In multicultural long-term care homes, many residents—especially those living with dementia—lose functional English and revert to their first language, making everyday communication with predominantly English-speaking staff extremely difficult. This leads to misunderstandings about care plans, unmet needs, and escalation of responsive behaviours because residents cannot clearly express discomfort, fear, or preferences. Staff currently rely on ad hoc workarounds such as searching for bilingual colleagues, phoning family members to interpret, or using generic translation apps, all of which are time-consuming, inconsistent, and not optimized for clinical conversations. These gaps increase staff stress, documentation errors, and potentially unsafe care conditions during tasks and Code White events. There is an urgent need for a reliable, healthcare-specific, real-time translation solution that supports person-centred communication across multiple languages at the point of care.
MIRA is an AI-powered auto-translation platform built for healthcare communication in long-term care homes. It delivers real-time voice-to-voice translation in 21+ languages through an easy-to-use voice–to–text–to–speech interface on tablets used by frontline staff. The system is privacy-conscious (PIPEDA compliant, HIPAA aligned) and can detect agitation, interpret tone and sentiment, and tailor prompts to the care environment and staff role to support safer interactions with residents who have cognitive impairment. In a 102-bed long-term care home, a 17-bed pilot unit is using MIRA to improve communication between staff, residents, and families, reduce work-related frustration, and decrease behavioural incidents linked to language barriers, with impact tracked through pre/post measures, usage logs, surveys, and focus groups. The project will produce practice-based evidence and a practical approach for scaling real-time multilingual translation across residential care.
