Email for more information

Company's other products

Email to a colleague

Printer friendly format

By Stuart Davie

Over the past few years, the FDA has increased pressure on the pharmaceutical and biotechnology industry to become more innovative in manufacturing. Through issuing new guidance to the industry in 2003 and 2004, the FDA encouraged the use of new quality measurement tools known as Process Analytical Technology (PAT) that greatly increase the need for data collection, data access and data analysis systems.

Implementing PAT not only helps ensure continuous compliance with FDA regulations, but such technologies offer bottom line business benefits to manufacturers who can now leverage greater process understanding. For example, time to revenue can be accelerated by shortening the tech transfer time and effort involved in the start-up of new product manufacturing. Process problems and failures can be minimized by using analytical technology to discern trends in the variability of critical process parameters and identify problems before they occur.

The move to a PAT approach is encouraging companies to re-think their systems strategy both for infrastructure and applications integration. The industry’s previous focus on validation of systems to ensure accuracy, repeatability and stability has certainly enhanced reliability but has come at a cost. The effort (ultimately read as “dollars spent”) to implement, change, upgrade and replace systems has been high and has, if not deterred companies, certainly slowed activity in this area. Once a system has been implemented and validated, there has been a general reluctance to change it unless absolutely necessary. In addition, concern over control or constraints on technology has sometimes precipitated separate networks for manufacturing systems and transactional systems such as enterprise resource planning (ERP). This, in turn, has challenged the management of these networks with regard to security, access, etc.

There is no doubt that the requirements for systems validation/qualification (IQ, OQ, PQ) are based upon sound principles of software development and implementation. However, the FDA and the pharmaceutical industry both know that it is time to raise the bar and implement changes that focus not just on the control of systems within a GxP environment but that also lead to greater process understanding and improvement.

In line with this we have seen the systems landscape change as a result of technology advances, industry standardization and vendor strategies. We currently see the following occurring:
• The realization that for an enterprise to manage the production environment effectively, shop floor data has to be made available to transactional systems, analytical systems and business intelligence tools.
• An obstacle to the free flow of data has been the myriad of proprietary systems, and we have seen the development and implementation of application program interfaces (APIs) within software packages based on industry standards such as ODBC, OLE-DB and OPC.
• Vendors operating at different ends of the systems spectrum are moving towards each other. For example, traditional shop floor control system companies such as Rockwell Automation are developing interoperability mechanisms to traditional enterprise business solutions such as ERP. At the same time, ERP vendors like SAP, known primarily for transactional processing systems, have purchased companies like Lighthammer (now re-badged as xMII) and are developing the linkage with shop floor systems such as MES, DCS, LIMS, etc.
• Many of the large systems and services companies have realized that they can position themselves at the center of any systems strategies by providing a hub or middleware layer to act as the “policeman” or director of information flow. We can see this with IBM (Websphere), Oracle (Fusion) and SAP (Netweaver).
• Network operating systems have developed much greater capabilities in the area of security, user authentication/management, inter domain operation, etc. The enhancement of Windows 2003 over Windows NT is clear evidence of this. Data transmission speeds continue to increase and compression techniques are better, allowing the movement of large amounts of data (e.g., continuous data measurements, both within LANs and across WANs). It can be argued that this may be of benefit to companies that operate with subcontractors or contract manufacturing organizations.
• The development of data historians provides a more practical mechanism to aggregate data from the controls of production equipment and, in turn, allows other tools such as analytics to easily access this data.
• On the hardware side, the advent of 64-bit CPUs, multi-core processors and faster disc speeds will enhance processing speeds which, in turn, should enable more effective processing of continuous data.

Many pharmaceutical companies are standardizing the desktop and back office environments not only for economic and management reasons but also for consistency and reduced validation efforts. Typically these standards have been born out of the more conventional transactional processing arena. As we move towards analyzing more process (continuous) data, we may find some challenges to these standards and modifications may be required.

One area of specific note is the existence of data warehouses and the potential that they bring to process understanding. Typically data warehouses are not real time and don’t contain continuous data. However, they can contain significant elements of data that will identify trends or directions. A PAT approach may be able to take advantage of this data/information even though it may be somewhat after the fact.

System interfacing/integration to address PAT requirements would start to look like this:

click the image to enlarge