The Benefits of Process Information Management in R&D - A Case Study

Capturing and managing key process data can reap significant benefits - as AstraZeneca has recognized. Part of the company's process research and development (PR&D) operation is located at its Loughborough (UK) facility, where new manufacturing processes are developed. The site also houses a pilot plant for primary (bulk active drug) manufacture.

In November 1999, AstraZeneca commissioned AspenTech (Cambridge, UK) to install its information management system (IMS) InfoPlus.21 at the pilot plant. The platform, which has been in use since August 2000, captures and analyses live and historic process data from the plant, and also gathers data from a number of other sources. The data are then available for secure, long-term storage whilst being easily accessible for analysis, allowing creation of a complete batch end report. Sources of data include

• distributed control systems

• manually entered data, such as batch data (yield, purity, finished product weights, comments) and campaign data

• Fourier transform infrared (FTIR) data (currently under implementation)

• vent header exhaust analysis system data (also currently under implementation).

Figure 1 illustrates a schematic of the system. Since implementation, the system has been used in many ways across a number of manufacturing and business processes. The current and potential benefits have been examined and analysed. Current benefits include predictive maintenance and reduced downtime; improved process analysis and time saving; and enhanced process development and technology transfer.

Predictive maintenance and reduced downtime

Process engineers use the IMS to highlight any deviations from default or validated operating norms, prompting corrective action to ensure the output product complies with registered quality specifications. Examples include the following:

Critical systems. Actively monitoring critical systems, such as the thermo fluid unit, gives engineers the information they need to maintain operations with significantly reduced downtime.

Effective preventive maintenance

The IMS can be used, for example, to monitor reactor vessel integrity 24 hours a day during test runs. Engineers can then predict any necessary maintenance and so avoid unexpected downtime, which traditionally can be up to several days per year.

These two benefits together increase plant availability by approximately 10–15 days annually (that is, by 4–6%). Other benefits include Control loops. The software monitors how well control loops are working and prompts corrective maintenance. If, for example, process parameters have deviated from compliance or validated limits, the system's analytical tools evaluate the divergence, allowing an early return to normal operation. Deviations must be proved not to have detrimentally affected product quality.

Figure 1: System schematic.

Overnight data. Gathering and monitoring of overnight test data is automatic, so analysis and trends can be available on a computer by the next morning.

Maintenance. Trend data can be accessed following any contracted out maintenance, for easy 'before' versus 'after' comparisons of equipment performance. This helps validate that process key performance indicators (KPIs) are unaffected. It also enables faster post-modification qualification testing and gives increased confidence in process consistency. It can avoid expensive batch failure and product write-off, as well as the detrimental impact on new product development.

Commissioning. Engineers can validate that a process or piece of equipment conforms to a design specification for commissioning a new plant in alignment with performance qualification activities.

The software can show that a plant is in control and fit for purpose. Effective preventive maintenance needs data capture and operational performance trending as a prerequisite, and the IMS automatically provides this. Additionally, process validation is quicker and process knowledge is enhanced.

Improved process analysis and time saving

The fundamental benefit to process chemists is that data are gathered automatically. The data generated are useful in assessing a batch's suitability - it can make the difference between scrapping a batch and proving that it is passable. Having access to real-time and historic process information leads to significant process improvement and savings in resource time, batch cycle time and the time to achieve key milestones in development projects.

Data gathering. By using the tool for historic and real-time process monitoring of all process parameters, data gathering time is reduced to virtually zero, allowing more time for process enhancement. Trend analyses that in the past were difficult can now be performed. More robust processes can result, and potentially large write-off costs avoided.

Deviation. In the event of deviation from process parameters, the process chemist must prove the batch integrity to QA standards. Process data captured by the system provides the evidence easily, so decisions can be made quickly.

Batch cycle time. The process of passing a batch is speeded up significantly, as all the data are readily available in real-time, and batch records can be kept up to date during manufacture. For example, process material may go into a 'hold' status (such as distillation), requiring continuous monitoring. The software allows the process to be monitored from the desktop, while generating parts of the batch record. Ultimately, there are time savings here both in terms of human resources and the overall batch cycle time. This may lead to faster release for clinical trial.

Enhanced process development and technology transfer

The IMS can be used as a development tool, as the earlier examples show. More robust and cost-effective manufacturing processes have been developed as a result of using the data capture and process trending tool. When a process is passed from the laboratory to the pilot plant, the plant chemist can see all the data and can make recommendations to improve cycle time and process robustness. For example,

• A process procedure stated 16 hours of flow hydrogenation. However, trend analyses proved that hydrogen uptake stopped after 4 hours, reducing the process time by 75% for that step.

• Another production process stated 10 hours drying in an oven. Monitoring the procedure showed this could be reduced to 4 hours without compromising product quality.

Bringing an IT system into an operations setting hopefully yields the expected benefits. Typically, though, other unforeseen opportunities emerge as users become more familiar with the system's capabilities. Some of these are readily achievable with little change; others may require tweaks to the system before they can be realized. The remainder of this article discusses the possible future benefits that encompass these two categories.

Readily achievable benefits

More accurate KPI measurement:

Measurement is the prerequisite to increasing plant availability and utilization. The necessary KPI measurement and calculation is currently a manual and labour-intensive process, using overall equipment efficiency (OEE) techniques. Also, data granularity is poor. The IMS continuously monitors the plant, and with only minor modification, the OEE can be automatically calculated and monitored with high accuracy.

Improved batch analysis: The batch functionality of the platform is already undergoing validation. When this is done, PR&D engineers are keen to make historical analyses of clean-in-place, start up and manufacture. Batch functionality allows users to compare different procedures so they can improve plant operations. Also, linking the software to quality assurance (QA) systems allows cross referencing and data analysis of many more batch pertinent parameters (such as QA test results trended against process performance or input raw materials). Additionally, batch-to-batch analyses allow the output quality to be compared with a 'gold batch' standard, and the impact and importance of different process parameters to be understood. This ultimately leads to more robust batch manufacture, more predictability in the shop floor schedule and, at the bottom line, a lower cost of goods in pilot plant and eventually in commercial manufacture.

Automated process monitoring and improved consistency: This is essentially already happening, but could be taken a step further by the relatively easy modification of linking automatic monitoring and alarms, signalling the need for maintenance (rather than waiting for an out-of-compliance alarm to occur). The most significant benefit would be fewer batch failures and, therefore, a lower annual cost of batch write-off.

Process change justification: As the wealth of data in the system increases, so does its ability to provide evidence for justifying process improvements that would otherwise be locked into difficult to access paper batch record systems.

Enhanced development and technology transfer: Changes in business process, such as technology transfer, can be made to stick and be more effective when applied simultaneously with system improvements or implementations. Numerous technology transfer dossiers, data and reports flow from development to commercial operations. These can be simplified, streamlined and made more accessible by using the software to automate parts of this process. The IMS can also be used for queries and will provide valid information in seconds rather than hours or days. The ability to refer back to historic batch data from a remote desktop at, say, a commercial operations facility will give a true insight into the manufacturing process that will make technology transfer both more effective and faster. Additionally, if the system was also used at the commercial manufacturing site, further rapid development based on real data trends is enabled at that site. Development staff can return to the pilot plant and remotely monitor the commercial batch manufacturing activity. The implications for smooth technology transfer, collaborative development, speed to market, cost of goods of the commercial product and cost of the development project, are considerable. In effect, using the software in this way provides a common process information management system from development to commercial manufacture.

Possible future benefits

Two particular possible benefits needing further application development are considered.

Advanced process monitoring: This would entail automated monitoring of key events, while also applying technologies such as multivariable statistical process control (SPC), inferential monitoring and neural networks. The benefits of this could be: early warning of potential process deviations; online inferred measurements of values that cannot be measured directly (such as colour); and derivation of relationships from the raw data to predict immeasurable inferred parameters, thereby explaining process cause and effect.

Batch record automation: Currently generated manually, this process is time consuming and can introduce errors. The automatic generation of the process data components of the batch record (21 CFR Part 11 compliance) would ease this burden for process chemists, and enable them to add more value in developing processes. This could be achieved using the IMS integrated with a full manufacturing execution system (MES), inclusive of electronic batch records to automatically produce the full and compliant batch record.

This illustrates the benefit potential for process improvement in the pharmaceutical industry by using information technology systems normally associated with more manufacturing-centric process industries. Pharmaceutical companies are starting to use these systems across commercial manufacturing sites producing both the bulk active drug and drug product, as well as in their pilot plants. The benefits will ultimately translate into bottom line cost reduction, and in the current aggressive business environment, that has to be a very welcome prospect.