Comment from Carey Johnson

I submit this comment to address the scope, interpretation, and potential misapplication of the FDA Guidance for Industry, Investigating Out-of-Specification (OOS) Test Results for Pharmaceutical Production (October 2006). While the guidance provides important direction for pharmaceutical manufacturers operating under current good manufacturing practice (CGMP), clarification is warranted to ensure it is not misconstrued or applied beyond its intended context. The guidance expressly states that it represents FDA’s “current thinking,” does not create or confer rights, and does not bind FDA or the public. It applies to chemistry-based laboratory testing of pharmaceuticals regulated by CDER and conducted under 21 CFR Parts 210 and 211. It is directed to manufacturing quality control and batch disposition decisions, not to forensic, clinical, or adjudicatory drug-testing contexts. Treating this guidance as a general evidentiary or scientific standard outside the CGMP framework risks misapplication and confusion. A core principle of the guidance is that every OOS result triggers a mandatory investigation. Such investigations must be timely, unbiased, scientifically sound, and fully documented, and must determine whether the OOS result arose from laboratory error or from the manufacturing process. Batch rejection does not eliminate the obligation to investigate, and subsequent testing does not negate the significance of the original OOS result. The guidance clearly prohibits the practice commonly described as “testing into compliance,” i.e., repeated testing until a passing result is obtained and then used to disregard an initial OOS result. Retesting must be predefined, scientifically justified, limited in number, and conducted pursuant to written SOPs approved by the quality control unit. All results, including both failed and passing retests, must be reported and evaluated. Invalidation of an OOS result is permitted only when a specific, documented, and scientifically credible assignable cause is identified, such as a demonstrated analytical or instrument error. In the absence of such an assignable cause, there is no scientific basis to disregard the initial OOS result in favor of subsequent passing results. This safeguard is fundamental to data integrity and prevents post-hoc rationalization of test outcomes. The guidance also places strict limits on statistical manipulation during OOS investigations. Averaging OOS and in-spec results is generally inappropriate because it can conceal variability and misrepresent product quality. Outlier testing does not identify root cause and may not be used to invalidate suspect results for chemical assays with low inherent variability. All individual results must be preserved, reported, and evaluated by the quality control unit. Throughout the guidance, FDA emphasizes transparency and recordkeeping. All raw data must be retained, and all test results—including initial OOS results and subsequent retests or resamples—must be reported. Selective disclosure or suppression of unfavorable results is inconsistent with CGMP principles. Given the increasing tendency for this guidance to be cited outside pharmaceutical manufacturing contexts, FDA should consider reaffirming that: (1) the guidance applies solely to CDER-regulated pharmaceutical CGMP decision-making; (2) it does not authorize dismissal of OOS results without documented assignable cause; and (3) it does not establish evidentiary or forensic standards for non-CGMP settings. Such clarification would protect the integrity of FDA’s guidance and prevent misuse. Respectfully submitted.