Electronic Health Records

Latest Health IT News

Google Patent to Build Predictive Timeline Using EHR Data, FHIR

Google application sought patent for aggregating EHR data into a patient medical history and predictor of likely health events.

Google EHR data mining with FHIR

Source: Google

By Kyle Murphy, PhD

- The United States Patent Office has published a patent application by Google with the goal of aggregating EHR data using the FHIR-based services to predict health events.

The system and method described in the application comprise three primary components: a relational database of de-identified electronic health records, a machine learning mechanism, and a device for displaying predicted clinical events.

“In the clinical setting, the management and presentation of information regarding a patient is an important aspect of patient care and healthcare decision making,” write Mossin et al.

“There is a need for systems and methods to assist healthcare providers to allocate their attention efficiently among the overabundance of information from diverse sources, as well as to provide predictions of future clinical events and highlighting of relevant underlying medical events contributing to these predictions in a timely manner,” the patent states.

In other words, Google is proposing clinical decision support to predict and influence patient behavior.

The patent application dating back to August 2017 shows Fast Healthcare Interoperability Resources (FHIR), the application programming interface (APIs) and specifications, as playing a critical role in transforming raw EHR data from a relational database into a timeline of medical events.

“The raw health records are patient de-identified and are transmitted over computer networks and stored in a relational database (RDB) and converted by a computer system functioning as a converter into a standardized format and stored in the memory,” the patent reads.

The standardized format is FHIR, and the bundles of time-sequenced resources form the basis of a patient’s medical timeline and future predictions.

“When converting data to a FHIR format, we did not harmonize variable names to a standard terminology but instead used the raw terminology provided by the health system, bypassing the traditional time-consuming harmonization of data,” the authors add.

In terms of the infrastructure necessary for supporting the service, the application leaves the door open for both on- and off-premise hardware and software approaches.

“The precise physical location and implementation of the predictive models and related computer or computer system may vary. In some instances it may be physically located at a medical system or hospital serving affiliated facilities, primary care physician offices, and related clinics, etc.” the patent states. “In other situations it may be centrally located and receive EHRs and transmit predicted future clinical events and related prior medical events over wide area computer networks and service a multitude of unrelated healthcare institutions in a fee for service, subscription, standalone product, or other business model.”

In either setup, rich data sets of aggregated EHRs will be necessary to yield reliable results from machine learning models and algorithms.

“The deep learning models are trained to predict one or more future clinical events and to summarize pertinent past medical events (e.g., problems, conditions, test results, medications, etc.) related to the predicted one or more future clinical events on an input electronic health record of a given patient,” note Mossin et al.

The patent speaks to “hundreds of thousands or even millions of patients” from “one or more institutions” that will supply the raw data necessary for its predictive models. The motivation for supplying this information is a difficult hurdle to overcome.

“A second limitation is that our approach relies on large datasets, powerful computing infrastructure, and complex algorithms, which require sophisticated engineering to replicate. However, this approach is what allows a single modeling architecture to achieve excellent predictive performance across a range of prediction tasks,” the authors admit.

With cloud computing and services growing in use across healthcare enterprises, Google appears eager to capitalize on the limitations of current EHR systems by putting its impressive infrastructure to work on preventive medicine.

X

EHRIntelligence

Sign up to continue reading and gain Free Access to all our resources.

Sign up for our free newsletter and join 60,000 of your peers to stay up to date with tips and advice on:

EHR Optimization
EHR Interoperability
MACRA

White Papers, Webcasts, Featured Articles and Exclusive Interviews

Our privacy policy


no, thanks

Continue to site...