DEATH CERTIFICATION IN TREATED CASES OF PRESENILE ALZHEIMERS-DISEASE AND VASCULAR DEMENTIA IN SCOTLAND

Introduction: although death certification data are commonly used in d ementia epidemiology, their reliability has been questioned. Methods: death certificates were available from the Registrar General for Scotl and for all patients with Alzheimer's disease/presenile dementia (AD P SD) or vascular dementia (VaD) who had died in Scotland up until 31 De cember 1994. Primary (immediate and underlying) and contributory cause s of death were noted as well as place of death. Occupations of male p atients were obtained from death certificates or from case notes and c lassified according to the Standard Occupational Classification. Bronc hopneumonia was considered a non-specific cause of death and specific causes of death were classified as: cardiac disease, dementia, cerebro vascular disease, neoplasms, other vascular diseases and other disease s. Place of death was recorded as psychiatric hospital, district gener al hospital, nursing home or private residence. Results: death certifi cates of 398 people who had been treated for AD PSD and 348 who had be en treated for VaD were identified. Bronchopneumonia was the most comm on immediate cause of death in the AD PSD group (70.9%) but less so fo r the VaD group (51.7%). For both groups place of death was associated with significant differences in pneumonia being reported as the immed iate cause of death as well as specific underlying and contributory ca uses of death. Dementia was recorded for 90.5% of AD PSD patients but for only 49.7% of the VaD group. Conclusions: Scottish death certifica te data significantly underestimate the prevalence of presenile VaD. C hanges in patterns of institutional care may affect dementia rates est imated from death certificate data.