| CPC G16H 50/20 (2018.01) [A61B 5/48 (2013.01); G06F 16/243 (2019.01); G06F 16/258 (2019.01); G06F 40/20 (2020.01); G06F 40/40 (2020.01); G16H 10/20 (2018.01); G16H 10/60 (2018.01); G16H 50/70 (2018.01); G16H 70/00 (2018.01)] | 13 Claims |
|
1. A method for detecting a recurrence of a medical condition in a subject previously treated for the medical condition by executing a plurality of machine-readable instructions of a non-transitory machine-readable medium, the method comprising:
receiving a first data set comprising first information of a first type relating to the medical condition of the subject corresponding to a first treatment;
receiving a second data set comprising second information relating to the medical condition of the subject corresponding to a second treatment, the second information being of a second updated type different from the first type;
generating, by a plurality of natural language processing (NLP) modules, a list of phenotypes from the first data set and the second data set;
extracting the first information from the first data set, by a first NLP module of the plurality of NLP modules;
extracting the second information from the second data set, by a second NLP module of the plurality of NLP modules;
reformulating, using a resource wrapper interface, the first information and the second information into structured data based on a standardized model and the list of phenotypes;
obtaining, from a user, a data retrieval request comprising a structured query adhering to a template and including a new phenotype, such that the new phenotype is received directly from the user within the structured query;
in response to obtaining the data retrieval request, automatically identifying the recurrence of the medical condition in the first information of the first type and the second information of the second updated type, the automatically identifying the recurrence comprising:
determining that the medical condition in the first information of the first type and the second information of the second updated type correspond to the new phenotype;
in response to determining that the medical condition in the first information of the first type and the second information of the second updated type correspond to the new phenotype, creating a domain specific template (DST) function corresponding to the new phenotype;
defining, by the DST function, the new phenotype, wherein the new phenotype is not included within the list of phenotypes generated by the plurality of NLP modules and is defined without receiving an updated list of phenotypes;
matching a first one or more pairs including data elements of the structured data based on the structured query, wherein each of the pair corresponds to the new phenotype, wherein at least one pair of the first one or more pairs includes a null entry corresponding to a missing data element in the structured data;
filtering the first one or more pairs to remove the at least one pair including the null entry, thereby resulting in a second one or more pairs, each pair of the second one or more pairs including two causally connected data elements from the structured data;
determining, by the DST function, the recurrence of the medical condition based on a pair in the second one or more pairs corresponding to the new phenotype, wherein the DST function comprises a decision tree;
marking the determined recurrence with a timestamp associated with at least one of the data elements in the pair; and
retrieving at least one of the pair of connected data elements in the second one or more pairs from the structured data in response to identifying the recurrence; and
providing, to a display, an output associated with the at least one of the connected data elements retrieved from the structured data, the result from determining the recurrence of the medical condition in the subject, and the timestamp.
|