This AHRQ-funded study estimated a 12% error rate in cancer diagnosis based on discrepant evaluation of two specimens from the same organ. Investigators compiled web-based pathologic data from four institutions and created a standardized system to establish correlation error frequencies between cytologic and histologic samples during a 6-month period (eg, bronchial washings and a lung biopsy). Findings suggested that error rates were dependent on the institution and an association existed between the institution and error cause, but agreement was lacking on whether these errors resulted from misinterpretation or poor clinical sampling. Disagreement also existed on the clinical significance of the errors, an issue that results from an undeveloped taxonomy in this arena. While a previous systematic review used autopsy findings to report on clinically significant errors, this study builds on that literature by employing a method that is not limited by the overall low rate of autopsies performed nationally.