Clinical information processing apparatus, method and program

A clinical information processing apparatus comprising:

• a registration case information obtainment unit that obtains registration case information for calculating a likelihood ratio including a multiplicity of registration cases about a plurality of comparison target patients, and to each of the multiplicity of registration cases a plurality of clinical-information items each of which is classifiable into a plurality of groups being correlated, and registration case information for calculating a degree of similarity including a multiplicity of registration cases about a plurality of comparison target patients, and to each of the multiplicity of registration cases a plurality of clinical-information items each of which is classifiable into a plurality of groups being correlated, and registration case information for estimating a key item including a multiplicity of registration cases about a plurality of comparison target patients, and to each of the multiplicity of registration cases a plurality of clinical-information items each of which is classifiable into a plurality of groups being correlated;
• a target case obtainment unit that obtains, as a target classification of a target clinical item, a classification of each of a plurality of clinical-information items about a target patient;
• a likelihood ratio information calculation unit that calculates, based on the registration case information for calculating a likelihood ratio, likelihood ratio information for each classification of a key item of the plurality of clinical-information items, and the likelihood ratio information correlating a likelihood ratio corresponding to each classification of at least one clinical-information item other than the key item to each classification of the key item included in the registration cases, and the likelihood ratio information being obtained by calculating a likelihood ratio between a likelihood of belonging to one classification of the key item and a likelihood of belonging to any classification of the key item other than the one classification of the key item when a case belongs to each classification of the at least one clinical-information item other than the key item;
• a key item estimation unit that tentatively estimates a classification of the key item to which the target patient is estimated to belong based on the target classification of the target clinical-information information item and the registration case information for estimating a key item;
• a weighting coefficient determination unit that determines, based on the estimated classification of the key item and the target classification of the target clinical-information item, weighting coefficient information in which a weighting coefficient corresponding to each classification of the at least one clinical-information item other than the key item is correlated to the estimated classification of the key item;
• and
• a degree-of-similarity calculation unit that specifies, based on the determined weighting coefficient information, the weighting coefficient corresponding to each classification of the at least one clinical-information items other than the key item for each of the registration cases included in the registration case information for calculating a degree of similarity, and that calculates a degree of similarity by using the specified weighting coefficient.

A clinical information processing method, the method comprising the steps of:

• obtaining registration case information for calculating a likelihood ratio including a multiplicity of registration cases about a plurality of comparison target patients, and to each of the multiplicity of registration cases a plurality of clinical-information items each of which is classifiable into a plurality of groups being correlated, and registration case information for calculating a degree of similarity including a multiplicity of registration cases about a plurality of comparison target patients, and to each of the multiplicity of registration cases a plurality of clinical-information items each of which is classifiable into a plurality of groups being correlated;
• obtaining, as a target classification of a target clinical item, a classification of each of a plurality of clinical-information items about a target patient;
• calculating, based on the registration case information for calculating a likelihood ratio, likelihood ratio information for each classification of a key item of the plurality of clinical-information items, and the likelihood ratio information correlating a likelihood ratio corresponding to each classification of at least one clinical-information item other than the key item to each classification of the key item included in the registration cases, and the likelihood ratio information being obtained by calculating a likelihood ratio between a likelihood of belonging to one classification of the key item and a likelihood of belonging to any classification of the key item other than the one classification of the key item when a case belongs to each classification of the at least one clinical-information item other than the key item;
• determining, based on the target classification of the target clinical-information item and the likelihood ratio information, weighting coefficient information in which a weighting coefficient corresponding to each classification of the at least one clinical-information item other than the key item is correlated to each classification of the key item;
• and
• specifying, based on the determined weighting coefficient information, the weighting coefficient corresponding to each classification of the key item and each classification of at least one clinical-information item other than the key item for each of the registration cases included in the registration case information for calculating a degree of similarity, and calculating a degree of similarity by using the specified weighting coefficient.

A non-transitory computer-readable recording medium having stored therein a clinical information processing program for causing a computer to function as:

• a registration case information obtainment unit that obtains registration case information for calculating a likelihood ratio including a multiplicity of registration cases about a plurality of comparison target patients, and to each of the multiplicity of registration cases a plurality of clinical-information items each of which is classifiable into a plurality of groups being correlated, and registration case information for calculating a degree of similarity including a multiplicity of registration cases about a plurality of comparison target patients, and to each of the multiplicity of registration cases a plurality of clinical-information items each of which is classifiable into a plurality of groups being correlated;
• a target case obtainment unit that obtains, as a target classification of a target clinical item, a classification of each of a plurality of clinical-information items about a target patient;
• a likelihood ratio information calculation unit that calculates, based on the registration case information for calculating a likelihood ratio, likelihood ratio information for each classification of a key item of the plurality of clinical-information items, and the likelihood ratio information correlating a likelihood ratio corresponding to each classification of at least one clinical-information item other than the key item to each classification of the key item included in the registration cases, and the likelihood ratio information being obtained by calculating a likelihood ratio between a likelihood of belonging to one classification of the key item and a likelihood of belonging to any classification of the key item other than the one classification of the key item when a case belongs to each classification of the at least one clinical-information item other than the key item;
• a weighting coefficient determination unit that determines, based on the target classification of the target clinical-information item and the likelihood ratio information, weighting coefficient information in which a weighting coefficient corresponding to each classification of the at least one clinical-information item other than the key item is correlated to each classification of the key item;
• and
• a degree-of-similarity calculation unit that specifies, based on the determined weighting coefficient information, the weighting coefficient corresponding to each classification of the key item and each classification of at least one clinical-information item other than the key item for each of the registration cases included in the registration case information for calculating a degree of similarity, and that calculates a degree of similarity by using the specified weighting coefficient.