One of the major advantages using an electronic document management system (EDMS) is retrieving a document quickly when it’s needed. Doing this efficiently depends a great deal on how the EDMS stores documents in its repository and how they are categorized. Most systems have one or more of the following ways to retrieve documents:
- structural search
- keyword and/or metadata search
- full text search
Structural search is used to find documents in a very structured environment. If you think of a doctor’s office or a financial advisor, they have a folder for each patient/client and all the documents for that patient/client are in the folder. To find a particular document, open the folder and scan the list of documents and find the one you need. There are several things that can be done to speed up the scan, such as categorizing the documents (Correspondence, Progress Notes, …). Doing this with an EDMS is much faster than burning a path to a physical filing cabinet. In general, this type of search is valuable when you know exactly what you are looking for. Structural search is highly dependent on the Graphical User Interface (GUI) used by the EDMS. A good GUI is intuitive and requires little or no training to search for and retrieve documents. If a user can visualize the filing structure and navigate to specific documents with minimum clicks and data entry, the GUI is probably most responsible.
An EDMS allows users to index documents with keywords. Those keywords can be entered later in a search field and a list of documents associated with a keyword or set of keywords will be presented. The more keywords associated with a document, the more specific the searches that can be performed. The fewer keywords, the more likely you would receive a longer list of documents returned by the search. This type of search is useful for finding all documents with specific key words. Something like “July vacation request”, would return a list of all July vacation requests (hopefully). This type of search is dependent on getting the keywords input correctly on the front end, but can be quite powerful when looking for specific document types.
Full Text Search (FTS) is yet another way to search for a document. FTS involves looking for a document based on a word or phrase that may be contained within the document. For example, if a clerk wants to find all prescriptions for “XYZ Drug”, they could type that into the search and get a list of all documents that contain “XYZ Drug”. FTS first and foremost requires that documents contain text. An EDMS that provides full text search indexes the text contained in all the documents within a database. Depending on the size of the document repository, the FTS database can become fairly large. As documents are filed and indexed, they become searchable using the EDMS FTS feature. This is a straightforward process for documents like emails, MS Word®, MS Excel® and other text-based documents. However, scanned documents do not contain text (a scanned document is an image) so they must be converted to a format that contains text and is searchable.
Both keyword searches and full text searches are very useful for finding lists of documents containing a particular phrase. The disadvantage to both of these is the search can possibly return long lists of documents. In addition, if the full text search is searching documents created with OCR, there may be issues with finding the text as OCR is not being 100% effective.
As you may have already concluded, no one search method is ideal for every business and every application. As with most things, there are compromises to be considered. For example, going with the easiest way to get files into an EDMS might mean it will take a little longer to retrieve. Careful, detailed indexing of every document may make it easier to find later, but there is a cost associated with this approach due to more time spent up-front filing documents. The best solution is to have the ability to run any of the discussed searches depending on the business process and the most efficient overall approach.
Jon Clark has written a more detailed white paper on this subject:
http://www.cabinetng.com/white-papers/document_search_methodologies.php