Moderated Poster Australian and New Zealand Stroke Organisation Conference 2025

Identification of in-hospital stroke episodes and the preceding associated hospital admission diagnoses using administrative data   (122576)

Serah Kalpakavadi 1 , Hoang Phan 1 , Sabah Rehman 1 , Kate Chappell 1 , Monique Kilkenny 2 , Lachlan Dalli 2 , Muideen Olaiya 2 , Dominique Cadilhac 2 , Lee Nedkoff 3 , Helen Castley 4 , Timothy Kleinig 5 , Seana Gall 1
  1. University of Tasmania, Hobart, TASMANIA, Australia
  2. Monash University, Melbourne, VICTORIA, Australia
  3. University of Western Australia, Perth, Western Australia, Australia
  4. Royal Hobart Hospital, Hobart, TASMANIA, Australia
  5. Royal Adelaide Hospital, Adelaide, SOUTH AUSTRALIA, Australia

Background

In Australia, we lack information about strokes that occur during hospitalization, and the primary condition causing hospitalization. We aimed to identify in-hospital strokes using admitted patients records and the primary causes of admission preceding these strokes. 

Methods

We used a dataset containing all Tasmanian public hospital admitted patients from 2007-2020. A validated high-specificity algorithm along with ICD codes distinguished in-hospital strokes from primary stroke admissions (community stroke). In-hospital strokes were identified using a secondary diagnosis of stroke flagged with an acute onset flag (e.g. ‘1’ I63), or a primary diagnosis of stroke within an episode of care preceded by a non-stroke event with transfer between medical specialties. We described the cohort demographics and the primary causes of hospitalization using ICD codes. 

Results

From 2007 to 2020, there were 973 in-hospital strokes (7% of n=14,235 total stroke admissions) with mean (SD) age 72.5 (15.6) and 53% male. Seventy-one percent (n=691) were identified through acute onset flag and 29% (n=282) through primary diagnoses following transfer between specialties. Most were ischemic strokes (58%), followed by unspecified (22%) and hemorrhagic (19%) strokes. The most common diagnoses for the primary reason for admission were circulatory system disorders (30%; most frequently cerebral aneurysm and myocardial infarction), injuries (14%; most often traumatic brain injury), and neoplasms (9%; most commonly brain malignancy). 

Conclusion

Proportionally few strokes occur during a hospital admission for another reason. The primary reason for admissions included vascular involvement or non-stroke brain disorders. Further analysis may identify high-risk subgroups for preventive interventions.