Paired cervical smears: a method of reducing the false-negative rate in population screening.

Cervical smears taken by the wooden Ayre spatula (widely used in the United Kingdom) and by a plastic spatula (Armovical) were compared for their ability to detect dysplastic cervical epithelium and endocervical cells. At the same time, the value of taking duplicate smears was assessed. Paired cervical smears were obtained at a single attendance from 21,352 women, using the 2 types of spatulas in 4 different sequences. Although individually the Armovical proved superior to the Ayre spatula for the collection of endocervical epithelial material, this advantage was not complemented by any marked improvement in its detection of cervical dysplasia. Paired smear sampling, however, was clearly more rewarding in this respect and also allowed the introduction of laboratory quality control, as each preparation of any given pair was examined by a different screener. The false-negative rate was 18.5%: it comprised 11.1% sampling error and 7.4% screening error. It is concluded that paired smear sampling in population screening would significantly reduce the false-negative rate.