Background: There is uncertainty regarding the impact of changes in stroke care and natural history of stroke in the community. We examined factors responsible for trends in survival after stroke in a series of population-based studies.
Methods: We used statistical models to assess temporal trends in 28-day and 1-year case fatality after first-ever stroke cases registered in 3 stroke incidence studies undertaken in Auckland, New Zealand, over uniform 12-month calendar periods in 1981-1982 (n = 1,030), 1991-1992 (1,305) and 2001-2002 (1,423). Cox proportional hazards regression was used to evaluate the significance of pre-defined 'patient', 'disease' and 'service/care' factors on these trends.
Results: Overall, there was a 40% decline in 28-day case fatality after stroke over the study periods, from 32% (95% confidence interval, 29-35%) in 1981-1982 to 23% (21-25%) in 1991-1992 and then 19% (17-21%) in 2002-2003. Similar relative declines were seen in 1-year case fatality. In regression models, the trends were still significant after adjusting for patient and disease factors. However, further adjustment for care factors (higher hospital admission and neuroimaging) explained most of the improvement in survival.
Conclusions: These data show significant downwards trends in case fatality after stroke in Auckland over 20 years, which can largely be attributed to improved stroke care associated with increases in hospital admission and brain imaging during the acute phase of the illness.