Comparison of disagreement and amendment rates by tissue type and diagnosis: identifying cases for directed blinded review.

We sought to determine whether a group of cases that was relatively high in disagreements and subsequent amendments could be identified and targeted for blinded review. During a 4-year period, 8,916 surgical pathology and nongynecologic cytology cases were subjected to blinded review; of these, there were 616 disagreements (6.9%), 69 (0.8%) had subsequent amendments issued, and 33 (0.4%) represented false-negative errors of blinded review. Tissues with the highest amendment rates were breast (4.4%), endocrine (4%), gynecologic (1.8%), and cytology (1.3%). Specimen types with highest amendment rates for specimens with more than 20 cases were breast core biopsies (4.0%) and endometrial currettings (2.1%). Diagnoses were divided into negative (64.7%), malignant (21.4%), nondiagnostic (1%), defined precursor lesions (ie, atypical ductal hyperplasia, tubular adenoma) (9.8%), and atypical or "suspicious" (3.1%). Amendment rates were highest for nondiagnostic material (5%) and atypical/suspicious (2.2%). Reviewing only nondiagnostic and atypical cases would have involved reviewing only 4.0% of cases and detected 14% of amendments. Reviewing all breast, gynecologic, nongynecologic cytology, and endocrine material would have involved reviewing 26.9% of cases and detected 88% of amendments. These data can be used to define material for directed blinded review that is relatively high in potential errors.