The estimation of false negatives in medical screening

Abstract

In a medical screening program for early detection of disease, one or more screening modes are administered to an apparently healthy population. Knowledge of the true disease status for all screened individuals would allow estimation of the false negative and false positive rates for each mode of detection and for the program as a whole. This paper develops capture-recapture methods applicable to programs as a whole. This paper develops capture-recapture methods applicable to programs when follow-up of individuals negative on screening is not performed or is incomplete. The methods require at least two independent modes of detection. Data from a breast cancer screening program illustrate the procedure. The results of four screening examinations at approximately one-year intervals and the long-term follow-up of all screened individuals support the usefulness of these methods in the evaluation of a screening program.