Cerner Innovation, Inc.

Christopher S Finn

David P McCallie, Jr

Margaret Cushing Kolm

Steven S Crooks

David J O'Larte

Thomas C Gifford

J D Tyler

Chad A Cole

David McCallie

Thomas Hanf

Savitha Venkat

Michael Alan Ash

Uttam Ramamurthy

James Gritter

Stephen Patel

Adam Sabaliauskas

Tonja McCord

Maulik Gandhi

Anubha Garg

Brandon Davis

David Dellsperger

Particular embodiments can update a deployed machine learning model with actual entity data depending on anomalies detected in stream data, which can be stored to a computer object, such as a journal. Various embodiments map particular subsets of a larger pool of raw input data to the particular models that need the input data and store the raw input data to computer objects so that the corresponding machine learning models can make predictions according to any suitable policy or triggering event on any of the data located in the computer objects. Such mapping allows each machine learning model to continuously make predictions based on the data it needs.

Intelligent updating and data processing for deployed machine learning models

A stateless, on-demand semantic indexing and search service for a health information exchange is provided that anticipatorily fetches data from one or more external sources, extracts semantic concepts from textual documents, and serves an in-memory semantic search index to users to ease finding of relevant patient data across a set of distributed health record data sources. Clinicians may search for not only text, but for semantic concepts, and the search index is able to return relevant results regardless of whether the source information is located inside or outside the local electronic health record. Predictive algorithms may be incorporated to pre-fetch, semantically process and index data for patients likely to be searched in an upcoming time window. Activity indicators for clinicians and patients may be utilized to pre-fetch, semantically process, and index data.

Semantic search for a health information exchange

Systems, methods, and user interfaces infer data organization for display. An inference system learns patterns of how various healthcare items are typically grouped on a display so data can be aggregated and displayed in appropriate groupings to a clinician user. From a user perspective, it appears as though the display has been manually edited or a mapping has been created so that items are displayed in appropriate groups, even though neither editing nor mapping has been performed.

Inferring semantic data organization from machine-learned relationships

In some examples, a word embedding is received. A word embedding includes a plurality of vectors in vector space. Each vector of the plurality of vectors represents a natural language word or other character sequence. Each vector is oriented in vector space based on a semantic similarity between each of the natural language word or other character sequence. A first distance is determined between a first vector and a second vector. A second distance is determined between a third vector and the second vector. Based at least in part on the first distance between the first vector and the second vector and the second distance between the third vector and the second vector, the third vector is moved closer or further away from the second vector. The moving is indicative of introducing a bias or removing a bias between the third vector and the second vector. This introducing or removing of bias supports more accurate and inclusive results from an application such as search of healthcare data, among other things.

Concept embeddings for improved search

Methods, systems, and computer-readable media are provided to detect similarities between two or more clinical documents. It is determined that a clinician is currently inputting data into a first clinical document that is associated with a patient. A selectable option is provided on a user interface into which the clinician is currently inputting the data. An indication is received that the selectable option has been selected by the clinician. An algorithm is applied to identify the second clinical document from a plurality of clinical documents. At least a portion of the inputted data in the first clinical document and at least a portion of data in the second clinical document are transformed to generate a new representation of the first clinical document that indicates the similarities that are potentially inaccurate or inappropriate between the first clinical document and the second clinical document.

Duplication detection in clinical documentation during drafting

Methods, computer systems, and computer storage media are provided for providing real-time analysis and annotation of clinical documents in a distributed system. A clinical transformation session is opened at a clinical transformation server maintaining sessions for one or more editors and agents operating on a clinical document. Sequences of operations on the clinical document are stored at a memory accessible by the server. At least a portion of the clinical document is analyzed in real-time to provide annotations and other document modifications to each of the one or more editors having a session at the server. Parallel annotations or modifications are resolved and a synchronized view of the clinical document is maintained based on operational transformation.

Real-time collaborative clinical document analysis and editing

Methods are provided for sharing patient information by way of an online collaboration system. A requesting clinician, through an online networking system, identifies one or more clinicians with whom the requesting clinician wishes to collaborate by sending at least one patient-focused clinical data element. For the clinicians who have been identified as not being authorized to access patient-identifying data, the at least one patient-focused clinical data element is de-identified such that it no longer includes any patient-identifying data. The de-identified patient-focused clinical data element is then communicated to the clinicians not allowed access to the patient-identifying information.

Clinical collaboration using an online networking system

Methods, systems, and computer storage media are provided for converting a source in a first terminology to a target in a different terminology. A source in a first terminology may be converted to a target in a second terminology using cross maps that map sources to all possible matching targets, probability data derived from frequency data illustrating a frequency of selection for targets in relation to specific sources, lexical matching algorithms indicating targets with highest lexical matches for specific sources, or a combination thereof. Any one, or a combination, of the data above may be used to identify a probabilistically most relevant target in a desired terminology.

Term classification based on combined crossmap

Methods, systems, and computer-readable media are provided to detect incremental changes in clinical documents associated with a patient. A user selection is received indicating that a user wants to compare two or more clinical documents that are associated with the patient or authored by the same clinician, enabling the user to become up-to-date on the patient's care. A first clinical document and a second clinical document are identified. For the first clinical document and the second clinical document, differences are algorithmically determined. The first clinical document is transformed into a new representation of the first clinical document. Further, the second clinical document is transformed into a new representation of the second clinical document. One or more of the new representations of the first and the second clinical documents are displayed, assisting the user to quickly find and understand the differences between the documents.

Duplication detection in clinical documentation to update a clinician

Methods, systems, and computer-readable media are provided to detect similarities in clinical documents that might be inaccurate or inappropriate. A first clinical document and a second clinical document that are to be compared are identified. This identification of the documents is based on times associated with the first and second clinical documents, an identity of clinicians who authored the first and second clinical documents, an identity of patients associated with the first and second clinical documents, a type of the first and second clinical documents, or contents of the first and second clinical documents. The first clinical document is compared to a portion of the second clinical document. A report is automatically generated, where the report indicates the similarities between the portion of the first clinical document and the portion of the second clinical document that are potentially inaccurate or inappropriate.

Duplication detection in clinical documentation

Computerized methods and systems for using user-attribute ontologies to calculate anonymity levels are provided. A semi-automated mechanism permits selective and optionally progressive exposure of demographic and clinical information to other members of a medical social network group that a member has joined. When a user decides to join a medical social network group (which may be based upon system suggestion or initiated of the user's own accord), he or she selects or creates a profile that has a statistically validated degree of anonymity, ranging from fully anonymous to fully transparent. Based upon a user's elected level of exposure, only certain data is exposed to the rest of the group's membership. Thus, the user is permitted to personally balance the tension between exposing a level of information that may render him or her personally identifiable and exposing enough information to obtain meaningful advice and/or camaraderie.

Using user-attribute ontologies to calculate user-anonymity degrees

A system and associated methods provide healthcare entities with distributed analysis capabilities for records of a patient population. The analysis seeks to find matches between a piece of clinical knowledge introduced to a healthcare entity and data contained in the patient population records. According to one method, a service generates executable code based on the piece of clinical knowledge. Through a communication with the service, the healthcare entity makes a determination as to what extent the clinical knowledge has relevance to a patient population of the healthcare entity. Based on this determination, the healthcare entity downloads at least a portion of the executable code. Then, the records of the patient population may be analyzed by the executable code to register matches of the clinical knowledge with data contained within the records. System activity may then be initiated based on the matches registered.

Growing community of inventors

Liberty, MO, United States of America

Christopher S Finn