Reliability of Blood Test Results in Korean Study
Author Information
Author(s): Yang Jae Jeong, Yang Ji Hyun, Kim Jimin, Cho Lisa Y., Park Boyoung, Ma Seung Hyun, Song Sang Hoon, Min Won-Ki, Kim Sung Soo, Park Man Suck, Park Sue K.
Primary Institution: Seoul National University College of Medicine
Hypothesis
Can clinical test values from different laboratories in the Korean Genome and Epidemiology Study be integrated through a statistical adjustment algorithm?
Conclusion
The two laboratories in the KoGES have good intra- and inter-laboratory reliability for ten chemical test values, and data can be integrated through algorithmic statistical adjustment using regression equations.
Supporting Evidence
- All correlation coefficient values indicated good intra- and inter-laboratory reliability, ranging from 0.842 to 1.000.
- Kappa coefficients were greater than 0.75, indicating substantial reliability.
- Regression models based on trial results had strong R-square values and zero sums of residuals.
Takeaway
This study shows that blood test results from different labs can be made to match up using math, so doctors can trust the results more.
Methodology
The study used quadruplicated standardized serological samples and external quality control data to assess reliability and constructed statistical compensation models using linear regression analysis.
Potential Biases
Potential discrepancies in test results due to differences in laboratory methods and instruments.
Limitations
The study focused on only two laboratories, and results for some tests like AST showed lower reliability.
Participant Demographics
Participants were part of the Korean Genome and Epidemiology Study, which focuses on major diseases among Koreans.
Statistical Information
P-Value
p<0.001
Statistical Significance
p<0.001
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website