Related Experiment Videos
Analytical goals for the estimation of non-Gaussian reference intervals
P Hyltoft Petersen1, E M Gowans, O Blaabjerg
1Department of Clinical Chemistry, Odense University Hospital, Denmark.
Scandinavian Journal of Clinical and Laboratory Investigation
|December 1, 1989
Summary
Establishing analytical quality goals ensures shared reference intervals across laboratories. These goals minimize population percentages outside reference limits, considering various biological distribution types for accurate health assessments.
Area of Science:
- Clinical Chemistry
- Laboratory Medicine
- Biostatistics
Background:
- Laboratories often require shared reference intervals for homogeneous populations.
- Defining analytical quality is crucial for consistent and reliable health data.
Purpose of the Study:
- To establish analytical quality goals for laboratories sharing reference intervals.
- To evaluate these goals based on acceptable population percentages outside reference limits.
Main Methods:
- Assessed goals assuming combined analytical bias and imprecision should not exceed 1.3-4.4% of the population outside reference limits.
- Evaluated goals for log-Gaussian and transformable distributions.
- Discussed goals for non-parametric distributions.
Main Results:
- For log-Gaussian distributions, maximum allowable bias and imprecision depend on the reference limit ratio.
- With a ratio of 1.25, bias <1.5% or coefficient of variation <3.5% is required.
- With a ratio of 10, bias up to 15% or coefficient of variation up to 34% is permissible.
Conclusions:
- Analytical quality goals are proposed for inter-laboratory reference interval harmonization.
- The study provides specific bias and imprecision limits based on population distribution characteristics.
- These findings support the development of robust quality standards in clinical diagnostics.