ABSTRACT
Due to the growing complexity of drug substances and formulations, developing a quality (1) drug product and a process that manufactures it in a reproducible manner becomes increasingly challenging and costly. Very often, even though a final drug product is successfully manufactured to meet the regulatory specifications, the drug product development is in part based on correlative methods, in which the limits on quality attributes and process parameters are derived empirically to ensure that the produced batches mimic the performance of batches tested clinically. This approach can guarantee consistent high-quality drug products only if the relationship between these limits and the drug performance, as well as the variability within the drug substance, excipient raw materials, and process characteristics, is well understood. Without this knowledge, the restrictions on quality attributes are established mainly based on the so-called zero tolerance criterion, which often leads to overspecification. In the worst-case scenario, other critical attribute(s) may not even be identified, monitored, and controlled. As a result, continuous improvement on product quality
and process, as well as risk assessment and management (2), would be difficult to carry out under this circumstance.
