The incidence of and mortality from cervical invasive carcinoma have decreased over the last 30 years in the developed world, a fact which has been attributed mainly to the implementation of cervical cytology screening programmes (Papanicolaou test). The accuracy of this screening has been questioned because of false negative reports, thus other technologies have been proposed. There are only 12 studies evaluating cytology screening in low prevalence populations. In these studies the specificity appears to be high (85-100%) but the sensitivity estimates are variable and generally much lower (22-99%). Sensitivity is increased if lower thresholds of disease are used. Sampling or laboratory errors are important causes of false negative reports. Cytological cervical screening does not demonstrate the characteristics of an optimal screening test since it is relatively insensitive to the presence of disease although highly specific. The true screening accuracy of cytology needs to be assessed so that we can introduce strategies to improve smear taking and interpretation.