Pharmaceutical production is at present characterized by static processes where quality is guaranteed by controlling the purity of the final product. Achieving better control throughout the process, as a means for improving product quality, is one of the objectives of the PAT initiative by the FDA. A data set consisting of 11 batches characterized by UV spectroscopy together with process data was used in this study. Design of experiments was used to introduce controlled process variation in test batches. The objective was to investigate possible advantages of MSPC using a combined data set, compared to separate models of the respective data sets. Individual models for the separate data sets show that they contain complementary information. A major advantage of combining spectroscopic and process data is that deviations that would go unnoticed using just an individual model can be detected and interpreted. All process manipulations were detected by the combined data set model. Implementation of these methods to batch processes in primary and secondary pharmaceutical production is feasible. An enhanced understanding of the process together with control tools should lead to a well-understood process and, ultimately, real time release. (c) 2006 American Institute of Chemical Engineers.