Abstract
An interactive dictionary driven system (Input Standardization System—ISS) has been designed and implemented. The function of the system is to improve the quality of data entered by medically untrained input clerks into the problem/manifestation and therapy descriptor fields in STOR, a computerized ambulatory care medical record system at UCSF. The ISS is able to (1) correct spelling errors, (2) expand abbreviations and truncations, and (3) suggest possible correct words to substitute for recognized misspellings and typographic errors. In a study of 2467 items processed by the ISS processor, 386 descriptors were correctly changed to more standard forms. Nine descriptors were incorrectly changed, six of which were judged to be potentially medically confusing. Only one of these errors would have occurred had the dictionary been more complete.