Abstract

[Please note: This article was also published in Quality Engineering with the QICID number of 11770. Do not order both articles.]

Sampling plans are used in decisions to accept or reject specific lots of products. To address FDA concerns that sampling plans are not "statistically valid", this paper prescribes a two-part process for selecting statistically valid sampling plans. First, the objective of the inspection is clearly defined, and second, the ability of the sampling plan to meet this objective is demonstrated.The Operating Characteristic (OC) curve describes a sampling plan's behavior. The OC curve is summarized by using the Acceptable Quality Level (AQL), which defines what the sampling plan will accept, and the Lowest Tolerance Percent Defective (LTPD), which describes what the sampling plan rejects. These standards should be documented for all sampling plans and all sampling plans should be periodically reviewed. Examples show how this process works.Using recognized standards does not ensure that a sampling plan is statistically valid. It is the responsibility of the user to make sure they are using an appropriate plan based upon the objective of the inspection. It is the actual AQL and LTPD of the sampling plan that describes its protection and determines its validity.