SOP on Evaluation of Out-of specification Results
1.0 Purpose: To laid down a procedure for the investigation of failures in the analytical work and evaluation of the out-of-specification results.
2.0 Scope: This SOP applies to Quality Control staff.
3.0 Responsibility: Quality Assurance Manager to ensure that the procedure is followed.
4.0 Materials and Equipment: As required.
5.0 Procedure: When a test result is out-of-specification (OOS), validity of the results should have been questioned. The possibility of an error in the analytical work should be the starting point of an investigation. In the case in question the following steps should have been followed :
5.1 All test preparations and standard solutions should be retained.
5.2 The analyst should check the data for compliance with specifications; if an out-of-specification result is obtained, analyst must report the result to the Quality Assurance Manager.
5.3 If an obvious and recognizable error has occurred during testing, no investigation is required. However, the testing should be stopped, the senior analyst notified, and the data invalidated. The analyst should document the error on the raw data record, and the senior analyst should initial and date the record. Where clerical errors such as transcription mistakes or calculations have occurred, the analyst should correct the error without obliterating the wrong entry.
5.4 The senior analyst must then immediately conduct an informal laboratory investigation of the result. At this stage Quality Assurance Manager review the analytical method and the raw data document that contained the OOS result, discuss the testing procedure along with any required calculations, also examine the working of the instrument used and then assess the accuracy of the result. If the result was invalid because of a specific identifiable error, that result can be ignored and a repeat analysis is performed. If the result of the retest meets the acceptance criterion, then the original OOS result is invalidated. The root cause or the assignable cause of the OOS result is documented. The investigation is then closed. In other cases, the investigation should proceed.
5.5 As part of this investigation the retained solutions, test units, and glassware used in the original measurements and preparations may be re-examined in order to ascertain their reliability of every individual value obtained.
5.6 If an assignable laboratory cause is identified and there is sufficient sample preparation remaining, the initial analyst should repeat the test in duplicate. If both results are within the specifications, the sample may be considered passes. If however, one or both of the duplicate results are outside the specifications, the likelyhood of analyst error should be examined.
5.7 The final decision must involve Quality Assurance manager and must apply its experience and knowledge of sample history. Persistent or frequent failures are indicative of inadequate analytical method validation or verification, inadequate training.