Asking those 'what if' questions

Successful automation can give companies the ability to ask more ‘what if?’ questions – but there are many considerations when planning and implementing pharmaceutical automation

An automated roller bottle incubation system for manufacturing a biopharmaceutical drug, developed by TTP LabTech

There is a constant pressure to increase efficiency and reduce costs across all industries. Pharmaceutical research and development laboratories are searching for new ways to increase the number of samples processed each year, and regard automation as a useful way of maintaining efficiency and low costs by allowing workers to do more. The view that automation leads to significant reductions in labour costs is now largely discredited but the idea lingers on.  In practice, those companies that have successfully automated probably employ the same number of staff as before but now have the luxury of asking many more “what if” questions, which is an important step towards finding new solutions to drive the pharmaceutical business forward.

The main driver to automation for pharmaceutical laboratories depends on the stage of the drug discovery process they are working on. In discovery and screening, it is the need for fast results. When a drug gets to development, accuracy is probably the key driver, as the quality of results is extremely important.

So what makes a process suitable for automation? Essentially, that the process is well understood and there are automatic measurements that can be used to form a feedback control of the parameters or assure the quality of the output.

Dr Philip Blenkinsop, Managing Director of TTP LabTech, explains: “If you’ve got a process that is implemented by manual methods and is stable and consistent then you stand a good chance that an automated equivalent will perform equally well. Unfortunately, we have met, from time to time, companies who believe that since machinery tends to be consistent and reliable, it will somehow remove the shortcomings of a process that is anything but”.

His advice is to take a walk around your existing process and try and gauge where human intervention and judgement is being applied. “There is a kind of inherent quality control built into repeated human actions which is often not recognised and can be the devil to replace with sensors and software. Automated systems are good at material transportation, weighing, heating, cooling, measuring and all the other basic lab processes but are not necessarily good at spotting subtle changes that indicate that a change of conditions is required.”

Unless the system provider can really understand the science that the process involves then what they will supply is mechanisation and not automation. Conversely, pure scientists do not make great automation engineers. This means that very specific skills are required to deal with those tricky process issues as they arise.

The obvious question companies need to answer before deciding to automate a process is: “What am I really seeking by automating?” Clearly defined goals are essential to the final success of any automation project, and they make success much easier to recognise – to promote internal acceptance as well as payback goals.

A second important question should be: “Have I captured all the information necessary to communicate the needs of the users to a system supplier?” Blenkinsop observes that “We [TTP LabTech] find that quite often it is what is left unsaid by the user that is critical to the viability of the project and have developed techniques of flushing this hidden information out before the design process starts. Simulation is an important tool to aid this process.”

TTP LabTech’s chemistry automation breaks down a process into standard steps such as heating, cooling, measuring and sampling, which allows every vessel to follow a different process

Developing new equipment should include ways of managing the risk involved. Blenkinsop agrees. “The first port of call for any company seeking to automate all or part of a laboratory process will be an equipment supplier or group of equipment suppliers. If you can buy proven equipment off–the-shelf that will do all, or nearly all, that you require then you should seize the opportunity.” However, in many cases you will find that a solution built from standard equipment will not perform in the way you want. The reasons are many. “Typically there will be gaps in the process where one vendor’s functionality will fail to overlap with the functionality of the next piece of equipment in the process,” Blenkinsop said. “That often leads to continuing with manual processes to cover the gap – what we call ‘manned automation’”. In other cases it will be the very specificity of the task that defeats standard equipment. Every company has a unique way of doing things and, whilst this can often be bent to match the working practices imposed by automation, the benefits of automation rapidly drop away once you start planning for the lowest common denominator.

Finally, there is the software. “Chemists and biochemists should be directing experiments not writing programming scripts or database queries” says Blenkinsop. “Writing application software that will meet the specific needs of the company, be sufficiently flexible that it will cope with future extensions, and be easily maintainable is a skilled job and not one that most companies who have tried to do it in-house have repeated.”

Deciding whether to outsource an automation project is not easy. A major consideration is whether your company has the time and resources to undertake a project that will effectively be a distraction for some while, even if the results will enhance your core business.

The main benefit of outsourcing automation is that you will get 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-house projects tend to start out poorly defined and suffer from expanding objectives. This is because users and providers are part of the same organisation so it is difficult to inculcate the professional distance needed to run the project on a tight rein.

Experienced systems integrators have the expertise to bridge the mechanical and software interfaces between vendors’ equipment. A less well recognised benefit is technology transfer – a custom automation company 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.

If you choose the right vendor, you will get a much broader range of skill sets and experience than if you tackle the task in-house – without the expense of supporting that expertise when you don’t need it. “The bulk of staff at TTP LabTech are mechanical , electrical engineers, software, and control engineers but we also have available a talented pool of scientists who are essential to progress our own range of products. In addition, we can call upon the expertise of 300 staff that make up the rest of the TTP Group – many of whom have the physics, chemistry, materials, and optics skills that a large project often requires,” said Blenkinsop.

It is important to 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 development of the project’s aim. As expected, progress reports on technical issues and demonstrations are provided regularly. TTP LabTech uses the same manager throughout the project duration – and encourages clients to do the same as this contributes greatly to the project’s success. Equipment is designed to conform to all relevant EU Directives, and appropriate QA and regulatory procedures are followed as required for pharmaceutical environments (e.g. GAMP and GxP).

Because bespoke automation is a big commitment, support is essential 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.

Successful laboratory automation will make a process faster or produce more accurate and repeatable results. But not all discovery is a deterministic process and often successful automation also provides the chance to explore wider chemistry space or obtain finer granularity of information which can lead to insights from staff that the day-to-day duties of the lab won’t permit.

 “The best projects are those where we sit down with the customer to define what appears to be a straightforward automated replication of a manual process and, by pooling our collective talents and a process of iteration, come up with a design that is considerably slicker, or simpler, or more flexible than that we might at first have conceived,” said Philip Blenkinsop.

“Turning that into a robust and reliable piece of hardware gives great satisfaction. I don’t want to suggest that there isn’t often heartache in between these two stages as prototypes don’t always work and alternatives have to be examined before a way forward can be found. But if there’s no challenge then it is likely that the benefits will be weak and the competitive advantage that automation can bring will be absent.”

By Chloe Milburn. Chloe joined TTP LabTech in 2002 as a technical author, working on customer communications and product user interface design.