familiarity, subjective frequency, age of acquisition), and report objective visual complexity and syllable count of the words. Naming labels were classified into words from the target language (i.e., Kannada), cognates (borrowed from/shared with another language), translation equivalents, and elaborations. The picture corpus had > 85% mean concept agreement with multiple acceptable names (1-7 naming labels) for each concept. The mean percentage name agreement for the modal name was > 70%, with H-statistics of 0.89 for nouns and 0.52 for verbs. We also analyse the variability of responses highlighting the influence of bi-/multilingualism on (picture) naming. The picture corpus is freely accessible to researchers and clinicians. It may be used for future standardization with other languages of similar cultural contexts, and relevant items can be used in languages from different cultures, following suitable standardization.

Cross-cultural application of the International Classification of Cognitive Disorders in Epilepsy (IC-CoDE): Cognitive phenotypes in people with temporal lobe epilepsy in India.

OBJECTIVE: Efforts to understand the global variability in cognitive profiles in patients with epilepsy have been stymied by the lack of a standardized diagnostic system. This study examined the cross-cultural applicability of the International Classification of Cognitive Disorders in Epilepsy (IC-CoDE) in a cohort of patients with temporal lobe epilepsy (TLE) in India that was diverse in language, education, and cultural background. METHODS: A cohort of 548 adults with TLE from Mumbai completed a presurgical comprehensive neuropsychological evaluation. The IC-CoDE taxonomy was applied to derive cognitive phenotypes in the sample. Analyses of variance were conducted to examine differences in demographic and clinical characteristics across the phenotypes, and chi-squared tests were used to determine whether the phenotype distribution differed between the Mumbai sample and published data from a multicenter US sample. RESULTS: Using the IC-CoDE criteria, 47% of our cohort showed an intact cognitive profile, 31% a single-domain impairment, 16% a bidomain impairment, and 6% a generalized impairment profile. The distribution of cognitive phenotypes was similar between the Indian and US cohorts for the intact and bidomain phenotypes, but differed for the single and generalized domains. There was a larger proportion of patients with single-domain impairment in the Indian cohort and a larger proportion with generalized impairment in the US cohort. Among patients with single-domain impairment, a greater proportion exhibited memory impairment in the Indian cohort, whereas a greater proportion showed language impairment in the US sample, likely reflecting differences in language administration procedures and sample characteristics including a higher rate of mesial temporal sclerosis in the Indian sample. SIGNIFICANCE: Our results demonstrate the applicability of IC-CoDE in a group of culturally and linguistically diverse patients from India. This approach enhances our understanding of cognitive variability across cultures and enables harmonized and inclusive research into the neuropsychological aspects of epilepsy.

The dynamics and control of a multi-lingual rumor propagation model with non-smooth inhibition mechanism.

In this paper, the dynamic behaviors and control strategies of a rumor propagation model are studied in multi-lingual environment. First, an S2E2I2R rumor propagation model is proposed, which incorporates a non-smooth inhibition mechanism. Meanwhile, the existence and stability of the equilibrium are analyzed, grounded in the spreader threshold of the government intervention. Finally, the optimal control and the event-triggered impulsive control strategies are proposed to mitigate the spread of rumors, and the comparison of their effectiveness is further presented by the numerical simulation and a practical case.

Multilingual end-to-end ASR for low-resource Turkic languages with common alphabets.

To obtain a reliable and accurate automatic speech recognition (ASR) machine learning model, it is necessary to have sufficient audio data transcribed, for training. Many languages in the world, especially the agglutinative languages of the Turkic family, suffer from a lack of this type of data. Many studies have been conducted in order to obtain better models for low-resource languages, using different approaches. The most popular approaches include multilingual training and transfer learning. In this study, we combined five agglutinative languages from the Turkic family-Kazakh, Bashkir, Kyrgyz, Sakha, and Tatar,-in order to provide multilingual training using connectionist temporal classification and an attention mechanism including a language model, because these languages have cognate words, sentence formation rules, and alphabet (Cyrillic). Data from the open-source database Common voice was used for the study, to make the experiments reproducible. The results of the experiments showed that multilingual training could improve ASR performances for all languages included in the experiment, except Bashkir language. A dramatic result was achieved for the Kyrgyz language: word error rate decreased to nearly one-fifth and character error rate decreased to one-fourth, which proves that this approach can be helpful for critically low-resource languages.

Analyzing political party positions through multi-language twitter text embeddings.

Traditional monolingual word embedding models transform words into high-dimensional vectors which represent semantics relations between words as relationships between vectors in the high-dimensional space. They serve as productive tools to interpret multifarious aspects of the social world in social science research. Building on the previous research which interprets multifaceted meanings of words by projecting them onto word-level dimensions defined by differences between antonyms, we extend the architecture of establishing word-level cultural dimensions to the sentence level and adopt a Language-agnostic BERT model (LaBSE) to detect position similarities in a multi-language environment. We assess the efficacy of our sentence-level methodology using Twitter data from US politicians, comparing it to the traditional word-level embedding model. We also adopt Latent Dirichlet Allocation (LDA) to investigate detailed topics in these tweets and interpret politicians' positions from different angles. In addition, we adopt Twitter data from Spanish politicians and visualize their positions in a multi-language space to analyze position similarities across countries. The results show that our sentence-level methodology outperform traditional word-level model. We also demonstrate that our methodology is effective dealing with fine-sorted themes from the result that political positions towards different topics vary even within the same politicians. Through verification using American and Spanish political datasets, we find that the positioning of American and Spanish politicians on our defined liberal-conservative axis aligns with social common sense, political news, and previous research. Our architecture improves the standard word-level methodology and can be considered as a useful architecture for sentence-level applications in the future.

An App-Based Physical Activity Intervention in Community-Dwelling Chinese-, Tagalog-, and Vietnamese-Speaking Americans: Single-Arm Intervention Study.

BACKGROUND: Physical inactivity is associated with adverse health outcomes among Asian Americans, who exhibit the least adherence to physical activity guidelines compared with other racial and ethnic groups. Mobile app-based interventions are a promising approach to promote healthy behaviors. However, there is a lack of app-based interventions focused on improving physical activity among Asian Americans whose primary language is not English. OBJECTIVE: This pilot study aimed to assess the feasibility and acceptability of a 5-week intervention using a culturally and linguistically adapted, evidence-based mobile phone app with an accelerometer program, to promote physical activity among Chinese-, Tagalog-, or Vietnamese-speaking Americans. METHODS: Participants were recruited through collaborations with community-based organizations. The intervention was adapted from a 12-month physical activity randomized controlled trial involving the app and accelerometer for English-speaking adults. Sociodemographic characteristics, lifestyle factors, and physical measurements were collected at the baseline visit. A 7-day run-in period was conducted to screen for the participants who could wear a Fitbit One (Fitbit LLC) accelerometer and complete the app's daily step diary. During the 4-week intervention period, participants wore the accelerometer and reported their daily steps in the app. Participants also received daily messages to reinforce key contents taught during an in-person educational session, remind them to input steps, and provide tailored feedback. Feasibility measures were the percentage of eligible participants completing the run-in period and the percentage of participants who used the app diary for at least 5 out of 7 days during the intervention period. We conducted poststudy participant interviews to explore overall intervention acceptability. RESULTS: A total of 19 participants were enrolled at the beginning of the study with a mean age of 47 (SD 13.3; range 29-70) years, and 58% (n=11) of them were female. Of the participants, 26% (n=5) were Chinese, 32% (n=6) were Vietnamese, and 42% (n=8) were Filipino. All participants met the run-in criteria to proceed with the intervention. Adherence to the app diary ranged from 74% (n=14) in week 2 to 95% (n=18) in week 4. The daily average steps per week from accelerometers increased each week from 8451 (SD 3378) steps during the run-in period to 10,930 (SD 4213) steps in week 4. Participants reported positive experiences including an increased motivation to walk and the enjoyment of being able to monitor their physical activity. CONCLUSIONS: This is the first pilot study of a multicomponent intervention and evidence-based mobile phone app to promote physical activity among Asian Americans who use apps in traditional Chinese, Tagalog, or Vietnamese, which demonstrated high feasibility and acceptability. Future work focused on multilingual mobile apps to address disparities in physical inactivity among Asian Americans should be considered.

Lost in Translation: Multilingual Analysis of Online Breast Cancer Information.

INTRODUCTION: Patients use the internet to learn more about health conditions. Non-English-speaking patients may face additional challenges. The quality of online breast cancer information, the most common cancer in women, is uncertain. This study aims to examine the quality of online breast cancer information for English and non-English-speaking patients. METHODS: Three search engines were queried using the terms: "how to do a breast examination," "when do I need a mammogram," and "what are the treatment options for breast cancer" in English, Spanish, and Chinese. For each language, 60 unique websites were included and classified by type and information source. Two language-fluent reviewers evaluated website quality using the Journal of American Medical Association benchmark criteria (0-4) and the DISCERN tool (1-5), with higher scores representing higher quality. Scores were averaged for each language. Health On the Net code presence was noted. Inter-rater reliability between reviewers was assessed. RESULTS: English and Spanish websites most commonly originated from US sources (92% and 80%, respectively) compared to Chinese websites (33%, P < 0.001). The most common website type was hospital-affiliated for English (43%) and foundation/advocacy for Spanish and Chinese (43% and 45%, respectively). English websites had the highest and Chinese websites the lowest mean the Journal of American Medical Association (2.2 ± 1.4 versus 1.0 ± 0.8, P = 0.002) and DISCERN scores (3.5 ± 0.9 versus 2.3 ± 0.6, P < 0.001). Health On the Net code was present on 16 (8.9%) websites. Inter-rater reliability ranged from moderate to substantial agreement. CONCLUSIONS: The quality of online information on breast cancer across all three languages is poor. Information quality was poorest for Chinese websites. Improvements to enhance the reliability of breast cancer information across languages are needed.

Family language policy in a transnational family living in Finland: multilingual repertoire, language practices, and child agency.

Each multilingual transnational family is unique and thus deserves to be carefully studied in terms of its family language policy (FLP). Speaker-centered approaches can provide a deeper understanding of linguistic diversity in a multilingual setting. The studied Russian-Italian family is raising a multilingual boy (8:2) in Finland. The multilingual repertoire includes Russian, Italian, Finnish, English, and Hebrew. In this case-study, an ethnographic approach is used to explore the multilingual family repertoire by presenting their lived experiences and language practices. I discuss the FLP and child's active role in shaping the family's linguistic practices (child agency). The following methods were combined: semi-structured interviews, language background surveys, written diary entries, self-recordings of interactions in the family, and a language portrait that depicts the child's multilingual repertoire. The interviews and other recordings were transcribed manually. The following research questions guided the study: (1) How do the family members describe their FLP? (2) How does the FLP evolve through everyday interactions (language practices)? (3) How does the child exercise his agency in the family setting? The results reveal that the family's language practices follow predominantly an one person-one language (OPOL) strategy; consequently, the child speaks a different language with each parent. However, the analysis of the language ideologies reveals positive attitudes toward both multilingualism and all the languages in the family's repertoire, which explains the multilingual practices having multiplicity and unexpectedness. FLP is shaping the family language practices. Evidence of language hierarchy can be explained by a number of family-external and family-internal social factors.

Adapting multilingual vision language transformers for low-resource Urdu optical character recognition (OCR).

In the realm of digitizing written content, the challenges posed by low-resource languages are noteworthy. These languages, often lacking in comprehensive linguistic resources, require specialized attention to develop robust systems for accurate optical character recognition (OCR). This article addresses the significance of focusing on such languages and introduces ViLanOCR, an innovative bilingual OCR system tailored for Urdu and English. Unlike existing systems, which struggle with the intricacies of low-resource languages, ViLanOCR leverages advanced multilingual transformer-based language models to achieve superior performances. The proposed approach is evaluated using the character error rate (CER) metric and achieves state-of-the-art results on the Urdu UHWR dataset, with a CER of 1.1%. The experimental results demonstrate the effectiveness of the proposed approach, surpassing state of the-art baselines in Urdu handwriting digitization.

Identifying developmental language disorder (DLD) in multilingual children: A case study tutorial.

PURPOSE: A long-standing issue in identifying developmental language disorder (DLD) in multilingual children is differentiating between effects of language experience and genuine impairment when clinicians often lack suitable norm-referenced assessments. In this tutorial we demonstrate, via a case study, that it is feasible to identify DLD in a multilingual child using the CATALISE diagnostic criteria, Language Impairment Testing in Multilingual Settings (LITMUS) assessment tools, and telepractice. METHOD: This tutorial features a case study of one 6-year-old Urdu-Cantonese multilingual ethnic minority child, and seven age- and grade-matched multilinguals. They were tested via Zoom using Urdu versions of the Multilingual Assessment Instrument for Narratives (LITMUS-MAIN), the Crosslinguistic Lexical Task (LITMUS-CLT), the Crosslinguistic Nonword Repetition Test (LITMUS-CL-NWR), and the Sentence Repetition Task (LITMUS-SRep). RESULT: The child scored significantly lower in the LITMUS tests compared to her peers in her best/first language of Urdu. Together with the presence of negative functional impact and poor prognostic features, and absence of associated biomedical conditions, the findings suggest this participant could be identified as having DLD using the CATALISE diagnostic criteria. CONCLUSION: The result demonstrates the promise of this approach to collect reference data and identify DLD in multilingual children. The online LITMUS battery has the potential to support identification of multilingual DLD in any target language.

Telemedicine for the Underserved Racial and Ethnic Minorities During COVID-19 and Beyond.

Objective: To demonstrate that a culturally and linguistically appropriate telehealth protocol can be implemented to improve the glycemic control of patients as an extension of regular clinical services and provide continuity of care. Methods: A telehealth platform was established during COVID-19 pandemic and from numerous telehealth encounters we sampled 498 patients who received telehealth intervention over a 12-month period for specific services: Rx refill, consultation for laboratory results, wellness evaluation and education, and acute or sick visits with appropriate referrals. This telehealth platform was integrated with a remote patient monitoring system utilizing a Bluetooth-enabled glucometer for patients with diabetes compared to their abnormal baseline hemoglobin A1C (HgA1C). The Blood sugar values were recorded at predefined intervals to monitor controls for diabetes. The ethnic diversity and level of education of patients required addressing the digital divide, language interpretation, and navigation at each monitoring step. Results: This method demonstrated that a culturally and linguistically appropriate telehealth protocol can be implemented to improve the glycemic control of patients in an intervention group compared with a control group. Validation of the glycemic control was based on 70 patients identified as eligible for participation based on the inclusion criteria: a HgA1C level of 7% or higher obtained within the last 10 months. Informed consent was obtained for 42 participants based on patient participation constraints during the COVID-19 pandemic. Conclusions: We conclude that telemedicine procedures utilized for patients with little or no prior knowledge of remote self-monitoring methods can support their treatment of chronic diseases, such as diabetes. The outcomes from the implementation of telemedicine services were observed in a well-defined group of underserved racial and ethnic minority patients at our clinic. We now have a protocol to expand this to other chronic diseases and used as a regular clinical procedure.

Assessing the impact of an online dementia awareness initiative co-created with and for English, Arabic and Vietnamese speaking communities: A case study.

BACKGROUND: Awareness and understanding of dementia remain limited in ethnically diverse populations in multicultural societies due to culturally inappropriate and inaccessible information. OBJECTIVE: To establish the impact, helpers and hinderers of an online multilingual dementia awareness initiative co-created with and for English, Arabic and Vietnamese speaking people. DESIGN: A case study using mixed methods to assess the impact and implementation of an information session on dementia knowledge. SETTING AND PARTICIPANTS: The study was conducted with English, Arabic and Vietnamese speaking individuals in Canterbury-Bankstown, Australia. INTERVENTION STUDIED: A dementia alliance co-created an online multilingual dementia information session, which was delivered synchronously in English, Arabic and Vietnamese by trained facilitators. MAIN OUTCOME MEASURES: In-session group discussions, quizzes and a postsession survey assessed the impact on dementia knowledge. A postimplementation focus group explored the factors that helped and hindered the initiative. RESULTS: The online dementia information session successfully supported participants understanding of dementia causes, impacts and care strategies. The initiative was hindered by competing priorities and limited accessibility to target audiences, while it was helped by the support of an established organisation and feedback mechanisms. DISCUSSION: Ongoing dementia education and awareness-raising campaigns that are culturally sensitive are needed in communities to promote dementia literacy and help-seeking. CONCLUSIONS: An online multilingual dementia information session can be an effective way to improve dementia literacy and advocate for change in multicultural communities. PATIENT OR PUBLIC CONTRIBUTION: English, Arabic and Vietnamese speaking members of the Canterbury Bankstown Dementia Alliance participated in the co-creation and evaluation of this initiative.

Multilingual Hate Speech Detection: A Semi-Supervised Generative Adversarial Approach.

Social media platforms have surpassed cultural and linguistic boundaries, thus enabling online communication worldwide. However, the expanded use of various languages has intensified the challenge of online detection of hate speech content. Despite the release of multiple Natural Language Processing (NLP) solutions implementing cutting-edge machine learning techniques, the scarcity of data, especially labeled data, remains a considerable obstacle, which further requires the use of semisupervised approaches along with Generative Artificial Intelligence (Generative AI) techniques. This paper introduces an innovative approach, a multilingual semisupervised model combining Generative Adversarial Networks (GANs) and Pretrained Language Models (PLMs), more precisely mBERT and XLM-RoBERTa. Our approach proves its effectiveness in the detection of hate speech and offensive language in Indo-European languages (in English, German, and Hindi) when employing only 20% annotated data from the HASOC2019 dataset, thereby presenting significantly high performances in each of multilingual, zero-shot crosslingual, and monolingual training scenarios. Our study provides a robust mBERT-based semisupervised GAN model (SS-GAN-mBERT) that outperformed the XLM-RoBERTa-based model (SS-GAN-XLM) and reached an average F1 score boost of 9.23% and an accuracy increase of 5.75% over the baseline semisupervised mBERT model.

Counseling in a different language: An analysis of experiences and resources in multilingual genetic counseling.

This study set out to investigate the experiences of bilingual/multilingual genetic counselors in the United States and Canada who have counseled in a non-English language and characterize their training experiences to identify potential areas for improvement. A total of 32 bilingual and/or multilingual genetic counselors completed online surveys. Approximately 83% of participants typically counsel patients in languages for which they believe their proficiency is at least good without the participation of an interpreter. Challenges to providing language-concordant care include insufficient patient-facing translation tools/resources, with roughly half reporting they have created their own resources out of necessity. For training programs, there was a strong desire for more supervision in bilingual/multilingual genetic counseling students' non-English language during training to help foster genetics-related language skills development.