Considerations in Monitoring and Controlling Pharmaceutical Manufacturing

Abstract

Monitoring and control of complex processes involve a number of variables whose interactions are necessarily complex. Unlike many other areas of pharmaceutical development, this has long been recognized by process engineers who have the task of guaranteeing the quality and performance of the product. A variety of statistical, physical, and mathematical approaches have been adopted depending on the needs of the assessment. As tools for this purpose have evolved from the ability to manage and store data, the concept of quality by design (QbD) has gained ground and is now a central theme for industry and government regulators. QbD requires significant preparatory consideration of any process by a team of qualified individuals to map out all of the known variables that might contribute to desired attributes of the product. Since there are many variables involved in manufacturing processes and some may not be known, or not subject to control, the mathematical approach to the complexity has included artificial neural networks which have the capacity to learn from data generated and to integrate that knowledge into a predictive approach to a range of activities from research to development. Indeed, advanced mathematical modeling and control tools will be needed as the industry slowly moves from batch to continuous processing.