The use of automation across the pharmaceutical industry has increased in recent years. Although the pressures on the industry to improve efficiency and reduce costs have driven this trend, technological advances together with an increased acceptance of “automation” have also contributed to it. Today, automation is not necessarily implemented to reduce labor costs, but rather to improve experimental accuracy and workflow efficiency. In drug development — the mid-stage between discovery and clinical trials — scientists are testing many more samples to measure a drug’s characteristics so the quality of results is crucial. Automation allows highly qualified scientists to focus on analyzing results or developing new areas for research rather than the laborious and repetitive manual steps of an experimental set-up.
Small automation projects can often be implemented in their entirety by an in-house team. Large, multi-equipment projects, however, almost inevitably involve the use of an outside vendor or development agency. Deciding whether to outsource an automation project is not easy. A major consideration is whether the company has the time and resources to undertake a project inhouse that will effectively be a major distraction for some while, even if the results will enhance the core business. One of the first steps in the project life-cycle is careful process evaluation, enabling the team to determine the limitations of a specific research area or task and investigate the benefits that automation could bring. To get the most benefit, the process to be automated should be stable and consistent. If the work-flows are also complex, repetitive, and therefore prone to human error then automation is likely to generate better quality data.
Next begins the search for an equipment supplier or, ideally, an off-the-shelf solution that will do all, or nearly all, the required functions. Typically there will be gaps in the process where one instrument’s functionality will fail to overlap with the functionality of the next piece of equipment in the work-flow and this commonly leads to the continuation of manual processes to cover the gap. This is often referred to as “manned automation.”
Companies like AstraZeneca have in-house teams, part of whose role is to evaluate the company’s technology portfolio. Mikael Arinder, Project Manager at AstraZeneca Strategy, Technology & Systems in Lund, Sweden explains, “We’re currently evaluating where there are gaps, and where could we use technology to create benefits. Then we map these gaps or needs towards technology on the market — commercially available systems — and if we see that there are gaps which don’t map against a commercial technique then we will probably go to suppliers to see if we can get this equipment built for us.”
Sometimes it will be the very specificity of the task to be automated that defeats the implementation of standard or offthe- shelf automation as each company will have its own special way of doing things. Compromises can be made to match the working practices imposed by automation, but the benefits of automation rapidly drop away once you start planning for the lowest common denominator.
It is at this point that the recruitment of outsourced automation expertise is most beneficial. Developing an automated solution to a pharmaceutical process that is optimized for specific company methodologies requires extensive scientific, engineering, and software expertise across a number of fields. It is unusual for this wide range of expertise to be available in-house. Furthermore, experienced systems integrators can avoid the pitfalls involved in bridging the mechanical and software interfaces between different suppliers’ instrumentation. Writing application software that meets the specific needs of the company, is sufficiently flexible that it will cope with future extensions, and easily maintainable is a skilled job and not one that most companies who have tried to do it in-house have repeated. Simplistically, time is money and while chemists or biochemists are writing programming scripts or database queries they are not using their expertise to develop and direct experiments.
The main benefit of outsourcing automation is that, with the right company, you will get a much broader range of skill sets and experience, including mechanical, electrical engineering, software, and control engineering expertise, than if you tackle the task in-house — without the expense of supporting that expertise when you don’t need it. You will also be presented with a dedicated team who will have a vested interest in meeting the user requirements in the simplest and most direct fashion, especially where there is a direct link between deliverables and the price of the job. In contrast, it is not uncommon for in-house projects to start out poorly defined and suffer from expanding objectives, mainly because, when users and providers are part of the same organization, it is difficult to inculcate the professional distance needed to run the project on a tight rein. Additionally, a less well-recognized benefit is technology transfer — a custom automation supplier can work across several areas of business and is in a better position to bring to your attention a technology that has been proven in another field but not yet applied to yours.
AstraZeneca has a taxing selection process because the right automation company is so crucial. “If we select the wrong vendor then we often go wrong in the end,” says Arinder. “We choose between 10 and 20, and then we funnel the list down to evaluating a few companies more thoroughly.” The list of points to assess is exhaustive. “We would look at the vendor’s ability in developing GMP compliant systems, look at Part 11 experience, look at quality assurance, at GAMP procedure — quality is important because the system that is being developed will be validated and when we have an FDA inspection in our labs they will also inspect the documentation on the development of the system.” AstraZeneca also checks standards for project management, written standards, key control documents, development documentation, testing, and of course the skills required. In particular, they are interested in process knowledge. As Arinder explains, “If there’s a tablet analysis system, then have you worked with tablet analysis systems before? We also look at track record — have you worked together with us before? How did that project go? Have you worked together with other pharma companies? Have you references? What came out of that? Then there are more things like, how stable is the company? How limited in size are you? Can you demonstrate value for money? How willing are you to agree on IP and such things? And finally, capacity to deliver and to support what’s been delivered.”
There’s a considerable risk involved with developing new equipment, or enhancing or integrating existing instrumentation. AstraZeneca manages this risk in several ways — some developments involve going to the original supplier for enhancements; in some cases new technology is developed in collaboration with a supplier and then handed over to a supplier to deliver and support the system. Arinder explains, “There are significant advantages to this because if they sell it to many customers the supplier will take care of the support and upgrades. And the support will be much better if there are many customers than if there is only one customer. There are other advantages, for instance, if you are developing something that is GMP compliant, legal authorities like the FDA are often very suspicious about automated solutions. If we automate it, we need to prove the automated solution gives exactly the same result as the manual. If it is launched on the market and many companies use it, the FDA is more inclined to accept it, so it becomes easier for AstraZeneca to use.”
To manage the risk associated with larger, bespoke projects, AstraZeneca adopts a phased approach where the risk is reduced by having several decision points. Arinder explains, “For one current project, the first thing we did was a design study, then we had a go/no go decision — should we proceed? Does it look promising? As it did, we then did prototyping for almost a year, then there was a new decision — should we proceed or not? Now we are building the first module. We will then implement it and test it. If it works, we will build an additional two modules and connect them. If we can’t connect them for some reason, then we can at least use them as stand alones.”
A successful automation supplier will follow a well defined process when consulting on the variety of projects they undertake. Good communication before, during, and after the project is essential with review meetings to discuss ideas explored, problems found, and any expansion of the project’s objectives. As expected, progress reports on technical issues should be provided regularly and consistency is key. Both customer and supplier should be encouraged to use the same manager throughout the project’s duration as this can significantly contribute to the success of the project. A professional supplier will also provide full training at project hand-over and ideally will be in a position to provide service and support teams to keep things running once the system has been commissioned. Bespoke automation is a big commitment for any company or laboratory, large or small, thus support and on-going development is an important part of the client relationship, whether it is for a new version of software or a mechanical tweak to the system.
From design to implementation
A first essential step in automating a specific process is the establishment of clear objectives. For example, not all drug discovery is deterministic and often the chance to explore a wider chemistry space or obtain finer granularity of information can lead to insights from your staff that the day-to-day duties of the lab won’t permit. The next step involves conveying these accurately to the automation supplier to determine the viability of the project — at this point, detail and thoroughness are essential and simulation is an important tool to aid the process.
If a manual process is stable and consistent then the company stands a good chance that an automated equivalent will perform equally well. The automation supplier really needs to understand the science that the process involves if they are to supply efficient automation with in-built systems and software for monitoring and tracking process parameters. If they do not fully understand the science, they will simply mechanize the manual process but without the benefits of flexibility or the inherent quality control that is built into human actions. A successful company that boasts a team of engineers together with scientists (chemists, biochemists, cell biologists, and physicists, for example) can take a more informed approach and very specific skills sets can be utilized to deal with those tricky process issues as they arise.
So, how do customers define successful automation? Arinder is clear, “When what’s been delivered meets the expectations in the business case. It is often important to have something that is really robust and really works, really meets the requirements — that’s really successful. All projects have to balance the demands of timescales, budgets, and quality. If we kept to the time plan and the budget but the result was unreliable, it would be useless — so quality is essential. However, we will not ignore the time plan and the budget to achieve it.”
From a supplier’s perspective, the most satisfying projects occur when the initial brief defines what appears to be a straightforward automated replication of a manual process but, by pooling the collective talents of the supplier’s and customer’s teams and a process of iteration, the end product is considerably slicker, or simpler, or more flexible than first conceived. There are often challenges along the way, but without challenges the benefits of the automation may be weak and the competitive advantage that automation can bring may be lost. For the customer, the delivery of a robust and reliable piece of hardware that meets or surpasses those initial objectives justifies the investment. The benefits of improving experimental accuracy and streamlining workflows can lead to a more efficient research program. Additionally, the development of a trusted long-term partnership between company and supplier can allow for an easy upgrade path as the company or laboratory grows and requirements change. When a new project arises, past experience provides assurance that a quality result will be delivered that can be developed and integrated with existing fully functional systems.
Philip Blenkinsop, Ph.D. is Managing Director, TTP LabTech Ltd, Melbourn Science Park, Melbourn, Royston, Herts SG8 6EE, UK; +44 1763 262626. U.S. Office: TTP LabTech Inc, One Kendall Square, Suite 341, Cambridge, MA 02139, U.S.; 617- 494-9794.