Abstract: Mechanisms are provided which configure a medical treatment recommendation system to implement a medical condition independent treatment recommendation model. The medical condition independent treatment recommendation model operates on medical condition independent scoring features that are independent of any specific medical condition. The mechanisms configure the medical treatment recommendation system to receive a medical condition cartridge providing at least one medical condition specific evaluation feature that is specific to a medical condition, and process patient information based on a combination of an application of the at least one medical condition specific evaluation feature, and an execution of the medical condition independent treatment recommendation model of the medical treatment recommendation system, to generate a treatment recommendation for a patient medical condition associated with the patient. The mechanisms then output the treatment recommendation for the patient.

Abstract: A method, computer program product, and computing system device for receiving, at a computing device, a plurality of candidate therapy options, a corpus of therapy data, and a set of therapy recipient cohorts. At least a portion of therapy data may be filtered from the corpus of therapy data based upon, at least in part, the plurality of candidate therapy options and the set of therapy recipient cohorts. A graph may be generated to include a plurality of nodes representative of the plurality of candidate therapy options and a plurality of connections between the plurality of nodes indicating a sentiment factor with respect to the plurality of candidate therapy options based upon, at least in part, the at least a portion of therapy data filtered from the corpus of therapy data. Using the graph, a ranked list of the plurality of candidate therapy options may be generated.

Abstract: A computer-implemented method, system and computer program product for understanding information using a document analysis technique. An initial corpus of information is formed by identifying a document(s) that match a search criteria. The initial corpus of information is expanded with a set of documents containing statements with a semantic meaning within a threshold degree of similarity to a semantic meaning of statements contained within the document(s) used to form the initial corpus of information. Viewpoint(s) are then extracted from the expanded corpus of information using a natural language processing technique. A new set of documents is analyzed by identifying the subject, assertion and context statements. Assertions in the new set of documents that are within a threshold degree of agreement or disagreement with the extracted viewpoint are highlighted to assist the user in understanding how information aligns with a viewpoint.

Abstract: Unstructured data are processed in response to a request for information to derive attributes. The attributes include a subset of variables utilized to determine the requested information. The attributes are ranked based on a set of relevance factors for determining the requested information. One or more attributes are selected for presentation to a user based on the ranking to receive modifications and produce modified attributes. The requested information is determined based on the set of variables including the modified attributes.

Abstract: Method and apparatus for performing feature engineering using negative inferences are provided. One example method generally includes identifying a plurality of concepts and analyzing a corpus of documents to determine a first co-occurrence rate for a first concept and a second concept in the plurality of concepts. The method further includes analyzing the corpus of documents to determine a second co-occurrence rate for the second concept and at least a third concept of a set of concepts related to the first concept and determining an inverse relationship between the second concept and the third concept. The method further includes generating test data for training a machine learning model including a negative inference between the second concept and the third concept and training the machine learning model using the test data.

Abstract: Techniques for cognitive analysis of documents are provided. A document corresponding to a clinical trial is received. Criteria are identified for the clinical trial based on processing the electronic document using natural language processing (NLP) techniques, where the criteria specify attributes of individuals that are eligible for the trial. Further, the criteria are analyzed using NLP techniques to determine a respective confidence level for each respective criterion, where the confidence level indicates a degree of certainty that the criterion is applicable to the trial. A knowledge graph is generated based at least in part on the clinical trial, where the confidence level assigned to each criterion is included in the knowledge graph for subsequent use. Finally, one or more therapies in the knowledge graph are selected to be recommended, based on a patient profile and the confidence level assigned to each of the criteria.

Abstract: Techniques for granular analysis of a knowledge graph are provided. A profile comprising a plurality of attributes is received, and a knowledge graph is analyzed to identify a plurality of therapies, based on the plurality of attributes. A document that is relevant to a first therapy of the plurality of therapies is identified, and a criterion stated in the document is determined. Further, an aggregate value is determined for the criterion, based on a plurality of participants associated with the document, wherein the first aggregate value represents attributes of the plurality of participants. A weight is generated for the document, based at least in part on the plurality of attributes and the aggregate value. A score is generated for the first therapy, based at least in part on the weight, and an optimal therapy is determined, from the plurality of therapies, based in part on the score.

Abstract: Techniques for unbounded therapy evaluation are provided. A request to suggest a potential therapy based on a patient profile is received. A plurality of accepted therapies is determined based on the patient profile, where the plurality of accepted therapies is based on stored definitions obtained from one or more subject matter experts. Next, a plurality of therapy components is identified based on the plurality of accepted therapies. A plurality of potential therapies is then identified based on the plurality of therapy components, where none of the plurality of potential therapies are included in the plurality of accepted therapies. A score is generated for a potential therapy of the plurality of potential therapies based on analyzing a knowledge graph, where the score indicates a suitability of the potential therapy for a patient associated with the patient profile. Finally, the potential therapy is provided, along with an indication of the score.

Abstract: Techniques for cognitively interpreting comparative statements are provided. A corpus of documents associated with treatment of a first medical disorder is received. Relative efficacies for a plurality of therapies are determined by, for at least a first respective document in the corpus of documents, identifying and extracting, using one or more natural language processing techniques, a conclusion specified within natural language text of the respective document, where the conclusion includes a comparison between two or more therapies of the plurality of therapies, determining a sentiment of the extracted conclusion, with respect to at least one therapy of the plurality of therapies, and identifying one or more cohorts of patients that are associated with the conclusion. A knowledge graph is generated based on the relative efficacies.

Abstract: Techniques for evaluating dynamically modified plans are provided. A selection of a treatment plan template is received, where the treatment plan template specifies a plurality of treatment stages, where each treatment stage defines a plurality of treatment options. A plurality of modifications to the treatment plan template is generated. It is determined, for each respective modification of the plurality of modifications, whether the respective modification is permissible, based on one or more predefined institutional criteria. Upon determining that a first modification of the plurality of modifications is permissible, a first treatment plan is generated based on the first modification to the treatment plan template. Further, a first predicted efficacy measure is generated for the first treatment plan based on analyzing a knowledge graph. Finally, the first treatment plan is provided, along with at least an indication of the first predicted efficacy measure.

Abstract: Techniques for identifying missing evidence are provided. A plurality of documents, each comprising digitally encoded natural language text data, is received. The plurality of documents is processed to determine a plurality of pair-wise comparisons between a plurality of therapies, where each of the plurality of pair-wise comparisons indicate a relative efficacy of at least one therapy in the plurality of therapies, as compared to at least one other therapy in the plurality of therapies. A knowledge graph is generated based at least in part on aggregating the plurality of pair-wise comparisons, and the knowledge graph is analyzed to identify one or more knowledge gaps within the knowledge graph. Finally, at least an indication of the identified one or more knowledge gaps is output.

Abstract: Mechanisms are provided for extending a medical condition base cartridge (MCBC) based on user modifications of the MCBC. The mechanisms generate a MCBC for a medical condition at least by performing cognitive processing of natural language content in a corpus to extract patient attributes and treatment information associated with the medical condition. Information stored in the MCBC is presented to a user via at least one interface and user input is received, where the user input indicates at least one of a modification of, addition to, or deletion of, a portion of the information presented to generate an extension to the MCBC. The extension to the MCBC is stored as an expert medical condition cartridge (EMCC) in association with the MCBC in a repository. The EMCC is input to a medical cognitive system to perform a medical cognitive operation on patient information based on the EMCC.

Abstract: Natural language processing includes processing unstructured data in response to a request for a conclusion to a line of inquiry to produce structured information that conforms to a computer readable information schema. Values obtained from the unstructured data are assigned to attributes of the structured information and determining a candidate conclusion to the line of inquiry based on the value-assigned attributes. A confidence score is computed that is indicative of a sufficiency of the candidate conclusion as the conclusion to the line of inquiry. One or more of the attributes are displayed to a user for modification responsive to the confidence score of the corresponding conclusion failing to meet a confidence criterion. The displayed attributes are ranked by relevance of the attribute to the corresponding conclusion. The conclusion to the line of inquiry is determined based on the modified values of the user-selected attributes.

Abstract: Auditing database access in a distributed medical computing environment may include receiving from a user a query of one or more databases within the distributed medical administration computing environment; determining by an audit manager in dependence upon audit policies for the medical computing environment whether any of the results of the query require auditing access; and if any of the results of the query require auditing access, creating an audit record and storing the audit record in an audit database.

Abstract: According to embodiments of the present invention, medical treatment outcomes are simulated. A system receives a request to determine a medical treatment pertaining to a medical condition of a patient. The request is applied to one or more models, via a processor, to simulate outcomes for a plurality of different medical treatments for the medical condition, wherein the one or more models account for impacts of the plurality of different medical treatments on use of medical treatments awaiting future approval. The medical treatment is determined from the simulated outcomes, via a processor, with a desired level of impact on the medical treatments awaiting future approval based on characteristics of the patient. Methods and computer readable media are also provided herein for simulating medical treatment outcomes.

Abstract: A method, computer program product, and computing system device for receiving, at a computing device, a plurality of candidate therapy options, a corpus of therapy data, and a set of therapy recipient cohorts. At least a portion of therapy data may be filtered from the corpus of therapy data based upon, at least in part, the plurality of candidate therapy options and the set of therapy recipient cohorts. A graph may be generated to include a plurality of nodes representative of the plurality of candidate therapy options and a plurality of connections between the plurality of nodes indicating a sentiment factor with respect to the plurality of candidate therapy options based upon, at least in part, the at least a portion of therapy data filtered from the corpus of therapy data. Using the graph, a ranked list of the plurality of candidate therapy options may be generated.

Abstract: Mechanisms are provided for training and operating a Question and Answer (QA) system pipeline. A corpus of information is received which comprises historical data to which one or more filter criteria are applied to extract filtered historical data relevant to a training objective for training the QA system pipeline. Attribute data, action data, and temporal characteristic data are captured from the filtered historical data. An answer key entry is automatically generated in an automatically generated training answer key data structure based on the attribute data, action data, and temporal characteristic data. The correct answer associated with the answer key entry is an action specified by the action data. The temporal characteristic data provides a historical context for the answer key entry. The QA system pipeline is trained using the automatically generated training answer key data structure.

Abstract: A method, computer system, and a computer program product for a dynamic question and answer (QA) process is provided. The present invention may include receiving an input by a user. The present invention may also include analyzing a user expertise level and an amount of experience based on the received input. The present invention may then include adjusting an expert recommendation to align with the analyzed user expertise level and the amount of experience based on the analyzed user expertise. The present invention may further include providing a plurality of feedback based on the adjusted expert recommendation.

Abstract: Provided are a computer program product, system, and method for determining user interactions with natural language processor (NLP) items in documents to determine priorities to present NLP items in documents to review. Information is received on user interactions with natural language processing (NLP) items in documents presented at least one computing device. Indication is made in a database of user interaction information on the user interactions with the NLP items indicated in the received transmitted information. The user interaction information in the database providing interaction information for NLP items in the selected document is used to determine an order of the NLP items in the selected document. Content of the NLP items in the selected document is transmitted to present to a reviewing user at a reviewing computing device in the determined order.

Abstract: Provided are a computer program product, system, and method for providing information on detected user interactions with natural language processor (NLP) items to a server to use to determine an order in which to render NLP items in a user interface. Interactions with natural language processing (NLP) items in documents are detected. For each interaction with an NLP item of the NLP items in the documents, determination is made of a context attribute value for a context attribute related to the interaction with the NLP item. Then interaction information is generated indicating the NLP item and the determined context attribute value. The interaction information is sent to the server to include in user interaction information in a database.