The information content of signs, symptoms, laboratory test results, and other events that constitute the medical record has been analyzed and it has been determined that the meaning of sign and symptom information can be described in a multiple-dimensional matrix. This multidimension meaning concept of signs and symptoms facilitates a logical structure for the storage of clinical information. This structure is not an arbitrarily imposed one, but an intrinsic, althoughnot commonly recognized, attribute of medical information. The utility of a medical record consisting of signs and symptoms coded in multiple dimensions is facilitated by regarding each sign or symptom as a point in an "n"dimensional hyperspace with "6' equal to the number of fields required to describe a sign or symptom. Sparse matrix techniques make it possible to store this conceptually very large matrix as a record of reasonable size. Access and reporting programs facilitate processing this information as if it were stored in a very large matrix. The feasibility of this matrix medical language concept has been demonstrated by developing modular computer programs for input, storage, and report by meaning of the signs and symptoms that constitute a patient's medica! history.
The "systems approach" and evolving techniques for computer processing of information facilitate new approaches to handling medical information that were not previously practical. Although the need is obvious, little has been done toward developing, defining, or even discussing a general language or structure for clinical information. This is particularly true at the level of detailed descriptors such as signs and symptoms. The standard medical classification schemes, such as the Standard Nomenclature of Diseases and Operations (SNDO) (I), the International Classification of Diseases-Adapted (ICDA) (2), and the Systematized Nomenclature of Pathology (SNOP) (3) are intended primarily for classification of diseases, operations, and pathological changes.
The primary objective of the Section of Medical Computer Sciences at Yale is the application of computer and information science to new methods for acquisition, storage, and reporting of clinical information (4). Basic to this effort is the development of a clinical descriptive language that reflects lay terms but is structured according to meaning. Initially this language is being developed and validated by