Use of Middleware Data to Dissect and Optimize Hematology Autoverification

Area of Science:

• Clinical laboratory science
• Medical technology
• Diagnostic hematology

Background:

• Hematology analysis is a high-volume clinical laboratory testing area.
• Autoverification of results using computer rules improves efficiency and targets manual review.
• Established autoverification rules were in place for over a decade.

Purpose of the Study:

• To analyze autoverification rules and identify opportunities for increased automation.
• To evaluate rules triggering manual review during a hematology instrument migration.
• To optimize autoverification rates in a large academic medical center laboratory.

Main Methods:

• Analysis of autoverification rules and flags associated with manual review.
• Evaluation of immature granulocyte (IG) flag and platelet (PLT) analysis methods (impedance vs. fluorescence).
• Adjustment of IG flag threshold from 2% to 5% and standardization to PLT-F analysis.

Main Results:

• Autoverification rates increased from 87.8% to 93.5% for CBC panels and 85.8% to 89.8% for individual CBC components.
• The IG flag alone prevented autoverification in 6.0% of samples at the 2% threshold.
• Raising the IG threshold to 5% and switching to PLT-F analysis were key to improving autoverification.

Conclusions:

• Detailed analysis identified specific rules (IG flag, PLT analysis) for optimization.
• Adjustments to autoverification parameters significantly enhanced automation rates.
• Improved software tools are needed for efficient and routine evaluation of autoverification parameters.