Tech giant Google is planning to build its own EHR (Electronic Health Record) for clinicians that collect patients’ medical records and then leverages Machine Learning to envisage clinical outcomes. For this, Google has applied a patent for a Deep learning system that amasses EHR data into a timeline.
The system and technique for foreseeing and summarizing medical events from EHRs are based on the popular data standard, called FHIR (Fast Healthcare Interoperability Resources), enabling Google to ingest data from a wide range of sources. The patent application first filed in August 2017, where the company argued that existing techniques of collecting and evaluating health data for prognostic purposes are insufficient, and needed much more time and effort to be scalable and repeatable. The 40-page application explains a new computer system, with a healthcare provider-facing interface which forecasts and summarizes medical events from EHRs. The application for patent explained a system that comprising three parts- a computer memory or database for storing amassed health records from millions of patients, a computer or processing unit to implement Machine Learning models on patients’ health records transformed into sole standardized data structure format, and an end-user device like computer terminal, tablet or smartphone, which shows healthcare providers the outcomes and relevant past medical events for patients.
The patent application has yet to be approved, and it is undefined whether the U.S. Patent and Trademark Office will move forward with the claim. If the patent is getting clearance, it will surely cause some stir among the thousands of healthcare providers working in predictive analytics, but it isn’t likely to disrupt the inexorable shift away from conventional analytics methods towards Deep Learning and other AI strategies.