Technology giant Google may be working on an EHR system that can predict patients' clinical outcomes and summarize patients' previous medical events, according to a patent application published Thursday by the U.S. Patent and Trademark Office, Politico's "Morning eHealth" reports.
How the application describes the new EHR system
The patent application, which Google originally filed in August 2017, lists 21 Google employees as inventors of the new EHR system, including Jeffrey Dean, head of Google AI.
According to the application, the EHR system consists of three parts:
- A computer memory that stores aggregated health records for millions of patients;
- A computer system that leverages machine learning to analyze health records in a standardized data structure format; and
- A computer interface for health care providers that predicts patients' future medical events and summarizes past medical events.
According to the application, the computer system would store a diverse range of patient data—including laboratory values, diagnoses, vital signs, medical notes, and medications—that would inform the system's predictions of future clinical events. The provider-facing computer system also would indicate the underlying medical data that contributed to its predictions, according to "Morning eHealth."
However, according to Advisory Board's Greg Kuhnen, media reports may be overstating the patent's aims. The figures in the patent application seem to suggest a module that sits outside institutions' EHR, which would be more accurately labeled Clinical Decision Support (CDS) technology than a true EHR system.
What the new system would mean for providers
The inventors wrote in their application that the new system would help health care providers "allocate their attention efficiently" to the most critical and timely information by helping them "addres[s] a pressing question facing the physician ... namely, which patients have the highest need for my attention now, and, at an individual level, what information in the patient's chart should I attend to."
In a hypothetical example cited in the application, a patient presents to the ED with abdominal pain caused by alcohol withdrawal—but the clinicians mistake his symptoms for septic shock and give him the wrong treatment. According to the inventors, the Google system could alert physicians that the patient was at risk for alcohol withdrawal symptom by "allocat[ing] their attention" to the patient's medical history, FierceHealthcare reports.
A Google spokesperson did not respond to Politico's request for comment on the patent application but instead referred to an article published in the journal Nature last May, in which Google scientists—in collaboration with researchers from UC San Francisco, Stanford Medicine, and the University of Chicago Medicine— described a similar project that would use machine learning to predict clinical outcomes.
In a blog post about the collaboration, the scientists said their model, which "picks up signals about the patient, their treatments, and notes written by their clinician, ... is more like a good listener than a master diagnostician" (Allen, "Morning eHealth," Politico, 2/4; Spitzer, Becker's Hospital Review, 2/5; Landi, FierceHealthcare, 2/4).
Cheat sheets: Understand interoperability, EHR optimization, and more
Download our cheat sheets so you can keep track of the fast-changing technologies and capitalize on opportunities for IT-powered innovation. Check out our guides for these topics and more: