Dysbacteriosis is a microscopical diagnosis. In women with dysbacteriosis, an overgrowth of coccoid bacteria and almost a complete absence of lactobacilli are observed in the (stained) vaginal smear. The aim of this study was to determine the accuracy of this microscopic diagnosis in clinical practice. The analysis concerned 342 consecutive cases in which the microscopy of the stained smears was performed by general practitioners trained in diagnosing dysbacteriosis. These smears were sent to the pathologist for confirmation of the microscopical diagnosis of the clinician. The cytological diagnoses of the pathologist, sometimes performed on restained slides when the quality of the staining was substandard, were considered as the "gold standard." In 92 of the 342 cases, dysbacteriosis was unequivocally established by the pathologist. Sensitivity and specificity of the microscopical diagnoses of the clinicians were 40% and 85%, respectively. There were 37 false-positive and 54 false-negative diagnoses of dysbacteriosis rendered by the clinicians. The most frequent reason for a false-negative diagnosis was an excess of lactobacilli in the smear. This study shows that even in stained smears it is difficult for clinicians to render a correct evaluation of the status of the vaginal flora.