Improving consistency in cervical cytology reporting

During the 1970s, the Papanicolaou method of classifying cervical cytology specimens and reporting diagnoses was replaced by more descriptive reporting systems. The plethora of reporting terms caused much confusion and a lack of standardization. To improve this situation, "The Bethesda System for Reporting Cervical/Vaginal Cytologic Diagnoses" was approved at a National Cancer Institute Workshop in 1988. In Australia, the Victorian Cervical Cytology Registry (VCCR) was established in 1989. Because of the absence of a standard format for reporting cervical cytology in that country, a coding schedule was developed by local cytopathologists. While the pattern of reporting smear diagnoses was found to be reasonably consistent within individual laboratories, substantial variation in reporting abnormal cervical smear diagnoses by 29 laboratories in Victoria, Australia, was observed. In 1992, a working party of the National Health and Medical Research Council of Australia proposed that a modified Bethesda System be adopted by Australian laboratories.

The aim of this study was to promote more uniform reporting of cervical/vaginal cytologic diagnoses by cytopathology laboratories in Victoria, Australia.

From the computer database, VCCR staff identified 80 slides that had been registered during the first half of 1991 and that covered the range of low-grade reports and negative reports. Each slide was identified by research number only. Two sets of 40 slides were compiled. Of the 29 laboratories that had worked with the VCCR during 1991, 22 agreed to participate in this study in 1992. One slide set was sent to each laboratory. An evaluation of the intralaboratory and interlaboratory consistency in reporting a set of 40 slides was undertaken. Analysis of the results compared the degree of consistency using current descriptive terminology that operates locally in Victoria with that which would pertain if the proposed Australian modification to the Bethesda System were adopted.

Intralaboratory agreement with previously reported slides was low on the squamous descriptor (49% agreement with original report) but higher on the human papillomavirus descriptor (76% agreement with original report) when the results were analyzed using the current terminology. Wide variation in reporting was apparent between laboratories; only 5% of the slides had agreement by all laboratories. Both intralaboratory and interlaboratory agreement improved substantially when results were grouped into the categories of the proposed Australian modification of the Bethesda Reporting System.

Substantial improvement in the consistency of reporting cervical cytology specimens would be likely if terminology incorporating the broad categories of the Bethesda System were adopted.