Introduction: Cervical cancer is the leading cause of cancer deaths in the female population in Colombia. The low impact of Papanicolaou smears in reducing cervical cancer mortality in some countries has been attributed to their low reproducibility and high rates of false negatives.
Objective: To evaluate the quality of the Papanicolaou smears in four regions of Colombia comparing the original report given by provincial cytologists or pathologists with a second report made by a team of expert pathologists from the Instituto Nacional de Cancerología of Colombia.
Materials and methods: A sample of 4,863 Papanicolau smears was selected by a simple stratified randomized sampling method. Three strata were defined according to the original cytological report as negative, positive and unsatisfactory. All slides were newly interpreted with Bethesda 2001 by two independent experienced pathologists blinded to the first results. Non-weighted kappa values were calculated for degree of agreement.
Results: Overall, the concordance between the two evaluations was very low with a kappa value of 0.03 (95% CI: 0-0.06). With regard to abnormalities in squamous cells, evaluation concordance was moderate with a value of 0.47 (95% CI: 0.41- 0.53); a trend was noted suggesting higher levels of concordance in the evaluations from the Tolima and Magdalena Provinces.
Conclusion: Problems related with the quality of Papanicolau smears or the accurate reading thereof may be factors that explain the low impact of massive screening in cervical cancer mortality in Colombia.