To compare the effectiveness of rapid screening of cervical smears as a method of internal quality control with 10% random rescreening. From June 5 to July 14, 1995 (6 weeks), all consecutive cervical smears received in the department (n = 8,800) were entered into the study and were prescreened for a duration of 30 seconds (n = 2,938), 1 minute (n = 2,925) or 2 minutes (n = 2,937) over a period of 2 weeks each. Rapid screening of all negative and unsatisfactory smears detected more cytologic abnormalities than would be expected with 10% random rescreening. Thirty-second rapid screening of all negative and unsatisfactory smears was more efficient in detecting false negatives than screening a proportion of the smears for longer periods of time. Rapid screening was also a very effective method of detecting severe cytologic abnormalities in unscreened smears, detecting over 90% of high grade lesions. Rapid rescreening of negative and unsatisfactory cervical smears is recommended as a very effective method of internal quality control. It is superior to 10% random rescreening in reducing the false negative rate. Thirty-second rapid rescreening is the most efficient period for which smears should be screened. Rapid screening of unscreened smears could be used as a means of selecting patients for prompt referral when a laboratory backlog exists.