Campaigns involving sending personally addressed letters to encourage women to have Pap smears increase Pap smear rates. The aim of this study was to assess whether this effect is maintained when campaigns are repeated regularly. In October 1992, a letter reminding women of the importance of screening was mailed to all women in three New South Wales postcode regions where a similar letter had been sent three years previously. The response was compared to the response in three regions receiving no earlier letter. The number of women attending for cervical screening during the three months after distribution of the letters was assessed from Health Insurance Commission claims for cervical cytology. These attendances were compared with expected attendances based on the attendance patterns over 28 pre-intervention quarters. Significant postintervention increases in attendance were observed in all three regions receiving an initial letter. However, in one region, the increase in attendances, around 1 per cent of eligible women, was not significantly greater than the increase in the control region (z = 0.15, P = 0.88). The second letter campaign had no measurable effect on attendances. No significant increase in screenings was observed in two of the towns. A significant increase was observed in one region, but this was not significantly greater than the increase in the control region (z = -0.05, P = 0.96). These results suggest that repeated direct-mail campaigns to promote screening for cervical cancer may be of no benefit. A one-off campaign may result in an increase in screenings in the short term.