LATENT CLASS ANALYSIS IN DIAGNOSTIC TESTS EVALUATION FOR CANINE LEISHMANIA INFANTUM INFECTION

Area of Science:

• Veterinary Diagnostics
• Parasitology
• Biostatistics

Background:

• Accurate diagnosis of canine leishmaniasis is crucial for effective treatment and control.
• Classical diagnostic accuracy assessment can be biased by imperfect reference tests.
• Leishmania infantum diagnosis in dogs often lacks a definitive reference standard.

Purpose of the Study:

• To compare Latent Class Analysis (LCA) with classical methods for evaluating diagnostic accuracy of canine leishmaniasis tests.
• To assess if LCA provides more reliable estimates of test performance, particularly for molecular assays.
• To address limitations of small sample sizes and confounding factors in diagnostic accuracy studies.

Main Methods:

• Compared Latent Class Analysis (LCA) with traditional 2x2 contingency tables.
• Utilized data from microscopy, IFAT, LAMP, and PCR tests on 75 suspected canine leishmaniasis cases in Tunisia.
• Microscopy was used as a reference standard in classical calculations, while LCA provided independent estimates.

Main Results:

• LCA broadly corroborated classical estimates for sensitivity and specificity.
• LCA provided higher sensitivity (68.7%) and specificity (86.2%) estimates for the LAMP cpb assay compared to classical methods (54.2% and 80%, respectively).
• LCA effectively controlled for limitations such as small sample size and confounding factors, yielding consistent test characteristic estimates.

Conclusions:

• Latent Class Analysis (LCA) is a valuable tool for evaluating diagnostic test accuracy in the absence of a perfect reference standard for canine leishmaniasis.
• LCA offers improved and more reliable diagnostic accuracy estimates, especially for molecular tests like LAMP.
• This statistical approach enhances the evaluation of diagnostic methods in complex epidemiological settings.