| Sumario: | The recent US Food and Drug Administration Process Validation Guidance has provided clear statements on the need for statistical procedures in process validation. FDA has redefined validation to include activities taking place over the lifecycle of product and process-from process design and development through ongoing commercialization. New applications have evolved as result of this guidance. Statistical applications should be used in process validation and related applications to improve decision-making. Development efforts should include statistically designed experiments to determine relationships and interactions between inputs and outputs. Manufacturers should understand the sources of variation, understand its impact on process and product, and control variation commensurate with the risk. Statistical methods should be used to monitor and quantify variation. Statistical methods should be used in support of sampling and testing in process qualification (PQ). Sampling plans should reflect risk and demonstrate statistical confidence. Validation protocol sampling plans should include sampling points, numbers of samples, sampling frequency, and associated attributes. Acceptance criteria should include statistical methods to analyze data. Continuing process verification data should include data to evaluate process trends, incoming material, in-process materials, and final products. Data should focus on ongoing control of critical quality attributes. FDA recommends that personnel with adequate and appropriate education in statistics should be used for these activities.
|
|---|