The Linguist

20 The Linguist Vol/65 No/3 ciol.org.uk/thelinguist FEATURES Could language technology improve patient outcomes in a failing maternity care unit, asks Renata Towlson The Nottingham and Nottinghamshire Local Maternity and Neonatal System came under scrutiny this year following the largest independent review of maternity services in NHS history. Led by Donna Ockenden, the review identified serious failings in maternity care. Among the concerns were inequalities experienced by women from ethnic minority backgrounds, and shortcomings in communication with women who have limited English proficiency (LEP) or whose first language isn't English. The Ockenden Review found that inequitable experiences and outcomes persisted across maternity services over a prolonged period from at least 2013 to 2021: 1 "Women from minoritised and marginalised populations frequently report racism, discrimination, unconscious bias, lack of poor [sic] culturally sensitive care, communication barriers, and unequal access to services. Migrant women, particularly those with precarious immigration status and those with LEP, face additional barriers relating to interpretation and fear of NHS charging, data sharing and immigration enforcement." As Senior Interpreting and Translation Services Innovation and Improvement Lead, I was asked to produce a report examining Nottingham University Hospitals NHS Trust's (NUH) linguistic support and a critical review of how NUH ITS (interpreting and translation services) are utilised to support women from ethnic minority backgrounds. We identified that preference was always given to a face- to-face interpreter and, when unavailable, reluctant use of a telephone interpreter. In urgent instances, like a sudden onset labour or clinical emergency, timely access to adequate linguistic assistance wasn't always possible. Consequently, some women experienced additional distress, uncertainty and reduced understanding of their clinical situation or care. We decided to trial alternative approaches while acknowledging the effort any trial requires from an organisation. We know how difficult it can be to convince staff and patients to do things differently and provide feedback. However, there are compelling real-world circumstances facilitating our research and operational decisions that we cannot ignore. Digital solutions In 2023-2024, NUH was asked to evaluate two complementary digital communication solutions that seek to improve communication where language barriers exist: CardMedic and Pocketalk. CardMedic is a clinically governed platform supporting effective communication in healthcare settings where language, cognitive, health literacy or other barriers exist. Pocketalk is a real-time, two- way translation system distributed by SCALE Innovation. Covering more than 84 languages, the handheld device is the size of a mobile phone – unique on the market in this respect. Methodology and strict evaluation criteria were designed to assess the platforms with a view to using them at NUH. We looked at their impact in four key areas: 1 Quality: improving staff and LEP patient experience and organisational culture. 2 Finance: using existing resources efficiently; reducing costs through proactive and confident use of AI and digital resources. 3 Equality, diversity and inclusion: strengthening anti-racism and inclusion capability; giving staff access to multiple resources to communicate meaningfully with LEP patients and their families. 4 Legal: supporting compliance with equality legislation and national guidance. CardMedic outcomes The evaluation of CardMedic provided strong, consistent evidence of enhanced safety, equity and efficiency across maternity, neonatal and community services. During the pilot, 88% of maternity staff were aware of CardMedic and 86% of those had registered to use the platform. Nearly half (48%) had used it in clinical practice and 81% of users rated it as extremely or somewhat useful. 2 Staff consistently reported increased confidence communicating with LEP women and families, improved ability to provide clear and timely explanations, greater efficiency during routine clinical interactions, and faster access to multilingual information. Its integrated call feature offers immediate, one- click access to a remote interpreter, making it quicker to connect to a (human) professional. System analytics demonstrated sustained growth across the pilot: 274 telephone interpreter calls totalling 1,648 minutes (in seven months), 5,800 pre-written card interactions, 75 BSL interactions and 4,489 Live Translate (automated translation) activities. After maternity staff received bespoke training and troubleshooting support, they began to use the technology more often and more confidently. 3 This suggests that, when introduced with the right support, CardMedic can improve communication equity, patient experience and operational efficiency while supporting the NHS's priority of reducing communication-related health inequalities. Can AI deliver?

• Cover