Developing an automated language-based cognitive assessment; CognoSpeak – working with an under-represented ethnic minority group

<h3>Background</h3> Current memory screening tools were developed based on normative data from almost exclusively white populations. This can unfairly disadvantage access to diagnoses for BAMER communities, and individuals who use English as an additional language. We collaborated with Israac, a Sheffield-based Somali community group, to investigate how Somali and white British individuals’ responses differ on both traditional cognitive screening tools and in our automated cognitive assessment tool, CognoSpeak™ . <h3>Methods</h3> We trained two research champions from Israac to recruit and assess Sheffield- based Somali participants using the Montreal Cognitive Assessment (MoCA), Rowland Universal Dementia Assessment Scale (RUDAS), Multicultural Cognitive Examination (MCE), and CognoSpeakTM. <h3>Results</h3> Preliminary analysis of 36 Somali and British-Somali participants found that 47.2% of participants scored below the MoCA threshold of 26/30. In contrast, only 5.56% of the participants scored below the MCE threshold of 70/100, and no participants scored below the RUDAS threshold of 23/30. <h3>Conclusion</h3> Cognitive tools must be co-developed for cross-cultural use. Training individuals from within the community to conduct research improved recruitment and community trust in researchers. We will present a larger data set from both the Somali and white British populations results’ on CognoSpeak and the three traditional cognitive assessments.