Linkage of hospital and death records increased identification of dementia cases and death rate estimates

Abstract

Background/aim: To use health record data linkage to improve case ascertainment and death rate estimates of deaths with dementia.

Methods: Retrospective population study. Western Australians older than 20 years who died between 1990 and 2005 with a dementia diagnosis in Western Australian hospitals and/or death records using the Western Australian Data Linkage System were classed as having a lifetime history of dementia. Cases with dementia documented on death certificates were classified as having dementia as a contributory cause of death (COD). Age-standardized death rates (ASDR) were estimated.

Results: 29,884 decedents were identified with a lifetime history of dementia. 88.2% had hospital records and 55.9% had death records indicating dementia. The ASDRs for dementia as a contributory COD increased from 50 to 81 per 100,000 person-years from 1990 to 2005. ASDR for lifetime history of dementia increased from 80 to 140 per 100,000 person-years over the same time period. In 2005, 50.1% (95% confidence interval: 47.2-53.1) of all females and 39.2% (35.8-42.7) of all males aged 85+ years died with dementia.

Conclusion: Data linkage nearly doubled case ascertainment of deaths with dementia because people hospitalized with dementia often lacked dementia documentation at death. These results have important implications for strategic healthcare planning. Data linkage methodology could improve studies of hospitalisation trends in dementia.