Adrian Howson

The FDA initiative —Process Analytical Technologies (PAT) — is slowly gaining momentum, creating a revolution in manufacturing and testing processes that aims to ensure product quality. Its growth will encourage faster testing techniques to bring analytical testing closer to on- and at-line testing during the product manufacturing process.  

In parallel with this, there will be a need for more off-line testing during development and — for the foreseeable future — testing for QC release in manufacturing. There has never been more pressure than now to transform and optimize these testing processes by the strategic application of automation and new technology.  

There are many automated systems available for drug discovery, but the automation of validation steps in the development process has not taken off so rapidly.  There are obvious benefits for end-to-end automation of development testing, and solutions are emerging that could radically reduce the burden of scale-up and method transfer to manufacturing. 

This article looks at automating development activity using integrated, high throughput approaches tailored to the needs of the individual organization. This style of automation has the potential to shorten the lead times during the development stage and get products to market earlier. 

The pharmaceutical industry is facing unprecedented challenges. Growth is slow, new product introductions are down, and generics continue to erode lifecycles. The upshot is that revenues can be lost at an alarming rate if new products are not launched quickly and efficiently.  			

Testing now constitutes a significant bottleneck in this process of launching drugs. As a result, organizations are increasingly outsourcing testing to increase throughput. However, this can sometimes be more expensive than the in-house option and can affect the quality of testing as methods may be performed differently by a separate organization, which may have conflicting pressures. 

This constraint — at the testing stage — is increasingly apparent, as discovery is already highly automated. Ten years ago, pharmaceutical companies started to invest heavily in new technologies, such as sample libraries, sample preparation systems and screening systems.  

Largely as a result of this level of automation, research is now delivering candidates for development in volumes that the development process struggles to cope with.  

It is therefore necessary to increase productivity at the development and manufacturing stages (Figure 1). The ability to test drugs more rapidly for toxicity, or for excipient compatibility for example, would make it easier to spot non-viable drugs early in the development cycle and kill them off.  

 Figure 1 As a new lead passes from research to development then manufacturing, the focus changes from innovation to one of efficiency and speed.			

Additionally, testing the viable drugs more efficiently at each stage of development would help companies get them on the market sooner, maintaining revenue growth, and hence shareholder value. 

Automation is the best way to alleviate the bottlenecks from tasks such as listed in Figure 2. It can deliver the opportunity to release staff — often highly paid PhD chemists — from routine tasks (checking out glassware, labelling, standards preparation) for more stimulating, and more value-adding work. Throughput, too, can be transformed, making batches available for further processing months sooner than in a manual set-up.  

Figure 2 Many routine tests must be performed on a new lead throughout its life.						

Many of the benefits of automation, which have already been demonstrated in the discovery process, are equally applicable to testing. Automated processes can accurately and reliably record exactly what happens at every stage (21CFR11 compliance), making the process much easier to control, and where appropriate allowing backtracking. 

Consistency is another major attraction of automation. The variation between results obtained by two analysts using the same standard piece of equipment can be significant; a good automated process does it the same every time. 

Another strong argument for automation is its impact on manufacturing. Automated test processes can often ease the transition of a drug from development to manufacture. 

There is no longer any need for the rather disorganized arrangements where, for example, development staff may carry out pilot testing while manufacturing expands its expertise in the requisite techniques and processes.  

Thus, provided testing has been automated in a way that can be scaled-up for manufacturing purposes, "Technology transfer" becomes exactly what the name implies — the hand-over of a system. 

Finally, the productivity benefits available to the area through implementation of automation are potentially even larger, especially when linked to the advantages arising from PAT implementation. There will be a need for fast, automated techniques for on- and at-line testing as well as the traditional end product testing that will be around for the foreseeable future. 

There are numerous reasons why automation has not yet managed to penetrate the development process to any great extent, when it is so well accepted in discovery. Foremost among these, perhaps, is the fact that development activities are subject to rigorous regulation in a way discovery is not. In the typical, tightly controlled, development environments, innovation and new technologies, including automation, are actively discouraged for fear of failing to satisfy the regulators — a failure that in the minds of managers, could jeopardise the launch of the next blockbuster. 			

Substances are measured and shaken according to a carefully prescribed recipe, in a daily repetition that can be mind-numbing for the highly educated individuals involved. Despite this repetitive nature, it should be noted that across teams, and certainly across departments and companies, there is a lack of standardization.  

The solvent and techniques used are anything but standard, each team believing it has the best way of tackling the problem. The proliferation of processes provides an additional barrier to automation. In discovery, by contrast, there has been a high level of automation, which has fostered standardization and has, in turn, facilitated further automation. 

A final explanation to why automation has failed to take-off within development is a disappointment with earlier automation initiatives. Existing equipment does little more than reproduce a specific manual task, and sometimes arguably, in a way inferior to the manual process. 

For example, some such machines are less efficient than a human tester, running only one test at a time, compared to the scientists ability to run several tests simultaneously.  

As part of the current good manufacturing practice (cGMP) initiative for the 21st Century, PAT is designed to encourage process improvement. Both efforts open the way for innovation, including automation, and also for the process of standardization that will make automation more feasible. 			

There are various approaches to automating testing. The low-benefit way is to mimic existing manual processes with a robotic arm, which can exploit the same techniques and tools as human analysts. While a process automated in this way may take as long as, or even longer than the manual version, benefits include, consistency, repeatability, and round-the-clock working.  

Much greater benefits, however, can be achieved if the process is analysed and innovated prior to automation. Such improvement is best achieved with independent assistance as test chemists may struggle to see how their tried-and-tested methods could be sharpened. Consideration of technologies and processes from other industries will help to highlight any possible innovations.  

An example of the type of process improvement that can be achieved, and the benefits of so doing, relates to the wide variety of media and solvents typically used in the extraction of a drug from a tablet for testing. Controlled release tablets, tend to need particularly aggressive solvents, which may in turn necessitate a special buffer and mobile phase. By innovating the extraction process, it becomes possible to reduce the number of solvents, mobile phases and buffers used with consequent efficiencies.  

Having optimized and standardized their processes, companies are well positioned to achieve the full benefits available from automation. Dedicated, high-throughput automation is the form that has been shown to deliver the biggest productivity improvements, principally because it produces a faster test speed overall (Figure 3).  

Figure 3 High-throughput inhaler testing automation.			

Although these high throughput systems are not currently available in packaged form, they are certainly now achievable. Solutions exist for each of the test stages, such as sample reception and division, test preparation, analysis, data crunching and reporting, there need be no gaps. With the aid of a piece of software known as a test scheduler, these solutions can be integrated into an end-to-end testing system, with automatic control of the flow of materials and allocation of resources. These systems are customised to accommodate the processes favoured by individual clients, and are implemented in close collaboration with the analysts concerned. 			

Productivity benefits of these systems can be spectacular. A 10-fold productivity increase can eliminate almost all human activity, including outsourcing. It can be seen, therefore, that the payback is highly attractive. By careful application of the 80:20 rule, several different testing methods can be accommodated by a single system, all within a validated environment. 

In view of the modest level of automation currently deployed in the typical analytical laboratory, this type of undertaking is a huge leap. To gain acceptance, it needs to be supported at the highest level within the company. In fact, the motivation for change is most likely to come from a high level, and not only because senior staff are aware of the pressure to improve productivity. Compared with grassroots staff, senior management are also more likely to be aware of the possibilities and advantages of automation as a result of their familiarity with activities elsewhere in the organisation. 

Automation initiatives require support and participation from other parts of the company, most notably manufacturing. One reason for its involvement is that, if development does adopt automated testing methods, manufacturing will have no reason not to adopt them too. 			

It, therefore, makes sense for the industry to participate in the specification of the new processes and systems, to ensure that they are also optimized for manufacturing purposes as well as for development. Principles such as Six Sigma and Lean Manufacturing should therefore be built into the system specifications. 

It should be noted that process improvements, of the kind we have discussed, can be applied to existing drugs as well as new ones, provided equivalence between the old and new processes can be demonstrated.  

In just 24 hours, an automated system can carry out a trial that would have taken 30 people a whole month in the old pre-automation days, with the additional advantage of complete, reliable recording of every step.  

The immediate gains from automated testing are in alleviating the development bottleneck, but manufacturing will be the area that achieves the greatest profit in the long run. 

is a principal consultant, both in product and process engineering at PA Consulting Group, UK.

Articles

The FDA initiative —Process Analytical Technologies (PAT) — is slowly gaining momentum, creating a revolution in manufacturing and testing processes that aims to ensure product quality. Its growth will encourage faster testing techniques to bring analytical testing closer to on- and at-line testing during the product manufacturing process.