Filepool to store 100 million pages on drugs approval administration and management

Voorburg 30 July 1999 The Dutch Institute for Drugs Approval and Administration (CBG) has selected the Filepool solution, an innovative way of document storage, to digitally manage the incredible amount of paper files. The CBG Institute is responsible for the final approval of each new drug that is introduced to the Dutch market by the pharmaceutical industry. The documentation for one single drug often exceeds the volume of three solid metres of contents. Since several parts of one file are distributed among different administrators, CBG is facing a gigantic file version management problem. To avoid mistakes in the various stages of document data handling, CBG uses Filepool software, a combined toolset to streamline every file transaction using digital fingerprint identification, or "e-Clipping".

Advertisement

The Dutch Institute for Drugs Approval and Administration (CBG) has selected the Filepool solution, an innovative way of document storage, to digitally manage the incredible amount of paper files. The CBG Institute is responsible for the final approval of each new drug that is introduced to the Dutch market by the pharmaceutical industry. The documentation for one single drug often exceeds the volume of three solid metres of contents. Since several parts of one file are distributed among different administrators, CBG is facing a gigantic file version management problem. To avoid mistakes in the various stages of document data handling, CBG uses Filepool software, a combined toolset to streamline every file transaction using digital fingerprint identification, or "e-Clipping".

The e-Clip constitutes a unique series of 34 signs to identify the contents of a specific file. If one small detail in a document is changed, the file will be assigned a new e-Clip. Given the fact that there exist over 2x10 to 69th different e-Clips, the possibility that the same e-Clip is assigned to two differing documents will practically be zero. The user only needs the e-Clip to request a document. It is no longer necessary to send the whole file to a colleague, since the e-Clip suffices to have the file searched by the document management system. The file name and date thus become irrelevant. As a result, the CBG database in fact is no document base but an e-Clips base with annotations relating to persons or processes. There is no central base to manage the storage location of the files corresponding to the e-Clips.

Via the e-Clip, all local document pools are activated by the request and the one that "recognizes" the e-Clip, sends the file to the user. It is of no importance whether the user is the author of the file or a colleague who is directed to the file through an e-mail or some other reference. The e-Clips have no semantic relation whatsoever with the file contents. Filepool has been developed by the Belgian Internet provider Planet Internet, which is a subsidiary of KPN Telecom. Three months ago, CBG has integrated Filepool into its Oracle database. Each incoming file is being divided into chapters. These are the smallest possible entities to be requested by the administrating experts. Every chapter acquires a digital fingerprint or e-Clip. The chapters are distributed among CBG collaborators attached to the university hospitals of Groningen and Nijmegen.

The experts' comments are collected and integrated into reports addressed to the drug designing companies in order to adapt, for instance, the matching medical description. The major pharmaceutical companies provide the CBG Institute with data in electronic format, such as Word documents, to speed up the approval procedure. Smaller companies however still deliver paper files which are scanned by CBG. The biggest problems consist in the size of the files and the great number of revised versions. The challenge lies in the exact identification of each document. This involves not only the contents but also the connections with relevant aspects of the business processing. Through the use of e-mail, electronic documents circulate in various copies. Users, independently from each other, are likely to make corrections and updates of the file.

These factors turn document names into poor identifiers. Therefore, a lot of organizations prefer to work with metafile-based archival systems which use lemmas. This kind of databases however also have their restrictions since users often do not know the lemma that corresponds with the requested file. If the document cannot be traced, this doesn't mean however that it doesn't exist. The idea to connect each document to a record in a database linking data and processes unequivocally to one another is not waterproof either. It still remains possible to make changes to a file without anybody noticing the file date change. The file date indeed indicates whether some user has been manipulating a document but people seldom are checking file dates. Until recently, even operating systems were overwriting without hesitation recent documents with older files carrying the same name.

CBG's Filepool fingerprint solution can solve these problems by means of specific techniques, such as replication. The same document may be stored in various locations but the slightest change is transmitted simultaneously to each site. The "broadcasting" technique allows to send a user's request all through the entire network. The first database which receives the request and is able to come up with an answer, deals with the question. Filepool integrates all these various techniques, together with file compression to save both bandwidth and space. As a result, there is one single system to secure the files against any changes, whether intentional or not. The Automatisering Gids has served as the news source to this article.


Leslie Versweyveld

[Medical IT News][Calendar][Virtual Medical Worlds Community][News on Advanced IT]