The effect of analyst training on fecal egg counting variability

Area of Science:

• Veterinary parasitology
• Diagnostic methods
• Biostatistics

Background:

• Fecal egg counts (FECs) are critical for managing parasitic infections in animals.
• Analyst variability in manual FEC methods has not been quantified.
• The impact of formal training on analyst performance in FEC is unknown.

Purpose of the Study:

• To quantify between-analyst variability (BV) and analyst precision (AP) for manual and automated FEC methods.
• To assess the impact of formal training on analyst performance.
• To compare the performance of modified McMaster (MM), modified Wisconsin (MW), particle shape analysis (PSA), and machine learning (ML) algorithms.

Main Methods:

• Three untrained analysts performed FECs using MM, MW, PSA, and ML on prepared slides across various egg per gram (EPG) levels.
• Analysts repeated the protocol after formal training.
• Calculated BV, AP, and the proportion of variance attributed to analysts.

Main Results:

• Total BV was significantly lower for MM post-training (p = 0.0105).
• AP variability and analyst variance decreased for manual MM and MW methods after training.
• MM exhibited the lowest BV pre- and post-training; PSA and ML showed minimal changes with training.

Conclusions:

• Automated FEC systems are valuable when formal training is limited.
• Formal training substantially enhances the accuracy and reduces variability of manual FEC methods.
• MM remains a reliable method with low variability, even without extensive training.