Decision analysis for best management of mildly dyskaryotic smear

Abstract

Should a woman with a mildly dyskaryotic cervical smear be referred for colposcopy or should the smear be repeated? One way to answer this question is to use decision analysis and compare the expected mortality and cost of each policy. Data for each component of the question were obtained from published work worldwide and were supplemented with an audit of mildly dyskaryotic smears in West Yorkshire, UK. 2 out of 1000 women with an initial mildly dyskaryotic smear will develop cancer if a conservative repeat smear policy is adopted in association with five-yearly cervical screening. This number can be reduced to 1·6 per 1000 if cervical screening is offered every three years. A policy of immediate referral for colposcopy is also associated with a subsequent cancer rate of 1·6 per 1000. Therefore, repeating the smear is almost as effective as an immediate referral to a colposcopy unit. Even if a five-yearly cervical screening programme is adopted, 2500 women with a mildly dyskaryotic smear will need to be referred for immediate colposcopy to save 1 additional cancer. A conservative policy is not financially cheaper: an average of six additional smears is required to save each colposcopy referral. Sensitivity analysis shows that the excess cost of the conservative policy increases exponentially as the risk of a subsequent cytological abnormality exceeds 60%. Local cytopathology laboratories should audit their recurrent dyskaryosis rate associated with borderline, mild, and moderate dyskaryosis before accepting the U-turn in the national recommendations.