A method is presented for estimating a discrete form of the lead time distribution of cases detected in a single-shot screening program. The procedure is based upon a three-state progressive disease model of disease natural history and screening which focuses on the age at entry into the preclinical disease state, the duration of the preclinical state, and the age at screening or observation. Estimation is accomplished by comparing observed incidence rates of the disease between a screened group and a randomized control group in successive follow-up intervals after the screen. No assumptions concerning the false negative rate of the screening test or the distribution of the duration of time spent in the preclinical disease state are required. An example is presented by means of breast cancer screening data.