Microscopic inter-observer reliability of intestinal parasitic infections in trained laboratory technicians, Mexico
Intestinal parasitic infections caused by Giardia lamblia (GL), Ascaris lumbriocoides (AL) and Entamoeba histolytica/dispar (Eh/Ed) are highly prevalent among indigenous groups in Mexico. In resource-constrained settings, direct microscopic fecal examination continues to be a common diagnostic method in spite its limited accuracy. This study aimed at illustrating the effect of training local laboratory technicians from a rural reference hospital located in a marginalized indigenous region of northern Mexico to assess the inter-observer reliability of GL, AL, and Eh/Ed diagnoses. Two experienced technicians working at the hospital were trained and standardized for two full weeks in the Parasitology Laboratory at the National Children’s Hospital from Mexico City. Diagnoses were made by microscopy of two serial stool samples processed using the modified Faust zinc sulphate centrifugal flotation technique to concentrate AL eggs and GL and Eh/Ed cysts. Observations were done independently, and the final diagnosis for each observer was established when at least one of the two samples resulted positive. Reliability analyses from serial stool samples were conducted using Cohen’s kappa correlation coefficient (ĸ) for each parasite. Agreement between observers reached 88.7, 72.4, and 80.5% for Eh/Ed, AL, and GL, respectively. Largest kappa coefficient was observed for GL (ĸ=0.55), followed by Eh/Ed (ĸ=0.30), and AL (ĸ=0.08). Prevalence of Eh/Ed, AL and GL according to observers 1 and 2 were 3.4 vs. 13.5%, 4.0 vs. 28.2%, and 32.2 vs. 33.3%, respectively. Except for GL, reliability was very low leading to major differences in prevalence estimates. These results question the value of training technicians, as intestinal parasitic microscopic diagnoses seemed to be very difficult to replicate between observers questioning their validity, leading to differences in clinical decisions, and in prevalence estimates.