Running chemical reactions in a continuous flow stream, rather than in conventional batches where materials are weighed and added to a batch unit operation, can deliver significant benefits to pharmaceutical manufacturers. Continuous flow chemistry applications can considerably shorten cycle times and costs as well as improve productivity, efficiency and throughput. At the same time, they enhance end product quality and consistency by using PAT-driven quality assurance strategies. These applications can slash the downtime associated with off-line quality control and testing, plus offer real-time monitoring capabilities.
To do so, the IT and automation infrastructure needs to provide complete visibility associated with the movement of raw materials throughout manufacturing in order to meet regulatory compliance and good manufacturing process. More precisely, it is crucial for pharmaceutical producers to be able to track the raw material lots that they use from process input to final batches. In other words, they must be able to identify with a high degree of confidence, the location of any material at any given time.
In addition to lot to end product traceability, the quality assurance data sets of the in-flight materials that have been measured by a PAT knowledge manager have to be tracked through the process and associated with the final batches. In this way, for each finished product it is possible to have complete sets of quality assurance data, together with information on the contained raw material lots. Yet further, the tracking system has to monitor the movement of any suspect or out-of-specification material as it progresses through the system, from the detection point to the sampling or rejection point. Also, it needs to drive the process valves or diverting system to acquire the target product. To achieve this level of functionality, the track and trace system requires real-time operation together with real-world connectivity.
Since manufacturing using flow chemistry is extremely different from batch production operations, they need alternative methods to ensure material track- and traceability. Traditional batch tracking solutions generally rely on the discrete, transactional movement of product batches or sub-batches, where mixing between these units seldom occur. Such track and trace systems are relatively straightforward to implement, and the movement of product can be defined with total confidence.
In continuous applications, a constant flow of materials that change state and structure travel through the various unit operations with varying levels of forward and back mixing. Therefore, the product movement is very much non-discrete and it is not possible to use conventional methods to track product.
Various track and trace methods and approaches can be taken to solve the problem. However, no matter what solution is chosen, the undeniable truth is that the tracking algorithm will be based on probability, as with continuous processes having total confidence the movement of any specific entity from one unit operation to the next is impossible irrespective of whether mechanistic, empirical or hybrid modelling is used. Thus businesses need to make sure that the levels of probability are acceptable in practical applications.
TRACK AND TRACE FOR CONTINUOUS FLOW
To address these challenges, one key element to look for is the ability of a system to effectively interact with PAT and automation systems to leverage the insight provided by real-time in-process data. This aspect is crucial for the system to dynamically respond to prevailing process conditions and, in turn, provide timely feedback to operators and automation systems on in-flight product quality, process conditions and the movement of product so as to enable the optimisation of product quality and process productivity. Where necessary, products for sampling or rejection can be removed from the line. This must be done while ensuring that only products meeting the quality criteria are output from the process. This combined automation, PAT and tracking configuration maximizes the efficiency and quality of production while also ensuring GMP and regulatory requirements.
An example of an effective system to increase supply chain visibility is the synTQ Dynamic Flow Modeller (DFM) by Optimal, used in conjunction with the company’s PAT knowledge management platform, synTQ. This combines a material tracking system with a PAT data manager which collects, processes, fuses and stores analytical information as well as predictions while communicating with the entire PAT environment. synTQ DFM provides a method to track products for sampling, rejection and to trace products, as well as their associated quality data, from starting lots to finished batches. This is achieved by configuring an off-line model of production lines and then in real-time, by integrating synTQ DFM into control, automation and PAT architectures.
Significant advantages are possible if the PAT platform can incorporate technologies to incorporate the real-time detection of low-concentration impurities. This is achieved, for instance, by communicating with on-line analysers such as gas chromatographs or high-performance liquid chromatographs which run continuously but in single-injection mode. Depending on the attribute being measured, the process reaction may be ‘steered’ to ensure that the creation of any unwanted, low concentration by-product is minimised, and in the worst-case scenario, material with too high a level of unwanted trace materials can be rejected from the system.
CHOICES
Businesses should favour a PAT knowledge management system that is cost-effective as well as easy to use, implement and maintain. This will improve return on investment while supporting operators in their daily tasks. In practice, this means selecting a platform that requires only minimal configuration, with no coding. This means the solution can be set up quickly and is easily adjusted, edited or updated by authorised staff who do not need programming skills.
In addition, companies should consider the cost/benefit advantages of using a process technique that relies in part on mechanistic modelling together with empirical modelling versus an approach that relies on empirical modelling only. Both have their place, and while a mechanistic approach has advantages, the current state of the art means that pure mechanistic modelling is seldom possible. Invariably, empirically-derived data and parameterisation will be used together with the mechanistic model to create a hybrid model. Purely empirical modelling, on the other hand, can be quicker to establish and set up. While practical testing is likely to be necessary for an empirical model, very often such practical experimentation is needed for parameterization of ‘mechanistic’ approaches too. Thus, an empirical system can in fact be faster and more cost-effective to implement.
An effective automated material and quality assurance management system can improve a company’s competitiveness and reliability while enhancing product quality and manufacturing sustainability.