A qualified (indeterminate) diagnosis (QD), such as "suggestive of malignancy," is thought to complicate patient management by heightening clinical uncertainty. We report that QDs increase the overall effectiveness of renal, thyroid, and breast fine-needle aspiration (FNA) biopsy and that the probability that a qualified diagnosis is negative (QDN) can be predicted by the formula QDN = number of QDs x (proportion of false-negative outcomes/disease prevalence expressed as a proportion). Results of renal (n = 24), thyroid (n = 163), and breast (n = 456) FNA biopsies performed from January 1992 through December 1998 were reviewed and correlated with results of tissue biopsies. For each body site, the FNA biopsies were placed into 1 of 2 diagnostic categories: unqualified diagnoses (UQDs) or QDs. Comparison of test performance characteristics for UQDs only and UQDs combined with QDs demonstrated that inclusion of UQDs increased FNA sensitivity and reduced FNA false-negative diagnoses. More important, the probability that a QD was negative could be predicted from test performance characteristics derived from UQDs.