The use of the variable-field-length (VFL) format for obtaining narrative medical data in a form suitable for computer input and subsequent storage, analysis, and retrieval has been described previously (see p 950).1 This report discusses the computer programs which have been produced, some data-analytical results thus far demonstrated, and the use of a large-scale computer facility for statistical analysis of narrative medical data. The group of programs is outlined by the use of a specific example. The advantages, limitations, costs, and possible future applications of computer processing of narrative medical data are also discussed.
The utility of the VFL format for the analysis of medical narrative data depends on the ability to produce computer programs that will perform medically meaningful operations on the data in a convenient and economical way. Such operations include, among others, the retrieval of sections of the records on the basis of content. For example,