A longitudinal methodology was adopted with the aim of investigating some of the bilingual linguistic factors which might operate in older adults in the early
stages of dementia. 120 referrals from General Practitioners in North Wales were obtained. Of the 41 which met the criteria at follow up, 25 completed data
collection. The Kendrick Cognitive Tests for the Elderly (KCT) and the Wechsler Memory Scale (WMS) were used to diagnose dementia in the study. 16 subjects
were diagnosed as suffering from dementia by the end of the study period. Subscales of the Boston Diagnostic Aphasia Examination Battery (BDAE) and the
Token Test were used to assess language skills. All tests were randomly administered in Welsh and English over a period of four days at six monthly intervals for between eighteen months and two years. A Language Use and Contact Questionnaire (LUCQ) was designed for the study as a semi structured interview schedule.
A significant performance deterioration in the dementia subjects was found.
As no method of measuring bilinguality could be identified from the literature it was hypothesised that performance change in dementia could distinguish
between a predominantly "Welsh" or "bilingual" -subject. It was then hypothesised that past language use and contact over six developmental life stages (as measured by the questionnaire developed for this study) would be related to present performance deterioration (measured in %change scores).
Findings do not support this hypothesis. Qualitative analyses of cases identified on• the basis of their past linguistic contact and dementia severity
identified Translation and Asides as strategies to facilitate response.
The study identified Subscales which were both sensitive and resistant to dementia and an order of decline in language skills was identified.
The main findings of the study were discussed in relation to methodological issues, clinical implications and future research directions.