H. Straatman et al., ESTIMATING LEAD TIME AND SENSITIVITY IN A SCREENING-PROGRAM WITHOUT ESTIMATING THE INCIDENCE IN THE SCREENED GROUP, Biometrics, 53(1), 1997, pp. 217-229
Early indicators of the effectiveness of a screening test for chronic
diseases such as breast cancer are the length of time the diagnosis is
advanced by screening, the lead time, and the sensitivity of the scre
ening test. This paper describes a model for simultaneously estimating
the mean lead time and the sensitivity when only the number of cancer
s detected at the successive screenings and the number of cancers occu
rring in the time interval between the screening examinations are know
n. This model is particularly useful in assessing the effect of screen
ing when the underlying cancer incidence in the screened group is unkn
own. The model is fitted to the data of 235 screen-detected breast can
cer cases and 146 interval cancers diagnosed across 6 screening rounds
of the program in Nijmegen. The maximum likelihood estimate for the m
ean lead time ranges from 1.3 years in the under age 50 group to 2.2 y
ears in the age 50-65 group, both estimates having large confidence in
tervals. The corresponding sensitivity estimates are 0.92 and 1.00.