OBJECTIVE: To identify variables that predict early rescreening after a negative Pap smear report. METHODS: Cohort study using the records of a statewide Cervical Cytology Registry in Victoria, Australia. The cohort comprised 31,082 women who had a negative Pap smear report during the first half of 1996 and who were rescreened within the subsequent 36 months. Early rescreening was defined as a further Pap smear within 21 months. RESULTS: The strongest predictor of early rescreening was a recommendation at the time of issuing the negative Pap smear report by the laboratory for retesting before two years (adjusted odds ratio = 3.81, 95% confidence interval (CI) 3.58-4.05). Mention of reactive or inflammatory change as part of the negative Pap smear report was also a powerful predictor (adjusted odds ratio = 1.67, 95% CI 1.50-1.85). Significant predictors associated with the women were young age, high socio-economic status and residence in the capital city. Significant predictors associated with the practitioner were if either the index or subsequent smear was collected by an obstetrician/ gynaecologist or a hospital-based clinic, or if the practitioner collecting the index smear was a female. The population-attributable risk per cent associated with the laboratory recommendation was 27%. CONCLUSIONS: This data suggests that a multifaceted strategy targeting pathology laboratories, practitioners and women may be needed to reduce early rescreening. IMPLICATIONS: Early rescreening is wasteful of health resources. New screening programs should be designed to avoid this problem.