article

# PAT: a comprehensive guide

Posted: 23 January 2008 | Steven Doherty, Affiliation: Eli Lilly; Richard Godec, Affiliation: General Electric Analytical Instruments; Matthew Smith, Affiliation: Hach Ultra; Petter Mörée, Affiliation: MKS Umetrics; Ingrid Maes, Affiliation: Siemens; Tony Slapikas, Affiliation: Ametek Process Instruments; Chris Hobbs, Affiliation: ABB | No comments yet

European Pharmaceutical Review presents a comprehensive guide to PAT addressing the challenges and advancements that are impacting upon PAT implementation in 2008 and beyond…

## Question 1: Since the inception of PAT by the FDA what level of impact do you think that the principles of PAT have had on Pharmaceutical manufacturing?

Hobbs: Pharmaceutical manufacturers recognise that there are benefits to be had by implementing PAT. The leading manufacturers have invested in single unit operational trials and have seen significant returns on investment, which has led to PAT being implemented on production sites.

Slapikas: There’s little doubt that the PAT initiative has provided the drive to let the production staff get a much better look into their processes than they’ve ever had. Under the PAT “Quality by Design” umbrella, PAT groups have been able to move instruments right into the production suites to give the operators a real-time look at their processes. At this stage, even though the instruments are not actively controlling the process or releasing the product, the insights gained by just the monitoring of the process is providing valuable information that is being used to plot the strategies for future process development.

Doherty: I principally support small molecule synthetic efforts, so some of my responses may be most directly applicable to that area. While the direct impact on manufacturing may not be readily apparent, particularly in terms of a large installed analyser base, I believe the environment is now more conducive to discussions about the prospective use of PAT. There seems to be more of an inclination to say “what could we do with PAT” rather than come up with reasons why not to pursue the opportunities.

Godec: The most obvious impact PAT and Quality by Design (QbD) principles have had on the pharmaceutical industry is to renew the focus on better understanding, the underlying critical quality attributes of the pharmaceutical manufacturing process. This improved knowledge enables companies to establish effective process control systems and to predict that product quality requirements will be consistently met. The improved visibility of these topics has enhanced the industry’s access to tools and approaches for achieving process control and predictability cost effectively.

A second important factor is that regulatory agencies have taken an active role in educating the industry about PAT and QbD concepts. The advantage for the regulators has been a renewed confidence that the companies actively improving their manufacturing processes are committed to maintaining a consistent and predictable final product quality. The advantages to the pharmaceutical company are lower regulatory risks and costs of a continuous compliance process.

Smith: Change has been evolutionary, not revolutionary and we have found that some companies are well ahead of others in researching and implementing online technologies. Some PAT technologies are very robust and easy to implement (online TOC, conductivity, temperature, pressure and flow sensors, for example), whereas others are more complex and require more thorough evaluation and development (online spectroscopy, for example). Trust and reliance on some of the more complicated technologies has been much slower than most had originally anticipated.

Mörée: The largest impact has probably been that Pharmaceutical manufacturing has started to actively look outside its own area and learn from other parts of manufacturing areas, for example, chemicals and food. In addition it has opened up the possibilities to actively use new technology.

Maes: As of yet the industry does not fully understand the impact it has on the current ways of working in the pharma industry. In fact it is revolutionising the way drugs are developed and also manufactured. Also, the industry is not yet truly understanding how PAT leads to the future focus. In the Keynote address at IFPAC last year, FDA’s Chief Medical Officer, Dr. Janet Woodcock, outlined the following:

1. Development and manufacturing should be integrated
2. Development of quality surrogates for clinical performance (link critical product attributes to clinical outcomes)
3. Rigorous, mechanistically based and statistically controlled processes

Closing the gap between development and manufacturing, supported by PAT is not greatly applied by the pharma industry. With regard to the second point raised by Dr. Woodcock, even fewer companies are realising the impacts and consequences. An important condition for being able to link clinical batch manufacturing and quality data to clinical outcomes, is the possibility to store data in the same format with the appropriate meta data / context data, so that these different data sources can be combined and evaluated. This also means that systems and data formats have to be aligned and be able to be merged. This puts the emphasis on common IT systems throughout the whole R&D process as well as on data mining tools, to be able to discover relationships. The benefit for the pharma industry is that companies will be able to gain insight and knowledge more quickly in therapeutic efficacies and mechanisms and how they are impacted by the manufacturing process. In fact this is nothing new, as the “voice of the customer” principles are already much longer applied in other industry, where customer satisfaction and product performance appreciation by customers is fed-back into process development and improvement cycles.

For the last point that Dr. Woodcock has mentioned it is obvious that true process understanding will offer the possibility to improve and optimise processes, based on cost aspects as well. Again this seems new for the pharma industry but is not for the other industries.

PAT / QbD provides great opportunities to the pharma industry in terms of reducing costs and time on regulatory strategies or approval aspects, especially for products for smaller market niches. Together with live-licensing options, the target market can be further enlarged, while the initial product has already been launched. This option also implicates that data sources have to be linked.

## Question 2: How does your company currently approach the implementation of PAT methodologies within your manufacturing process?

Doherty: I think PAT methodologies are appropriately seen as a tool, not a panacea. The majority of our installations are propagating out of development with new chemical processes. In some instances PAT is viewed as a tool to facilitate tech transfer, since the first time at scale is when you really get the most value and gain the most insight into the process. Once demonstrated successfully, it then becomes a business decision whether to retain or remove such PAT measurements.

Organisationally, we have a corporate group that straddles the R&D/manufacturing interface. That group serves as a resource for both functions, and can also use the development environment to test out and make our PAT platforms more robust prior to manufacturing deployment.

## Question 3: In your opinion, what are the major obstacles which can arise in the implementation of PAT methodologies to the Pharmaceutical Manufacturing process?

Hobbs: PAT is still an emerging technology and as with all new technologies there is risk. It is therefore important to ensure that the fundamentals are as good as they can be. A clear business case is required which is well defined with clearly defined goals and expectations that can be easily understood by all. Failure to implement this will lead to a lack of clarity and an inability to realise the business benefits. Equally this should be supported by a validation strategy which ensures that the implementation and its boundaries are well understood and tested.

Choosing and specifying the requirements for the technology needs to be carefully thought through. How does the PAT data need to be managed? How will the PAT data be used? Are all the technologies to be used compatible with each other? One of the major technological obstacles today is the lack of a clear standard for data exchange between PAT instrument, Data Management Platforms and Control Systems. This assumes that you have already done the work to identify the Critical Quality Attributes and the associated Critical Process Parameters necessary to control them.

Slapikas: The principles of PAT were already well established at the PR&D and pilot plant levels before the initiative actually began. These groups had been working with on- or at-line technologies for quite some time and were already aware of the gains that could be had by moving analyses closer to the process. Once the FDA formalised the PAT concept, a door appeared to let the R&D and pilot folks share what they knew with the manufacturing folks. Once the door was opened, the manufacturing group had only 3 questions:

1. What is the ROI if I use the technology?
2. How much harder will I have to work to use it?
3. How reliable is it and what will it take to keep it that way?

Doherty: To be honest, I think many practitioners of PAT have done us a disservice by focusing principally on complex, higher order technologies in publications and talks. Consequently, I think one obstacle is the perception that PAT has to be complicated. While that may be the case for PAT tools in development, the focus needs to be on providing the simplest tool that can accomplish the control objective for our manufacturing colleagues.

Another related obstacle is that the business case has only recently become a conspicuous part of the conversation. In that respect, I think pharma PAT folks could have done a better job of leveraging the learnings from other process analytical markets, where long-term cost of ownership, ROI, and NPV are well characterised and a fundamental part of the discussion.

Godec: It is valuable to have a leadership team committed to the top-down implementation of PAT and QbD principles and willing to invest the time and money to acquire the knowledge and experience necessary for continued success. A lack of top management commitment to PAT or QbD requires the additional development of a business justification to overcome the resistance. Often the required knowledge and experience is absent. It is important to invest in the development of engineering tools, measurement metrics, methods, and experience to predict and achieve a cost-efficient level of process understanding and process control. This is not to say that the economical advantages of these principles cannot be demonstrated to upper management through bottom-up small-scale projects.

Smith: In our observation, the greatest obstacle to the implementation of PAT methodologies is the change control process required in order for pharmaceutical companies to meet regulatory requirements. Since existing processes are approved, there is often little momentum to change. Thus, we see PAT principles being driven in some companies through the development process in new drugs, but resistance to changing existing processes. We have found that some of this resistance stems from the fact that the product release process is currently ‘owned’ by the Quality department, whereas online instrumentation is ‘owned’ by the Engineering or Manufacturing departments. Thus, changeover to an online release process requires a very close alliance between engineering and quality that may not currently exist in some pharmaceutical companies.

Mörée: A major obstacle is the inertia in big pharmaceutical organisations as well as the regulatory bodies. To change the beliefs and thoughts around how to ensure high quality is a major challenge for the management, especially as this may mean that some departments will need to be strengthened whilst others will have to be reduced.

Maes: The pharmaceutical industry has been slow to adopt PAT in the past, in large part because of regulatory requirements based on off-line batch inspection, slowed progress towards more innovative manufacturing. Now, however, the FDA is engaged on a major drive to encourage the spread of PAT and Quality by Design, as part of its cGMP’s for the 21st century and Quality initiatives.

One of the biggest obstacles for truly implementing PAT / quality by Design that pharma companies are facing is the lack of multidisciplinary knowledge, or typical engineering skills. With PAT, new technologies for the pharma industry have been introduced (for example, data mining, process analytics and so on).

Furthermore, a lot of companies have just started by purchasing a couple of process analysers, but haven’t investigated how they want to build a PAT system, or how to integrate this system into the development or manufacturing architectures, which supports the company objectives and which carries benefits.

Another obstacle, often misused is the regulatory aspect. Often RA departments are not open or willing to think differently in terms of a regulatory strategy, based on new approaches that FDA is offering. The unknown is used as a risk argument and inhibits the application of new process development methodologies.

Thirdly, working cross divisional is rather new for a lot of pharma companies. PAT requires collaboration between development and manufacturing, in cross divisional teams.

## Question 4: What do you believe can be done in order to minimise these obstacles?

Hobbs: Firstly, there is the need to ensure that the implementation team processes the correct skill sets. Partnerships between instrument suppliers, control system suppliers and manufacturers allow additional skills to be added to the team. The importance of training and understanding cannot be understated.

Secondly, it is necessary to have a clear validation plan. A poor validation approach will lead to spiralling costs. If possible, the vendors should be included in this process as a joint approach can often lead to reduced cost.

Thirdly, there must be clear requirements on how the system will be used and that all ‘Use Cases’ are covered, along with a good understanding of how the PAT data will be managed: in particular if more than one instrument is required or you intend to have more than one PAT application, then the addition of a data management system will allow data to be extracted, synchronised and stored efficiently.

Fourthly, it is important to ‘Define standards for Data Exchange’. Initiatives such as the recently announced OPC Foundations ADI (Analyzer Device Integration) group should be supported and encouraged.
Finally, be prepared for organisational issues including:

• Division of responsibilities for PAT between Development, production and Quality and so on
• Creation of routes for communication between functions/divisions/disciplines
• Definition of new development/production and quality roles to take account of QbD and PAT tasks
• Recruitment of new resources or planning of training for existing resources if PAT driven competency requirements are not met by current resources

Slapikas: Coming up with accurate ROI numbers is absolutely critical in getting production sign-off on any new technology. This part of the job resides with the actual pharmaceutical manufacturer since only they know what the true cost is for time and materials involved in the process. In Ametek’s case, since we deal a lot with dryer end-point detection instruments, the ROI numbers will include the cost of the realised time savings; the saving in reduced utilities costs; the initial installed cost of the analyser; and the long-term cost of operating the analyser, just to name a few. Coming up with the numbers takes time and requires bringing together a number of various process owners in order to get the complete picture.

Making the technology easy to operate and maintain, rests solely with the technology provider. I haven’t seen many PAT-type technologies that weren’t fairly complex. If we look at what we see on a regular basis: mass spectrometers; optical spectrometers; acoustical techniques; high-end imaging, none of these would be considered “simple” instruments. While “process hardened” variations of these instruments have been around for a long time, most of them have never seen the inside of a pharmaceutical manufacturing site or have dealt with batch processing. The workflow can be completely different from, say, a typical hydrocarbon processing facility and any new technology introduced into pharmaceutical manufacturing is bound to have teething problems as the technology is tuned to the process.

Doherty: Nothing sells like success, so the two things that I think would be the most beneficial to advancing the cause would be to make sure that the applications we focus on initially are successful, and then to document and advertise those successes. It also helps to have the R&D and manufacturing folks engaging in healthy and active discussion, using real data from the process to guide our decisions.

Godec: An important way to minimise these obstacles is to initiate and participate in actual PAT or QbD projects. The real projects provide the practical knowledge and experience needed to produce accurate project cost analysis and project management predictability, both leading to lower business risk. It is valuable to utilise the manufacturing standards being issued by the ASTM International E55 Committee on Pharmaceutical Manufacturing Standards. Participation in a PAT or QbD project should also involve early communication and discussions with the appropriate regulatory agencies, as appropriate depending on the scope and potential impact to the product quality. This will minimise the regulatory risk of the project.

Smith: Companies that have successfully incorporated PAT principles on a widespread basis have made PAT part of their company culture. This has allowed traditional barriers between quality and engineering to be broken down and has allowed for the organisation to work towards the common goal to reduce cost and waste. It is incumbent on each pharmaceutical company to find the path that leads to the greatest savings. As an instrument vendor, we need to work with both quality and engineering to demonstrate that the instruments are reliable and qualified to allow the customer to achieve real-time release.

Mörée: We think that the different organisations need to find their own answer to this.
Maes: Knowledge and experiences from other industries can speed-up the transfer of knowledge about PAT approaches and systems in the pharma industry. I believe also that vendors can play an active role in this, as most of the vendors of PAT tools are active outside the pharma industry as well and can bring this experience to the pharma industry.

Furthermore, pharma companies have to think about their future R&D and manufacturing vision (based on their actual products and pipeline products) and investigate how PAT can play a role in this. This will avoid companies investing in PAT tools that are not supporting this vision, but that are purchased for the sake of technology alone.

## Question 5: What real term benefits do you believe the implementation of PAT methodologies can have for the Pharmaceutical industry?

Hobbs: In manufacturing, the use of PAT on single unit operations has led to reduced cycle times and better process understanding. A simple example would be none invasive quality measurements by predicting quality attributes in line as opposed to having to sample and perform lab measurements. This means less lab time in sample taking and preparation for measurement.

If you now ask yourself what can be achieved by implementing PAT on multiple unit operation within a production process, it is possible to see the reduction of batch times, reduced lab testing and with more in line quality testing comes the benefit of a more consistent product. This is without even considering the fringe benefits of better process understanding leading to the development of better manufacturing techniques or perhaps even the use of PAT in cleaning operations to optimise the cleaning process and of course the ultimate goal of ‘real time release’ of the product.

Reduced cycle time should lead to the greater utilisation of equipment which leads to an improvement in the Overall Equipment Effectiveness (OEE).

For example: – A leading Pharmaceutical company after implementing an ABB PAT system on a drying process achieved the following:

• Reduction in cycle time fro 40-48Hrs to 24-30Hrs
• PAT provided additional information on filter problems.

### Observations:

In terms of PAT implementation, maximum benefits were achieved with a broad PAT definition. This means looking at the full opportunities offered by PAT, as outlined in the FDA PAT Guidance (for example real-time product release, manufacturing performance improvement, quality consistency improvement and regulatory flexibility). This was preferable to a “limited PAT” approach based only on the implementation of an on-line sensor. We found that the feasibility of a broad PAT enabled manufacturing process, could be demonstrated with much more certainty.

_________________________________________________________________________________________

## Richard Godec

Richard Godec is the New Product and Market Development Manager at General Electric Analytical Instruments, a division of the Water and Process Technologies Business Group. He is responsible for the development of new analyzers to measure TOC and TOx in water, trace boron in water, trace urea in water, and organic sulfur in petrochemicals. He holds over 36 International Patents in these fields. Mr. Godec is currently an active member of American Chemical Society, AIChE, ITRS (Semiconductor Ultrapure Water Roadmap Committee), ASTM International’s E55 Pharmaceutical Manufacturing Standards Committee, and the ISPE “PAT Community of Practice” Education Committee. He holds a Chemical Engineering degree from the University of Colorado, USA.

## Petter Mörée

Petter Mörée is currently Director of On-line products for Umetrics. As Umetrics’ major customer base is within Pharmaceutical production his work has a dual focus.

For Umetrics’ collaborations to implement multivariate on-line technology, the SIMCA product series, with systems suppliers such as Siemens, GE and ABB is an important way to provide easier ways to implement solutions for the pharmaceutical industry. Petter is also responsible for sales to large Pharmaceutical companies in the Germanic market.

In both of these areas, Petters strong technical background with a M.Sc. in technical chemistry with a specialisation towards chemometrics is a large benefit. After his M.Sc. Petter joined Umetrics as an application specialist and has for the last 5 years, worked towards PAT with increased responsibilities.

## Steven Doherty

Steve Doherty is currently in the corporate PAT group at Eli Lilly and Co. in Indianapolis. Steve received his B.A. in Chemistry at Wabash College in 1981 and his Ph.D. in Analytical Chemistry at Indiana University in 1989. His current position involves the deployment of process analytical tools in development and manufacturing environments to increase process understanding, and partnering with instrument vendors to drive the implementation of new technologies. He is past Chairman for the Center for Process Analytical Chemistry’s Industrial Advisory Board, serves on the Scientific Board of the International Forum on Process Analytical Sciences, and is on the Steering Team of the Pharmaceutical Process Analytical Roundtable.

## Ingrid Maes

Ingrid is responsible for innovative technologies, including Process Analytic Technology (PAT), within the Siemens Headquarter Competence Centre Pharma, located in Antwerp (Belgium). She obtained a Master degree in Chemical Engineering, and in biotechnology and medicinal chemistry, from the University of Brussels. She has worked, for the past 15 years, in Process Analytics and Multivariate Data Analysis as Marketing and Sales Manager, and for developing new application fields for Process Analytics and control, in many industrial branches. She is author of many presentations at international conferences. She is also involved in various PAT related organisations, such as Executive Committee member of ASTM E55 (Pharmaceutical Manufacturing) and the ISPE PAT Interest Groups (SIG). She has presented at the FDA inspectors and reviewers training in Washington.

## Tony Slapikas

Mr. Slapikas has been involved in the design, application, sales, and servicing of mass spectrometers for close to 30 years. He is currently the Product Manager for Process Mass Spectrometry at Ametek Process Instruments with special emphasis on developing on-line analytics and applications for the pharmaceutical industry.

## Matthew Smith

Matthew J. Smith is the Life Sciences Marketing Manager for Hach Ultra Analytics, manufacturer of Anatel TOC analyzers, Orbisphere and Polymetron electrochemical sensors, Met One and HIAC particle counters. Prior to joining Hach Ultra in 2006, Matt’s PAT expertise was developed in the field of infrared, near-infrared and Raman spectroscopy. In this capacity, Matt served as Director of Sales for Axsun Technologies and spent 17 years in sales and marketing with the Nicolet product division of Thermo Fisher Scientific. Matt earned his Ph.D. in Physical Chemistry from Duke University in 1988.

## Chris Hobbs

Chris Hobbs is responsible for ABBs PAT Main Execution Centre (MEC) for Europe based in St Neots, Cambridgeshire UK. Chris is one of the founder members of ABBs global team for PAT and brings with him more than 20 years of practical experience of DCS deployment over many types of industries. Over the last few years Chris has moved from taking a leading technical role with Process Control, MES and ERP products to a strategic solution deployment role for ABB’s PAT solution developed by ABB’s global PAT team in partnership with the pharmaceutical industry.

### Related people

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Send this to a friend