Reliability of Stroke Registry Data
Author Information
Author(s): Reeves Mathew J, Mullard Andrew J, Wehner Susan
Primary Institution: Michigan State University
Hypothesis
This study aims to assess the inter-rater reliability of individual data elements in the Michigan prototype of the Paul Coverdell National Acute Stroke Registry.
Conclusion
The study found excellent reliability for many data elements, but poor reliability for time-related variables, indicating a need for improved data collection methods.
Supporting Evidence
- Excellent reliability was observed for many registry variables including age, gender, and race.
- Poor reliability was noted for time-related variables such as stroke onset time and initial brain imaging time.
- The study highlights the need for improved data definitions and training for abstractors.
Takeaway
The study checked how well different people record the same stroke data, finding that some data is recorded very reliably, while other important times are often missed.
Methodology
Data was collected from 15 hospitals over 6 months, with trained staff abstracting charts and an audit nurse re-abstracting a sample of cases to assess reliability.
Potential Biases
There were systematic differences between hospital abstractors and the audit abstractor, indicating potential bias in data recording.
Limitations
The study had limited resources for chart abstraction and relied on retrospective data for the audit, which may have increased discrepancies.
Participant Demographics
{"mean_age":70.9,"gender_distribution":{"female":59,"male":41},"race_distribution":{"white":77,"black":17,"other":1,"not_documented":5}}
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website