A method, in a data processing system comprising a processor and a memory, the memory comprising instructions that are executed by the processor to configure the processor to implement a patient summary generation engine, the method comprising:
- receiving, by the patient summary generation engine, a request for generation of a summary output for a specified clinical data category, of a plurality of clinical data categories, specified in the request;
- retrieving, by the patient summary generation engine, stored clinical data category specific knowledge resource data structures specifically corresponding to the specified clinical data category, wherein the clinical data category specific knowledge resource data structures specify recognizable patient EMR elements for the specified clinical data category, and wherein different clinical data category specific knowledge resource data structures are associated with different clinical data categories in the stored clinical data category specific knowledge resource data structures;
- configuring, by the patient summary generation engine, computer natural language processing logic to perform natural language processing of patient electronic medical record (EMR) data specifically for the specified clinical data category, based on the retrieved clinical data category specific knowledge resource data structures, wherein the retrieved clinical data category specific knowledge resource data structures corresponding to the specified clinical data category specify at least one of terms, phrases, or medical codes corresponding to the specified clinical data category, and wherein the computer natural language processing logic is modified to configure the computer natural language processing logic to extract, from content of the patient EMR, portions of the content corresponding to the at least one of terms, phrases, or medical codes specified in the retrieved clinical data category specific knowledge resource data structures;
- parsing, by the configured computer natural language processing logic, a patient EMR to extract a plurality of instances of a medical concept based on the clinical data category specific knowledge resource data structures, wherein at least two instances of the medical concept utilize different representations of the medical concept in the patient electronic medical record;
- performing, by the patient summary generation engine, a similarity analysis between a plurality of combinations of the instances of the medical concept to thereby calculate, for each combination of instances in the plurality of combinations of instances of the medical concept, a similarity metric value, wherein the similarity analysis combines a lexical similarity analysis measuring similarity of text, and a semantic similarity analysis measuring a similarity of meaning based on the clinical data category specific knowledge resource data, for each combination of instances to generate the similarity measure value for the combination of instances;
- clustering, by the patient summary generation engine, the instances of the medical concept based on the calculated similarity metric values for each combination in the plurality of combinations of the instances of the medical concept to thereby generate one or more clusters;
- selecting, by the patient summary generation engine, a representative instance of the medical concept from each cluster in the one or more clusters based on user specified selection criteria, wherein different user specified selection criteria are used for different users, and wherein the user specified selection criteria indicates user preferred characteristics of medical concepts in a cluster to use to select a representative instance;
- and
- generating, by the patient summary generation engine, a summary output of the patient EMR comprising the selected representative instances of the medical concept from each cluster.