[Performance of a field test as proof of automatic cell analysis]

Area of Science:

• Medical diagnostics
• Biomedical engineering
• Clinical pathology

Background:

• Manual screening methods are standard but can be labor-intensive.
• Automated cell analyzers offer potential for increased efficiency and throughput.
• Comparing the accuracy, specifically false negative rates, between manual and automated methods is crucial for clinical adoption.

Purpose of the Study:

• To establish criteria for justifying the use of automated cell analyzers over manual methods.
• To develop a statistical framework for comparing false negative rates.
• To provide formulas for calculating the required sample size for such comparisons.

Main Methods:

• Utilizing a sign test to evaluate the difference in false negative rates between automated (A) and manual (M) methods.
• Designing study protocols to identify discordant results (positive by one method, negative by the other).
• Deriving formulas to determine the necessary sample size (n) for statistically significant comparisons.

Main Results:

• The core finding is that automated analysis is validated only when its false negative rate (beta (A)) is demonstrably lower than the manual method's (beta (M)).
• The proposed design allows for direct comparison of discordant cases.
• Statistical methods are provided to support the comparison.

Conclusions:

• Automated cell analyzers can be reliably implemented if they demonstrate a lower false negative rate than manual methods.
• The study provides a statistically sound approach for validating automated screening tools.
• Sample size calculations ensure the robustness of validation studies.