Indirect method for validating transference of reference intervals

Area of Science:

• Clinical Chemistry
• Laboratory Medicine
• Health Informatics

Background:

• Transference of multicentre reference intervals (RIs) is crucial but often relies on limited verification methods.
• Current verification techniques, like using reference individuals or single serum samples, have inherent limitations.
• Retrospective methods using historical data exist, but a near prospective approach is desirable for new method implementations.

Purpose of the Study:

• To develop and evaluate an indirect, near prospective method for validating the transference of reference intervals (RIs).
• To utilize data mining of laboratory information systems to calculate and verify RIs.
• To compare the results of this novel indirect method with established verification techniques.

Main Methods:

• A data mining experiment was conducted using historical laboratory information system data.
• Request patterns of patients with assumed healthy characteristics were identified to extract laboratory results.
• New RIs were calculated using an indirect method and compared to transferred RIs verified by a reference serum.

Main Results:

• The indirect data mining method produced results generally similar to the established reference serum verification method for RIs.
• The method successfully generated results across all patient stratifications.
• Minor unexplained incongruences were observed for specific stratifications and one analyte compared to the NORIP RIs.

Conclusions:

• Patient request patterns can serve as a surrogate for good health status in validating RIs.
• The proposed data mining method offers a viable approach for validating RI transference.
• This method is potentially applicable in healthcare systems with similar laboratory information infrastructure.