| CPC G16H 50/20 (2018.01) [G06F 40/20 (2020.01); G16H 10/60 (2018.01); G16H 15/00 (2018.01); G16H 50/70 (2018.01)] | 11 Claims |
|
1. A diagnosis and treatment support system for re-training a previously trained model, comprising:
a storage apparatus that stores the trained model, which infers information related to a state of a patient from an examination value for a predetermined examination item, wherein the trained model was previously trained using information from patients other than a target patient; and
processing circuitry configured to:
calculate, based on correlations between examination values for plural examination items included in diagnosis and treatment information on plural patients, a conversion function that enables statistical derivation of a value from an examination value or values of one or plural examination items, the value being a possible examination value for another examination item;
store, in an electronic memory, the conversion function in association with an index value, the index value quantitatively representing a condition as to the examination items used for calculation of the conversion function;
generate, based on an inference obtained by inputting an examination value for a first examination item included in diagnosis and treatment information on the target patient into the trained model, support information for supporting diagnosis and treatment of the target patient;
in a case where examination items included in the diagnosis and treatment information on the target patient do not satisfy a requirement for the predetermined examination item,
derive an examination value for a second examination item not included in the diagnosis and treatment information on the target patient from an examination value for the first examination item included in the diagnosis and treatment information on the target patient, by using one conversion function of stored conversion functions, the one conversion function outputting the derived examination value for the second examination item, and
select the one conversion function for use based on the associated index value when there are two or more conversion functions allowing derivation of the examination value for the second examination item;
cause the trained model to infer the information related to a state of the target patient by together inputting the derived examination value for the second examination item and the examination value for the first examination item into the trained model; and
re-train the previously trained model, based on the derived examination value for the second examination item and the examination value for the first examination item so that the previously trained model is additionally trained using the derived value of the second examination item related to the target patient.
|