Three firms have come together to help the Pfizer pharmaceutical company with the development of a cost-effective solution for the conversion of documents to electronic formats. The so-called DocCon application will be used for the internal review of new drug applications as well as for their submission to the US government's Food and Drug Administration (FDA). To this purpose, Platform Computing has delivered the LSF Suite analysis and workload management software to harness the power of Compaq's Windows NT based workstations' cluster, whereas Infodata Systems has provided vast expertise in organizing the workflow for both electronic document creation and management. The DocCon solution will allow Pfizer to process 60.000 pages in 24 hours.
The conversion of documents from traditional paper formats into Adobe's Portable Document Format (PDF) is a time-consuming, complicated, expensive, yet essential procedure, since Adobe PDF constitutes the selected format by the FDA for electronic submission of new drug applications. Up till now, Pfizer has been applying the conventional method of scanning paper documents into an electronic format such as TIFF. Next, an index of the document is generated by means of an optical character recognition (OCR) tool. The electronic pages have to be converted from TIFF to PDF formats after which the individual PDF pages are stitched together into the original document. Finally, the OCR file is reaffixed to the PDF formatted document. Instability however often causes the existing system to crash and slow down the conversion process.
The newly provided solution allows Pfizer to separate the huge documents into individual pages in order to distribute their conversion over multiple central processing units (CPUs). Afterwards, the pages are reassembled into one single, cohesive document. For this matter, Platform's LSF Suite has been implemented into the Compaq Windows NT based workstation cluster to enhance both the multiprocessing speed and the inherent scalability of the computing resources. The shorter the conversion process, the sooner the drug can be submitted to the FDA for approval, and the less financial loss there will be, due to weeks or months of unprotected product patents. For popular drugs, the lost revenue could add up to more than $1 million per day.
The current workstation cluster is comprised of 18 Professional Workstation 8000s, equipped with four 200 MHz Pentium Pro processors each. They have been stacked into racks that are linked together. In a later stage, the Pfizer system will be extended to more than 100 workstations with four processors each, thus turning it into the largest workstation cluster ever for a dedicated use in commercial industry. Every single processor that is joint to the cluster will add another page for conversion in parallel with the rest of the document. A typical drug filing for submission to the FDA ranges from 250.000 to two million pages. The DocCon application is able to handle 60.000 pages in 24 hours.
The implementation has required a joint effort from all the partners involved. Adobe offered the conversion tool; Infodata provided the workflow solution; Platform conducted the management of the distributed computing resources as well as the technical consultancy while Compaq's Enterprise Consulting Services (ECS) team combined the separate elements, in order to customize the application for Pfizer's specific needs. The DocCon solution however can be leveraged for various other industries and markets, relating to broadcast and video, finances, computer-aided engineering, geographical information systems, chemistry, insurance business, since these are all dealing with a huge amount of data that has to be processed.