Related Experiment Video
Updated: Aug 12, 2026
Dual-Task Stroop Paradigm for Detecting Cognitive Deficits in High-Functioning Stroke Patients
Published on: December 16, 2022
Validity of stroke diagnosis on hospital discharge records in Saskatchewan, Canada: implications for stroke
1Department of Community Health and Epidemiology, University of Saskatchewan, Saskatoon, Canada.
Insights
This study assessed the accuracy of stroke diagnoses in Canadian hospital records. Tertiary hospitals showed high accuracy (90%), but community hospitals had lower validity (78%), limiting regional stroke surveillance.
Area of Science:
- Neurology
- Public Health
- Health Informatics
Background:
- Accurate diagnosis coding on hospital discharge records is crucial for epidemiological studies and public health surveillance.
- Previous research has highlighted potential inaccuracies in administrative data for specific medical conditions.
Purpose of the Study:
- To evaluate the validity of stroke diagnoses recorded in hospital discharge data in Saskatchewan, Canada.
- To compare the accuracy of stroke diagnosis between tertiary-care and community hospitals.
Main Methods:
- A retrospective review of 1,494 hospital discharge records with a primary diagnosis of stroke or stroke-related condition.
- Utilized the 1980 USA National Survey of Stroke clinical algorithm as the gold standard for validation.
- Compared positive predictive values (PPVs) between different hospital types and regions.
Main Results:
- The positive predictive value for a primary stroke diagnosis in tertiary-care hospitals was approximately 90%.
- In community hospitals, where many stroke cases were coded under ICD9 436, the PPV was 78%.
- Significant regional variations in diagnostic accuracy were observed.
Conclusions:
- While stroke diagnoses in tertiary-care hospitals are largely valid, lower accuracy in community hospitals poses challenges.
- The variability in diagnostic coding across regions limits the utility of hospital discharge data for comprehensive stroke surveillance in Saskatchewan.
Abstract:
This study examines the validity of the diagnosis of stroke on hospital discharge records in Saskatchewan, Canada. In total, 1,494 records with a discharge diagnosis of 'stroke' or a 'stroke-related condition' were reviewed. The clinical algorithm of the 1980 USA National Survey of Stroke was considered the 'gold standard'. The positive predictive value of a primary diagnosis of stroke in the tertiary-care hospitals was about 90%. In community hospitals the majority of stroke cases were coded as ICD9 436 in which the positive predictive value was 78%. The variation between regions would limit the use of hospital discharge data for stroke surveillance.
