Managing demand for the direct antiglobulin test with a big data-derived predictive equation

Area of Science:

• Clinical Chemistry
• Haematology
• Laboratory Medicine

Background:

• The direct antiglobulin test (DAT) is crucial for diagnosing autoimmune haemolytic anaemia.
• Indiscriminate ordering and methodological issues can reduce clinical laboratory efficiency.
• There is a need for optimized DAT usage while maintaining diagnostic quality.

Purpose of the Study:

• To develop and validate a predictive equation for identifying negative DAT results.
• To optimize the use of DAT in clinical practice.
• To improve clinical laboratory efficiency and resource management.

Main Methods:

• Utilized laboratory information system (LIS) data from 1155 patients for DAT analysis.
• Performed multiple logistic regression analysis to establish a predictive model.
• Validated the predictive equation using a separate cohort of 164 patients.

Main Results:

• Developed a predictive equation: z = -2.884 - (0.373 x Haptoglobin) + (0.312 x %Ret).
• The equation achieved a sensitivity of 81.6% and a negative predictive value of 95.8%.
• The model demonstrated an area under the curve of 0.812, with potential to reduce DAT testing by 61.6%.

Conclusions:

• The developed equation effectively predicts negative DAT results.
• Its integration into LIS offers a practical screening tool for optimizing DAT demand.
• The equation supports efficient clinical laboratory practice and resource management.