A standardized test for visual analysis of human sperm morphology

Objectives To develop a method to train and test simultaneously a large number of observers in the practice of visual sperm morphology analysis. Design Photographs of fixed and stained sperm were prepared. Fields of suitable sperm images were selected and individual images were numbered on each negative. Two tests, which contained a total of 100 sperm images, were created. Thirty images in each test consisted of three repeats of 10 images, while 70 images in each test were unique. The tests were administered to individuals participating in an American Fertility Society postgraduate course. Sperm images were projected on a screen and participants classified each sperm using the method that was used in their own laboratory. Setting Postgraduate course of The American Fertility Society. Results The majority of individuals participating in the tests used some version of the World Health Organization method. The group using the Strict method reported a lower value for the percentage of normal sperm than the groups using the other methods. The variability for the percentage of normal sperm was highest for the Strict method. The degree of classification reversal, i.e., classifying a sperm as normal during one repeat and then reversing the classification during another repeat, was high for all groups (26% to 44% of the classifications). Some degree of improvement was seen from test 1 to test 2. Conclusions It is possible to develop efficient and inexpensive methods to train observers to perform the sperm morphology assay. Such methods also enable the objective measurement of the acquisition of proficiency, comparison between different classification methods, and identification of specific differences between observers. Such methods will become more important with implementation of the Clinical Laboratory Improvement Act.