Abstract
For the management of patient databases where the data items differ from patient to patient, the notion of conditional existence (CE) is introduced to capture the relevant items for each individual while placing every individual under the same set of items. Assuming a table consisting of patients (rows) and items (columns), a CE is expressed in the form f(X)→E(Y) and is interpreted as “a set of items Y is meaningful for those who satisfy the condition f(X)”. Based on CE's, the partition of the items is derived to form a basis for describing relevant items for any individual. Some practical applications are shown including the input procedure with the capability of identifying relevant items dynamically, and the usefulness of the methodology in patient data management practice is discussed.